Power supply testing q's...

[GibsonGM]

Ok now I've done it.  Computer PC speaker amp hack, using a tillman as a preamp, sounds really good with a strat 

But the power supply now has a nasty 120Hz buzz because I over-cranked it for too long...the LED on the power amp module was dimming as I jammed.  Amp module sounds fine with another PSU, so it's something in the stock one.

12V, 1A Wallwart, ~16v unloaded, non-switching type, the 4 diodes check out ok. The 2200u / 25v cap shows a steadily climbing resistance as it charges thru meter.  What other things can one do to test these things?    What dies when you ask for it to supply more current than it should - does the trafo start to short? That's all I can come up with since diodes/caps look ok.  Would that affect the 'not filtering' aspect, tho? 

Maybe just pop a new cap in there and see what happens...this is more curiosity than need, I'll probably just find another one in the junk box...I thought it was interesting tho.

[R.G.]

Is the 120hz hum constant, or does it get worse when you use more current?

If it were mine, I'd replace both cap and all diodes as opposed to spending debug time, but yes, it's an interesting problem. If you have an o'scope, look at the output voltage as you load it. If you get increasing ripple with load, it's probably the cap going high resistance internally. If the ripple is mostly 16mS between peaks, a diode has opened or gone high resistance. If the voltage just sags a lot, it could be that a thermal fuse in the transformer has started to open and only got as far as high resistance. Could be something related to ground traces, but that is a darker horse.

[GibsonGM]

Thanks R.G., things I wouldn't have thought about!   It does the same buzz regardless of what's happening.   I'd estimate the thing to be about 20 yrs old anyway, so maybe the cap has had it.  Couldn't take 'going to 11'.   

I'll play with it a little more, and if a quick fix isn't happening then will just get a new one.   I get curious about how things fail under stress - didn't know diodes could 'go high resistance'.

[anotherjim]

Some PSU's have a protection device buried in with the transformer windings. These can be one-time fuses 
Maybe some have polyfuse. Those have uncertain reset times. Did you let it cool right down? Might be hotter than you think where it's buried (if there is one).

[mozz]

 I don't know about a resettable fuse, I've seen thermal fuses in there. Maybe a diode is shorted? How much ac ripple do you have?

[GibsonGM]

Hey Jim...yeah, I've seen those fuses (and have 'bypassed' them a time or 2, lol)...don't think anything 'went' on this one. It's outputting ~16VDC unloaded (with a lot of ripple obviously)...the amp WORKS using it, just very very noisy.   

Mozz: I haven't measured the ripple...I have a Tektronix CRO from 1982, 60MHz...is it as easy as just applying a resistive load and 'zooming' on the AC riding on the DC?

I ordered up some 2200u/25V caps for the junk box and to pop in this PS brick, kind of has to be that or the diodes, no? Tho each diode did read .5Vf (I will check reverse resistance next...)

Just in case: Replacement diodes on a 1A rated supply should be 1.5A, right?

[R.G.]

Just in case: Replacement diodes on a 1A rated supply should be 1.5A, right?

Not needed for full wave bridge. See https://www.hammfg.com/electronics/transformers/rectifier in the full wave bridge, capacitor input model. 1A diodes are OK for a 1A rated dc output. The transformer secondary RMS is however 1.6 to 1.8 times the DC current out, depending on the filter capacitor's size compared to the load current.

[ElectricDruid]

You said it's not a switching supply, right? so it's got a transformer in there?

It sounds to me like physical noise from the transformer. There's a powerful magnetic field in there that's getting power from one set of windings to the other. On old transformers, the laminated core can get loose and vibrate in the field. It's typically worse with more current drawn, and you can set it off by really hammering the poor tranny, but once they're loose, they're loose and it'll do it all the time at any load. I had an old benchtop power supply that I built out of junkbox parts that was like this. In the end, I had to sling it because the buzz was awful.

[GibsonGM]

Thanks again R.G....the diodes appear to be something a bit larger than 1N4001...4002 size perhaps.  Number is obscured at the moment. Full wave, discrete diodes, non-switching so just tranny, diodes, 1 cap.   There are more filter caps on the board.

Tom, that is interesting. I think I may have *heard* of that physical noise phenomenon, never experienced it in audio work.   This thing buzzes ALL the time, the same 'level' if you will, no matter what you do to the amp (quiet, loud, with the tillman input section disconnected...).    The wallwart itself has no 'buzz' or hum going on.    I HAVE heard one doing that before, tho it wasn't something used for audio.

I think it's worth changing the cap - easy to get to - and if that doesn't fix it, it gets replaced.   I'm just really curious in what ways increasing the input signal caused it to fail, rather than just sag and making the amp module distort like crazy...it was TRYING to keep up. I guess there is not a lot of bomb-proofing built into something like this - it was designed on that 'thin line' and if you go much past it, it smokes.   I could just get a 2A supply and use a regulator in there

The only reason I'm messing with this is for mic'd amp recording at low volume/pushed amp tone.  It sounds pretty good really.

[Rob Strand]

I like RG's idea.  Replace the bridge and cap and see if it fixes it.  Then take it from there.

Having the diodes packed deeply inside a wall wart certainly doesn't help them keep cool.  If they cook AC would get through giving a buzz.   It would also stress the transformer in that state so don't leave it on while testing.

A 12VA transformer is hard to blow-up but it's not impossible.  You might even see low voltage if it's got a thermal fuse which is half  fried.

[GibsonGM]

Something like what I'm doing might work better if I add another FET stage to the 'preamp' for drive, cut some low end, and avoid diming the thing to get that nice 'power amp distorion'     I think too much low end and the move from PC speakers (3") to an 8" guitar speaker may have pushed things too far...

[MrStab]

120Hz... could movement from the speaker have slightly loosened a transformer which was quite prone to vibrating to begin with, from either its own shielding or from the PC speaker chassis?

[GibsonGM]

120Hz... could movement from the speaker have slightly loosened a transformer which was quite prone to vibrating to begin with, from either its own shielding or from the PC speaker chassis?

That would be a good guess Grant, sure, but it's a wall-wart, so it's not getting hit by the vibes.  I do have a Hot Rod Deluxe doing the 'crackle' thing a little, and that may well be related to vibration!

[Rob Strand]

Something like what I'm doing might work better if I add another FET stage to the 'preamp' for drive, cut some low end, and avoid diming the thing to get that nice 'power amp distorion'     I think too much low end and the move from PC speakers (3") to an 8" guitar speaker may have pushed things too far...

The only way to save the supply is to limit the amplifier output.    You can shave off some output power with series diodes to the power amp and (power) resistors in series with the speaker.    You trade that off against loss of output power.

You could just try a 12V 2A+ switchmode from an old router.   (Watchout though the DC jack of these things are usually reversed from effects pedal adaptors.)

[GibsonGM]

Thanks for the tip, Rob. That was bothering me - 'fix it, then do it again...'.    First thought was add another FET preamp section, and set up some preamp distortion with it to sort of replace hitting power amp too hard.   Then set the output of the preamp to the power amp so as to be in a 'safe' region, and use a 'master' pot with limiting resistors to make that the max you can crank it.

If one increases the ampacity of the power supply, what's the 'down side'?  Heat? Frying the chip amp?

[antonis]

If one increases the ampacity of the power supply, what's the 'down side'?  Heat? Frying the chip amp?

Nothing of these..

Current is drawn according to demand..
On the contrary, bigger ampacity ensures less voltage sag..

Or am I missing something concering your particular query, Sir..??

[GibsonGM]

If one increases the ampacity of the power supply, what's the 'down side'?  Heat? Frying the chip amp?

Nothing of these..

Current is drawn according to demand..
On the contrary, bigger ampacity ensures less voltage sag..

Or am I missing something concering your particular query, Sir..??

I am curious if, because I am boosting the INPUT of this modular power amp quite a bit...can I input so much that I cause it to overheat and destroy itself or some such thing, LOL?  I do know that there would be a maximum input voltage the amp can handle, I appear to have not exceeded that yet.  I may have exceeded the power the transformer can output.  I'm aware that current is drawn rather than 'given'.      Can I make the amp 'runaway' by too large an input signal and a power supply that is rated 2x what is called for...? 

It's just something I've never considered. 

[Rob Strand]

If one increases the ampacity of the power supply, what's the 'down side'?  Heat? Frying the chip amp?

Anything can happen.  Fried PSU or fried amp chips.

If you started with a clean sheet of paper you would choose an amp chip (which provides the power you need).   Then design the heat sink (or PCB heatsink) so it the chip wouldn't fry under abusive conditions (as defined in the application).   Then choose a PSU which will handle it.    Even then you have the option of having low power, or high power where the chips is right in the edge of temperature and current.     The more you push things the higher the risk of frying the amp chip.

When you have a fixed amp with given heatsink and PCB (something already built)  there will be a upper limit on the power supply voltage, which sets the output power, and will not fry the amp chip.   From that you can choose a power supply with a corresponding voltage and current rating.

The issue with a lot of audio amplifiers, especially "HiFi",  is they aren't designed to be run flat-chat for long periods of time.   For music tracks the sound would be so distorted at high levels that they sound bad and that makes people back off the level.     Designers of audio amps often take this into consideration to cut costs on component ratings, heatsinking and power supply ratings.  That comes about to some degree from the high peak to average levels in music.

Deliberately distorted guitar pushes things well beyond the normal "HiFi' bounds.    So prolonged abuse has a risk of taking something out.

So the way to fix an off the shelf product is to back-off the output power (as per prev post).   There's not many practical ways to do it.   Obviously you don't want to back the power off to the point where it's no better than a transistor radio.    However, bullet proofed designs against failure under prolonged worst case conditions often end-up with oversized designs which look far too large for what they are but engineering-wise that's what it has to be.

FWIW,  in industry you might see pumps, grinders, welders with a duty-cycle rating which means the units can only be run under full load for a specified time - typically to stop them from overheating.   If you are in a factory where the device had to operate continuously for weeks on end without failure then it would be a considerably larger device.    For home products you might see warnings about running certain devices more than 10 mins or 15 mins.

[GibsonGM]

Ah, ok. Duty cycle on a welder.  That's basically what the question was about. I can design a tube preamp and understand ground-up (still working on power tubes, but you get the drift).  Since this little thing is already built, was curious how abuse would be tolerated, esp with a supply that can output *way* more power than the design calls for.  Clearly...the tolerance is probably not much! ha ha.  Good point about 'backing off', yes, clearly music would sound horrible cranked way up and already boosted thru something like this.  It also has 'fat caps' - too much low end to compensate for the orig. tiny speakers - they gotta be downsized.

I think if doing something like this, maybe the best route is to get your 'saturation' off-board as much as possible, and let the power amp part do its thing without a lot of stress.   I'm shooting for a little cleanish but Strat-dirty thing to record with (if that makes sense...'strat-o-blaster' territory, twang, not metal).    Apparently using the power amp module for 'power amp distortion' might be a bad idea, ha ha! 

[Rob Strand]

I think if doing something like this, maybe the best route is to get your 'saturation' off-board as much as possible, and let the power amp part do its thing without a lot of stress.   I'm shooting for a little cleanish but Strat-dirty thing to record with (if that makes sense...'strat-o-blaster' territory, twang, not metal).    Apparently using the power amp module for 'power amp distortion' might be a bad idea, ha ha! 

If you can put a resistor in series with the speaker without losing tone then that would reduce all the ratings issues.   In fact if you can just use a 100 ohm or even 1k resistor then it would start to be more of a preamp.   There's some LM386 designs which pretty much do this.  The output cap would need to be tweaked so the LF roll-off stayed the same, or is set to the desired frequency.

Suppose that actually sounds OK.  Now you take that preamp signal and feed it into the original power amp and crank it to near full. The power amps probably going to get nearly as stressed since the idea is the same final output signal makes it's way to the speaker.

The biggest abuse saver would be to roll-of the low-end.  Either with the output cap or with a high-pass filter.    There's a lot of "abusive" energy in the low end and when you filter it off the stresses go down and the sound only loses a small amount of character in the scheme of things.