Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Synth clips awfully on speakers/soundcards but NOT on headphones? Halp (Read 13872 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Google seems to disappoint when it comes to quite vague combinations of terms like these—at least if you don’t already have a degree in electronics! So I would greatly appreciate advice from our electronics-/engineering-savvy members. I think this thread is roughly on-topic.

I just bought an old synth. I quickly became disappointed and, just as much, baffled when I realised that any time it’s playing more than one note, it generates horrendous clipping in any speakers or soundcards to which it is connected. The signal appears to be clipping, but it unfortunately doesn’t seem quite that simple, because…

I then discovered this problem does not occur on headphones, which reproduce the synth just fine, as I expected it to sound on the other devices. Also, it sounds pretty quiet there, and indeed, this synth has been claimed to have a low output before. So, the clipping isn’t intrinsic to the outputs, but they’re definitely doing something that speakers, etc. don’t like, causing said devices to read a signal that clips hugely. Because of the low output on headphones, I can’t see how the problem could be due to the outputs being simply too loud/hot, and so I think some other electronic factor must be at work.

What in the world is going on here? Is there any way to assess, and preferably to fix, it? Bear in mind that my knowledge of electronics is, to be charitable, extremely slight. A friend and I think it’s most likely an issue with impedance, perhaps due to the higher-impedance(?) speakers and soundcards drawing a more powerful signal than they should due to some load(?) introduced earlier in the circuitry. However, we can’t make a single guess beyond that and can’t see any component in the pre-amp that is different from what the schematic says it should be. I seem to recall hearing that it may as well be impossible to calculate the impedance introduced by the entire circuit, including the majority that lies beyond the preamps.

Another confusing aspect of this is that the clipping seems to occur only when I try to play more than one voice (read: note) at a time. This makes me wonder if the pre-amp was somehow ‘normalised’ to the maximal no-clip gain for a single voice without any regard for what would happen when something as basic as polyphony was thrown into the mix.

I don’t suppose it’s as simple as shoving a large resistor into the circuit somewhere. :/ Any advice that might help me to salvage this unit will be hugely appreciated. I hope someone might recognise this symptom and be able to suggest the most likely cause—and, with any luck, a feasible solution. Thanks!

Edit: To be clear, this isn’t an intrinsic incompatibility of this type of synth with speakers or soundcards, for I have another one that works just fine. And, again, we couldn’t find any differences between the two that could logically have any affect upon the preamp whatsoever.

Edit again: It has stereo outputs, which naturally have seemingly identical preamps, and both exhibit this problem.

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #1
FWIW, a higher impedance load draws less power.  Probably what happens is that your device has a relatively high output impedance and when discharged into a high impedance load, the voltage exceeds some maximum and clips. A low impedance load creates a voltage divider which prevents something from clipping.  No idea whats going on that would cause this (maybe a failing transistor somewhere in the output stage???), but if putting a load prevents clipping you can fix that by getting a headphone splitter and connecting a resistor on one of the outputs (or even a pair of headphones).

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #2
Thanks for the reply!

FWIW, a higher impedance load draws less power.  Probably what happens is that your device has a relatively high output impedance and when discharged into a high impedance load, the voltage exceeds some maximum and clips. A low impedance load creates a voltage divider which prevents something from clipping.
Although it’ll take me a while to digest, I think this is very helpful.  AFAIR, we found, predictably, that the headphones whose specifications we could find had a considerably lower impedance than my monitor speakers, and I presume the same would apply to the other speakers and soundcards on which I tested the synth and observed the same problem.

Quote
No idea whats going on that would cause this (maybe a failing transistor somewhere in the output stage???)
Now this is where I get very interested! Each of the two stereo preamps contains a transistor, and both seem to be controlled by a circuit on another part of the board, which splits off towards the preamps at a third transistor. We couldn’t fathom what these transistors might be doing, but again, and especially speaking for myself here, we’re by no means experts on this subject. What do transistors normally do in preamps, what might be going wrong with this function, and is a problem with impedance a likely result? I’d really appreciate a detailed and noob-style explanation here: you’ve converged on something we already considered, so it might be a very promising direction to investigate. Excellent!

Quote
but if putting a load prevents clipping you can fix that by getting a headphone splitter and connecting a resistor on one of the outputs (or even a pair of headphones).
See, I thought about this. My logic was that I should connect a splitter to one output and only then connect it to one of the speakers or soundcards that exhibit the problem. That should demonstrate whether the problem is conclusively attributable to impedance, right?

Anyway, thanks again, a lot!

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #3
Since it happens to both stereo channels, I guess it would not be a transistor in the amp itself (since that would only clip one channel).  Probably its either made like this for some reason (I know nothing about synths...), or possibly the power supply voltage is off meaning that the full dynamic range of circuit cannot be used.  Maybe someone else here has more experience. 

Quote
AFAIR, we found, predictably, that the headphones whose specifications we could find had a considerably lower impedance than my monitor speakers, and I presume the same would apply to the other speakers and soundcards on which I tested the synth and observed the same problem.


If they're powered, they're basically a line in (1-10kohm typical).  If they're unpowered like most headphones, 16-300 ohms is common.  If the output impedance of the device is 100 ohms, hooking up a 100 ohm pair of headphones would reduce the maximum voltage through the output stage by 2.  10kOhm line in would reduce it by 10k/(10k+100) ~~ 1.    So its plausible that having a load could help you reduce clipping. 

I would get a splitter and put a load on it and see if the other output into line in is ok. 






Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #4
So its plausible that having a load could help you reduce clipping. 

I would get a splitter and put a load on it and see if the other output into line in is ok.

Indeed, I managed to put together a splitter by combining all the adaptors I could find, haha, and it does significantly reduce the clipping on the speaker when the headphones are plugged in compared to not. I’m not sure the clipping is gone entirely, but it’s a lot better when the headphones are in the other channel of the splitter.

So, this indicates that impedance is indeed the problem, right? Now the questions become: Why would this synth be doing this but my other one not? What sort of things could cause the impedance to be different and outwith the usable range on this particular unit? And, in any case, what are the options to fix it? Loading the outputs manually is not really practical, and as I said, I’m not sure it completely solves the problem, at least not with the unpowered headphones I have here to test (whose impedance I don’t know). So I hope there’s a plausible reason for the raised impedance and, correspondingly, a way to address it.

Since it happens to both stereo channels, I guess it would not be a transistor in the amp itself (since that would only clip one channel).  Probably its either made like this for some reason (I know nothing about synths...), or possibly the power supply voltage is off meaning that the full dynamic range of circuit cannot be used.  Maybe someone else here has more experience.

I guess the transistors take a backseat now due to the above, but the transistors in both outputs are linked back to a single other transistor elsewhere in the board, so is it possible that could still be involved? Could a fault in that common transistor raise the impedance? I warned you that I know almost nothing about this stuff!  Anyway, you say you know nothing about synths, but given that this is unlikely to have been a design choice, your or anyone else’s knowledge about audio circuits in general seems applicable and invaluable.

Cheers again!

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #5
Like saratoga mentioned, check your power supply for proper voltages.

If it turns out everything is working as intended, and it just can't handle a high impedance load, it wouldn't be hard to stick a resistor and opamp at the output to get a more useful output.

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #6
Quote
Like saratoga mentioned, check your power supply for proper voltages.
Yeah, I hot-swapped in the PSU from the working unit (since it just enters the mainboard in one big cable), and the clipping is still present when the outputs are at normal load.

Quote
If it turns out it's working as intended, and just can't handle a high impedance load, it wouldn't be hard to stick a resistor and opamp at the output to get a more useful output.
I struggle to see how it could be working as intended when my other unit has no such problem and no obvious electronic differences that could account for the much higher impedance of its problematic counterpart. Might this somehow be attributable to problems with the existing opamps or, again, the transistors?

The only other thing I considered was that maybe something is messing with the output of the DAC and its associated circuitry, perhaps its buffer amp. But as usual, I’m in no position to judge the veracity of this suggestion. Although my gut tells me that such a departure from the normal operating conditions might be more likely to fry the DAC than to alter the eventual impedance of the outputs.

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #7
From the symptoms I would guess that the problem is leaky output capacitors. Typically the output stage will be at a DC voltage, which is coupled through an electrolytic capacitor to block the DC from the load. If the load is low impedance, such as headphones, then the leakage results in a relatively small DC voltage across the load. If instead you have a high impedance, such as the input to a power amplifier, then a relatively large DC voltage will be generated, saturating the amplifier's input stage.

The solution is to replace the old electrolytics. In the mean time placing an extra load, such as your headphones, across the output will be a temporary solution.

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #8
This sounds very promising! Thanks a lot.

Even with the working unit, I was already planning to replace some components in the circuits for other reasons, and some of the caps will probably be on that list. Within each preamp, there are two capacitors that are electrolytic. One is clearly functioning in a first-order HPF just before a second-order LPF, and the other is after the LPF doing something else that I can’t guess*. My plans were, as said, based upon the working unit, and I never implicated the electrolytic caps among the ones that might benefit from being replaced. Am I reading you correctly that it’s only the electrolytic ones that will be responsible, if any capacitors are? And only the ones near the jack, not earlier in the circuit?

Also, what would be the likely cause of the failure? It seems highly coincidental that both outputs exhibit the same problem, to a perceptually equivalent degree, and it seems unlikely that they would have degraded at exactly the same rate with age, so my money would be on a previous voltage surge that went through both preamps. Either that or a really terrible batch, maybe?

I hope not too many caps need replaced, but I guess we’ll start nearest the outputs and test at every step, working our way backwards if need be. If, as I asked, this could only be due to the minority that are electrolytic, that’ll narrow things down a lot.

Thanks again!

*This is a steep learning curve, but I can now see that the second electrolytic cap, nearer the output, is acting in another HPF-style circuit with a resistor, presumably as the coupling/DC-blocking capacitor that you supposed. As I said earlier, in my post below (poly hime taradox), the signal then goes through another few resistors and a small film capacitor before getting to the jack.

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #9
Tantalum caps can fail too, but they tend to short-circuit and/or explode, so you'd probably know if one died. Electrolytic caps die in all sorts of ways, I had one in the power supply for a headphone amp leak out the top vent a tiny bit and fail open-circuit with virtually no capacitance remaining. Other times they can keep their capacitance but have their ESR shoot way up, making it even harder to tell what the problem is unless you have an ESR meter.

Is it actually a stereo device, or just mono split into two channels for output?

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #10
Tantalum caps can fail too, but they tend to short-circuit and/or explode, so you'd probably know if one died. Electrolytic caps die in all sorts of ways, I had one in the power supply for a headphone amp leak out the top vent a tiny bit and fail open-circuit with virtually no capacitance remaining. Other times they can keep their capacitance but have their ESR shoot way up, making it even harder to tell what the problem is unless you have an ESR meter.
Well, FWIW, I can’t see any physical leakage from the vents, which are still flat and unbroken, or any blow-out at the bottom: no visible symptoms that I can detect. Of the other capacitors in this area, most are ceramic, and a few are made of metallised film. There are other electrolytic caps, but they’re quite far away and seem to have other roles not related to the preamp.

Only the one in the HPF and another after the opamp/LPF are ceramic electrolytic, and although the latter seems most likely to be contributing to the excessive impedance of the outputs if I’ve interpreted pdq correctly, the signal still has to go through a few resistors and a small film capacitor before hitting the jacks.

[edit] I just noticed that two of the other electrolytic capacitors are functioning in the circuit around the DAC: one is definitely interacting with the buffer amp that it uses to verify/condition its output before actually sending it as sample-and-hold, whereas the other comes between the chip’s VCC, AGND, and DGND. Might either of these capacitors be relevant? Like the others, they have no visible symptoms of failure. [/edit]

Quote
Is it actually a stereo device, or just mono split into two channels for output?
It’s properly stereo. Like I said, this makes me rather apprehensive that the fix won’t be as simple as pdq’s elegant suggestion. :/

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #11
I’ve replaced three pairs of electrolytics: two near the outputs, two further back in the preamps, and two around the DAC. The outputs are still clipping on speakers as before. The consolation prize is that I managed to jam an ill-fitting new battery holder in there whilst I was on a roll with my soldering.

So, the problem is not with those electrolytic caps. I’ve since noticed, however, that the electrolytic capacitors that are near the outputs geographically are not so near them electronically. Rather, each output is directly preceded by a small capacitor made of metallised film and with a capacitance of 2000 pF (2 nF). So, to reiterate a previous question:

[to pdq:] Am I reading you correctly that it’s only the electrolytic ones that will be responsible, if any capacitors are? And only the ones near the jack, not earlier in the circuit?

In other words: is it possible for these smaller caps to be functioning as the coupling/DC-blocking capacitors? Sure, all caps block DC, but if I remember rightly from my research on all this, it’s much less likely that these film caps would have failed due to a surge, age-related deterioration, or something, isn’t it? The most plausible explanation I can think of is that someone plugged something with phantom power into the outputs, blasting them with 48 unexpected volts. Could that break film caps as well as electrolytic ones? In other words, what’s the typical surge tolerance of the former type? I couldn’t find any specific discussions about this when searching.

I’m really hoping pdq is correct in the extremely promising diagnosis that the problem is caused by unfiltered DC saturating the higher-impedance speakers/soundcards (but not the lower-impedance headphones). However, I really need an updated professional opinion based upon this new info and whether the film capacitors could still be implicated. I already struggled to understand how both outputs could have failed in the same way:

Quote
Also, what would be the likely cause of the failure? It seems highly coincidental that both outputs exhibit the same problem, to a perceptually equivalent degree, and it seems unlikely that they would have degraded at exactly the same rate with age, so my money would be on a previous voltage surge that went through both preamps. Either that or a really terrible batch, maybe?

And now I’m even more confused.

Thanks again in advance to anyone who has any thoughts on this.

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #12
Film capacitors are not likely to be the cause, both because they very rarely fail, and they would not likely cause your symptoms.

I probably should have asked you first to do some testing. If you have access to a voltmeter, measure the DC voltage at the outputs, both with and without the headphones connected. If the DC voltage is near zero (i.e. hundredths of a volt or less) then this is working correctly and my original guess was wrong.

There is some possibility that the output stage is bipolar (with both positive and negative supplies) and there is no blocking capacitor. A problem with one of the supplies would then cause a DC offset and the kind of symptoms you are seeing.

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #13
Thanks for replying!

Film capacitors are not likely to be the cause, both because they very rarely fail, and they would not likely cause your symptoms.
As I suspected! :/ Some consolation to know my research was fairly accurate, at least, haha. But I probably ordered a pack of 5 earlier today for no pressing reason. Between those and all the ones I replaced earlier, which probably aren’t culpable either, I’ll have more spares than I’ll ever need at this rate!

Quote
I probably should have asked you first to do some testing. If you have access to a voltmeter, measure the DC voltage at the outputs, both with and without the headphones connected. If the DC voltage is near zero (i.e. hundredths of a volt or less) then this is working correctly and my original guess was wrong.
I don’t have a voltmeter, but I’ll look into whether I can borrow one to check this.

Quote
There is some possibility that the output stage is bipolar (with both positive and negative supplies) and there is no blocking capacitor. A problem with one of the supplies would then cause a DC offset and the kind of symptoms you are seeing.
Another very interesting suggestion. You’re just full of great ideas!  Is the bi-polar nature, perchance, hinted at by the transistors that I mentioned earlier? There’s one in each channel of the pre-amp, and both link back to a transistor on another part of the board. Could you explain the advantages of a bi-polar output stage? I don’t know whether there’s any relevant info in this patent I found on a search:
Quote
http://www.google.com/patents/US5179293[/url] ]When the output stage is an analog amplifier, in active mode it amplifies its input signal. In either arrangement [digital or analogue], in the "inhibit mode", the output stage transistors are turned off by reverse-biasing their base-emitter junctions relative to the output node of the circuit, to provide a floating high-impedance output node. In the inhibit mode, additional means are provided for substantially cancelling leakage current at the output node.
FWIW, I already bought a bunch of spares of all the types of transistors I could see, partly in case they were involved and partly for future-proofing; so, if it’s possibly attributable to any of those, I can do some swapping ASAP.

Again, thanks a lot for the suggestions here. This problem is way beyond my knowledge and abilities, so I need all the informed ideas I can get.  It’s been quite a learning curve since I started poking around on this board! I hope it can be saved, somehow.

Edit: The schematic shows the op-amps as having positive and negative inputs at 5A, if that’s relevant to the question of whether the outputs overall are bi-polar.
Son of Edit: But in terms of the signal, the positive input gets the signal, whereas negative receives the feedback from the output of the opamp. Yeah, I have no idea how this stuff works.

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #14
The big advantage of a bipolar output stage is that it eliminates the large DC blocking capacitor (which adds expense) and the loss of low frequencies that can result from a too-small capacitance into a low impedance load.

The disadvantage is more power supply parts.

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #15
Well, in lieu of any answers on whether/how transistors might be involved, I went ahead and replaced the pair of transistors near each output and the one earlier in the circuit that leads to both. And, crossing all digits and touching all available wood, that seems to have done the trick!

I’d be very interested in an explanation as to what might have been wrong with the transistor(s). The single one that links forward to the pair near the jacks is located near a couple of links to the RESET pins on the two main processors, so I thought it might be involved in switching those and the outputs between powered-off and -on modes; I wonder whether a fault in this function could have caused a DC bias on the outputs via their respective transistors. Alternatively, the latter two might both have coincidentally undergone one type of failure and caused it. As you can tell, I’m just throwing out guesses here, haha. This problem has been quite fun in its own way, though!

I suspect the problem was in the single, common transistor and that I should have just tried that first to avoid unnecessary extra work, but oh well! I already replaced all those capacitors earlier that seemingly were blameless. Part of me thinks I should just go wild and replace all the other caps I have spares of, just for the heck of it, but I’m going to walk away whilst I have the chance and just live with this patchwork of old and new components.  It’s louder than my other one, too, which is quite handy, although I’ll have to monitor the volume and reduce it when using lots of notes at high volumes, or else it’ll be clipping again.

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #16
I don't think I can shed any further light on why replacing that transistor fixed the problem, but bravo for persisting until it was fixed. Most people would have given up before then. 

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #17
Heh, well, I tend to get stubborn with things like this.  But yeah, I might not have been so persistent if you hadn’t suggested some plausible explanations to keep me optimistic, so thanks again! Same to saratoga and Rotareneg, too.

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #18
You're lucky your headphones still work. I've been told DC can destroy certain headphones in no time.
Music: sounds arranged such that they construct feelings.

 

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #19
This happened again! twice!  Which did help me to conclusively narrow down the cause, so I guess it wasn’t completely without benefits.  In both cases – and hence presumably the original one – the issue was not the single, common transistor but rather the per-channel muting transistors: a 2SC2878 on each of the two outputs. I presume their failure seriously alters the output impedance to create the symptoms we discussed last year: harsh clipping but ameliorable by connecting headphones in parallel.

Even more intriguingly, in both cases, replacing each channel’s respective transistor was not sufficient: only when I replaced both transistors would either channel sound proper. I guess they form an intrinsic link of some variety between the two channels. I did notice some bleed-through between the two channels not too long before the second failure, so that might have been a type of leakage indicating that sort of link.

As to the cause of their failure, perhaps the PSU has ripple or other variation a bit above spec that does not bother any other components but is considered irksome by the 2878s? or maybe one of my cables is dodgy and causing ephemeral shorting of the output jack that somehow fries one or both of the muting transistors?

In any case, by all indications, the 2878 seems to be a very sensitive device and prone to dying at the slightest prod. Where this gets somewhat worrying is that 2878s, although designed specifically for audio and muting it on power-up to avoid DC thump, are discontinued and won’t be available forever! I think I have about 8 left at present. I should probably do something about that PSU…

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #20
Quote
I don’t suppose it’s as simple as shoving a large resistor into the circuit somewhere.
I'd try 10K resistors from the outputs to ground.     Or, maybe 1K.  If the problem is a leaky* capacitor, and the leaky capacitor is not defective, the leakage resistance should be in the megohms and 10K should be plenty low enough.

If there is a capacitor in series with the output, even a new-good electrolytic capacitor will "leak" to some extent (if the DC "load" impedance is high enough).      A line-level output should normally have a resistor across the output to discharge the capacitor.  But for a "headphone output", they might leave out the resistor.

Maybe in your case, it's the transistor that's "leaking".  In any case, if it is  supposed to be a headphone output, it may be designed for a DC load (a path for DC current).  Your speakers & soundard probably have a blocking capacitor at their input, which means there is no DC current path on the synth's output.

I've also seen a case where the lack of a resistor on an input can result in the input voltage drifting-up through the input capacitor, causing distortion.  But, since you've tried this with more than one soundcard & monitor, it's obviously the output of the synth, and your soundcard/speaker inputs are not the problem.


I'd be surprised if those transistors are completely dying, but I wouldn't be surprised if the design is marginal (or marginal when there is no headphone load).  ...There is a quick-and-dirty way to check transistors with a multimeter.  It won't measure the gain of the transistor,or tell you if it's "leaky", but it will tell you if the transistor is "blown' or "dead". 




* I'm talking about electrical DC current leakage... not chemicals/fluids leaking out. 

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #21
…Did you notice my latest post just above yours?

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #22
You might try the EEVblog, you're bound to find someone there who can help you figure out what's going on.

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #23
Well, again, I know exactly what’s going on proximally – the 2878s keep failing and are seemingly highly sensitive – but not the physical mechanism or why it keeps recurring.

Anyway, that site seems interesting, so thanks for the link.

Synth clips awfully on speakers/soundcards but NOT on headphones? Halp

Reply #24
on the face of it its highly unlightly that a solid state NPN tansitor is a "highly sensitive device"

Did you check the voltages against the data sheet, available from serching google

and , and  also on the face of it,  before it blows it would need  to get hot

are you sure they are blown, because if they blew without getting hot that would be odd. They are  not cmos chips after all.