distortion in 40m linear amplifier

Thread Starter

Joao Goncalves 1

Joined Feb 17, 2019
7
hello. i am joao, and i am a radio amateur. i have built several small projects before, from antennas to small am/ssb transmitters, receivers, and small amps.
i am trying to correctly build a push pull mosfet amp. i am using a yeasu ft757 as an exciter for both am or ssb. i would like to build this amp and then a bigger one so i could use it on 80m am nets or maybe sometimes on 40m as a ssb amp, for a few more db's. i have built an amp like in the schematic, but i tried to test it with only only mosfet, out of curiosity. with the drain connected to a 100n capacitor and to a dummy load, i can get around 40, 50w max, with different mosfets. at around 2.5v bias voltage and 38v dc supply. when transmitting, the voltage rises on the bias potentiometer, as if rf is coming into the bias dc line. and the audio on am sounds distorted, kind of fuzzy. with a modulated input ranging from 0.25w, to around 2watts.
thanks for reading. any help is very appreciated.

joao g.
 

Attachments

UweX

Joined Sep 2, 2020
12
Are the inductors rated for the power? Size of the core, thickness of the wires ?
Are the transistors selected as matched pairs ? ( or at least from the same lot, so they can be assumed to be similar). A small resistor in the sources is a good idea. It should be small, but the the individual bias current can be checked. If the transistors are unmatched you have higher distortion for sure.
The circuit shown here has no feedback. The advantage of that is , that the gain is the highest possible gain. The disadvantage is higher distortion, but the filters at the output should reduce any harmonics.
 

sagor

Joined Mar 10, 2019
512
Is the attenuator good enough to prevent overdriving the MOSFETs? Also, the 5.1V zener will drift a bit as it warms up, which is normal for a Zener. That may affect the bias voltages a bit. You should use a proper 5V regulator instead of a zener. If the bias load is small, you may be able to get by with a 78L05 regulator. Check the input voltage to it, to make sure you don't exceed maximum input voltage.
It may be possible that L1,L2 and R3,R4 are picking up some RF. I would put an extra bypass cap between the L and the R, though RF should not get thru L1 and L2 in the first place, but no harm in doing so.

Good Luck
 

Thread Starter

Joao Goncalves 1

Joined Feb 17, 2019
7
thanks guys. good advice. i will try to improve the build, taking that advice in consideration. i will let you know how it works out.
thanks again, and best regards.

joao
 

UweX

Joined Sep 2, 2020
12
Maybe it is also a good idea to publish a photo from your setup. This may already clarify some points.
A small attenuation at the input maybe necessary to avoid a high SWR for the controlling radio. What attenuation did you choose ?
On the first glance I didn't recognize the separate bias settings for each transistor. Taken that into account, you can adjust the bias current for each transistor easily. But be aware, when the gate voltage for both transistors is much different ( my guess > 100mV at the same current) , the transistorss are not matched.
 

Thread Starter

Joao Goncalves 1

Joined Feb 17, 2019
7
hello friends. i made a few changes to the amp. Since i could not get it to sound right, I went back a bit and tried to use only one transistor ( based on similar designs ), and i think it finally is working properly. distortion is minimal on AM, i think it it's fully modulated since i am using the 757. i ended using a pair irf740 for the finals (parallel), running on 40vdc, and it works well. it sounds nice at 25w carrier, modulated. now to get more power i will build another amp. maybe the one i meant to build or a more powerful one so i can use it on am nets. the design is not very clean, i know, but it was experimental. thanks everyone, for the help.
 

UweX

Joined Sep 2, 2020
12
Just some comments to the picture published here: I can't recognize the relay. If you use a wrong one, you may have some capacitive coupling between input and output.
The general construction looks good with the big groundplane underneath. The points important to me are:
Even if this PA should only work at 80 or 40 m, the transistors have a higher frequency response, so the construction needs to work well up to that frequency. The cooling construction looks good for something like 5 W, maybe with this free airflow 10 W per transistor. At higher power I expect it to become hot. I also don't recognize a grounding of the heat sink. The transistors are isolated by the isolation pad, but that is only a DC-isolation. At higher frequencies you still have a capacitive coupling and that may give some unwanted radiator just in the middle of the circuit. So keeping both heat sinks grounded is one advice.
The wiring in the input and gate biasing circuit looks very spacious. I would put the components closer together and this way reduce the chance of any spurious coupling at higher frequencies.
You should also watch the signals at the drain with a scope, not only during adjusting DC-bias, but also with some RF applied.

Even if you want to finish a PA on your own, take a look on https://www.banggood.com/100W-SSB-L...tsink-CW-AM-FM-C4-005-DIY-KITS-p-1586430.html . There are some differences, like the T/R switch is directly controlled and not through the rectifier as in you circuit, but it is a compact construction. The size is mainly determined by the heat sink. If you target 80W, the heat sink should be sized for 50 W power dissipation at least. I prefer passive cooling at that power level, it avoids a noisy fan.

vy 73, Uwe
 
Top