UK Microwave Group Wiki - User contributions [en-gb]

"Franco" SU-02 10GHz PCB

2024-02-14T13:35:02Z

M0JUR: /* Example of tunable fixed bias */

[[File:Franco PCB.png|400px]]

Ex equipment low noise amplifer PCB, featuring PTFE substrate and 4 independent amplifiers (labelled AH, AV, EH and EV) based on the ultra-low-noise pseudomorphic HJ FET NE32584C.

These PCBs have been sourced from RF Microwave shop in Italy (originally run by Franco Rota, I2FHW - hence the nickname of the board) and can be used to make a very good 5.6GHz LNA, 10GHz LNA, 24GHz doubler, or just a source of the NE32584c devices or 50-Ohm microstrip lines.

Original DUBUS article + the NE32584C datasheet: [[:media:Franco PCB.pdf]]

== Circuit diagrams ==

<pdf height="600">File:Su02 schematics amp.pdf</pdf> <pdf height="600">File:Su02 schematics psu.pdf</pdf>

Published with thanks to Chris, G3WIE for having reverse-engineered the circuits of the Franco board posted the his findings to on the UK Microwave Ops mailing list.

== Power supply and biasing ==

Each Franco board comes with a power supply capable of providing both positive and negative voltage to a single amplifier section at a time, with the input voltage range from approx. 8V (the limiting factor here being the operational range of the first voltage regulator) to about 20V (above which one risks frying the on-board IC). Furthermore, each amplifier section features an active bias section configured to maintain nominal drain current on the NE32584C given the voltages provided by the on-board PSU. For details, see below.

Should you want to use your own PSU and/or use fixed bias instead, keep in mind the following:
* the NE32584C is rated for the absolute maximum drain-to-source voltage of '''4.0V''' and the absolute maximum gate-to-source voltage of '''-3.0V''',
* its nominal drain current is 10mA,
* as usual with a FET, care should be taken keep the magic smoke in by ensuring Vds is present ''if and only if'' Vgs is.

=== On-board PSU ===

The power-supply section of the board produces the supply voltages of +3.1V and -3.3V for the amplifier sections as well as controlling which of the amps is powered up. The main workhorse here is the hex Schmitt-trigger inverter 74HC14.

To generate supply voltages, energise the "14V" pad near the large voltage regulator. Do '''not''' apply more than 20V here, it will damage some of the inverter gates. On the other end of the scale, 8V has been shown to suffice.

==== Output switching ====

Switching between AV/EV and AH/EH is achieved by varying the voltage applied to the "14V" pad - anything under approx. 15.5V enables the V pair, anything above that threshold enables the H pair. In the terms of the schematics above, varying the input voltage governs which of the lines labelled A and B is at 0V (which allows the amp connected to it to switch on) and which one is at 5V.

As for switching between AH/AV and EH/EV, which pair receives power is governed by the voltage applied to the pads: "10V" and "RX12". By default, ''i.e.'' with both of these pads unconnected, the PSU directs Vdd/Vbb to the E pair. With 12V applied to the "RX12" pad (or 8.5-9V to the "10V" pad), the A pair is powered up instead. In the terms of the schematics above, this affects which of the two bipolar transistors: Q1 and Q2 is switched on.

=== Active bias ===

Each of the four amplifier sections features a BJT-based active bias circuit with resistor values chosen so that for the Vdd/Vbb value of 3.1V and the Vgg (bias) voltage of -3.3V, both as provided by the on-board PSU, the NE325 ends up at Vds = 2V and Id = 11mA. The same circuit also serves as an on-off switch - in order for the section in question to operate the "A or B" must be at 0V.

G3WIE has created a spreadsheet which can be used to calculate resistor values for given supply voltages and desired FET parameters, or vice versa: [[File:Hemt active biasV3.ods]]

Warning: if you choose to retain the active bias circuit, its is ''critical'' for Vdd and Vbb to be tied together as per the original design. If they are not and you suddenly lose Vbb but not Vdd, it's bye-bye FET.

=== Example of tunable fixed bias ===

While there are of course many ways of biasing a FET, a common approach is to install a trimmer resistor between the source of Vgg and the FET gate so that bias voltage, and by extension drain current, can be tuned for best signal-to-noise ratio. One way of doing this on a Franco board is as follows:
* Remove Q3 and R19. You might also salvage R15 and R16, as they serve no function with Q3 gone
* Cut the track between R19 was, and R20. Where exactly to do this depends on the amplifier in question; for AV and EH the best place is probably the short length of track between C16 and where the collector of Q3 was, in case of AH and EV aim for the short diagonal track between C16 and C17
* Grab a small trimpot, with maximum resistance of 10k or so. Both PC-pin and SMD should work here so choose whichever you feel more comfortable soldering, that said if you opt for the pins make sure you do not have them short-circuiting other parts of the amp. Install it as follows:
** the first (input) contact - between where R19 was and the cut you have just made
** the second (variable-resistance) contact - between R20 and the cut you have just made
** the third (fixed-resistance) contact - to ground. The pad grounding C17 is quite convenient for this
* IMPORTANT: Before doing anything else, turn the trimpot so that it offers high resistance between R20 and where R19 was. You want to start with ''low'' Id
* Connect the wire which will provide Vgg to where R19 was
* Grab a meter and confirm you have got everything wired connectly. You want:
** (high, adjustable resistance + 100R) between the source of Vgg and the drain,
** 10k between the source of Vgg and ground, and
** ''no'' short between R20 and ground (if you do measure a short here, before disassembling everything turn the trimpot a little bit away from maximum and try again).

== 10GHz ==

With the board in question having been designed for 10GHz operation it is perhaps no surprise that the choice of variants is rather wide here.

=== Single-stage LNAs by Gerald, F6CXO ===

Probably the simplest variants (there is one for SMA input and one for direct waveguide input) out there, as bias circuitry and socket fitting aside the only modifications needed is removing a resistor from the RF input line + replacing the resistor on the RF output line with a small DC-blocking capacitor. Claimed performance is 14dB gain, 0.7-0.8dB noise figure. For details see the VHF Communications 3/2004 article on pages 8-10 in the PDF above; the circuit diagram has only been annotated in French but the comments are simple enough to understand for an English speaker).

=== Two-stage LNA by Pete, GM4BYF ===

In the aforementioned article Gerald, F6CXO mentions the possibility of combining two Franco amplifier sections into a two-stage device. Indeed, the file above includes (on pages 4-7) an Italian-language article - unfortunately with no translation, and a very low-resolution circuit diagram - featuring photos of two devices which appear to have done just that. However, it has been observed that merely connecting the output of amp one to the input of amp two can cause stability problems.{{Citation needed}}

In order to avoid the stability issues Pete, GM4BYF has constructed a two-stage version in which the two amplifier sections are laid out back to back, and connected with a short length of semi-rigid coax fed through a specially drilled hole. Details can be found here: [[File:Franco 10GHz LNA GM4BYF.pdf]].

TODO: add performance data, if/when available

=== Dual two-stage LNA by Geoff, GI0GDP ===

This variant uses ''two'' two-stage amplifiers running in parallel to produce around 75mW output from 2.5mW input. In order to avoid excessive amounts of cutting and reassembling the PCB, the two input-stage FETs are turned upside down.

[[File:Franco 10GHz four-amp LNA GI0GDP.jpg|800px]]

Modification details:
* cut two sections out of a complete Franco PCB: one with both E-pair amps (which will serve as the input stage), one with the A pair (for the output stage)
* join the sections as show in the photo, making sure the back plane is nicely joined (possibly with copper foil + some solder) as well. Make sure the RF lines are ''not'' connected, though
* connect the two sections' RF lines using ATC capacitors. Geoff has used 2pF ones here but in his own words, the exact capacitance doesn't really matter
* desolder and turn around the two E-pair FETs
* relocate the E-pair source and drain resistors (yes, all four of them!) to reflect the new orientation of the FETs. When in doubt what goes where, consult the photo
* replace the two 68R drain resistors in the ''A-pair'' (''i.e.'' output) amps with jumpers. Note that the photo shows 1R0 rather than 0R chips owing to the fact the author did not have any of the latter at hand at the time
* remove and jumper the four resistors from the RF lines
* make your own biasing arrangements as desired, bearing in mind that with E-pair FETs having been turned around the original active bias circuit will ''not'' work without extensive modifications

Nb. Disregard the two through-hole ceramic capacitors, they are only there to provide mechanical support and extra soldering surface.

=== Two-stage "NE325+MGF1601" LNA by Geoff, GI0GDP ===

In Geoff's latest take on the subject, a single two-stage assembly has its output-stage NE325 replaced with an MGF1601 clone ("the case is not quite right for a genuine Mitsubishi version but that isn't a show stopper" -- GI0GDP) readily available from Far Eastern suppliers. Reasonably expected performance is 100mW out for 5mW in, that said it will vary depending on the details of your assembly, what kind of enclosure you use (and if it has a lid on or not), and how much time you are willing to spend snowflake-tuning the NE325 gate. Do keep in mind that the two FET types have got quite different voltage and biasing requirements.

The project in question was described in the November/December 2023 issue of Scatterpoint, available to UKuG members.

== 5.6GHz ==

[https://wiki.batc.org.uk/5.6_GHz#Receive_Pre-amplifiers The BATC Wiki] shows modifications made by Daniel, DL3IAE for the use of a single SU-02 amp in this band.

== 24GHz ==

TBA

"Franco" SU-02 10GHz PCB

2024-02-14T13:34:08Z

M0JUR: Add instructions for (one way of) replacing active bias with a trimpot

[[File:Franco PCB.png|400px]]

Ex equipment low noise amplifer PCB, featuring PTFE substrate and 4 independent amplifiers (labelled AH, AV, EH and EV) based on the ultra-low-noise pseudomorphic HJ FET NE32584C.

These PCBs have been sourced from RF Microwave shop in Italy (originally run by Franco Rota, I2FHW - hence the nickname of the board) and can be used to make a very good 5.6GHz LNA, 10GHz LNA, 24GHz doubler, or just a source of the NE32584c devices or 50-Ohm microstrip lines.

Original DUBUS article + the NE32584C datasheet: [[:media:Franco PCB.pdf]]

== Circuit diagrams ==

<pdf height="600">File:Su02 schematics amp.pdf</pdf> <pdf height="600">File:Su02 schematics psu.pdf</pdf>

Published with thanks to Chris, G3WIE for having reverse-engineered the circuits of the Franco board posted the his findings to on the UK Microwave Ops mailing list.

== Power supply and biasing ==

Each Franco board comes with a power supply capable of providing both positive and negative voltage to a single amplifier section at a time, with the input voltage range from approx. 8V (the limiting factor here being the operational range of the first voltage regulator) to about 20V (above which one risks frying the on-board IC). Furthermore, each amplifier section features an active bias section configured to maintain nominal drain current on the NE32584C given the voltages provided by the on-board PSU. For details, see below.

Should you want to use your own PSU and/or use fixed bias instead, keep in mind the following:
* the NE32584C is rated for the absolute maximum drain-to-source voltage of '''4.0V''' and the absolute maximum gate-to-source voltage of '''-3.0V''',
* its nominal drain current is 10mA,
* as usual with a FET, care should be taken keep the magic smoke in by ensuring Vds is present ''if and only if'' Vgs is.

=== On-board PSU ===

The power-supply section of the board produces the supply voltages of +3.1V and -3.3V for the amplifier sections as well as controlling which of the amps is powered up. The main workhorse here is the hex Schmitt-trigger inverter 74HC14.

To generate supply voltages, energise the "14V" pad near the large voltage regulator. Do '''not''' apply more than 20V here, it will damage some of the inverter gates. On the other end of the scale, 8V has been shown to suffice.

==== Output switching ====

Switching between AV/EV and AH/EH is achieved by varying the voltage applied to the "14V" pad - anything under approx. 15.5V enables the V pair, anything above that threshold enables the H pair. In the terms of the schematics above, varying the input voltage governs which of the lines labelled A and B is at 0V (which allows the amp connected to it to switch on) and which one is at 5V.

As for switching between AH/AV and EH/EV, which pair receives power is governed by the voltage applied to the pads: "10V" and "RX12". By default, ''i.e.'' with both of these pads unconnected, the PSU directs Vdd/Vbb to the E pair. With 12V applied to the "RX12" pad (or 8.5-9V to the "10V" pad), the A pair is powered up instead. In the terms of the schematics above, this affects which of the two bipolar transistors: Q1 and Q2 is switched on.

=== Active bias ===

Each of the four amplifier sections features a BJT-based active bias circuit with resistor values chosen so that for the Vdd/Vbb value of 3.1V and the Vgg (bias) voltage of -3.3V, both as provided by the on-board PSU, the NE325 ends up at Vds = 2V and Id = 11mA. The same circuit also serves as an on-off switch - in order for the section in question to operate the "A or B" must be at 0V.

G3WIE has created a spreadsheet which can be used to calculate resistor values for given supply voltages and desired FET parameters, or vice versa: [[File:Hemt active biasV3.ods]]

Warning: if you choose to retain the active bias circuit, its is ''critical'' for Vdd and Vbb to be tied together as per the original design. If they are not and you suddenly lose Vbb but not Vdd, it's bye-bye FET.

=== Example of tunable fixed bias ===

While there are of course many ways of biasing a FET, a common approach is to install a trimmer resistor between the source of Vgg and the FET gate so that bias voltage, and by extension drain current, can be tuned for best signal-to-noise ratio. One way of doing this on a Franco board is as follows:
# Remove Q3 and R19. You might also salvage R15 and R16, as they serve no function with Q3 gone
# Cut the track between R19 was, and R20. Where exactly to do this depends on the amplifier in question; for AV and EH the best place is probably the short length of track between C16 and where the collector of Q3 was, in case of AH and EV aim for the short diagonal track between C16 and C17
# Grab a small trimpot, with maximum resistance of 10k or so. Both PC-pin and SMD should work here so choose whichever you feel more comfortable soldering, that said if you opt for the pins make sure you do not have them short-circuiting other parts of the amp. Install it as follows:
** the first (input) contact - between where R19 was and the cut you have just made
** the second (variable-resistance) contact - between R20 and the cut you have just made
** the third (fixed-resistance) contact - to ground. The pad grounding C17 is quite convenient for this
# IMPORTANT: Before doing anything else, turn the trimpot so that it offers high resistance between R20 and where R19 was. You want to start with ''low'' Id
# Connect the wire which will provide Vgg to where R19 was
# Grab a meter and confirm you have got everything wired connectly. You want:
* (high, adjustable resistance + 100R) between the source of Vgg and the drain,
* 10k between the source of Vgg and ground, and
* ''no'' short between R20 and ground (if you do measure a short here, before disassembling everything turn the trimpot a little bit away from maximum and try again).

== 10GHz ==

With the board in question having been designed for 10GHz operation it is perhaps no surprise that the choice of variants is rather wide here.

=== Single-stage LNAs by Gerald, F6CXO ===

Probably the simplest variants (there is one for SMA input and one for direct waveguide input) out there, as bias circuitry and socket fitting aside the only modifications needed is removing a resistor from the RF input line + replacing the resistor on the RF output line with a small DC-blocking capacitor. Claimed performance is 14dB gain, 0.7-0.8dB noise figure. For details see the VHF Communications 3/2004 article on pages 8-10 in the PDF above; the circuit diagram has only been annotated in French but the comments are simple enough to understand for an English speaker).

=== Two-stage LNA by Pete, GM4BYF ===

In the aforementioned article Gerald, F6CXO mentions the possibility of combining two Franco amplifier sections into a two-stage device. Indeed, the file above includes (on pages 4-7) an Italian-language article - unfortunately with no translation, and a very low-resolution circuit diagram - featuring photos of two devices which appear to have done just that. However, it has been observed that merely connecting the output of amp one to the input of amp two can cause stability problems.{{Citation needed}}

In order to avoid the stability issues Pete, GM4BYF has constructed a two-stage version in which the two amplifier sections are laid out back to back, and connected with a short length of semi-rigid coax fed through a specially drilled hole. Details can be found here: [[File:Franco 10GHz LNA GM4BYF.pdf]].

TODO: add performance data, if/when available

=== Dual two-stage LNA by Geoff, GI0GDP ===

This variant uses ''two'' two-stage amplifiers running in parallel to produce around 75mW output from 2.5mW input. In order to avoid excessive amounts of cutting and reassembling the PCB, the two input-stage FETs are turned upside down.

[[File:Franco 10GHz four-amp LNA GI0GDP.jpg|800px]]

Modification details:
* cut two sections out of a complete Franco PCB: one with both E-pair amps (which will serve as the input stage), one with the A pair (for the output stage)
* join the sections as show in the photo, making sure the back plane is nicely joined (possibly with copper foil + some solder) as well. Make sure the RF lines are ''not'' connected, though
* connect the two sections' RF lines using ATC capacitors. Geoff has used 2pF ones here but in his own words, the exact capacitance doesn't really matter
* desolder and turn around the two E-pair FETs
* relocate the E-pair source and drain resistors (yes, all four of them!) to reflect the new orientation of the FETs. When in doubt what goes where, consult the photo
* replace the two 68R drain resistors in the ''A-pair'' (''i.e.'' output) amps with jumpers. Note that the photo shows 1R0 rather than 0R chips owing to the fact the author did not have any of the latter at hand at the time
* remove and jumper the four resistors from the RF lines
* make your own biasing arrangements as desired, bearing in mind that with E-pair FETs having been turned around the original active bias circuit will ''not'' work without extensive modifications

Nb. Disregard the two through-hole ceramic capacitors, they are only there to provide mechanical support and extra soldering surface.

=== Two-stage "NE325+MGF1601" LNA by Geoff, GI0GDP ===

In Geoff's latest take on the subject, a single two-stage assembly has its output-stage NE325 replaced with an MGF1601 clone ("the case is not quite right for a genuine Mitsubishi version but that isn't a show stopper" -- GI0GDP) readily available from Far Eastern suppliers. Reasonably expected performance is 100mW out for 5mW in, that said it will vary depending on the details of your assembly, what kind of enclosure you use (and if it has a lid on or not), and how much time you are willing to spend snowflake-tuning the NE325 gate. Do keep in mind that the two FET types have got quite different voltage and biasing requirements.

The project in question was described in the November/December 2023 issue of Scatterpoint, available to UKuG members.

== 5.6GHz ==

[https://wiki.batc.org.uk/5.6_GHz#Receive_Pre-amplifiers The BATC Wiki] shows modifications made by Daniel, DL3IAE for the use of a single SU-02 amp in this band.

== 24GHz ==

TBA

"Franco" SU-02 10GHz PCB

2024-02-14T12:45:08Z

M0JUR: Replace hand-drawn circuit diagram with KiCad versions (G3WIE-approved), reorganise the power section a bit

[[File:Franco PCB.png|400px]]

Ex equipment low noise amplifer PCB, featuring PTFE substrate and 4 independent amplifiers (labelled AH, AV, EH and EV) based on the ultra-low-noise pseudomorphic HJ FET NE32584C.

These PCBs have been sourced from RF Microwave shop in Italy (originally run by Franco Rota, I2FHW - hence the nickname of the board) and can be used to make a very good 5.6GHz LNA, 10GHz LNA, 24GHz doubler, or just a source of the NE32584c devices or 50-Ohm microstrip lines.

Original DUBUS article + the NE32584C datasheet: [[:media:Franco PCB.pdf]]

== Circuit diagrams ==

<pdf height="600">File:Su02 schematics amp.pdf</pdf> <pdf height="600">File:Su02 schematics psu.pdf</pdf>

Published with thanks to Chris, G3WIE for having reverse-engineered the circuits of the Franco board posted the his findings to on the UK Microwave Ops mailing list.

== Power supply and biasing ==

Each Franco board comes with a power supply capable of providing both positive and negative voltage to a single amplifier section at a time, with the input voltage range from approx. 8V (the limiting factor here being the operational range of the first voltage regulator) to about 20V (above which one risks frying the on-board IC). Furthermore, each amplifier section features an active bias section configured to maintain nominal drain current on the NE32584C given the voltages provided by the on-board PSU. For details, see below.

Should you want to use your own PSU and/or use fixed bias instead, keep in mind the following:
* the NE32584C is rated for the absolute maximum drain-to-source voltage of '''4.0V''' and the absolute maximum gate-to-source voltage of '''-3.0V''',
* its nominal drain current is 10mA,
* as usual with a FET, care should be taken keep the magic smoke in by ensuring Vds is present ''if and only if'' Vgs is.

=== On-board PSU ===

The power-supply section of the board produces the supply voltages of +3.1V and -3.3V for the amplifier sections as well as controlling which of the amps is powered up. The main workhorse here is the hex Schmitt-trigger inverter 74HC14.

To generate supply voltages, energise the "14V" pad near the large voltage regulator. Do '''not''' apply more than 20V here, it will damage some of the inverter gates. On the other end of the scale, 8V has been shown to suffice.

==== Output switching ====

Switching between AV/EV and AH/EH is achieved by varying the voltage applied to the "14V" pad - anything under approx. 15.5V enables the V pair, anything above that threshold enables the H pair. In the terms of the schematics above, varying the input voltage governs which of the lines labelled A and B is at 0V (which allows the amp connected to it to switch on) and which one is at 5V.

As for switching between AH/AV and EH/EV, which pair receives power is governed by the voltage applied to the pads: "10V" and "RX12". By default, ''i.e.'' with both of these pads unconnected, the PSU directs Vdd/Vbb to the E pair. With 12V applied to the "RX12" pad (or 8.5-9V to the "10V" pad), the A pair is powered up instead. In the terms of the schematics above, this affects which of the two bipolar transistors: Q1 and Q2 is switched on.

=== Active bias ===

Each of the four amplifier sections features a BJT-based active bias circuit with resistor values chosen so that for the Vdd/Vbb value of 3.1V and the Vgg (bias) voltage of -3.3V, both as provided by the on-board PSU, the NE325 ends up at Vds = 2V and Id = 11mA. The same circuit also serves as an on-off switch - in order for the section in question to operate the "A or B" must be at 0V.

G3WIE has created a spreadsheet which can be used to calculate resistor values for given supply voltages and desired FET parameters, or vice versa: [[File:Hemt active biasV3.ods]]

Warning: if you choose to retain the active bias circuit, its is ''critical'' for Vdd and Vbb to be tied together as per the original design. If they are not and you suddenly lose Vbb but not Vdd, it's bye-bye FET.

== 10GHz ==

With the board in question having been designed for 10GHz operation it is perhaps no surprise that the choice of variants is rather wide here.

=== Single-stage LNAs by Gerald, F6CXO ===

Probably the simplest variants (there is one for SMA input and one for direct waveguide input) out there, as bias circuitry and socket fitting aside the only modifications needed is removing a resistor from the RF input line + replacing the resistor on the RF output line with a small DC-blocking capacitor. Claimed performance is 14dB gain, 0.7-0.8dB noise figure. For details see the VHF Communications 3/2004 article on pages 8-10 in the PDF above; the circuit diagram has only been annotated in French but the comments are simple enough to understand for an English speaker).

=== Two-stage LNA by Pete, GM4BYF ===

In the aforementioned article Gerald, F6CXO mentions the possibility of combining two Franco amplifier sections into a two-stage device. Indeed, the file above includes (on pages 4-7) an Italian-language article - unfortunately with no translation, and a very low-resolution circuit diagram - featuring photos of two devices which appear to have done just that. However, it has been observed that merely connecting the output of amp one to the input of amp two can cause stability problems.{{Citation needed}}

In order to avoid the stability issues Pete, GM4BYF has constructed a two-stage version in which the two amplifier sections are laid out back to back, and connected with a short length of semi-rigid coax fed through a specially drilled hole. Details can be found here: [[File:Franco 10GHz LNA GM4BYF.pdf]].

TODO: add performance data, if/when available

=== Dual two-stage LNA by Geoff, GI0GDP ===

This variant uses ''two'' two-stage amplifiers running in parallel to produce around 75mW output from 2.5mW input. In order to avoid excessive amounts of cutting and reassembling the PCB, the two input-stage FETs are turned upside down.

[[File:Franco 10GHz four-amp LNA GI0GDP.jpg|800px]]

Modification details:
* cut two sections out of a complete Franco PCB: one with both E-pair amps (which will serve as the input stage), one with the A pair (for the output stage)
* join the sections as show in the photo, making sure the back plane is nicely joined (possibly with copper foil + some solder) as well. Make sure the RF lines are ''not'' connected, though
* connect the two sections' RF lines using ATC capacitors. Geoff has used 2pF ones here but in his own words, the exact capacitance doesn't really matter
* desolder and turn around the two E-pair FETs
* relocate the E-pair source and drain resistors (yes, all four of them!) to reflect the new orientation of the FETs. When in doubt what goes where, consult the photo
* replace the two 68R drain resistors in the ''A-pair'' (''i.e.'' output) amps with jumpers. Note that the photo shows 1R0 rather than 0R chips owing to the fact the author did not have any of the latter at hand at the time
* remove and jumper the four resistors from the RF lines
* make your own biasing arrangements as desired, bearing in mind that with E-pair FETs having been turned around the original active bias circuit will ''not'' work without extensive modifications

Nb. Disregard the two through-hole ceramic capacitors, they are only there to provide mechanical support and extra soldering surface.

=== Two-stage "NE325+MGF1601" LNA by Geoff, GI0GDP ===

In Geoff's latest take on the subject, a single two-stage assembly has its output-stage NE325 replaced with an MGF1601 clone ("the case is not quite right for a genuine Mitsubishi version but that isn't a show stopper" -- GI0GDP) readily available from Far Eastern suppliers. Reasonably expected performance is 100mW out for 5mW in, that said it will vary depending on the details of your assembly, what kind of enclosure you use (and if it has a lid on or not), and how much time you are willing to spend snowflake-tuning the NE325 gate. Do keep in mind that the two FET types have got quite different voltage and biasing requirements.

The project in question was described in the November/December 2023 issue of Scatterpoint, available to UKuG members.

== 5.6GHz ==

[https://wiki.batc.org.uk/5.6_GHz#Receive_Pre-amplifiers The BATC Wiki] shows modifications made by Daniel, DL3IAE for the use of a single SU-02 amp in this band.

== 24GHz ==

TBA

File:Su02 schematics amp.pdf

2024-02-14T11:59:47Z

M0JUR: Digitised and extended version of the G3WIE circuit diagram of the "Franco" SU-02 PCB. Part 2/2: amplifier (one of four). Reviewed and approved by the author.

Digitised and extended version of the G3WIE circuit diagram of the "Franco" SU-02 PCB. Part 2/2: amplifier (one of four). Reviewed and approved by the author.

File:Su02 schematics psu.pdf

2024-02-14T11:59:20Z

M0JUR: Digitised and extended version of the G3WIE circuit diagram of the "Franco" SU-02 PCB. Part 1/2: power supply. Reviewed and approved by the author.

Digitised and extended version of the G3WIE circuit diagram of the "Franco" SU-02 PCB. Part 1/2: power supply. Reviewed and approved by the author.

"Franco" SU-02 10GHz PCB

2024-02-08T14:09:15Z

M0JUR: At a glance on the Web, BC807-40 is the only PNP BJT in the SOT-23-3 package using the device code "5C"

[[File:Franco PCB.png|400px]]

Ex equipment low noise amplifer PCB, featuring PTFE substrate and 4 independent amplifiers (labelled AH, AV, EH and EV) based on the ultra-low-noise pseudomorphic HJ FET NE32584C.

These PCBs have been sourced from RF Microwave shop in Italy (originally run by Franco Rota, I2FHW - hence the nickname of the board) and can be used to make a very good 5.6GHz LNA, 10GHz LNA, 24GHz doubler, or just a source of the NE32584c devices or 50-Ohm microstrip lines.

Original DUBUS article + the NE32584C datasheet: [[:media:Franco PCB.pdf]]

'''WORK IN PROGRESS. If you have time and know more details about any of the projects below, or even a new project, please add them!'''

== Power supply and biasing ==

The NE32584C is rated for the absolute maximum drain-to-source voltage of '''4.0V''' and the absolute maximum gate-to-source voltage of '''-3.0V'''. The nominal drain current is 10mA. As usual with a FET, care should be taken keep the magic smoke in by ensuring Vds is present ''if and only if'' Vgs is.

[[File:SU-02_circuit_schematics.jpeg]]

Chris, G3WIE has reverse-engineered the circuits of the Franco board and produced the schematics show above. All bipolar transistors feature the same markings and have been confirmed as PNP; given the markings, the package and the measured hFE they are most likely BC807-40. The element marked with a question mark in the schematics has since been identified as a Zener diode.

=== PSU ===

The power-supply section of the board produces the supply voltages of +3.1V and -3.3V for the amplifier sections as well as controlling which of the amps is powered up. The main workhorse here is the hex Schmitt-trigger inverter 74HC14.

To generate supply voltages, energise the "14V" pad near the large voltage regulator. Do '''not''' apply more than 20V here, it will damage some of the inverter gates. On the other end of the scale, 8V has been shown to suffice.

==== Output switching ====

Switching between AV/EV and AH/EH is achieved by varying the voltage applied to the "14V" pad - anything under approx. 15.5V enables the V pair, anything above that threshold enables the H pair. In the terms of the schematics above, varying the input voltage governs which of the lines labelled A and B is at 0V (which allows the amp connected to it to switch on) and which one is at 5V.

As for switching between AH/AV and EH/EV, which pair receives power is governed by the voltage applied to the pads: "10V" and "RX12". By default, ''i.e.'' with both of these pads unconnected, the PSU directs Vdd/Vbb to the E pair. With 12V applied to the "RX12" pad (or 8.5-9V to the "10V" pad), the A pair is powered up instead. In the terms of the schematics above, this affects which of the two bipolar transistors: Q1 and Q2 is switched on.

=== Active bias ===

Each of the four amplifier sections features a BJT-based active bias circuit with resistor values chosen so that for the Vdd/Vbb value of 3.1V and the Vgg (bias) voltage of -3.3V, both as provided by the on-board PSU, the NE325 ends up at Vds = 2V and Id = 11mA. The same circuit also serves as an on-off switch - in order for the section in question to operate the "A or B" must be at 0V.

G3WIE has created a spreadsheet which can be used to calculate resistor values for given supply voltages and desired FET parameters, or vice versa: [[File:Hemt active biasV3.ods]]

Warning: if you choose to retain the active bias circuit, its is ''critical'' for Vdd and Vbb to be tied together as per the original design. If they are not and you suddenly lose Vbb but not Vdd, it's bye-bye FET.

== 10GHz ==

With the board in question having been designed for 10GHz operation it is perhaps no surprise that the choice of variants is rather wide here.

=== Single-stage LNAs by Gerald, F6CXO ===

Probably the simplest variants (there is one for SMA input and one for direct waveguide input) out there, as bias circuitry and socket fitting aside the only modifications needed is removing a resistor from the RF input line + replacing the resistor on the RF output line with a small DC-blocking capacitor. Claimed performance is 14dB gain, 0.7-0.8dB noise figure. For details see the VHF Communications 3/2004 article on pages 8-10 in the PDF above; the circuit diagram has only been annotated in French but the comments are simple enough to understand for an English speaker).

=== Two-stage LNA by Pete, GM4BYF ===

In the aforementioned article Gerald, F6CXO mentions the possibility of combining two Franco amplifier sections into a two-stage device. Indeed, the file above includes (on pages 4-7) an Italian-language article - unfortunately with no translation, and a very low-resolution circuit diagram - featuring photos of two devices which appear to have done just that. However, it has been observed that merely connecting the output of amp one to the input of amp two can cause stability problems.{{Citation needed}}

In order to avoid the stability issues Pete, GM4BYF has constructed a two-stage version in which the two amplifier sections are laid out back to back, and connected with a short length of semi-rigid coax fed through a specially drilled hole. Details can be found here: [[File:Franco 10GHz LNA GM4BYF.pdf]].

TODO: add performance data, if/when available

=== Dual two-stage LNA by Geoff, GI0GDP ===

This variant uses ''two'' two-stage amplifiers running in parallel to produce around 75mW output from 2.5mW input. In order to avoid excessive amounts of cutting and reassembling the PCB, the two input-stage FETs are turned upside down.

[[File:Franco 10GHz four-amp LNA GI0GDP.jpg|800px]]

Modification details:
* cut two sections out of a complete Franco PCB: one with both E-pair amps (which will serve as the input stage), one with the A pair (for the output stage)
* join the sections as show in the photo, making sure the back plane is nicely joined (possibly with copper foil + some solder) as well. Make sure the RF lines are ''not'' connected, though
* connect the two sections' RF lines using ATC capacitors. Geoff has used 2pF ones here but in his own words, the exact capacitance doesn't really matter
* desolder and turn around the two E-pair FETs
* relocate the E-pair source and drain resistors (yes, all four of them!) to reflect the new orientation of the FETs. When in doubt what goes where, consult the photo
* replace the two 68R drain resistors in the ''A-pair'' (''i.e.'' output) amps with jumpers. Note that the photo shows 1R0 rather than 0R chips owing to the fact the author did not have any of the latter at hand at the time
* remove and jumper the four resistors from the RF lines
* make your own biasing arrangements as desired, bearing in mind that with E-pair FETs having been turned around the original active bias circuit will ''not'' work without extensive modifications

Nb. Disregard the two through-hole ceramic capacitors, they are only there to provide mechanical support and extra soldering surface.

=== Two-stage "NE325+MGF1601" LNA by Geoff, GI0GDP ===

In Geoff's latest take on the subject, a single two-stage assembly has its output-stage NE325 replaced with an MGF1601 clone ("the case is not quite right for a genuine Mitsubishi version but that isn't a show stopper" -- GI0GDP) readily available from Far Eastern suppliers. Reasonably expected performance is 100mW out for 5mW in, that said it will vary depending on the details of your assembly, what kind of enclosure you use (and if it has a lid on or not), and how much time you are willing to spend snowflake-tuning the NE325 gate. Do keep in mind that the two FET types have got quite different voltage and biasing requirements.

The project in question was described in the November/December 2023 issue of Scatterpoint, available to UKuG members.

== 5.6GHz ==

[https://wiki.batc.org.uk/5.6_GHz#Receive_Pre-amplifiers The BATC Wiki] shows modifications made by Daniel, DL3IAE for the use of a single SU-02 amp in this band.

== 24GHz ==

TBA

"Franco" SU-02 10GHz PCB

2024-02-07T22:41:41Z

M0JUR: Fix link to GM4BYF instructions

[[File:Franco PCB.png|400px]]

Ex equipment low noise amplifer PCB, featuring PTFE substrate and 4 independent amplifiers (labelled AH, AV, EH and EV) based on the ultra-low-noise pseudomorphic HJ FET NE32584C.

These PCBs have been sourced from RF Microwave shop in Italy (originally run by Franco Rota, I2FHW - hence the nickname of the board) and can be used to make a very good 5.6GHz LNA, 10GHz LNA, 24GHz doubler, or just a source of the NE32584c devices or 50-Ohm microstrip lines.

Original DUBUS article + the NE32584C datasheet: [[:media:Franco PCB.pdf]]

'''WORK IN PROGRESS. If you have time and know more details about any of the projects below, or even a new project, please add them!'''

== Power supply and biasing ==

The NE32584C is rated for the absolute maximum drain-to-source voltage of '''4.0V''' and the absolute maximum gate-to-source voltage of '''-3.0V'''. The nominal drain current is 10mA. As usual with a FET, care should be taken keep the magic smoke in by ensuring Vds is present ''if and only if'' Vgs is.

[[File:SU-02_circuit_schematics.jpeg]]

Chris, G3WIE has reverse-engineered the circuits of the Franco board and produced the schematics show above. All bipolar transistors feature the same markings and have been confirmed as PNP. The element marked with a question mark in the schematics has since been identified as a Zener diode.

=== PSU ===

The power-supply section of the board produces the supply voltages of +3.1V and -3.3V for the amplifier sections as well as controlling which of the amps is powered up. The main workhorse here is the hex Schmitt-trigger inverter 74HC14.

To generate supply voltages, energise the "14V" pad near the large voltage regulator. Do '''not''' apply more than 20V here, it will damage some of the inverter gates. On the other end of the scale, 8V has been shown to suffice.

==== Output switching ====

Switching between AV/EV and AH/EH is achieved by varying the voltage applied to the "14V" pad - anything under approx. 15.5V enables the V pair, anything above that threshold enables the H pair. In the terms of the schematics above, varying the input voltage governs which of the lines labelled A and B is at 0V (which allows the amp connected to it to switch on) and which one is at 5V.

As for switching between AH/AV and EH/EV, which pair receives power is governed by the voltage applied to the pads: "10V" and "RX12". By default, ''i.e.'' with both of these pads unconnected, the PSU directs Vdd/Vbb to the E pair. With 12V applied to the "RX12" pad (or 8.5-9V to the "10V" pad), the A pair is powered up instead. In the terms of the schematics above, this affects which of the two bipolar transistors: Q1 and Q2 is switched on.

=== Active bias ===

Each of the four amplifier sections features a BJT-based active bias circuit with resistor values chosen so that for the Vdd/Vbb value of 3.1V and the Vgg (bias) voltage of -3.3V, both as provided by the on-board PSU, the NE325 ends up at Vds = 2V and Id = 11mA. The same circuit also serves as an on-off switch - in order for the section in question to operate the "A or B" must be at 0V.

G3WIE has created a spreadsheet which can be used to calculate resistor values for given supply voltages and desired FET parameters, or vice versa: [[File:Hemt active biasV3.ods]]

Warning: if you choose to retain the active bias circuit, its is ''critical'' for Vdd and Vbb to be tied together as per the original design. If they are not and you suddenly lose Vbb but not Vdd, it's bye-bye FET.

== 10GHz ==

With the board in question having been designed for 10GHz operation it is perhaps no surprise that the choice of variants is rather wide here.

=== Single-stage LNAs by Gerald, F6CXO ===

Probably the simplest variants (there is one for SMA input and one for direct waveguide input) out there, as bias circuitry and socket fitting aside the only modifications needed is removing a resistor from the RF input line + replacing the resistor on the RF output line with a small DC-blocking capacitor. Claimed performance is 14dB gain, 0.7-0.8dB noise figure. For details see the VHF Communications 3/2004 article on pages 8-10 in the PDF above; the circuit diagram has only been annotated in French but the comments are simple enough to understand for an English speaker).

=== Two-stage LNA by Pete, GM4BYF ===

In the aforementioned article Gerald, F6CXO mentions the possibility of combining two Franco amplifier sections into a two-stage device. Indeed, the file above includes (on pages 4-7) an Italian-language article - unfortunately with no translation, and a very low-resolution circuit diagram - featuring photos of two devices which appear to have done just that. However, it has been observed that merely connecting the output of amp one to the input of amp two can cause stability problems.{{Citation needed}}

In order to avoid the stability issues Pete, GM4BYF has constructed a two-stage version in which the two amplifier sections are laid out back to back, and connected with a short length of semi-rigid coax fed through a specially drilled hole. Details can be found here: [[File:Franco 10GHz LNA GM4BYF.pdf]].

TODO: add performance data, if/when available

=== Dual two-stage LNA by Geoff, GI0GDP ===

This variant uses ''two'' two-stage amplifiers running in parallel to produce around 75mW output from 2.5mW input. In order to avoid excessive amounts of cutting and reassembling the PCB, the two input-stage FETs are turned upside down.

[[File:Franco 10GHz four-amp LNA GI0GDP.jpg|800px]]

Modification details:
* cut two sections out of a complete Franco PCB: one with both E-pair amps (which will serve as the input stage), one with the A pair (for the output stage)
* join the sections as show in the photo, making sure the back plane is nicely joined (possibly with copper foil + some solder) as well. Make sure the RF lines are ''not'' connected, though
* connect the two sections' RF lines using ATC capacitors. Geoff has used 2pF ones here but in his own words, the exact capacitance doesn't really matter
* desolder and turn around the two E-pair FETs
* relocate the E-pair source and drain resistors (yes, all four of them!) to reflect the new orientation of the FETs. When in doubt what goes where, consult the photo
* replace the two 68R drain resistors in the ''A-pair'' (''i.e.'' output) amps with jumpers. Note that the photo shows 1R0 rather than 0R chips owing to the fact the author did not have any of the latter at hand at the time
* remove and jumper the four resistors from the RF lines
* make your own biasing arrangements as desired, bearing in mind that with E-pair FETs having been turned around the original active bias circuit will ''not'' work without extensive modifications

Nb. Disregard the two through-hole ceramic capacitors, they are only there to provide mechanical support and extra soldering surface.

=== Two-stage "NE325+MGF1601" LNA by Geoff, GI0GDP ===

In Geoff's latest take on the subject, a single two-stage assembly has its output-stage NE325 replaced with an MGF1601 clone ("the case is not quite right for a genuine Mitsubishi version but that isn't a show stopper" -- GI0GDP) readily available from Far Eastern suppliers. Reasonably expected performance is 100mW out for 5mW in, that said it will vary depending on the details of your assembly, what kind of enclosure you use (and if it has a lid on or not), and how much time you are willing to spend snowflake-tuning the NE325 gate. Do keep in mind that the two FET types have got quite different voltage and biasing requirements.

The project in question was described in the November/December 2023 issue of Scatterpoint, available to UKuG members.

== 5.6GHz ==

[https://wiki.batc.org.uk/5.6_GHz#Receive_Pre-amplifiers The BATC Wiki] shows modifications made by Daniel, DL3IAE for the use of a single SU-02 amp in this band.

== 24GHz ==

TBA

"Franco" SU-02 10GHz PCB

2024-02-07T14:41:56Z

M0JUR: Add photo of the GI0GDP dual two-stage 10GHz LNA

[[File:Franco PCB.png|400px]]

Ex equipment low noise amplifer PCB, featuring PTFE substrate and 4 independent amplifiers (labelled AH, AV, EH and EV) based on the ultra-low-noise pseudomorphic HJ FET NE32584C.

These PCBs have been sourced from RF Microwave shop in Italy (originally run by Franco Rota, I2FHW - hence the nickname of the board) and can be used to make a very good 5.6GHz LNA, 10GHz LNA, 24GHz doubler, or just a source of the NE32584c devices or 50-Ohm microstrip lines.

Original DUBUS article + the NE32584C datasheet: [[:media:Franco PCB.pdf]]

'''WORK IN PROGRESS. If you have time and know more details about any of the projects below, or even a new project, please add them!'''

== Power supply and biasing ==

The NE32584C is rated for the absolute maximum drain-to-source voltage of '''4.0V''' and the absolute maximum gate-to-source voltage of '''-3.0V'''. The nominal drain current is 10mA. As usual with a FET, care should be taken keep the magic smoke in by ensuring Vds is present ''if and only if'' Vgs is.

[[File:SU-02_circuit_schematics.jpeg]]

Chris, G3WIE has reverse-engineered the circuits of the Franco board and produced the schematics show above. All bipolar transistors feature the same markings and have been confirmed as PNP. The element marked with a question mark in the schematics has since been identified as a Zener diode.

=== PSU ===

The power-supply section of the board produces the supply voltages of +3.1V and -3.3V for the amplifier sections as well as controlling which of the amps is powered up. The main workhorse here is the hex Schmitt-trigger inverter 74HC14.

To generate supply voltages, energise the "14V" pad near the large voltage regulator. Do '''not''' apply more than 20V here, it will damage some of the inverter gates. On the other end of the scale, 8V has been shown to suffice.

==== Output switching ====

Switching between AV/EV and AH/EH is achieved by varying the voltage applied to the "14V" pad - anything under approx. 15.5V enables the V pair, anything above that threshold enables the H pair. In the terms of the schematics above, varying the input voltage governs which of the lines labelled A and B is at 0V (which allows the amp connected to it to switch on) and which one is at 5V.

As for switching between AH/AV and EH/EV, which pair receives power is governed by the voltage applied to the pads: "10V" and "RX12". By default, ''i.e.'' with both of these pads unconnected, the PSU directs Vdd/Vbb to the E pair. With 12V applied to the "RX12" pad (or 8.5-9V to the "10V" pad), the A pair is powered up instead. In the terms of the schematics above, this affects which of the two bipolar transistors: Q1 and Q2 is switched on.

=== Active bias ===

Each of the four amplifier sections features a BJT-based active bias circuit with resistor values chosen so that for the Vdd/Vbb value of 3.1V and the Vgg (bias) voltage of -3.3V, both as provided by the on-board PSU, the NE325 ends up at Vds = 2V and Id = 11mA. The same circuit also serves as an on-off switch - in order for the section in question to operate the "A or B" must be at 0V.

G3WIE has created a spreadsheet which can be used to calculate resistor values for given supply voltages and desired FET parameters, or vice versa: [[File:Hemt active biasV3.ods]]

Warning: if you choose to retain the active bias circuit, its is ''critical'' for Vdd and Vbb to be tied together as per the original design. If they are not and you suddenly lose Vbb but not Vdd, it's bye-bye FET.

== 10GHz ==

With the board in question having been designed for 10GHz operation it is perhaps no surprise that the choice of variants is rather wide here.

=== Single-stage LNAs by Gerald, F6CXO ===

Probably the simplest variants (there is one for SMA input and one for direct waveguide input) out there, as bias circuitry and socket fitting aside the only modifications needed is removing a resistor from the RF input line + replacing the resistor on the RF output line with a small DC-blocking capacitor. Claimed performance is 14dB gain, 0.7-0.8dB noise figure. For details see the VHF Communications 3/2004 article on pages 8-10 in the PDF above; the circuit diagram has only been annotated in French but the comments are simple enough to understand for an English speaker).

=== Two-stage LNA by Pete, GM4BYF ===

In the aforementioned article Gerald, F6CXO mentions the possibility of combining two Franco amplifier sections into a two-stage device. Indeed, the file above includes (on pages 4-7) an Italian-language article - unfortunately with no translation, and a very low-resolution circuit diagram - featuring photos of two devices which appear to have done just that. However, it has been observed that merely connecting the output of amp one to the input of amp two can cause stability problems.{{Citation needed}}

In order to avoid the stability issues Pete, GM4BYF has constructed a two-stage version in which the two amplifier sections are laid out back to back, and connected with a short length of semi-rigid coax fed through a specially drilled hole. Details can be found here: {{File:Franco 10GHz LNA GM4BYF.pdf}}.

TODO: add performance data, if/when available

=== Dual two-stage LNA by Geoff, GI0GDP ===

This variant uses ''two'' two-stage amplifiers running in parallel to produce around 75mW output from 2.5mW input. In order to avoid excessive amounts of cutting and reassembling the PCB, the two input-stage FETs are turned upside down.

[[File:Franco 10GHz four-amp LNA GI0GDP.jpg|800px]]

Modification details:
* cut two sections out of a complete Franco PCB: one with both E-pair amps (which will serve as the input stage), one with the A pair (for the output stage)
* join the sections as show in the photo, making sure the back plane is nicely joined (possibly with copper foil + some solder) as well. Make sure the RF lines are ''not'' connected, though
* connect the two sections' RF lines using ATC capacitors. Geoff has used 2pF ones here but in his own words, the exact capacitance doesn't really matter
* desolder and turn around the two E-pair FETs
* relocate the E-pair source and drain resistors (yes, all four of them!) to reflect the new orientation of the FETs. When in doubt what goes where, consult the photo
* replace the two 68R drain resistors in the ''A-pair'' (''i.e.'' output) amps with jumpers. Note that the photo shows 1R0 rather than 0R chips owing to the fact the author did not have any of the latter at hand at the time
* remove and jumper the four resistors from the RF lines
* make your own biasing arrangements as desired, bearing in mind that with E-pair FETs having been turned around the original active bias circuit will ''not'' work without extensive modifications

Nb. Disregard the two through-hole ceramic capacitors, they are only there to provide mechanical support and extra soldering surface.

=== Two-stage "NE325+MGF1601" LNA by Geoff, GI0GDP ===

In Geoff's latest take on the subject, a single two-stage assembly has its output-stage NE325 replaced with an MGF1601 clone ("the case is not quite right for a genuine Mitsubishi version but that isn't a show stopper" -- GI0GDP) readily available from Far Eastern suppliers. Reasonably expected performance is 100mW out for 5mW in, that said it will vary depending on the details of your assembly, what kind of enclosure you use (and if it has a lid on or not), and how much time you are willing to spend snowflake-tuning the NE325 gate. Do keep in mind that the two FET types have got quite different voltage and biasing requirements.

The project in question was described in the November/December 2023 issue of Scatterpoint, available to UKuG members.

== 5.6GHz ==

[https://wiki.batc.org.uk/5.6_GHz#Receive_Pre-amplifiers The BATC Wiki] shows modifications made by Daniel, DL3IAE for the use of a single SU-02 amp in this band.

== 24GHz ==

TBA

"Franco" SU-02 10GHz PCB

2024-02-07T14:39:54Z

M0JUR: Almost complete (still need to add photo of Geoff's four-amp variant) the list of 10GHz LNAs

[[File:Franco PCB.png|400px]]

Ex equipment low noise amplifer PCB, featuring PTFE substrate and 4 independent amplifiers (labelled AH, AV, EH and EV) based on the ultra-low-noise pseudomorphic HJ FET NE32584C.

These PCBs have been sourced from RF Microwave shop in Italy (originally run by Franco Rota, I2FHW - hence the nickname of the board) and can be used to make a very good 5.6GHz LNA, 10GHz LNA, 24GHz doubler, or just a source of the NE32584c devices or 50-Ohm microstrip lines.

Original DUBUS article + the NE32584C datasheet: [[:media:Franco PCB.pdf]]

'''WORK IN PROGRESS. If you have time and know more details about any of the projects below, or even a new project, please add them!'''

== Power supply and biasing ==

The NE32584C is rated for the absolute maximum drain-to-source voltage of '''4.0V''' and the absolute maximum gate-to-source voltage of '''-3.0V'''. The nominal drain current is 10mA. As usual with a FET, care should be taken keep the magic smoke in by ensuring Vds is present ''if and only if'' Vgs is.

[[File:SU-02_circuit_schematics.jpeg]]

Chris, G3WIE has reverse-engineered the circuits of the Franco board and produced the schematics show above. All bipolar transistors feature the same markings and have been confirmed as PNP. The element marked with a question mark in the schematics has since been identified as a Zener diode.

=== PSU ===

The power-supply section of the board produces the supply voltages of +3.1V and -3.3V for the amplifier sections as well as controlling which of the amps is powered up. The main workhorse here is the hex Schmitt-trigger inverter 74HC14.

To generate supply voltages, energise the "14V" pad near the large voltage regulator. Do '''not''' apply more than 20V here, it will damage some of the inverter gates. On the other end of the scale, 8V has been shown to suffice.

==== Output switching ====

Switching between AV/EV and AH/EH is achieved by varying the voltage applied to the "14V" pad - anything under approx. 15.5V enables the V pair, anything above that threshold enables the H pair. In the terms of the schematics above, varying the input voltage governs which of the lines labelled A and B is at 0V (which allows the amp connected to it to switch on) and which one is at 5V.

As for switching between AH/AV and EH/EV, which pair receives power is governed by the voltage applied to the pads: "10V" and "RX12". By default, ''i.e.'' with both of these pads unconnected, the PSU directs Vdd/Vbb to the E pair. With 12V applied to the "RX12" pad (or 8.5-9V to the "10V" pad), the A pair is powered up instead. In the terms of the schematics above, this affects which of the two bipolar transistors: Q1 and Q2 is switched on.

=== Active bias ===

Each of the four amplifier sections features a BJT-based active bias circuit with resistor values chosen so that for the Vdd/Vbb value of 3.1V and the Vgg (bias) voltage of -3.3V, both as provided by the on-board PSU, the NE325 ends up at Vds = 2V and Id = 11mA. The same circuit also serves as an on-off switch - in order for the section in question to operate the "A or B" must be at 0V.

G3WIE has created a spreadsheet which can be used to calculate resistor values for given supply voltages and desired FET parameters, or vice versa: [[File:Hemt active biasV3.ods]]

Warning: if you choose to retain the active bias circuit, its is ''critical'' for Vdd and Vbb to be tied together as per the original design. If they are not and you suddenly lose Vbb but not Vdd, it's bye-bye FET.

== 10GHz ==

With the board in question having been designed for 10GHz operation it is perhaps no surprise that the choice of variants is rather wide here.

=== Single-stage LNAs by Gerald, F6CXO ===

Probably the simplest variants (there is one for SMA input and one for direct waveguide input) out there, as bias circuitry and socket fitting aside the only modifications needed is removing a resistor from the RF input line + replacing the resistor on the RF output line with a small DC-blocking capacitor. Claimed performance is 14dB gain, 0.7-0.8dB noise figure. For details see the VHF Communications 3/2004 article on pages 8-10 in the PDF above; the circuit diagram has only been annotated in French but the comments are simple enough to understand for an English speaker).

=== Two-stage LNA by Pete, GM4BYF ===

In the aforementioned article Gerald, F6CXO mentions the possibility of combining two Franco amplifier sections into a two-stage device. Indeed, the file above includes (on pages 4-7) an Italian-language article - unfortunately with no translation, and a very low-resolution circuit diagram - featuring photos of two devices which appear to have done just that. However, it has been observed that merely connecting the output of amp one to the input of amp two can cause stability problems.{{Citation needed}}

In order to avoid the stability issues Pete, GM4BYF has constructed a two-stage version in which the two amplifier sections are laid out back to back, and connected with a short length of semi-rigid coax fed through a specially drilled hole. Details can be found here: {{File:Franco 10GHz LNA GM4BYF.pdf}}.

TODO: add performance data, if/when available

=== Dual two-stage LNA by Geoff, GI0GDP ===

This variant uses ''two'' two-stage amplifiers running in parallel to produce around 75mW output from 2.5mW input. In order to avoid excessive amounts of cutting and reassembling the PCB, the two input-stage FETs are turned upside down.

Modification details:
* cut two sections out of a complete Franco PCB: one with both E-pair amps (which will serve as the input stage), one with the A pair (for the output stage)
* join the sections as show in the photo, making sure the back plane is nicely joined (possibly with copper foil + some solder) as well. Make sure the RF lines are ''not'' connected, though
* connect the two sections' RF lines using ATC capacitors. Geoff has used 2pF ones here but in his own words, the exact capacitance doesn't really matter
* desolder and turn around the two E-pair FETs
* relocate the E-pair source and drain resistors (yes, all four of them!) to reflect the new orientation of the FETs. When in doubt what goes where, consult the photo
* replace the two 68R drain resistors in the ''A-pair'' (''i.e.'' output) amps with jumpers. Note that the photo shows 1R0 rather than 0R chips owing to the fact the author did not have any of the latter at hand at the time
* remove and jumper the four resistors from the RF lines
* make your own biasing arrangements as desired, bearing in mind that with E-pair FETs having been turned around the original active bias circuit will ''not'' work without extensive modifications

Nb. Disregard the two through-hole ceramic capacitors, they are only there to provide mechanical support and extra soldering surface.

=== Two-stage "NE325+MGF1601" LNA by Geoff, GI0GDP ===

In Geoff's latest take on the subject, a single two-stage assembly has its output-stage NE325 replaced with an MGF1601 clone ("the case is not quite right for a genuine Mitsubishi version but that isn't a show stopper" -- GI0GDP) readily available from Far Eastern suppliers. Reasonably expected performance is 100mW out for 5mW in, that said it will vary depending on the details of your assembly, what kind of enclosure you use (and if it has a lid on or not), and how much time you are willing to spend snowflake-tuning the NE325 gate. Do keep in mind that the two FET types have got quite different voltage and biasing requirements.

The project in question was described in the November/December 2023 issue of Scatterpoint, available to UKuG members.

== 5.6GHz ==

[https://wiki.batc.org.uk/5.6_GHz#Receive_Pre-amplifiers The BATC Wiki] shows modifications made by Daniel, DL3IAE for the use of a single SU-02 amp in this band.

== 24GHz ==

TBA

File:Franco 10GHz four-amp LNA GI0GDP.jpg

2024-02-07T14:16:00Z

M0JUR: Photo of the four-amp 10GHz LNA based on "Franco" SU-02 PCBs, by Geoff GI0GDP

Photo of the four-amp 10GHz LNA based on "Franco" SU-02 PCBs, by Geoff GI0GDP

File:Franco 10GHz LNA GM4BYF.pdf

2024-02-07T13:47:54Z

M0JUR: Assembly instructions for a two-stage 10GHz pre-amp with using two sections of a "Franco" SU-02 PCB joined back to back, by Pete GM4BYF.

Assembly instructions for a two-stage 10GHz pre-amp with using two sections of a "Franco" SU-02 PCB joined back to back, by Pete GM4BYF.

"Franco" SU-02 10GHz PCB

2024-02-07T13:17:10Z

M0JUR: The "DUBUS article" file includes the FET datasheet

"Franco" SU-02 10GHz PCB

2024-02-07T13:12:44Z

M0JUR: Write up PSU reverse-engineering

"Franco" SU-02 10GHz PCB

2024-02-06T17:01:38Z

M0JUR: Link the G3WIE spreadsheet

[[File:Franco PCB.png|400px]]

Ex equipment low noise amplifer PCB, featuring PTFE substrate and 4 independent amplifiers based on the ultra-low-noise pseudomorphic HJ FET NE32584C.

These PCBs have been sourced from RF Microwave shop in Italy (originally run by Franco Rota, I2FHW - hence the nickname of the board) and can be used to make a very good 5.6GHz LNA, 10GHz LNA, 24GHz doubler, or just a source of the NE32584c devices or 50-Ohm microstrip lines.

Original DUBUS article [[:media:Franco PCB.pdf]]

'''WORK IN PROGRESS. If you have time and know more details about any of the projects below, or even a new project, please add them!'''

== Power supply and biasing ==

The NE32584C is rated for the absolute maximum drain-to-source voltage of '''4.0V''' and the absolute maximum gate-to-source voltage of '''-3.0V'''. The nominal drain current is 10mA. As usual with a FET, care should be taken keep the magic smoke in by ensuring Vds is present ''if and only if'' Vgs is.

[[File:SU-02_circuit_schematics.jpeg]]

Chris, G3WIE has reverse-engineered the circuits of the Franco board and produced the schematics show above. All bipolar transistors feature the same markings and have been confirmed as PNP. The currently unidentified element, marked with a question mark in the schematics, is suspected to be a Zener diode.

=== PSU ===

TBA

=== Active bias ===

Each of the four amplifier sections features a BJT-based active bias circuit with resistor values chosen so that for the Vdd/Vbb value of 3.1V and the Vgg (bias) voltage of -3.3V, both as provided by the on-board PSU, the NE325 ends up at Vds = 2V and Id = 11mA. The same circuit also serves as an on-off switch - in order for the section in question to operate the "A or B" point must be connected to ground.

G3WIE has created a spreadsheet which can be used to calculate resistor values for given supply voltages and desired FET parameters, or vice versa: [[File:Hemt active biasV3.ods]]

Warning: if you choose to retain the active bias circuit, its is ''critical'' for Vdd and Vbb to be tied together as per the original design. If they are not and you suddenly lose Vbb but not Vdd, it's bye-bye FET.

== 10GHz ==

TBA

* single-stage variants from the DUBUS article
* two-stage variant from the DUBUS article (unstable!)
* two-stage upside-down variant by Pete, GM4BYF - UKmicrowaves file
* dual two-stage variant by Geoff, GI0GDP - TBA
* two-stage NE325+MGF1601 variant by Geoff, GI0GDP - Scatterpoint

== 5.6GHz ==

[https://wiki.batc.org.uk/5.6_GHz#Receive_Pre-amplifiers The BATC Wiki] shows modifications made by Daniel, DL3IAE for the use of a single SU-02 amp in this band.

== 24GHz ==

TBA

File:Hemt active biasV3.ods

2024-02-06T17:00:02Z

M0JUR: Spreadsheet created by Chris, G3WIE. It contains two calculators meant to facilitate modifications of active bias circuits for HEM FETs like the NE32584C: the first will help you choose resistors in an active bias circuit given supply voltages and a de...

Spreadsheet created by Chris, G3WIE. It contains two calculators meant to facilitate modifications of active bias circuits for HEM FETs like the NE32584C: the first will help you choose resistors in an active bias circuit given supply voltages and a desired Vds, the second will calculate Vds and Id for given supply voltages and resistor values.

The initial parameters of the calculators have for the NE32584C-based amplifiers on the "Franco" SU-02 board.

"Franco" SU-02 10GHz PCB

2024-02-06T16:54:13Z

M0JUR: G3WIE notes on the active-bias circuit of the amp sections

[[File:Franco PCB.png|400px]]

Ex equipment low noise amplifer PCB, featuring PTFE substrate and 4 independent amplifiers based on the ultra-low-noise pseudomorphic HJ FET NE32584C.

These PCBs have been sourced from RF Microwave shop in Italy (originally run by Franco Rota, I2FHW - hence the nickname of the board) and can be used to make a very good 5.6GHz LNA, 10GHz LNA, 24GHz doubler, or just a source of the NE32584c devices or 50-Ohm microstrip lines.

Original DUBUS article [[:media:Franco PCB.pdf]]

'''WORK IN PROGRESS. If you have time and know more details about any of the projects below, or even a new project, please add them!'''

== Power supply and biasing ==

The NE32584C is rated for the absolute maximum drain-to-source voltage of '''4.0V''' and the absolute maximum gate-to-source voltage of '''-3.0V'''. The nominal drain current is 10mA. As usual with a FET, care should be taken keep the magic smoke in by ensuring Vds is present ''if and only if'' Vgs is.

[[File:SU-02_circuit_schematics.jpeg]]

Chris, G3WIE has reverse-engineered the circuits of the Franco board and produced the schematics show above. All bipolar transistors feature the same markings and have been confirmed as PNP. The currently unidentified element, marked with a question mark in the schematics, is suspected to be a Zener diode.

=== PSU ===

TBA

=== Active bias ===

Each of the four amplifier sections features a BJT-based active bias circuit with resistor values chosen so that for the Vdd/Vbb value of 3.1V and the Vgg (bias) voltage of -3.3V, both as provided by the on-board PSU, the NE325 ends up at Vds = 2V and Id = 11mA. The same circuit also serves as an on-off switch - in order for the section in question to operate the "A or B" point must be connected to ground.

G3WIE has created a spreadsheet which can be used to calculate resistor values for given supply voltages and desired FET parameters,

Warning: if you choose to retain the active bias circuit, its is ''critical'' for Vdd and Vbb to be tied together as per the original design. If they are not and you suddenly lose Vbb but not Vdd, it's bye-bye FET.

== 10GHz ==

TBA

* single-stage variants from the DUBUS article
* two-stage variant from the DUBUS article (unstable!)
* two-stage upside-down variant by Pete, GM4BYF - UKmicrowaves file
* dual two-stage variant by Geoff, GI0GDP - TBA
* two-stage NE325+MGF1601 variant by Geoff, GI0GDP - Scatterpoint

== 5.6GHz ==

[https://wiki.batc.org.uk/5.6_GHz#Receive_Pre-amplifiers The BATC Wiki] shows modifications made by Daniel, DL3IAE for the use of a single SU-02 amp in this band.

== 24GHz ==

TBA

"Franco" SU-02 10GHz PCB

2024-02-06T15:56:20Z

M0JUR: First batch of notes on power. Add link to the 5.6GHz mod on BATC Wiki

File:SU-02 circuit schematics.jpeg

2024-02-06T15:10:11Z

M0JUR: Reverse-engineered circuit schematics of an SU-02 board, produced "in Biro-CAD" by Chris, G3WIE.

Reverse-engineered circuit schematics of an SU-02 board, produced "in Biro-CAD" by Chris, G3WIE.

Main Page

2024-02-06T14:38:53Z

M0JUR: Update the link to the SU-02 page

[[file:ukuglogo.jpg|left|80px|middle]]

<big><big><big>'''Welcome to the''' '''''[http://www.microwavers.org UK Microwave Group]''''' '''Wiki'''</big></big></big>

The place to find information on equipment and projects for Amateur Microwave operation on bands above 1 GHz.
 

 

 

This wiki is designed to be a reference library, developed by the UK Microwave Group for use by the worldwide microwave community. Any information which you feel is of use to others can be put on the Wiki, but please observe any copyright restrictions on your material you use. However, before you can add or edit content you will need to [[Registering|register for an account]].

== Contents ==
* [[Registering]] on the UK Microwave Wiki
=== New to Microwave operation? ===
* Get the basics in our collection of [[Getting Started|Getting Started guides]]
* Glossary and index of terms and TLAs used in Amateur Microwave operation [[Glossary of terms|Glossary of terms]]

=== The UK Microwave Group Information and services===
* [[Joining UK Microwave Group]]
* [[Beaconspot]]
* [[Contest Calendar]]
* [[UK Microwave Group on Twitter]]
* [[UK Microwave Group Youtube Channel]]
* [[Chip bank|The UK Microwave Chip Bank]]

=== Scatterpoint Magazine ===
* Full details inc archive at: [http://scatterpoint.org/ scatterpoint.org]

=== Microwave Roundtables ===
* [[Presentations]] Presentations from Roundtables.
* Recordings from Roundtables can be found on the UK Microwave Group Youtube channel.

===Microwave SDR projects===
* [[Hayling project]] The UK Microwave Group SDR transceiver
* [[Langstone Project]] The prototype microwave SDR project
* [[Langstone V2]] Latest version with additional support for the Lime SDR and Lime RFE

=== Getting on the Microwave Bands ===
The Low Bands:
* [[1.3 GHz]]
* [[2.3 GHz]]
* [[3.4 GHz]]

The middle bands:
* [[5.7 GHz]]
* [[10 GHz]]

The High bands
* [[24 GHz]]
* [[47 GHz]]
* [[76 GHz]]

The Millimetre Bands
* [[122 GHz]]
* [[134 GHz]]
* [[241 GHz]]

Terahertz
* [[275+ GHz]]

=== Operating ===
* [[Awards and Trophies]]
* [[Digital modes using WSJT]]
* [[Digital modes using Opera]]
* [[Talkback for microwave operation including ON4KST]]
* [https://www.google.com/maps/d/u/0/edit?mid=1VeLnRRbLwloDTL2i9-HE2sxqL0c&ll=52.95536632179757%2C-0.49709419999999227&z=7 Google map] showing possible portable operating sites
* [[Mapping tools]] - including finding the IARU LOCATOR of a site, UK postcode to Lat and Long conversion and a topographic overlay for Google maps
* [[Propagation tools]]
* [[EMF]] - guidance on EM Field exposure

===Microwave EME===

*[[Microwave_EME]]
* Flight Refuelling ARS 10 GHz EME Project History - G4RFR: [[:File:FRARS EME Project.pdf]]

=== Measurement techniques ===
* [[Measuring sun noise]]
* [[Realtime signal power plot]] Software from G4JNT

=== Construction projects ===
* [[PE4302]] - PE4302 variable attenuator project
* [[ADF435x PIC]] - PIC controllers for Chinese ADF4350/4351 boards
* [[F6BVA 3cm transverter]] - F6BVA 3cm to UHF transverter project

=== G4BAO's Bodger's Guides ===

* Modifying AFL 900MHz Hybrids for 23cms: [[:File:900MHz Hybrids.doc]]
* Modifying Mini Circuits ZAPD1 splitters for wideband: [[:File:Wideband splitter.doc]]
* 13cm PA using a G4BAO 23cm board and an MRF19085: [[:File:MRF19085.doc]]
* Bodging 1900MHz QRO amps for 13cm: [[:File:1900AMPS.doc]]
* Using a Lucent ILam QRO SSPA on 13cm: [[:File:ILAM_Mods1.doc]]
* Using a ceramic MRF9045 in the G4BAO PA PCB: [[:File:ceramic.doc]]
* A control board for a GaAsFET PA: [[:File:GaAs PA control.doc]]
* A 70cm converter: [[:File:70cm converter.doc]]
* Dishal's method for tuning up filters: [[:File:Dishal.doc]]
* Simple PIC controller to use as a beacon with the G4JNT synthesiser board: [[:File:Beacon_PIC.doc]]
* A small, high current Stepdown PSU after a design by G3WDG: [[:File:stepdown_PSU.pdf]]

=== Useful Circuits and notes ===
* [[ADF series of synthesizers]]
* [[Filters]] - Collection of links and designs for that most critical but over looked component!
* [[Pre-amps]] - Notes, suppliers and circuits of preamplifiers
* [[Power amplifiers]] - Notes, suppliers and circuits of power amplifiers
* [[LeoBodnar GPS Settings]]
* [[Waveguide]] - An introduction to Waveguides
* [[G3WDG Microwave Designs]] - Construction Notes for the G3WDG Series of Kits
* [[Microwave signal source]] - How to generate low power test signals on all bands up to 122GHz
* [[Waveguide Slot Array calculator]] - Updated design sheet with standard w/g sizes and single-side array support
* [[Hybrid Networks|Hybrid Networks and their Uses]] - A description what hybrids do and how they can be used
* [["Franco" SU-02 10GHz PCB]] The very useful 10GHz amplifier PCB from RF Microwave in Italy

=== Miscellaneous Equipment manuals and schematics ===
* [[Mutek]] Circuits and documentation for the Mutek range of equipment
* [[Microwave modules]] Circuits and documentation for the Microwave modules range of equipment
* [[Cellflex]] Data sheets for Cellflex cables
* Spec and technical drawings of standard, anti-cocking and precision [https://flann.com/wp-content/uploads/2015/09/Waveguide-and-Flange-Information.pdf Flann flanges]
* [[Kuhne]] Electronic Microwave Components

=== Test equipment manuals ===
For HP and Agilent equipment see the UKMicrowaves Group files at
[https://groups.io/g/UKMicrowaves/files/Test%20Equipment%20-%20Manuals]
and [https://groups.io/g/HP-Agilent-Keysight-equipment groups.io] which has replaced the old
[https://groups.yahoo.com/neo/groups/hp_agilent_equipment/info Yahoo group]

UKuG thanks BATC for hosting this facility

"Franco" SU-02 10GHz PCB

2024-02-06T14:37:11Z

M0JUR: M0JUR moved page Franco 10GHz PCB to "Franco" SU-02 10GHz PCB: RF Microwave is no longer run by Franco, and it probably helps to mention the code name too

Franco 10GHz PCB

2024-02-06T14:37:11Z

M0JUR: M0JUR moved page Franco 10GHz PCB to "Franco" SU-02 10GHz PCB: RF Microwave is no longer run by Franco, and it probably helps to mention the code name too

#REDIRECT [["Franco" SU-02 10GHz PCB]]

"Franco" SU-02 10GHz PCB

2024-02-06T14:34:57Z

M0JUR: Slightly reorganise the preamble, add sections for different project types