144MHz LDMOS 1kW Amplifier

This is the story of my home built 1kW LDMOS Amplifier.
I am not a quick builder, and the reason I think is that I sometimes tend to overdo my research.
This project took me about a year realize.

I have been searching for some photos, but they are not as many as I hoped.
Some of theme seems to be lost.
Anyhow I hope that you may find the story inspiring and interesting.
I will not only write about the success but also the miss happenings that occurred.

So here it is.......

The design
The construction is based on DUBUS 4/2010 by F1JRD and I managed to get my hands on maybe one of the last kits sold from RFHAM.COM.
I don't know what happened, but RFHAM.COM is no longer accessible and no more kits are to be found.
If you have some information, please let me know in the comments below.

F1JRD has a short description of the amp on his page at qsl.net. F1JRD amplifier description.

My construction criteria

• I wanted a robust and simple amplifier easy to build and maintain.
• My space is limited, so the smaller the better.
• Water cooled with low acoustical noise.
• Good looking exterior.
• As few external buttons as possible using a touch display instead.
• Software controlled Bias, fan regulation and sequencer.

The cooling block and cupper pallette
I was very lucky to find a cooling block and a piece of cupper plate in the scrap at work.
The cooling block was leaking, and the cupper plate came from a scrapped rectifier.
I told the production manager about my project and asked him if I could take the scrapped cooler
and the cupper. No problems there. If you are nice to people, they will often be nice in return.


The raw copper plate material.


A friend of mine helped me with the milling of the surface. This was needed to get a clean plane surface and a good heat transfer to the cooling block.


The finished copper plate.


The amplifier kit was easy to build and had just a handful of components.
A low pass filter was ordered from dutchrfshop https://dutchrfshop.nl/en/
The LDMOS was soldered to the palette using the method described by Jim Klitzing W6PQL.
He has made a good video that he posted on his homepage and on YouTube.


Low pass filter to the left and amplifier board to the right.


Added circulation fans to the cooling block.
Amplifier is hooked up for initial test.



I measured all the cables thoroughly.
Everything was set. Power supply on. Bias voltage set. No smoke. Very good!
But..... No output power from the amplifier, but it is pulling current from the power supply.
Gaaah! What is wrong !? Checking the setup again, Hmmm.... Dummy load is cool, and the
cable between the amp and the dummy load is warm. I measured the cable again with my DMM.
WTF! Cable is broken! No contact with centre conductor!

New cable, second attempt.
1000 +Watts with 3W of drive. Happy face :) Moving on.


I needed a metal case for my amplifier.
There are a lot of audio stores on the internet selling suitable chassis for home builders, but boy
are they expensive. So I started to search for alternatives and this nice little chassis came up.
It is a chassis for PC home builders in Mini-XTI format made by Fractal Design.
It was in a perfect size to fit my amp and wallet. It needed some internal rework though.
The price today for this model called "Core 500" is about 70$. Money well spent.
The Core 500 is prepared to fit a radiator with cooling fans and comes with a 120mm fan mounted
at the back of the chassis.

Fractal design Core 500 https://www.fractal-design.com/products/cases/core/core-500/black/

Ok, so the rf amplifier works, but to make it complete a lot of components needed to be obtained.
A shopping list was written.

Shopping list

• Nextion touch display.
• Arduino UNO board.
• Relay board to turn on fans, pump etc.
• RF Relays.
• Main power high current relay.
• Step down converter from 48/15V DC. To control fans pumps and relays.
• Second step down converter 15/9V to power the Arduino and display.
• Circulation pump for cooling liquid.
• Cooling radiator with fans.
• Temperature sensors to measure water and LDMOS temperature.
• DAC for Bias control.

I chose to work with the Nextion display because it is built like an industrial HMI system which I am
familiar with through my work as an automation engineer.
On a nearby HAM and electronics auction I found some stuff that I thought might be useful in my
project. Including a Velleman relay board and a couple of Vellleman DC/DC converters.



The Velleman K8056 relay board.

This board has an PIC16F630 microcontroller which has a software that controls the relay board.
I was thinking about to replace the PIC controller and use my Arduino to control the relays.
The board used BC547 transistors to switch the relays.
They were all replaced with BC337:s because they have a higher collector current which I needed
to switch my coax relays. The first to relays RY1 and RY2 were removed and I connected my coax
relays in their positions instead.


Velleman voltage regulators. Left one modified for higher current.


A switched step down DC/DC converter was ordered from china.
The idea was to convert the DC voltage in two steps 48V to 15V and then from 15 to 9V to power
the Arduino and the Nextion display.


48/15V DC/DC step down converter.



An Arduino Uno from the junk box.


Nextion 3.2" TFT touch display.



Cooling radiator with fans to the left and radiator top view to the right.


A combined tank and cooling pump capable of circulating about 300l/h.

An MCP4725 DAC circuit was ordered to control the Bias voltage of the LDMOS.
Even though Microchip recommends it to be used in LDMOS circuits I have not seen anyone using
it in amateur radio applications. I think using this DAC with an MCU is a great way to control the
Bias voltage, because then the amplification class can be software controlled.
Meaning I can program the Arduino in a way that amplification class A, B, AB, C etc can be user
selectable through an options menu.
The MCP4725 puts out a very clean DC signal and the voltage is set via the I2C Bus.


Picture shows the small DAC board with the MCP4725 chip in the centre.


All the boards and components mouted in the box.
An 8db attenuator board for RF input was purchased from Digit@lion Technologies.
This will dampen the input signal to a more suitable level for most 2m transceivers on the market.
About 20W input for 1kW+ output from the amp.
It is the little blue board in the bottom right of the picture.



The on/off buttons on the top of the front cover that is normally used to turn a PC on or off
was rewired, to power on/off my amp. Also, the built in indicator lamps in the buttons were rewired
and used. One USB port on the front was cut and rewired to my Arduino board.

Eeeehh.... I was not to happy with the wiring but, it would be ok for testing.
The same with the boards mounted in the front.
Not the best looking installation but good enough for tests.......NO it was not!
I would later find out the hard way, when increasing the output power, the Arduino could not
handle the RFI, and it would shut down and reboot. More about that later.




The Software
It was time to start writing the software for the Arduino and the Nextion display.
The functions I wanted to implement via the Arduino were the following.

• Amplifier On/Off.
• Circulation pump.
• Temperature controlled fan speed regulation.
• Temperature controlled Bias regulation.
• Over temp protection.
• SWR protection.
• RF relay sequencer.
• Selectable amplifier class via touch display.
• QRO or QRP via touch display.
• Show water and heat sink temperatures on touch display.

Reading the temperature
To be able to read the water temperature and the heat sink temperature I purchased two
NTC 10k thermistors and a voltage divider.
Below is the Arduino test function I wrote using the Steinhart equation to calculate the temperature.

// Read Heatsink temperature
void ReadSinkTemp()
{
  double Temp;
  int nSinkTemp;
  nSinkTemp = analogRead(ai0_SinkTemp);
  Temp = log(((10240000 / nSinkTemp) - 10000));
  Temp = 1 / (0.001129148 + (0.000234125 + (0.0000000876741 * Temp * Temp )) * Temp );
  dSinkTemp = Temp - 273.15; // Convert Kelvin to Celcius
  } 

Ok, now I had the temperature readings, but how about the Bias voltage vs the temperature?
This called for some research.


Bias Control
The Freescale LDMOS requires 2500mA Idq@50V for class AB operation.
In order to maintain consistent linearity over a wide temperature range, I needed to regulate the
Bias voltage when the temperature was changing.
My thoughts about this was to use a micro controller to do the job, but what did the temperature
curve look like? Searching the internet I could not find any temperature curves.
Anyhow, the best way to know is always to measure. So, I turned of the cooling, applied the
supply voltage and set the drain current at 2500mA at room temperature 24°C.

As temperature increased, I adjusted the Bias voltage in 10mV steps and when Idq reached 2500mA
I made a note of the temperature. I stopped when the temperature reached to about 90°C.
The table below shows the actual values that were measured.

Measurement	Bias mV	Temp°C
1	2784	24
2	2770	30,5
3	2760	36,1
4	2750	40,3
5	2740	45,2
6	2730	49,4
7	2720	53,4
8	2710	57,4
9	2700	62,2
10	2690	66,1
11	2680	70,2
12	2670	74,2
13	2660	78,3
14	2650	82,5
15	2640	86,7
16	2630	90,5

When putting the table into excel and then creating a diagram I was happy to see that it was a
linear curve. That would make the programming so much easier.
I could just interpolate the min and max values with a small equation.



// Calculate output bias y2 with interpolation.
x1 = 24;              // Low point temperature C
x3 = 90.5;           // High point temperature C
x2 = SinkTemp; // Current Sinktemp (Close to LDMOS)
y1 = 2784;         // High point bias voltage mV
y3 = 2630;         // Low point bias voltage  mV

y2 = ((x2 - x1) * (y3 - y1) / (x3 - x1)) + y1;

The same measurements were also made with 200mA of Idq for class C operation.

Fan speed regulation
Now that I had temperature reading it was easy to implement fan speed regulation.
One relay on the relay board was removed, and the fans were wired in its place.
Then a PWM output from the Arduino was used to control the speed via the BC337 transistor that
previously was wired to switch the removed relay.



Nextion display
I looked around on the internet for a suitable display that would suit my needs and I stumbled acrossthis wonderful display from the Chinese supplier Itead.
It is not just a dumb touch display, it is an HMI (Human Machine Interface), which meens that
most of the configuration, pictures, touch buttons etc are already available in the free configuration
tool Nextion Editor. It can be downloaded at https://nextion.tech/nextion-editor/

Below is a screen shot from the development environment ver. 0.53.



There was on downside with this editor though.
I could never make the fonts look good. The spacing between the letters was uneven.
I don't know the cause of this problem, but it could be solved using pictures instead of plain text.
The development tool has been reworked and I have not yet had the time to check if this
problem is fixed.

Never the less, if you decide to use this display in some of your projects you will find that it is very
easy to learn how to use. I can really understand why it has become increasingly popular over
the last couple of years.

Testing the amplifier

Earlier I mentioned a little about some RFI problems that occured.
It turned out to be a more serious problem than I thought.
When putting out about 600W+ from the amplifier the Arduino would shut down and reboot.
Another serious problem, and one that I should have forseen was RFI from the Arduino to my
2m receiver.
When I turned on my amplifier the S-meter reading showed about S5 across the whole band.

CRAP!!

Maybe some of you more experienced builders are laughing now, but I fixed it :)
It took a lot of work though.

Shielding the Arduino board
It was obvious that I needed some shielding of the Arduino micro controller but the space was
limited. The solution was to buy two RF shield enclosures from Teko and a DB25 D-Sub filter
contact with 4.7nF filter capacitors.


Arduino and temperature boards in shielded enclosures.



The completed RF shield work.

It took me a couple of days to rewire an resolder all the connections but it had to be done.
I must admit that I was a little nervous when turning on the amplifier and the 2m receiver
again for the first time after shielding the Arduino. Would it work?

YES! YES! YES!

S-meter reading with a 50Ω dummy load attached, is ZERO.
Some oscillator noise could still be heard at 144.000Mhz at about -120dbm.


Noise floor at about -140dbm

Ok, good.
But how about the RFI that shut down the Arduino?
I tested the amplifier at full output power 1200W. The Arduino was still happy and so was I!

It was time to test the amplifier in the Nordic Activity Contest also known as NAC.
Although it was not yet completly finished I was egor to try it out for a litlle longer period of time
and see how it would perform.

The amplifier worked but there were some minor issue that I would like to fix.
When calling long CQ:s, the cooling system showed to be a little undersize causing the heatsink and
cooling liquid slowly heat up to unexceptable temperatures.
The weak part was the cooling pump which has a max rated temperature of 60° Celsius


A new pump with a higher flow has now been ordered, and I am waiting for it to arrive.
Hopefully it will do the job.

One more thing that still needs to be done is to add SWR protection.

An RF coupler board was purchased on ebay that I will use for high SWR protection.
It also has the ability to show output power.

For now I will have to settle for an external SWR/Power meter and keep a close eye on it during
transmission.


Side view of the interior. Cooling pump and coaxial relays.



Amplifier side by side with my Yaesu FT-726R


Hpe cu agn.
73's de sm7tkr.