Push-pull IRF740's for an intermediate QRP amp

[ka1tdq]

I'm starting a new linear amplifier project to be an intermediate amp between a pw AM transmitter and an also up-and-coming single 3-500 linear amp.  I've attached a picture of my parts and pieces. 

The mosfets are IRF740's and the power transformer was scrapped from an old organ.  Output from the transformer is 36 vac, but that will come down with a full wave bridge and 25mH choke. 

Yesterday I welded together and painted a massive heatsink behind the chassis.  I will mount the mosfets directly to it through a hold in the die cast box. 

I'm looking to get the carrier level to around 15 watts with headroom for modulation peaks.  This would be a nice level to drive a larger linear amplifier.

Jon
KA1TDQ

[ka1tdq]

Attached is a copy of the schematic.  I've borrowed parts and pieces from other similar designs floating out there.

It has an RF sensing circuit to switch DC power to the mosfets and to bypass the linear in receive mode.

Jon
KA1TDQ

[W3GMS]

Jon,
Looks basically ok.  You may want to have a resistor  across the secondary of the input transformer.  With the very high impedance of the FET's its sometime nice and produces better stability to have the secondary of the transformer terminated.  Then the reflected impedance would be very predictable back to the primary based on the turns ratio of the input transformer.  Based on your turns ratio and the secondary termination resistance you could adjust things so the primary sees 50 ohms.  Also especially in linear service you may find you need some negative feedback around each RF device.   

Just some tweaks as you go through the build.

Have fun....
Joe, W3GMS     

[AB2EZ]

Each IRF740 FET has 1400pf of gate to source capacitance...  so the input impedance is not very high (or very low) on 75m (-j29 ohms). A more elaborate impedance matching network may be required if you want the input impedance to be close to 50 ohms (resistive), and if you don't have a lot of drive power to waste as heating power in swamping resistors or on reflected input power.

You may need a center tap on the secondary of the input transformer to provide an RF path directly to ground.

The drain current v. gate to source voltage characteristic of these FETs is not very linear (see the specification for drain current v. gate to source voltage)... so, as Joe points out, you may need about 0.5 ohms or 1 ohm of non-inductive source to ground resistance (negative feedback) on each FET to obtain linear operation.

You might want to consider the "hf packer" design... which uses inexpensive IRF510's (much lower gate to source capacitance), and might provide enough power to drive your high power linear amplifier. I used one to interface a 250 mW (carrier) AM driver to a 3-500Z linear amplifier putting out 200W of carrier with plenty of headroom for 100 percent modulation peaks.

[Tom WA3KLR]

I'm concerned about the basic filtered power supply.  If you are going to proceed with that supply rather than a nice clean regulated supply, you should have a relatively heavy value bleeder resistor to load the supply, and much larger filter capacitance (easy to achieve today).  I hope that the choke resistance is low relative to the application current and voltage.

I have not reviewed the r.f. aspects of the amp.

GL OM.

[ka1tdq]

The original designs that I was browsing called for IRF510's, so it seems like that's the way to go.  I have some coming in a few days anyway.

Someone also suggested to me that the toroid on the output may be a little small and that a BN-43-7051 would be a little more hefty.  I'll make that change and also put a small resistance in series with each gate.  10 ohms each seems like a value that wouldn't hurt. I guess this aids in stability and linearity. 

Along those same lines, paralleling the secondary of the input transformer with a 1k resistor seems alright.

I checked the resistance of the choke and it is about 1 ohm (significant at up to 5 amps current draw). If I regulate the voltage down to 28 volts using a zener/680 ohm resistor/NPN transistor, this would allow for the DC voltage swing down during heavy current draw.  A steady DC voltage would probably be best anyway.

Jon
KA1TDQ

[ka1tdq]

Ah, and Stu, point taken on the input drive/impedance.  Attached is screen shot of the IRF510 linear which is on the web.  It uses a better matching network for the input, and I don't need to waste power that I don't have.  I'm probably going to actually need the attenuator pad which the article calls for when using 5 watt peak transmitters.

As mentioned in a previous thread, when my transmitter's collector voltage is set to 15 vdc, I'm getting about 5 watts out considering 150% positive peaks.

I'll stick with this input design, a regulated 28 volt supply and use a larger ferrite on the output.

I'll need a few more small parts/pieces to finish this amp.  It's probably going to need to wait a couple weeks though.  Right now I'm attempting to drill holes in the chassis and heat sink in the garage without waking my wife upstairs.  It sounds like a small deal, but these things can get serious.

[AB2EZ]

Jon

I suggest that you take a look at the construction manual for the HF Packer: http://site.hfprojectsyahoo.com/yahoo_site_admin/assets/docs/ManHFPAV4.20454748.pdf

When you finish building it, and you are ready to set it up for operation... I suggest that you carefully follow the bias adjustment instructions. In particular, start with the pots set for zero gate-to-source bias (zero drain current in both transistors). Adjust one of the pots to obtain an increase in total current drawn from your 28V power supply of 4 mA*. Then adjust the other pot to increase the total current drawn from your 28V power supply by another 4 ma*.

*Note that the HF Packer includes an internal B+ supply that produces around 30 volts that is supplied to the drains of the IRF510's. But the adjustment procedure in the manual measures the current drawn from the external 12.6V supply that feeds the HF Packer's internal 30V supply. Therefore 10mA of current drawn from the external 12.6 volt supply of the HF Packer (as called for in the manual) corresponds to around 4mA drawn from the HF Packer's internal 30V B+ supply

My experience is that if you don't adjust the two IRF510 drain currents to be equal, and approximately 4mA each... then you may end up with a parasitic oscillation that destroys one or both transistors. With proper adjustment of the bias, my HF Packer (I sold it recently because it was surplus to my current needs) was very stable and reliable.

Stu

[ka1tdq]

That's quite the manual, thanks! I browsed it very quickly and it has some other helpful tid bits too. I definitely did note your power supply to current conversion though. I'm not going to try re-inventing the wheel. This design seems to work well. My goal is just linearity at 15 watts. A few more small parts and I'm there.

Jon

[ka1tdq]

I finished the outside chassis today and mounting the heat sink.  Once all the parts come in I can do the wiring.  I've also attached a picture of how it will look in the shack.  No tubes!

Jon
KA1TDQ