Anyone know how to find the switching frequency of a PSU?

Take a look at the gate/base of the power switching devices. I suggest doing this with a PSU not connected to the rest of the PC since the probe capacitance may alter the switching characteristic. You'll need to load the PSU with a few fans and fiddle the 'power on' pin.
You'll be able to see if anything goes awry with the voltage when you probe, if it's OK (I suspect it will be) you may want to reconnect a PC system as the switching frequency could be variable.
Do we assume this has something to do with audio?
 
weird_dave said:
Do we assume this has something to do with audio?
Click to expand...
Yep, I'm planning to power an an audio amp module with the PSU. The datasheet for it recommends a PSU with a high switching frequency.

I forgot to mention this is not a traditional PC PSU, I think it's designed for charging 24v batteries, which means turning it on with no PC is much easier. ;) Here's a pic:
22012010404.jpg


What you say base/gate of the switching devices, am I right in assuming it's the middle pin?
 
Lol, this really is general hardware :D
The middle legs are often the collector/drain. Really need to know what the devices are. Can you remove the clamps and read off the device numbers? Got circuit diagrams by any chance?
Failing that, probe each leg until you see a waveform. Also, you could put a hefty load on and try to spot the ripple, this would be the switching frequency :)

Edit:
Just had a thought, if it is for charging batteries, the output might be awful. You might be able to see ripple with no load.....
 
Last edited:
weird_dave said:
Edit:
Just had a thought, if it is for charging batteries, the output might be awful. You might be able to see ripple with no load.....
Click to expand...
There's only one little capacitor on secondary so you can bet that thing has some ripple.
 
weird_dave said:
Lol, this really is general hardware :D
The middle legs are often the collector/drain. Really need to know what the devices are. Can you remove the clamps and read off the device numbers? Got circuit diagrams by any chance?
Failing that, probe each leg until you see a waveform. Also, you could put a hefty load on and try to spot the ripple, this would be the switching frequency :)

Edit:
Just had a thought, if it is for charging batteries, the output might be awful. You might be able to see ripple with no load.....
Click to expand...
Ok, excuse my late reply but I've had to improvise a little here.

I tried it out with no load, and I could smell something was getting a bit too hot. Turns out R37 (the resistor by the outputs) was roasting hot. I now have a blister on my finger now to show for it! :o :D

So, I figure some sort of dummy load would be best, if only take a little heat away from the resistor... Enter the car bulbs!

Wired in series means they'll take up to 28v no problem. Here's my awesome wiring... ;)
28012010413.jpg


It does the job however, even if they are blindingly bright, even with 20v...
28012010412.jpg


Finally, with a dummy load the resistor now runs warm, the the PSU is doing some work. Bulbs pulling around 5A I think...

Anyway, I managed to get a few shots of the scope, probe in 1 hand, cameraphone in the other... The switching device turned out to be the biggest heatsinked IC.

28012010417.jpg


After all that, 66Khz... Awesome, much nicer than my PSU I'm using right now. Quite a funky looking waveform too, I was expecting a normal square wave.
EsaT said:
There's only one little capacitor on secondary so you can bet that thing has some ripple.
Click to expand...
Remember this is a SMPS, not a linear PSU, so the smoothing cap can be lower in value. 1000uF is fine and I'm getting around 10mV of ripple with 5 amps of load, which very low.
 
Hmm, that is a bit strange. I assume 0v is the middle line? Scope input on DC? The display isn't saying much, where's the timebase and v/div? x1 probe or x10 (and is the scope set for it)?
 
weird_dave said:
Hmm, that is a bit strange. I assume 0v is the middle line?
Click to expand...
Yep, I thought the wave was a bit of a strange shape.... 0v is the middle line.
weird_dave said:
Scope input on DC?
Click to expand...
When I had the scope probes on the DC outputs, I was getting the usual horizontal straight line. I should have mentioned this in my previous post... :o

I was trying to find the switching frequency when I took the photo of the scope, so in the pic I posted, I had it connected like this:
The black scope probe is on the 0v DC output on the PSU.
The red scope probe is on one of the outside legs of the switching device.
weird_dave said:
The display isn't saying much, where's the timebase and v/div? x1 probe or x10 (and is the scope set for it)?
Click to expand...
Unfortunately, I don't have a proper scope, they're not cheap, so I'm using a multimeter which has a scope function which is better than nothing. :) It only has 1 input and tends to automatically set the ranges which can be a blessing and a curse, tho I can tweak scales somewhat.

The scales are displayed down the right hand side,
The vertical scale is 29v per division,
The horizontal scale is 4.2µS per division.

I'm assuming the peaks are around 40v, and I guess when they are smoothed out, 20v on the output seems like a reasonable figure.
 
More like 50V peak. Long day, brain hurts!
Looks like it should do the trick anyway :)
You might be able to see the ripple with your meter if you put a large cap in line with the positive probe :)
 
Music at last! :cool:
28012010422.jpg

Very little ripple and very little change in voltage when the music dynamics change, I'm more than happy it's all up and running and it sounds fine with no noise or any other unexpected things which can be caused by a bad PSU.

The resistor (R37) gets a little hot (70ºC) when the volume is quiet, as there is very little load, but I plan to replace this resistor with a higher wattage one, which should solve this problem.
weird_dave said:
No probs, it's nice to see a bit of electronics outside work :)
Click to expand...
It's only a hobby for me right now, but I do enjoy it and I'm learning new things all the time. :)
 
Excuse my late reply, this thread dropped off the bottom of my userCP.
weird_dave said:
That resistor should be fine as it is.
Click to expand...
I switched it out for a 25W resistor in the end, I don't like things running that hot... the 25W resistor still gets warm but not hot unless I take the output voltage above 30v.

There was plenty of room for it also:
30012010426.jpg

weird_dave said:
Interesting audio PCB there.....
Click to expand...
I don't think Sure are a competitor, here's a page with more info:

http://www.sureelectronics.net/goods.php?id=1044

It's a very nice sounding amp, but it sounds a lot nicer with a few upgraded parts:
eeeedddiiiiiittttt.jpg


If now sounds better than some amps costing ten times as much... A good PSU really helps the sound as well, tho it working surprisingly well from batteries. :)
 
Last edited:
I would expect batteries to be quite good, no switching noise :)
I have to admit, I didn't realise class D had come down in price so much, I'll have to start looking at them again :)
 
weird_dave said:
I would expect batteries to be quite good, no switching noise :)
I have to admit, I didn't realise class D had come down in price so much, I'll have to start looking at them again :)
Click to expand...
Again, apologies for the belated reply, I've been busy trying to finish the case and connections, it always takes longer than expected... :o

Batteries are better in that respect, they're a very clean source of power. I do find the internal resistance can let them down a little, I notice the voltage can start to drop when the current goes up. This is not normally too much of a problem if a decent capacity battery is used however.

To save taking this thread any more off topic, I thought I'd post more info about the amp in the DIY audio thread, linky: http://forums.overclockers.co.uk/showpost.php?p=15942877&postcount=1316 :)

In short, any class D amp with a Tripath IC has a very nice sound, it's a shame the company went under but they IC's are still being produced in the fast east it seems! ;)
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom