Well, how I go on to get the amplifier on level,....???  Date is January 13, 2020.

I was thinking what caused a low current. Maybe the impedance of the output transformer was to high. So, I changed the output transformer to a smaller one. Look at the picture in the previous pages. I put the smaller transformer at the edge of the box so you can see what transformer I put in now. Well, the result was that I got a higher current at a lower Voltage (about 4 amps, at 15 V). But the output was disappointing, only 15 watt. I changed the input power a bit, no effect. I lowered the feed back resistor of 470 ohm to a bit more and a bit less, no significant change,......I changed the choke coil and decreased the number of winding's from 12 to 10. No effect at all,......I got desperate.

I was getting to distrust my power supply. It was quite an old unit, and certainly not calibrated.

So I changed everything back to the begin state and started to think about a good power supply. The first idea was getting a car battery. But I had a few old 12V batteries from a APS unit. I am recharging the batteries now,.......and go to bed.

Tomorrow, I start using the batteries. First one, and if that goes OK, two in series, with a big resistor in series to avoid accidents,....

 

See you tomorrow,...

 

Well, the batteries were not good anymore, so I disposed them in a car repair shop nearby.

I picked one of my good car batteries and put it in series with a power supply of 13.8 Volt, 10 amps. In series with the amplifier I put 13 resistors of 13 ohm 5 watt. Look at the picture in the right side of this page. I made two banks of 13 R's. And I start using one bank.

So if the amplifier is drawing 1 Amp, the voltage over the resistors is 1Volt.

I had 26 Volts of my power supply. And I pushed the PPT,......of course I had the dummy load connected.

I measured 20 Volt on the V+ of the amplifier (the red plug at the upper right side of the amplifier box). And the voltage over my 1 ohm series R was 6 Volt. So my amplifier was drawing 6 Amps. The end stage (two FETs) was doing about 5.4 Amps, the rest was flowing through the electronic relais switch components.

5.4 Amps x 20 V = 108 Watt DC input.

The output was 55 Watt !!!! It was beyond my expectations.

Problems,......

• one FET was warmer than the other, 37 degr celcius, versus 48.
• At the entry of the upper ferrite tube, the temperature was going rapidly to 90 degrees !!!
• My antenna, which did on 14.100 an swr of 1:1.2 had great effect on the current. Instead of 6 amps on the dummy load, I got over 8 Amps on the antenna
• The wave forms on the drains were absolutely horrible,.....see foto

 

So, I got power,.....but there was some challenge to improve,.......

 

Time to stop when you have a good feeling,.......next time we are going to solve the issues.

January 15 2020

 

So, with the good feeling I started to experiment a little further. I got my 55 Watts, so why dont we try to remove the 1 ohm resistor in the power line and look what happens,.....

And then, with 25 Volts V+ and about 6 Amps, I got 70 Watts out RF, fantastic,...............but not for long,........the power supply stopped working and the reason was that the fuse which is between the car battery and the amplifier was blown. But why,.....I found out that the upper FET was burned. The drain-source resistance was only 1 ohm. I had to replace it, and I did. I also have a feeling why the FET was gone. If you have the  bias voltages not correct, and start turning it, the signal on the drains can get really HORRIBLE. It means a lot of harmonics are created, and they can go nowhere, because the filter is blocking them to go into the dummy load. And I think that I turned a bit to much on the biases and created a lot of power but a great deal was HARMONIC power, and killed the FET. The Voltages were swinging to over 100V,.....above spec.

What I learned is that if you have to tune the amplifier you have to start low and keep track of the biases and the input power/signal waves. If all is OK, than increase the V+ and again check the wave forms on the drains with a scope, which is capable of seeing the harmonics. I had in loan a digital scope of 200MHz, which is perfectly OK for our measurements.

 

OK, again I connected my variable power supply (0-30V, 4 amps), and only one 1 ohm resistor bank. The following measurements were done:

Power supply      V+               Amps     RF power output

      13V             10.1 Volts          2.7A               12W

      15V             11.3 Volts          2.8A                16W

      removed the 1 ohm resistor

      15V             14.1 Volts          3.2A                23W

Why 0.9 voltage drop at the DC power entrance,.........?

Well, put the scope on it and you see ! RF voltage on the DC power input,......So, another decoupling C op 22nF connected to the DC input. The RF signal was gone for 90%. And use thick wires. I changed from 0.5 to 1.5 mm2.

And now I see:

      15V             15 V                    3.4 A              26W

Again a problem solved.

Now I want to discuss the wave forms of the drains, see the picture I put in the right part of this page.

In the picture at the right you see a yellow and a purple wave. Left and right channel of the scope, and upper and lower FET drain.

You can see that the basic 14 MHz waveform is there. But we see other wave forms.

Dirty higher frequency. We can imagine that this high frequency power can not pass our fantastic output filter,........ but where does it go,........well into heat somewhere in the FETs and in the other components. The high frequent energy is bouncing around the circuit until it is all transformed into heat. So why are these higher frequency components in there ?????? My first reaction,....I really don't know.

Let's put up the DC input voltage a bit.

     16V              16V                      3.5A              26W

     17V              17V                      3.4A               28W 

     18V              18V                      3.5A               32W

      19V             19V                      3.8A               38W

      20V              19.9V                  3.9A               41W

 Let's again look at the wave forms on the drains: See picture at the right.

The yellow one is pretty ok, but the purple one is still pretty messy. The question is why.

Well I measured the middle point of the primary winding of the output transformer, and this should be rf-dead. 0 Volts at he best. But it isn't. I measure 3V peak-peak. Part of it is probe antenna effects, but part of it is certainly that the earth of the mid is not 0. So it will be added to the drain voltages and will give all kinds of complex effects. And I have not the tools to exactly find out what it is and why.

So let it be,............

The point is now, how do we go on. I think, that 20V is a pretty stupid voltage to work with. And certainly with about 5 amps. I would really like to work with 24 Volts, or even a bit more 24.8V, so I can use in the fields two car batteries. But I have to be careful,........my measurements give me a FACT, and that is that the voltages over the drains are going way over 60 Volts, IF YOU HAVE NOT TO MUCH HIGHER HARMONICS in your signal. And what happens if you forget  to connect your antenna,.......I think that the voltages are going over 100V and that is the end of the FET specifications,......

Yes, if you push it, you can go to 70 watts RF, but than you have to make sure that the amplifier is tuned 100%, and you certainly have to make a high-SWR module, which turns down the input or puts the bias to 0.

Why were my voltages on the Drains were to high,...?????

Well, I found out the reason. The transformation factor of my transformer was not matched to the output impedance of the FETs. I changed the transformer from 2:4 wnds to 2:3 wnds. That was not a good idea. It got worse. Than I changed it to 2:5 wnds. That appeared to be very helpful.  The voltage after the transformer got higher and the current lower. Before the transformer the drain voltages got lower and the current higher. That was exactly what I was looking for. A far more better impedance match.

Drain voltages had now the level of about 60 Volt. From 0-60 Volt. So, peak value 30V. The current was about 3.5 average, about 4A peak, that gives an output impedance of 7.5Ohm. Multiply by (2/5) to the power 2 = 25/4 = 6.25 gives an output impedance of 7.5x6.25 = 46.8 Ohm. That happens to be very close to the input impedance of the filter (50ohm) . Very nice match !!

 

The following measurements were done:

20Volt DC

Drain voltages max 60 Volt, current average 3.7A with 42 Watt RF output after the low pass filter !!. Look at the pictures 

21Volt DC

Drain voltages max 60 Volt, current a bit more and 43 Watt

22Volt DC

Drain voltages again around 60 Volt, current again a bit more 3.9A with about 47W

23Volt DC

Drain Voltage again around 60 Volt, current goes a bit above 4A with about 50W

24Volt DC,

Drain Voltage a bit over 60V, current a bit more over 4A,  with 53 Watt.

Here my 30V,4A power supply gets in trouble. At more than 4 A the voltage is      pushed back a bit.

But let me try 25Volt DC and optimize the wave forms. I got again a bit more than 4 A and the power was about 53 Watt. So increasing the DC voltage further was not very helpful, but the wave forms got better !!!  Less harmonics, so less useless power transformed into heat.

 

I also put the scheme into Simetrix, a scheme simulation program, which is free for a certain functionality, which enough for my purpose. The details of the results are given in a separate chapter, see table of content.

 

I decide that here I end the push for power with the following lesson learned:

• Use the right components, especially for the cores and the ferite, and certainly the capacitors. The 680p capacitor is in my amp a SilverMica type.
• I used a cooling block which is far oversized for 40Watt, good to know.
• I have to put more energy into a good earth. RF-cold should be really RF-cold. No unknown currents, and strange voltages on places where it can create damage.
• Watch the signals,.....with a good scope, 200MHz worked ok for me.
• The power supply should be able to deliver the current easily, and on the DC input it should be RF-stone cold.
• Measure temperatures of all components. Bad components tend to be lossy and get warm. Replace it asap.

Now I understand the basics of the amplifier, I am going to build it from scratch on a much smaller cooling block, and because now I now exactly which components I need the print will be much better designed, with a LOT of RF decoupling capacitors.

 

So, I am going to do the following to end the proto1:

• Build a box with input of two car batteries.
• To work with 50Watt output I connect the amp directly to the battery. But I know that is risky. I only have a 10A fuse as security measure. I have to add a BAD SWR feed back device, to protect my amp
• So, I also build in a 20V, 5A module for the amp, for only 40Watt output, with a LM338. It is secured at 5Amperes, which give me a more secure feeling.
• Build in a 15V, 1 Amp module for the QCX CW Transciever

Go into the field, put my 20m GP up and start calling DX,......