Linear PA power supplies

[Wendymott]

Hi Peeps. Just a quick question. To get a reasonably stable PSU with good current capability are Switch mode PSU's a solution?
Obviously not to be used on receive, but in TX mode is it permissible?Hhas anyone used this type of PSU?

In receive mode the PSU would need to be shut down, dunno if anyone has looked at this.

I would add this would be for solid state PA's, not valve.

[Tyso_Bl]

Yep, they have to be well filtered to stop RF from interfering with their regulation, looking the otjer way, any switching noise needs to be kept out of the transmitter, otherwise your PA will radiate amplified and cross modded signals tohether with the intended output.

I've succesfully used a stack of three seies connected computer atx supplies to give me 15v at 75A, worked om receive too, builtinto an old metal screened PC case.

[Radio Wrangler]

Switchers for linear power amplifiers can be fine. There's one by my right elbow.

RFI proofing is an obvious need, but less obvious is that linear amps running SSB and CW take current in surges/pulses. The control loop dynamics of the switcher have to be able to handle this without dips or worse overshoots. It can be dne, but not all switchers are well enough designed.

David

[Andrew B]

I suppose what you mean by solid state PA's. Modern MOSFET's can run quite high voltages, say 60v at 4A would be easier with a conventional supply.

[Wendymott]

Thanks guys....It was just an idea to reduce the weight of a high capacity PSU.. yes Andrew... it will be mosfet. Im doing "baby steps" at the moment..sounding out what is permissible.
I have a torroidal Transformer but quite heavy, plus the smoothing caps and series pass transistors.
I will post again later.

[kalee20]

Like Radio Wrangler says, the transient response of switchers (the time it takes them to recover from a change in load, to re-establish their stable output voltage) may be an issue. It can never be as good as a linear - a switcher's control circuit can do nothing at all till it's time for the next switching pulse!

But, if you have a representative PSU, try it, while scoping the output. You may find that a few big electrolytics across the output are all that is needed.

[BottleMan]

I've used off-the-shelf SMPUs in audio amplifiers for some years. Mostly enclosed types of about 150/200 watts at 48V. You get PFC thrown in as well courtesy of CE regs. I've paid between £60 and £80. Never had any trouble with stability due to rapid load changes, but I have selected those that can operate happily on no load. Hope this helps.

Graham

[Radio Wrangler]

Many years ago Linn went over to SMPS for their power amplifiers simply to have PFC to give active cancellation of the harmonic currents created by rectifiers/reservoirs.

Done well, they are great things.

David

[Wendymott]

Im not looking at "super" power..... maybe up to 100W PEP..... I would rather concentrate on a multiband version, after I have ironed out the bugs in this one. I will leave the "Shouty" signals to the "I" ones...

[Argus25]

Originally Posted by Wendymott

Thanks guys....It was just an idea to reduce the weight of a high capacity PSU.

That is a reasonable idea. Many "off the shelf" switch-mode PSU's do have objectionable high frequency noise on the DC outputs in the mV or 10's of mV range, but it is easily ameliorated with a choke in series with their output. Something simple like 20 or 30 turns of 0.7mm dia enameled wire wound on a 3/4 inch dia dust iron toroid, then a filter electrolytic of a few thousand uF and a 0.1uF ceramic cap. That way you can knock the noise off the DC voltage source.

However the switch-mode psu itself radiates electromagnetic signals from its transformers and even its circuit tracks & wiring. So it helps if it is in a well shielded enclosure. In fact most of the psu's radiation happens this way. This is why if you put a medium wave radio within a few feet of a computer monitor, TV plasma/lcd panel or computer psu, this blasts the MW and HF bands with noise & buzz.

So it is a better proposition in fact to use a switch-mode psu for a transmitter power supply, if you add additional filters to its DC output. And perhaps a less satisfactory proposition to use one to power a sensitive HF radio receiver.Hence the saying that a lot of radio operators won't have a switch-mode psu in their shack.

Any digital switching is not good for a sensitive receiver because the switching transients have Fourier components all across the HF band.

Communications radios with digital tuning systems are notorious for producing birdies on the bands they tune.

[Wendymott]

I was given a 5 channel audio amp yesterday with a whopping "C" core transformer. When I figure out the windings I will dump the SMP's idea and join a body building club.

Thanks for all your comments.

[Andrew B]

Most of these SMPS chips have an "inhibit" function, could you use that ?

Press the microphone to to "enable", releasing the key switches driver to inhibit?

[G6Tanuki]

There's nothing wrong with using a SMPS to drive a SSB transmitter/amplifier: I use a 24V 3A "Mascot"-brand SMPS to power my Clansman PRC320 on both transmit and receive - putting a 'scope on the supply rails shows no significant over/undershoot even when 'talking it up' to full-aggression (the PRC320 has a Plessey SL600-series VOGAD speech-processor built in: it's well-known for driving the PA transistors rather hard).

I've had no issues with RF getting into the PSU's regulator - but there again the "Mascot" SMPS were regularly specified for use in first-generation cellphone base-stations so come with "DC-to-light" input/output filtering and a nice all-enclosing shielding case.

[Radio Wrangler]

Originally Posted by Argus25

Any digital switching is not good for a sensitive receiver because the switching transients have Fourier components all across the HF band.

I have an Icom IC765 HF radio about 27 years old and even in that era they did a switch-mode power supply that was carefully enough screened not to cause any discernible problems. It's a very good radio indeed. I also have its modern DSP-based grandson, the IC7700 and it has no problems with the receiver, but its spectrum display does have a fixed-offet spur. Again it's all powered from an SMPS.

I have a couple of HP spectrum analysers with SMPS, and they can't afford to have SPMS spurs detectable, or else people will be chasing them when trying to debug other equipment.

It could be argued that the Agilent Noise Figure Analyser is the most sensitive receiver of all, it measures the noisiness of RF devices, amplifiers or full receiving systems right down to the thermal noise floor. Yup, there's an SMPS running in it.

They can be used in sensitive gear, and they have been, but it does take effort to get it right.

David

[BottleMan]

I'm glad to hear that SMPSUs have not been universally rubbished here. Especially as I spent most of my professional life associated with them. As with most products there are good and bad.
However, slightly OT, I decided to design a pre-amplifier for a moving coil (sub mV) cartridge for use with HI-FI (snake oil territory). For this I used off-the-shelf DC to DC converters with an SMPS Plug top unit. This really is a giant no-no for that fraternity. Do you know what? No one has noticed so far and it has received glowing private reviews.

I agree, a good SMPS can be a lot better than a badly designed linear one (all that diode switching noise). We (global we) seem to be fixated on the cheap SMPS without any RFI suppression components (to save a few pence), they are rubbish.

So Wendy, a good SMPS will do for both TX and RX, I have a Mydel 10A 13.8V job, can't hear it on any of my rigs.

[Wendymott]

Hyas Merlin..... Please can you send me a link or where to look for Mytel .....The internal Output module is now supplied by a Linear PSU, with a LM350K (3A) regulator. But I still need a good PSU for a bit of "Umph" in the external PA.

[G6Tanuki]

"Mydel" gear is available from several of the generic ham-radio stores.

Not sure if they do 24V models though ~ I'm only familiar with their 12V ones.

[Wendymott]

No wonder I couldn't find any info.... "Typo" LOL... Grrr THanks guys.