Page 11 of 14

Re: Thought I'd try one of these.....

Posted: Fri May 03, 2019 7:55 pm
by Albert H
The 9180 is still working with 4 Watts in and slight over-bias - normally this would be enough to kill the FET. It's delivering about 195 Watts out, with a nasty amount of third harmonic. It's been warming up the work bench for almost three days.

I'm looking forward to trying some of the newer devices - the latest generation look like they're almost indestructible!

Re: Thought I'd try one of these.....

Posted: Sat May 04, 2019 5:52 am
by Lubomir73
That's fantastic Albert. Three days... 195 watts out! That's almost 100 watts each half!!! I assume you are using both sides of a new, dual 9180 when your said 195 watts out. So you can teach an old dog new tricks. I feared that your schottky input diode might give out at this level. It's only rated at 200mw. And the excess power has to go somewhere, either to the diode or the FET.

Hey, third harmonic... Tune up your LPF to kill the 3rd harmonic.

May I please suggest that this device is really good for much more. If you want to continue the stress test, parallel a couple of diodes at each gate. Match them... Simply use a digital ohmmeter to check their forward resistance and choose accordingly. Run the bias so that you get Iq at least 700 ma, (just as the specs say), and then ramp the input up to whatever level you want . I think you may be surprised at what this device may be really capable of. Pay close attention to heatsinking. (nice big hunk of metal, sparing on the heat sink compound, let the two surfaces touch as much as they can, Keep a fan on it.) Next, you may be telling us that your have continuous 300 or 400 watts out. Hey... $9.00 USD in, 300 watts out, thats worth it to me even with an old device like this one.

Re: Thought I'd try one of these.....

Posted: Sun May 05, 2019 3:04 am
by Albert H
My PA isn't optimised at all. It's just a thrown-together junk of cheap Chinese rubbish.
There's now two Schottky Diodes paralleled at each gate to ground. It's still doing around 200 Watts out, and I've put another exciter into it at 101 MHz. The third is reduced since the filter is starting to have more effect at this frequency. It's still being deliberately overdriven a little. The excess drive is dissipated as heat in the FET.

LB delivered a strip of big FETs this afternoon, and I have the board layout on acetate ready to photo-etch later on today. He claims that they are good for 600 Watts and I intend to try out the gate protection idea. We need some cheap 500 Watt PAs in a hurry, so these might provide an ideal solution. They're so cheap now that the expensive parts are the heatsink and the power supply!

This is almost getting easy!

Re: Thought I'd try one of these.....

Posted: Sun May 05, 2019 4:13 am
by nrgkits.nz
Albert you could bolt it to the underside of a cold water tank again for heatsinking then that’s atleast one expensive part not needed anymore, then whoever replaced gotts’ will flood the top of the block again when they collect it.

Re: Thought I'd try one of these.....

Posted: Sun May 05, 2019 11:47 pm
by Albert H
I'm not going to be putting rigs up tower blocks any more!

Re: Thought I'd try one of these.....

Posted: Mon May 06, 2019 8:17 am
by Albert H
I spent several hours building my first PA with the new FETs. These are "house marked" devices, but are meant to be good for 600 Watts. Initially, I biased the pair until I got 450 mA each side, and gave the board 500mW of drive. The output was around 95 Watts and everything was stone cold. The output purity - without a lowpass filter, just with the match up to 50R - was pretty good. The second was down around -50dBc and third was about the same. Increasing the bias a bit increased the gain, and I found that the bias became increasingly touchy. The purity could be optimised by slightly offsetting the bias! This suggested that the FETs in the package I was trying had a slight imbalance!

I wound the bias up to about 900mA a side (as recommended in the spec sheet) and increased the drive to 1.5 Watts at 98 MHz. 400 Watts and -48dBc second. A tiny tweak on the bias of one of the sides got the output up by 60 Watts and the second harmonic down to -56dBc.

I was concerned that I'd cocked up the transformer, so powered everything off, and measured my coax lengths and checked for braid damage. In the end, I wondered if the FET was dodgy, so put another one in with a different date code.

I ran through the same tests and got similar results, but with the slightly increased bias in the other side of this one! I was glad that I had used 10-turn pots for the bias trim.

I messed around for some time getting the bias right, and eventually got a gain and unfiltered purity that was good. The PA was delivering exactly 500 Watts for 1.7 Watts of drive and the purity was very good. Adding the output filter lost me about 1.5 Watts, so a minor tweak to the drive got the power back to where it should be. I left it running for an hour or so while I had dinner, then checked it over again when everything was warmed - nothing got hot! The bias was offset by about 65mV when the gain and purity was optimised.

These devices are meant to be MMRF1016 types - anyone tried these? I was surprised at the bias offset, but eliminated board layout or component issues by trying out three devices - all with slightly different results!

Setting up the bias is really important, and tweaking for best gain against purity is probably the right way to set them up. I'm glad that my exciter can be tweaked very precisely, and has output levelling across the band!

Sweeping the frequency through the PA just confirmed the lowpass filter shape! The gain remained the same pretty much across the band with slight droop at each end. Gain is around 26dB and efficiency calculates at 72%. These things remind me of the old twin tetrode bottles!

Re: Thought I'd try one of these.....

Posted: Mon May 06, 2019 8:21 am
by Albert H
Has anyone else found side-to-side differences with power FETs?

Re: Thought I'd try one of these.....

Posted: Tue May 07, 2019 12:08 pm
by nrgkits.nz
I haven’t noticed any side to side differences with the devices I’ve been using, although I haven’t tried tweaking the bias on one side - I must give this a try and see if it improves the power out and harmonics. I still managed to achieve the power out and efficiency as stated in the data sheets.

Have you got a photo of your PA Albert? I’d be interested in how you did the LPF, did you connect it directly after the 50R output match on same PCB or connect it on a separate PCB with a short piece of 50R coax?

What are you using for a load also? I’m using a bunch of 100R RF resistors paralleled up on a large heatsink and sitting it in a tray of water under the bench but not covering the resistors - I’ve had 400w continuous into it without issues. I use RG213 from the PA to a telewave wattmeter then to the load, the wattmeter also couples a sniff of the RF into the analyser so I can measure harmonics.

Re: Thought I'd try one of these.....

Posted: Wed May 08, 2019 11:50 pm
by Albert H
I'm travelling at the moment - home on Sunday. I'll snap a couple of pictures. My dummy load is a commercial one, made in the 60s by a company called "Blyth". It's oil filled and is supposed to be good for 5kW in short bursts or 2kW continuously. I've used it up to 1800 Watts without problems. It has a 20dB coupler built in so getting signal for the analyser is easy.

The LPF is on a separate board with a tinplate box around it. It's connected to the PA with about 18cm of WF103. The output socket (N-type) mounts through the filter PCB, then through the main case.

I played around with the PA for a while before I left and decided to see just how far I could push it. I stopped at 850 Watts, since I was running out of drive. The mains transformer got quite toasty, so giving up was probably a good idea!

LB tried out one of these PAs and got similar results, with minute bias offsets for best performance! He tried one out on 2m SSB and said that he got the best two-tone results he's ever seen in a solid-state PA. He said it was like the old valved gear.

Re: Thought I'd try one of these.....

Posted: Thu Jun 27, 2019 7:51 am
by zeljko
Can someone tell me why can't see any .jpg on forum?

Re: Thought I'd try one of these.....

Posted: Fri Jun 28, 2019 3:06 pm
by BriansBrain
zeljko wrote: Thu Jun 27, 2019 7:51 am Can someone tell me why can't see any .jpg on forum?
Hi Mr.zeljko ...

You should be able to see the jpgs now you have joined the Forum :roll:

Re: Thought I'd try one of these.....

Posted: Fri Sep 06, 2019 11:18 pm
by Fgbokoiii
Hi kits,
Thanks of your reply helps a lot. Sorry for late reply...

Fgbokoiii

Re: Thought I'd try one of these.....

Posted: Mon Sep 09, 2019 10:56 pm
by Albert H
I played with a few more of the "house marked" dual FETs. I think that these may be rejects, since their gate parameters differ side-to-side! They're ridiculously cheap - the big costs are the output capacitors, power supply and heatsink arrangement.

A friend of mine has been making homebrew high voltage small value capacitors for VHF / UHF output stages. I'm looking forward to trying some of these out since they cost little to make and I'm running out of Arco capacitors!

I've also found a supply of "Telecomms" 48V power supplies. These are available up to 200A, so powering these new PAs will be easy since the PSUs are designed for use in electrically noisy environments.

My latest exciter design is truly minimalist - a low noise "power oscillator" going into a broadband buffer and delivering 500mW up to 5 W. Mod linearity is good, and frequency stability is almost good enough to leave out the PLL! The output purity is good enough to go straight into an aerial, but it's designed to drive the new PAs. I've eliminated the microphonic VCO coil and have replaced it with about 17.5 cm of RG174 coax. The PLL is my usual 5-chip cheap 74HC logic job with a 4MHz crystal. The whole thing fits into a tinplate box 15 X 7.5 X 3cm and runs on 18V. The PLL division ratio is set by solder-blobs between tracks. The mod sensitivity is 1V p-p for 75kHz deviation pretty much right across the band. We're entering a new period of really cheap rigs!

Re: Thought I'd try one of these.....

Posted: Tue Sep 10, 2019 2:43 am
by nrgkits.nz
I bought a bag of 3.2MHz crystals out of China which I think were also rejects as the frequency tolerance of each one was always out by a certain amount, much more than the expensive crystals I previously bought from DigiKey. However there were no issues as I included a “fine tune” adjustment trimmer on my design and each one aligned on my analyser using the counter function.

Re: Thought I'd try one of these.....

Posted: Wed Sep 11, 2019 6:48 pm
by Lubomir73
Wow, I thought this thread was dead a couple of months ago.
Guess not! We need to keep it going... when we get something new!
I am also amazed at the amount of information that i am learning going thought this and all the other forums.

Re: Thought I'd try one of these.....

Posted: Thu Sep 12, 2019 11:16 am
by nrgkits.nz
Here’s one of my PA’s, this one will do 150w quite easily, 200w max and uses the MRFE6VP5150N. Drive power is max 1.5w, and I’ve also deliberately over driven it upto 4 watts without any issues. It’s also been run into a dead short and survived.

This design is based on the data sheet design, however I had to recalculate the stripline traces used for matching the drains to the output balun, the data sheet design uses Rogers PCB 0.75mm thickness, I’ve redesigned it for standard FR4 1.6mm thickness, hence thicker atripline traces to achieve the same impedance. It’s easy to do this using AppCad and Excel with some formulas, then simulate it in SimSmith.

Without a low pass filter, the second harmonic is at -35dBc or thereabouts at 88MHz, and -40dBc at 108MHz.
46279530-EFE8-4CD1-A8D4-F1131C3285A4.jpeg

Re: Thought I'd try one of these.....

Posted: Thu Sep 12, 2019 12:09 pm
by thewisepranker
Nice.

Re: Thought I'd try one of these.....

Posted: Wed Apr 15, 2020 10:21 pm
by rigmo
Image

Re: Thought I'd try one of these.....

Posted: Wed Apr 15, 2020 10:31 pm
by rigmo
MRF9180 170W tuned with network analyser input 1.5W
protection : lighting, SWR double, overheating doable.

Re: Thought I'd try one of these.....

Posted: Thu Apr 16, 2020 5:43 am
by teckniqs
I'm confused, why does it say 42.1 on that screen? ....It can't be the amps and it certainly can't be the volts, so what is it?