MRF101 Band 2

[mikroman]

Suono from Italy use MRF300 AN and BN mosfets in there transmitters.
They run a single MRF300 on 250W and a double on 500W and it's runs cool for 24/7 dat in and out.
I have some MRF300 transmitters running also 24/7 and no problem
But overdrive or mismatch on the Gate will be a problem and will blow your MRF (never overdrive more then 1dB)

According to the service manual the low power transmitter lines 50-100W all use MRF300 AN but with 24v instead of 48v. I had some on my workbench but I didn't inspect the output stages. I will definitely take a closer look at it when it is on my desk again.

[Krakatoa]

So it seems these devices need to be way under run to get 24/7 reliability.
Maybe the 100W and 300W specs were thought to be achieved using the devices in pairs?

[Albert H]

EXACTLY!!! These devices just cannot get rid of the heat if they're run close to their rated output powers. They are a totally worthless design, and I predict that Mororola / NXP will drop them in favour of versions in sensible packages. The whole reason that they're cheap is because they made a few million of them before they realised that they wouldn't be thermally reliable, and they want to offload the stock so they can move to quality packaging.

In the meantime, I've been building a few 120 Watt PA boards with 2SC1971 into MRF317. They're utterly reliable, and the little batch (8 pieces) were shipped to a customer in Eastern Europe. The most difficult part of the construction was winding the output end coils (they have 16 swg wire). Each board was tested at 88 MHz, 98.4 MHz and 107.8 MHz, without adjustments. All of them provided exactly 120 Watts out into 50Ω and ran virtually cold. They have larger heatsinks than most people would use, because they're designed to be used without forced air cooling, and they're going to end up in a very hot and dusty country. They all exhibited the same harmonic response (no products worse than -68 dBc at any in-band frequency) and they remain stable into my "poor SWR" test loads up to an SWR of 2.5 : 1. The mains supplies to power them were more expensive than the rest of the PAs put together!

One of my colleagues was experimenting with an MRF101-based amplifier, on a board similar to the "Enigma" design. He tried three devices and managed to fry them all doing the same tests that I did with the MRF317 amps. We're not going to bother with those parts until they're re-released in a sensible package.

[yellowbeard]

Sounds like they'd be ideal for things like DAB where you run an amplifier at 5 or 10% of its rated power to keep it ultra linear. That might be an idea for any of you that have some and feel ripped off. All the development kits by the manufacturer come bolted to a really fat aluminium heat spreader so they certainly knew about the heat issues during development.

[EFR]

Hows about Ampleon BLF series? Price is about 0.20€ for an watt if using only 1000W out of BLF188X, quite rugged, better case desing, but still if you want run it on full power around the clock, it needs quite big copper heat spreader.

[Albert H]

The BLF188X is the way to do it. I would expect the need for a big heat spreader - I've always used one from about 100 Watts upwards. At ~1 kW, the '188 dissipates around 120W of heat, but has the package to do it.

The first RF Power Transistors of useful power I worked with were the TRW TP 9383 and its series. a pair of '9838s could do >300W at 28V supply, with about 30W of drive (these were bipolar transistors, not FETs). Again, TRW had built them into nice, big, rugged cases, with large flanges to get rid of the heat. The 300W pallet modules I developed in Bordeaux in the early 80s are still widely used, because they're rugged!

[EFR]

I have run 188 on HF and build couple 2m amps as order, but never on FM.

On HF, these cheap AliExpress BLF188X work like they should.