Mike-Delta wrote: ↑
Fri Aug 10, 2018 7:27 pm
Thank you for your answer.
I was surprised why the Mosfet works first and suddenly dieed without reason. But yes, that seems to happen more often.
Do you think it would be better to remove the directional coupler if I do not need it?
The output power was fine in my opinion. ~ 0.8W in 100W out. Which terminating resistor do you mean? Thanks for the other tips. But there are no obvious reasons for the loss of the Mosfet?
I do not believe that self oscilation was the reason but I had no obvious explanation.
I was driving it with an yaesu handheld radio on 144 MHz.
I will drive it next time with the exciter on the radio band.
Sorry for any grammatical mistakes. I am no native speaker.
I'm sure your English is far better than my attempts at your language would be - unless you're Dutch
Your observations are about right, I'd say. The amps I've built with MRF9180 easily achieve 150W with around 1.1W of drive in band II. I just found that the transistors supplied with the Chinese kits are fragile. When I first discovered these circulating on eBay I bought a couple of each to look at the design and to work out how they were doing it so cheap. It's a while ago now but here are my experiences as far as I can remember.
First board I built was a 100W effort which came with an XRF186. Worked OK and I was easily getting 100W with a couple of Watts in. FET was robust. I managed to crank it up to 120W very briefly. It's still alive although I don't use the board, I wasn't happy with the harmonics performance even after sorting out the filter response, but I'd never intended to put the thing on air. I was impressed enough to buy a couple of extra second hand Chinese MRF186's though, to use in an improved version of the amp.
On the next 100W board I popped an MRF186 by overdriving it, trying to get a feel for the device parameters on Band II (data sheet is obviously geared to 800MHz). I was using a driver board with an RD06 and was getting careless with the twiddler - I was being lazy when I should have been using a pad instead of trusting my instincts with too many parameters in play. Don't know if a "proper" MRF186 would have survived this, although I suspect the XRF186 in the first board might have done.
So, obvious precaution. The 186 in particular doesn't like being overdriven. I forgot to ask also, how much quiescent bias current you were using but from what you say, it doesn't sound like you were overbiasing and it looks like the 9180 can take a fair Idq anyway (1400mA?). This is the quoted figure from their tests on UHF though and you may find you need less. I wasn't using anywhere near 1400mA as I recall, and gain/harmonics were as expected.
The first board I built with the MRF9180 ran great for a few seconds and then died. Subtle fault, and completely invisible to the eye. What had happened was, the solid outer of the 50Ω cable had sheared all the way around. A tiny gap meaning the outer wasn't connected to the board at one end. I'd probably been over zealous with the wire strippers and bending it into the loop had finished it off. Obviously therefore, a serious mismatch killing the FET. Others have reported the newer Freescale MRF9180s standing up to this kind of abuse, but I expect that without realising there was a problem, they would have died eventually - perhaps not as quickly as mine. I killed one of my spare Chinese MRF186's before I diagnosed the fault too, it died even quicker than the 9180. Lesson learned - check the output transformers and matching carefully with a multimeter even if they look fine!
By the way, every time I popped a FET, one side died.
There's no harm in the directional coupler. If you don't need it, you don't have to populate the components on the coupled lines (diodes, capacitors and resistors). It won't affect the amplifier performance - as far as the output of the amplifier is concerned it's just going through a small 50Ω microstrip transmission line. For this type of PCB, the dimensions of that central line look about right (around 3mm width for 50Ω). The terminating resistors I was talking about are the resistors to ground at one end of each coupled line (the tracks running parallel to the microstrip carrying the RF output). Each coupled line (forward and reverse) can be considered part of a 2 port network with the main transmission line. Optimal power transfer to the sensing network will occur when the line is terminated with its characteristic impedance, which, in the case of my board, the designer had assumed to be 50Ω. I assumed not, as the tracks were somewhat narrow, and I obtained more output from the coupler by increasing the termination resistance to 150 if memory serves. This allowed me to drive my control board properly (yes, I admit I built a rig with one of these boards
). All kinds of factors affect not only microstrip, but this kind of coupler circuit so if you do want to use the coupler, you might find you can optimise it by experimenting and the value of that terminating resistor is significant, so it's a good place to start.
As I said, I've never had one of these boards self oscillate under any circumstances, but it's easy enough to spot (still getting power out after you drop carrier). I'd be surprised if you're getting self oscillation using a good 2m handheld and assuming a proper patch lead, connectors etc. Your construction looked spot on!