5 watt driver amp T1 position

[Robert]

I installed all the parts with soldering paste and a frying pan. !!!  Crazy idea but it worked. !!

Now I have the parts installed, I was reading about T1 being a 4:1 transformer. what way is it supposed to be installed? The only position indicator I can see is a little dot above the number 4. so, I have it installed with the 4 with the dot on top, then R in the middle and 7 on the bottom. hard to explain. !!!!   It reads 4r7 horizontal,  and I rotated it 90 degrees to the right and soldered. I have one chance of getting it right. Is the position I have it in correct??????

Robert KH2BR

[jcve...@gmail.com]

Hi Robert,

A frying pan doesn't seem too crazy, I use a toaster oven. T1 is a couped inductor that has 2 identical inductors on the same core. It doesn't matter which way you put it on the board; interconnections on the board connect the inductors in a 4:1 configuration.

73,

Jim WA2EUJ

[vu3...@gmail.com]

I am beginning to build the 5W amp. I am not yet brave enough to use a frying pan, considering the fact that it takes ages for replacement parts to reach my side of the world.

I am wondering if I can use a hot air gun instead to install just the main ldmos transistor Q1 and may be U1. Is that doable?

73

Ram VU2JXN

[jcve...@gmail.com]

U1 is easy to do with an iron an tweezers, hold the device in place and solder one of the legs (not the middle one because it's connected to the ground plane). Once you have it in place, solder the other two and the tab. 

Q1 shouldn't be too bad if the heatsink is detached, two soldering irons would be my first choice to flow the area under the device but you should be able to use hot air as well.

73,

Jim WA2EUJ

[jcve...@gmail.com]

What kind of heatsink do you have attached to the PCB. I recommend a Wakefield 567-94AB but other units may be acceptable. You cannot use the board without a heatsink.

Jim.

[vu3...@gmail.com]

I don't have a heatsink installed yet. Thought I would do that after bias adjustment. At 100mA, it gets very very hot.

Is the heatsink required even during the bias adjustment?

73

Ram

[vu3...@gmail.com]

I just fixed the recommended heatsink. At 25-30mA, the heatsink and the Q1 gets a bit warm, after that some kind of thermal runaway happens, the current across R5 (I sense test points) keeps jumping up steadily and the ldmos device gets very hot.

I wonder what I did wrong here.

73

Ram

[jcve...@gmail.com]

At 100 mA and 13.8V the device is dissipating 1.38W. A heatsink is required at this power level.

It sounds like you amp is oscillating as soon as the bias gets high enough to turn the amp on. Make sure to terminate the input and output in 50 ohm loads and make sure all of the 0.1 caps are in place. Als check the inductor in series with the gate.

With U1 and Q1 installed, the circuit has about 40 dB gain so you need to take precautions to prevent oscillation.

Jim.

[vu3...@gmail.com]

Thanks Jim. I am building the non-mmic version. I have verified that all the 0.1 caps are in place and the 68nH inductor in series with the gate. All the parts are from mouser. The input and output are terminated with 50 ohms. I will insert a tap and look at the spectrum analyzer output tonight to verify that it is oscillating.

Anything else that I can do to arrest the oscillations?

Ram VU2JXN

[jcve...@gmail.com]

Things to check:

1. Make sure that the solder under Q1 flowed completely and there is a solid ground connection to all parts of the MOSFET's ground pad. This is also necessary for good heat transfer.

2. Make sure that the RF connections are 50 ohm coax with center and shield connections; use RG-316 or RG-174.

3. Use a will regulated 13.8V power supply.

4. If you are not using the the MMIC pre-driver remove all of the biasing components.

I have only seen the circuit oscillate when the pre-driver is installed.

73,

Jim WA2EUJ

[vu3...@gmail.com]

I did a quick experiment. I inserted a tap on the output between the amp and the 50 ohm dummy load (with rg-316 cable) and hooked up the tinySA. Input terminated with 50ohm using a coax (rg-316). Increased the bias to around 80mV and I started seeing the current jump up steadily. However no oscillations seen on the SA. SA was set to 1-400MHz bandwidth. There is a very tiny blip near 100MHz, but I am not sure if it counts as oscillation. Even without any input, I see outputs of that level. The power supply used to power the PA is the same commercial power supply I use with my other rigs.

I guess the item 1 in your list is the one to investigate now. I had applied thermal paste as well to the heatsink/pcb interface surface. When the device gets very hot, the heatsink gets warm as well, but perhaps the LDMOS device is not making enough contact with the PCB. I will remove the device with a hot air gun and redo the soldering and check.

Who knew there would be so much of learning from this tiny amplifier!!

Thanks again, Jim

73

Ram VU2JXN

[jcve...@gmail.com]

The act that the part is getting hot but the heatsink is only getting warm is a sue sign that you aren't getting good heat transfer. After you solder the part, make sure the backside of the PCB is  smootd and doesn't have any solder lumps you don't want any air gaps.

The AFT05MS016 is designed to work up to 1000 MHz so sweeping up to 400MHz might be missing the oscillation.

73,

Jim WA2EUJ

[vu3...@gmail.com]

tinySA has 800MHz bandwidth. I tried it with 800MHz as well, but do not see any oscillation in the output. I will use the hot air gun tomorrow and get the part out of the board and clean the board and re-do the soldering. May be I will not use solder paste this time and just use regular solder. Perhaps there wasn't enough paste applied during my first attempt. I was afraid, it would spread and short the ground and the gate terminal on the board.

73

Ram VU2JXN

[vu3...@gmail.com]

Just to conclude this thread: I found the culprit. It was "operator error" (in combination with a typo in the schematic pdf). The bias current we are aiming for is 100mA, so the voltage across the 0.1 ohm resistor should be 10mV (the schematics incorrectly says 0.1V across the Isense terminals, which would be 1A bias current). Instead of blindly following instructions, I should have done the calculations (100mA times 0.1 ohm == 10mV or 0.01V, not 0.1V). Posting for the benefit of future builders.

73

Ram VU2JXN