PSU Rating Questions

[reverend]

Question for you Necks...

If I am running 100W using an MRF101 at 48 Volts. Assuming it's running at about 50% efficiency then I only need a 4A PSU, though it would be running flat-out, and a bit more would be useful in-hand so maybe a 5A PSU.

I currently have a 10 Amp PSU which I feel is a bit of overkill.

Here are the questions...

1. What rating of 48V PSU do you use to power a 100 Watt transmitter
2. What device do you use to step this down to 12 Volts to power the exciter (noting that the LM2596 has a maximum input voltage of 40V)

Thanks in advance,

Rev

[EFR]

6A if FM rig and SMPS, if just old regular transformer, little bit more, depending about regulation on 48V line.

For 12V I just use second smaller PSU, or just some random transistor and zener from junkbox.

[shuffy]

Generally agree with EFR, here are some of my experiences with the "50V PA, everything else is lower" scenario to back it up.

1. Buck converter(s) from 50V for exciter etc. As you say, most of the cheap converters available on Aliexpress etc are OK but can't handle the 50V in. I tried the "HW-636" which certainly used to be based on the LT3800 and worked OK, stepping down from 48V to 12V at a couple of hundred mA, but needed additional output filtering. I've just had a look and there's now a HW-636B which looks like a significantly different circuit which I've no experience of. Generally however, people don't seem to like these as there's no protection on the board and the smoothing caps are underrated and are known to fail, but the use cases I saw were "asking for it"...

2. I've built portable MRF101 rigs where the thing runs from 24V and the PA is driven using a 400W boost converter (step up) with the rest of it driven by the ubiquitous LM2596 buck converters (step down). Works OK for a few hours at a time but the boost converter needs additional heatsinking/fan.

3. Separate PSU (usually cheap Chinese switch mode with additional filtering) for the exciter. You could do this linear but more expensive.

The PAs I build these days, if I ever build a rig that is, all want 50V and the exciter supply voltages are getting progressively lower - the one I use on larger designs, for example, wants 7.2V at up to 1.4A. I therefore use a separate supply. For your MRF101, if you're using a 1W exciter wanting 12V at say 200mA, then the linear voltage dropper with Zener and a suitably rated transistor will be fine.

IIRC these MRF101a amps are quite efficient on VHF and draw a little over 3A for 110W or so out at 48V. I usually use a PSU rated at double the required input.

I tend to try and balance size, efficiency, cost etc when coming up with arrangements for powering this sort of gear. Not ideal but if you need to be prepared to lose rigs then cost has to be a factor, so usually I've got a SMPS plus some form of boost or buck converter in the mix somewhere. Most of the popular ones are OK if you don't push them too hard but there are always problems to overcome - noise, poor quality components, poor layout etc. There's usually plenty of evaluation needs doing when a new module ends up on my desk.

[King Croccy]

For a 100w tx I would probably go with a 250W smps, The manufacturers usually state to de-rate the output for continuous usage. Generally you would run @ 80% of its output.

So 250W x 0.8 = 200W

200W x 0.5 (50% Efficiency) = 100W (for P.A)

In reality a modern day device like the mrf101 is going to have a much higher efficiency than 50%. It will probably be in the 70-80% region so you will have plenty of headroom. I've seen most guys running them at 120W but if you run 100W by lowering the smps voltage you will most likely get better efficiency.

With regard to the exciter power... there's two ways I'd go (and it usually depends on the circumstances and who you are selling it too ). If I'm setting the rig up myself, I usually go with a programmable zener and pass transistor bolted onto the same p.a heatsink. I knock the 48v down to 28V with this device then feed into lm317 regulators for 15v out (exciter power). It does make the heatsink run a little bit warmer but if you have low swr, it really wont matter. Also it will be a lot better than using "buck boost" convertors.

If you're selling to a "n.e. guy" I would probably go with the 2nd psu as you have no idea what swr he will have and you need to take into account that it is likely that the rig may run hotter due to the additional swr that may need to be dissipated!

[jvok]

I've measured about 3.3a at 48v when running near the bottom of the band. Enigma recommend this supply for their mrf101 module but 200w might be cutting it a bit fine at the top of the band, maybe 250w would be better. https://enigma-shop.com/products/single ... ply-module

LM2575HV modules will work on 48v, Enigma sell a couple of cheap ones for 12v or 15v out. Tbh I've only ever used them for fans, I usually have a transformer/linear reg for drivers and audio/rds stages. https://enigma-shop.com/products/low-no ... 4100069666

For higher power 300w+ I found its cheaper to put a couple of 24v 300w smps in series than a single big supply. Then you can take 24v off the centre tap through a dropping resistor to a 7812/7815 for drivers

[reverend]

Thanks for all the information. Sounds like something in the range of 5A/6A is the way to go then.

For the 12V supply, using a simple zener/transistor arrangement, and assuming around 200mA of current consumption for the driver (and more for the stereocoder etc...) the transistor is going to be generating around 8W of heat which is what I was trying to avoid. Point taken on the noisiness of SM buck converters - they always need extra filtering. Particularly like JVOK's idea of 2 x 24V psu in series as long as the SMPS concerned don't try and ground the outputs.

Definitely some food for thought - much appreciated.

[Albert H]

One other thing to consider is filtering.the outgoing eails. The cheao PSUs from China often produce plenty of "ripple". With linear supplies, used to add 1000µF per Amp drawn. It's also worth considering the fact that some of the ripple will be at a few hundred kilohertz, so could add to the noise mixed into the carrier.....

[jvok]

One thing to think about when talking about de-rating is how long does it really need to last? At the day job I'll use generous safety margins on component ratings, but thats because we're making serious industrial gear thats expected to run 24/7 for a decade plus. Meanwhile a pirate rig is lucky to make it 6 months before the bods take it (or it gets nicked). Putting in a 300w supply when you only needed 200w is throwing money down the drain at that point.

[reverend]

In the end, for the 12V/15V feed for the driver, I got hold of a batch of LM2596HVS buck converters, which were 5 for 10 quid, set the output at about 18V and then used a 7812 or 7815 (which I have an almost unlimited bucket of). All runs nice and cool, doesn't need enormous filter caps, and is clean too! I could probably have just used the buck converters and put a big cap on the output but having done this in the past there's still too much crap on the output for the driver to be clean enough.