Linear vs Switching Power Supply

[WB9YBM]

I run Astron switching "SS" series power supplies here. No noise on HF-800 MHz. If you buy a cheap china one from Amazon, expect cheap filtering. Buy a name brand and spend a bit more and you will be happier.

"cheap" doesn't always have anything to do with how clean it is; rather if the power supply switches at the same time that the sine-wave on the input (i.e. pre-rectification/filtering) as is crosses the 0V point, the supply will have less "hash" on the output. At least, that's what I've noticed on supplies coming across my test bench.

 

[jonwienke]

Switching power supplies typically switch in the tens or hundreds of kilohertz to keep ripple low without requiring hideously large output filter inductors and capacitors. How does one sync a 50 KHz switch frequency to a 60 Hz input sine wave?

 

[WB9YBM]

Switching power supplies typically switch in the tens or hundreds of kilohertz to keep ripple low without requiring hideously large output filter inductors and capacitors. How does one sync a 50 KHz switch frequency to a 60 Hz input sine wave?

Rough guess: at 3 KHz (i.e. 50 Hz x 60 Hz) (I say "rough guess" because it's late at night and I'm half asleep )

 

[W5lz]

Typically, if a supply of either sort has been on the market for a couple of months, and if it has noise problems, and if you monitor any of the ham sites, you will hear about them. I've had at least one of almost every supplier of power supplies over the years. The ones I've had the most trouble with have been one of the 'name-brand's. Wasn't difficult to cure, but it was a pain for a while. I happen to have had a few 'off brand' supplies (switching type) that have done as well as any supply I've ever heard of. Any brand power supply can have problems, big name or ones unheard of before.

 

[jonwienke]

Rough guess: at 3 KHz (i.e. 50 Hz x 60 Hz) (I say "rough guess" because it's late at night and I'm half asleep )

It's a trick question. The correct answer is that there isn't any meaningful way to correlate a switching power supply's switching so that switch events correlate with zero crossings of the input waveform. The switch frequency, whatever it is, has to be pulse width modulated based on load conditions, not on whether the AC input waveform is crossing zero. Post 21 is incorrect.

Monitoring zero crossing events IS important when trying to connect an AC generator to an AC power bus, to ensure the connection does not occur when the generator output is 180 degrees out of phase with the power bus. That tends to be hard on the generator and the contactor. Examples include a home solar array connecting to utility power, to reduce or eliminate utility power consumption. The solar array has to phase and frequency match utility power, or connecting the two will let the magic smoke out of something.

 

[R8000]

"cheap" doesn't always have anything to do with how clean it is; rather if the power supply switches at the same time that the sine-wave on the input (i.e. pre-rectification/filtering) as is crosses the 0V point, the supply will have less "hash" on the output. At least, that's what I've noticed on supplies coming across my test bench.

What I experienced in our staging area and test lab over the years, you get what you pay for. Switching supplies used by reputable companies have always given good service and worth every penny.

 

[prcguy]

I've used a number of similar 12V server power supplies from Dell, Cisco, etc, and they all make various amounts of RFI within the HF bands but they are usually not so bad at VHF and UHF. Sometimes its the AC input side that needs some ferrite or an AC line filter and other times the 12V output can be tamed with ferrite.

I have a home brew 1.2kW HF amplifier that has an Elpac brand 3,000W server power supply (53V, 57A!) that trashed the HF bands and I've almost got it quieted down after a month of experimenting with ferrites and bypass caps on its input and output lines. When I'm done the filtering will be almost as large as the power supply itself.

I used one of these types with no discernable problems.

Dell Poweredge Server Power Supply Rack D670P-S0 0MY064 670W Case Computer | eBay

Max Output - 670W. Pulled from Working Poweredge Server. Didn't find what you were looking for?.

www.ebay.com

 

[WB9YBM]

What I experienced in our staging area and test lab over the years, you get what you pay for. Switching supplies used by reputable companies have always given good service and worth every penny.

I agree, although I'd like to add that especially with top-end equipment a percentage of what you're paying is for the privilege of having that "big name" on the front panel.