Faulty ATX PSU in Dr. Dre Beatbox - any suggestions as to cause. (Not ATX)

[Old Sparky]

Having identified several faults in this PSU, replacingl parts as requitred, including several electrolytic capacitors and a s/c schotky rectifier diode in part of the circuit which supplies + & -7v DC to the amplifier board, I powered up the PSU only to find that there is no DC at the point which supplies an estimated 50v (The smoothing caps. on this point are 470u/63v) to the amp. I'm  puzzled.  This DC supply is derived from a transformer fed from the incoming AC mains  via a bridge rectifier and two smoothing caps.  There is ca. 340vDC across these caps., but the voltage being fed to the aforementioned transformer fall to very low levels (<3 or 4v) after less than a minute.  THe associated transistors checkout OK 'cold', but I guess they could be breaking down in use.  Other parts, including the transformer in question check out OK.  No spares or service info. seems to be available for the 'Beatbox', which doesn't help.  I've traced out most of the circuit, which uses a TL494 DIL IC, but haven't drawn it out in a suitable form to post.  Maybe I should just replace the transistors in the primary circuit and see what happens?

[Nick]

The TL494 is a very very common smps regulator. It's quite possible that the PSU goes into shutdown as a failsafe because one of the output rails is out of band.

Note that many smps units don't do well without a load.

[Old Sparky]

Fair comment, Nick, but the (presumed) 50vDC in this PSU should be present when mains is applied, so that the associated power amplifier PCB, which also contains the switcing line, is powered. From there, via a momentary switch, the PSU is controlled, and outputs +7 & -7v to the (as far as I can tell) the preamplifier, which is on the same PCB as the power amp. (That's how I understand the circuit anyway) I don't have any bench PSU which goes above 24v, but can use that as a substitute for the main DC supply, but will have to make up or buy a dual rail 7v supply to replace the other one. This will prove if the amplifier works, at least! BTW, I have a long running thread over on UKVRRRDF about this problem, and included in it is a link to a similar ATX PSU.

[Alistair D]

http://www.danyk.cz/s_atx_en.html

This is a site with various computer power supply schematics. You might be lucky and find something similar to yours.

Al

[Mark Hennessy]

If this unit really is an ATX power supply, why on earth are you trying to fix it? ATX power supplies can be found for less than £20.

But no ATX power supply that I'm aware of provides 50V or +/-7V, so I assume that it's not an ATX power supply at all. Please clarify this, because this is very misleading. I will remove "ATX" from the thread title if appropriate.

Please also do us the courtesy of linking to the thread on UKVRRR. Not all of us are members there, and not all of us want to have to search for it. Before any of us here respond, we do need to see what you've already been told so that we're not wasting our time repeating advice.

In other words, help us to help you.

As a rule, putting a voltmeter across the primary of the transformer in a switched-mode power supply is a rather dangerous thing to do. Depending on the topology used, there could easily be spikes on that primary of over 1000V. Not a nice way to treat a DVM, and it could also upset the operation of the supply. Don't! If you must check a switched-mode power supply "live", then you need an isolation transformer and an oscilloscope - and if you wish to probe the waveform on the primary of the transformer, you need x100 probes rated to a couple of thousand volts. It's far safer to put the 'scope across one of the secondary windings.

[Old Sparky]

I first describet as a SMPSU, but, unlike those it doesn't feature an opto-isolator. Point taken, Mark, about what may happen to a DMM used to measure voltages on such a supply. I don't have an isolation transformer, so maybe I should invest in one. As you know, I do have a good 'scope, for which many thanks. As to linking to the UKVRRDF thread(s), apologies if I've asked the question before, but I don't see how this is done here on GVR. Point me in the right direction, and I'll gladly do so.

[Mark Hennessy]

There are many switched-mode power supplies that manage without an opto-coupler for feedback. Using one should result in better regulation of the output voltage, but for non-critical applications, other techniques are adequate. But I've never seen an ATX power supply without an opto-coupler, so if this mystery supply doesn't have opto-couplers, then it's even less likely to be an ATX type...

For avoidance of doubt, the acronym "ATX" refers to a particular form-factor used by PCs. The power supply specification is a sub-set of that - more here: https://en.wikipedia.org/wiki/ATX - so I really doubt that this boombox incorporates an ATX PSU.

Including a link in a post is as simple as just typing it in. Using copy-and-paste is more convenient and less error-prone. There's no special procedure needed. At the risk of teaching you to "suck eggs", simply return to the thread on UKVRRR, click on the web address and press Ctrl+C. Then back to GVR, and paste (Ctrl+V) in the message. If you prefer, right-click on the web address and choose "Copy" from the menu that pops up, and then when you're typing the message back here, right-click followed by "Paste".

In the meantime, I will say that switched-mode power supplies are extremely complicated beasts, and no-one I know has a 100% success rate with them. And I'm talking about people who specialise in repairing them. They are a bit of a special interest for me, and I run a 4 day course about repairing them, but despite that, I reckon I can only fix "most" rather than "all" - a success-rate of perhaps 80-90%. And that's with the thing in front of me - not remotely via a forum. So in short, it might help your sanity with this sort of repair job to swallow a healthy dose of reality - part of which is learning to turn down these sorts of repair jobs.

You can download an old version of the notes for my course here: https://www.dropbox.com/s/sjswuzqxbkanw2...s.pdf?dl=0 - ignore the numerous typos - it's been updated a lot since this version. It obviously has a broadcast bias, and covers a lot of basics that you can almost certainly skip over, but it's a fairly readable introduction to the topic with minimal maths. With power supply repairs, the only way to improve your success rate (and personal safety) is to understand how they work and what typically goes wrong. Note that there is a section in there entitled "When not to attempt a repair"!

The final thing to say: is this power supply definitely faulty, or is it being pulled down by a fault elsewhere? Have you tested the power supply in isolation (into suitable dummy loads), or have you tested the rest of the unit using external power supplies? Sorry if these are obvious questions, but there's no point getting into the specifics of the PSU until we've answered this.

[Nick]

Wot he said. Great notes too, Mark.

[boater sam]

I think your assumption that its a 50v power pcb based on the caps being 63v may be way off the mark.
63v is a very common voltage for electrolytics used in a very wide range of applications at varying voltages.

The transformer is mains fed?
Diode bridge and caps @340v DC follow?
But then you say that the INPUT to this transformer falls to <3 to 4 v after a time, that's not logical, what am I missing?

[Mark Hennessy]

We don't know which model this actually is - there are many possibilities - and they don't appear to give specifications for the power output, but I could well believe that 50V is used. Assuming that is a single 50V rail, a decent single-ended amplifier powered from 50V could deliver around 35 watts into 8 ohms. Lower impedances or BTL operation could yield considerably more. These "beat boxes" are known for being loud and aggressive (and rather expensive!), so a big PSU is not a surprise. Of course, the +/-7V rails are for op-amps, etc.

It's a switched-mode supply, so yes, diode bridge and smoothing caps straight off the mains. No idea what topology it uses from this point, but the fact there is no voltage across the primary of the transformer means that the supply isn't oscillating for whatever reason. From the limited descriptions given, it's not possible to determine if there is a start-up fault or an overload on the secondary side. If the latter, that could be because of a fault in the PSU, or it could be because of a faulty amplifier loading it down - no information has been given to enable us to eliminate that as a cause as yet.

If the folks at UKVRRR were unable to help, I'm doubtful that I can add much to this. It seems to me that the PSU either needs to be sent away to someone who specialises in PSU repair work - which, as soon as you're past the "changing knackered caps" level becomes a highly specialised field of work - or the unit written off as BER. I'd take a look, but I'd charge a proper rate for my time, with no guarantee of a result (I don't do no-fix, no-fee work for anything modern and disposable).