50W 222 Amplifier and Modifications - K7MDL2/IC-705-BLE-Serial-Example GitHub Wiki

17 January 2025

I ordered a 50W 200-300MHz amp off AliExpress. Most of these target the jammer market and come with a NE555 chip sweeping a VCO over some range of frequencies. In jammer config the VCO is active and there is no input SMA connector. Some come with the VCC removed, or at least disabled (no output cap, maybe no power to the VCO). and an input SMA connector installed. Some adds show limited info for a range of frequencies, even though you are only buying one amp that covers one the ranges. The ad may have a correct title, but the details are cloned from a different band. My 900Mhz amp had details for 2.4G WiFi.

To run a VCO they need to accommodate around 0 to 10dBm drive so these have a lot of gain compared to most amateur market amps. 40 to 46dB gain for the 50W category. Most of these size amps run on 28V, the ones I have mostly use discontinued and/or active NXP parts. A MMIC type 1st stage, a 4W drive 2nd stage, then the final. They come in very compact milled boxes and require a heat sink. Some have just GND and V+, others have a 3rd wire for PA_En. The PA enable takes 5VDC and turns on the amp, a small LED will turn on inside.

This 222MHz amp I thought would have the input SMA installed, VCO disable or removed like the other 3 similar amps I received. Nope. I bench tested it with a wattmeter and spectrum analyzer and it puts out 40W at 24V and 50W at 28V and sweep the whole segment 200-300MHz. There is another emission at 380-420 MHz or so.

To use this amp it requires some drilling and tapping for the SMA connector, VCO removal, soldering in a 0603 5pF cap.

Below is basically a step by step with pictures and some text showing how I modified mine. At the end I got 50W out as expected at 222MHz. Expect this to take several hours total. Probably less for you since I had to spend time figuring out how to disassemble it, how to remove the VCO, and the best way to mount the SMA. I found several of the M1.5 lid screw very tight or stripped. The screws holding the thin Teflon PCB are pretty tight also, one was stripped. the screws in teh middle pass through a brass or copper heat spreader and are on the short side engaging only a few threads so they are easy to strip, and one was stripped.

Putting it back together you might feel like they are too short or stripped, use only 2 fingers to tighten. You also do not want to crush the Teflon PCB. Initially the new heat sink grease you apply will squeeze down and you need to periodically circle around and retighten each screw, this lets the loose ones grab some threads. I ended up cutting down a new long M2.5 screw to engage all the threads for stripped screw.

Here is a look at the amp as I received it. No SMA, VCO installed. 50W out as advertised. It is 4..375" long by 1.5" wide.

A close up the VCO and RF input area. The NE555 is on a small board mounted on 3 header pin directly above the VCO. It ramps a voltage controlling the VCO. You can just see the RF Input trace to the far left side. On the right side is a NXP M6S004 3GHz 4W driver amp used in several of these amps.

The 28V power and 5V PA_En (PA Enable) wires. These small gauge with silicone jackets. Very rubbery. SMA bulkhead connectors use M2.5 screws. The single threaded hole you see here is M3.

The wires pass through a milled slot instead of using feed through caps and are soldered to the PCB. Note the +28 and EN silkscreen labels in case there was any doubts.

Now to take it apart. I removed the 2 output SMA screws then desoldered the center pin sliding the part out.

I desoldered the 3 pins holding the NE555 board. The V0204-10 VCO is below, it is a small PCB module soldered to the main PCBV with a metal can over the top. I also started to remove the PCB screws.

This unit was advertised as a differentiator to have an RF output isolator which I wanted. You can see here it that it is not installed.

I soldered a piece of wire to the non-existent isolator 50ohm load solder pad. This gives you something to help lift the PCB out of the case. It will be stuck down by the heat sink grease. I used the wires and NE555 board at the other end to slowly break it loose from the case.

Out of the case!

Now to size up where to install the input SMA connector. You could put it high or to a side and run a small jumper wire from the center pin to the PCB input trace. The SMA mounting holes are 1/4" radius from the center pin. Cannot use the M3 and it is too far to the left anyway. Cannot restrict the wire holes. Maybe a feedthrough cap can fit in a M3 hole? If you horizontal mount the 2-hole flange version, the right-side hole will hit the lid screw located there. I decided to do the diagonal mount like the output has.

Used a 4-40 screw pushed through the M3 hole to see how things lined up accurately. Too far off, squeezes the wires, wrong size hole. Nope.

Since the amp will be inside a box with limited access to the ends I decided to use right angle SMA connectors and save on space and 2 less right angle adapters needed. Same hole spacing, 1/4" radius.

I first drilled the center pin hole. I start with a 1/16th drill to locate the holes more accurately. My drill press tends to move the right after starting the motor and I often end up with holes not exactly where I intended them. Then I moved up to 1/8" and finally a 3/16". I should have used a 7/64", the 3/16 is too large.

After I drilled and installed this, I realized it is about a 1/16" too high. You can use the output end as the centerline reference and get it right. It is close enough though, even better if here is some play in the screw holes. Since I do not have a M2.5 tap so I drilled and tapped for 4-40 screw. They are a tad larger and require drilling out the SMA flange holes.

Now to remove the VCO. Used a small screwdriver and pried up one end of the metal shield. Ther are same gaps around the side. You can destroy the VCO module, no matter, just don't hit the main PCB parts. You can see there is no input coupling cap installed on the RF input trace since there was a VCO installed. I ended up installing a 5.0pF 0603 size cap there.

Next, I used a new knife blade to saw through the solder joints under the module. Solder is soft and some sawing and rocking motion with blade pressure will eventually work though the soft solder. You do not want to lift much or you will risk lifting the PCB foil.

Victory! I intentionally broke off the corner of the VCO module to better see what was under it. Mostly ground plane foil. Pictures a little fuzzy but you can see I knocked loose a small cap above the knife. I resoldered it back on but I think it filters the VCO 5V Vcc and is not required.

A closer look. The Input RF trace just passed under the corner of the VCO module then on to the 1st stage driver.

Here you can see the cleaned up milled aluminum case.

There is an input attenuator to the first stage. I am considering increasing the attenuation later to match transverter output. Right now the UR3LM 222 transverter puts out 8W. I am thinking I will remove the final stage amp transistor and see what I get, if it is low enough all the attenuation can be done here. The cap at the far left is the newly installed RF input coupling cap.

At this point I put on new heat sink grease over the whole bottom (not too much) and put all the screws back in. As mentioned at the start, one screw was stripped and a few others would not get started until the grease was fully compressed.

I put it on the bench into a dummy load and wattmeter again. I initially used a 100pF input coupling cap but only got 30W out. It was very touchy, possibly ancient HP sig gen output switch getting flaky or a bad cable. Wiggled cables, put my finger around the input, put the lid on, messed with the slug, not much change except when I move my arms around the box. Maybe oscillation with 35-40dB gain here, but don't think so. I replaced the input cap with a 5.0pF and replaced the input test cable. Now it got 42W with the 50W slug. Not the 50W I measured before, same slug and output cable. This wattmeter has a push switch for peak reading and after toggling it I got >50W out in normal mode. Problem solved.

Next up is spectrum analyzer tests and evaluate the need for an output Low Pass Filter. First is the baseline, amp off. Looking at a 1800MHz wide spectrum. SA center set to 1000MHz. 30dB external attenuator and 30dB internal attenuation set. 47dBm out into a 30dB 100W attenuator leaves 17dBm at the input of the SA. Using a HP 8640B signal generator at 15dBm out at 222.1MHz. 50W out, 28V, and 2.2A.

Here is our baseline, amp off. Pretty quiet.

Now turn on the amp

Our 222 signal is at 47dBm. Making that our refence point of 0

Fundamental @ 222	2nd @ 444	3rd @ 666	4th @ 888	5th @ 1110	6th = 1332
0db	-17dB	-29dB	-37dB	-52dB	-42dB

Cannot see 7th or higher harmonics.

Not pretty. Since this was designed to be a jamming amp I am sure they did not care about harmonics, the more the better for them. This will need some serious filtering. Maybe a diplexer type where all that harmonic energy is routing into a dummy load instead of reflected back to the amp or burned as heat in the LPF components. Hopefully it can be made small enough to fit inside my box.

Next up is to measure amplifier linearity. That will take some more work to properly setup. I can use my HL2 (SDR low level output and Thetis SDR app to generate a two-tone signal at 28Mhz IF into the 222 Xvtr. But first I have to attenuate the 8W down o 10-15dBm. At idle there is no current draw suggesting it is set up for Class C operation. There is a pot that when turned the idle current goes up but the power out and current draw drop significantly. At some settings I can adjust the drive level so the 2nd harmonic drops way down while the 3rd stays about the same. Increasing the drive causes output to fall off around this point. Need to do a Pin/Pout plot. Much left to test and adjust still.