Author Topic: Could a consumer product really be designed this badly?  (Read 7334 times)

0 Members and 1 Guest are viewing this topic.

Offline DJPhilTopic starter

  • Frequent Contributor
  • **
  • Posts: 511
  • Country: 00
Could a consumer product really be designed this badly?
« on: May 13, 2010, 10:27:11 am »
This is a request for a second opinion, in case I've missed something large and obvious in my attempt to troubleshoot a guitar amp. A friend of mine complained that his little practice amp (Peavey Backstage 2 2004 model) would hum incessantly. I packed up 80% of my lab and headed out to his house for a few days to visit and tinker with it to see what I could find out.

I arrived to find this amp in excellent shape, it was nearly brand new. I tested everything I could think of with a multimeter, and generally satisfied myself that there was no simple solution like a bad ground. There was only one major anomaly that I could find: the (only) opamp in the circuit was being fed a nice, relatively clean supply of +/- 24Vdc (47.99Vdc across the power pins) with a 30mV rms ripple. The schematic lists the chip as '4558', and the board contains a Texas Instruments RC4558P opamp. The datasheet lists an absolute maximum supply voltage of +/- 18Vdc. The schematic lists a part number for the transformer, which matches the part number on the transformer label, and it's rated (and tests at) 32Vdc CT. I've tested the schematic against the board itself and as far as I can tell it agrees with the schematic, both in part values and circuit tracings. All the electrolytic caps were removed, tested, and resoldered. All were fine, and the two filter caps I couldn't test (too large for meter) were replaced in turn with my one spare. All that in spite of the fact that I could not find more than 30mV rms ripple anywhere I expected DC. Near as I can tell from the schematic and the board, there's been no attempt to drive the opamp with less than the full split rail voltage that's used to drive the output transistors, which is exceeding the maximum rating on the datasheet by a combined 12Vdc.

I'm not ready to give up yet, as I'm sure I've missed something simple, but I'm left with some questions that I'd like the community's opinion on.

1. Is it normal for manufacturers to horrifically abuse parts as was done with this opamp? I mean, how can they expect to feed an extra 12Vdc to every opamp in every unit they produce and not suffer a huge failure rate?

2. Assuming everything else is normal, working, and fine with this board, could an overpowered opamp be causing a volume variable hum with nothing plugged in to the amp? I suppose anything's possible, but has anyone seen an opamp's PSRR drop to . . I guess it would be a negative value . . with an overvoltage supply?

3. Is there anything other than sheer luck that keeps this opamp from melting and exploding?

I've attached PDFs of the amp schematic and 4558 datasheet for reference. I'm really hoping this is just me being dumb.  ???
 

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18085
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Could a consumer product really be designed this badly?
« Reply #1 on: May 13, 2010, 11:44:48 am »
sounds like a very bad design, yes manufacturers can be this stupid. We are fitting circuitry to our air con stuff at work designed to work on a 24 V vehicle system with a 25V input cap....
 

Offline Pyr0Beast

  • Contributor
  • Posts: 47
Re: Could a consumer product really be designed this badly?
« Reply #2 on: May 13, 2010, 11:56:30 am »
Does the unit have 110/220V switch ? - Check input voltage and rating for the device.
Wrong datasheet ? There are similar parts with similar voltages, however I doubt 18V opamp would survive 24V for any longer than a couple of minutes.

1. Usually no

2. If gain is large enough you'll pick up all sorts of things.

3. If your readings are correct and it is what PDF says it is, than yes, there is nothing that will hold him back when it decides to blow itself up.
 

Offline DJPhilTopic starter

  • Frequent Contributor
  • **
  • Posts: 511
  • Country: 00
Re: Could a consumer product really be designed this badly?
« Reply #3 on: May 13, 2010, 08:10:15 pm »
A response, and some more info as I think of it.

Does the unit have 110/220V switch ? - Check input voltage and rating for the device.
There's a different model available for 220V, and this one is labeled properly and has the right parts for 110V. I did think that line voltage was involved while I had it out at my friend's, as his house has a nominal line voltage of about 126Vrms. It behaves similarly here at my house with 117Vrms, and although I think the hum might be slightly quieter I've got no data.

Wrong datasheet ? There are similar parts with similar voltages, however I doubt 18V opamp would survive 24V for any longer than a couple of minutes.
I did spend some time looking at other datasheets, and I took a spin through the table in Art of Electronics opamp tables. It was extremely rare for an opamp to allow 48Vdc supply voltage without specifically being rated as 'High Voltage', though a few were rated at 44Vdc. I suppose it's possible that they contacted TI asking about overvoltage and were told, "Sure, it'll probably work, sort of, at degraded performance."

1. Usually no
That's a relief. The budget guitar amp world is rife with problems and cost cutting, I just didn't think it would get this bad.

2. If gain is large enough you'll pick up all sorts of things.
I had a feeling that with that much voltage I could reasonably expect nearly anything to happen. I don't think the gain is especially high, I'll go back through and try to work it out. I'm not sure I'm reading the schematic right, but I think one half is used for summing the cd input with the guitar and the other is used as a buffer or low gain driver for the output transistor pair.

3. If your readings are correct and it is what PDF says it is, than yes, there is nothing that will hold him back when it decides to blow itself up.
I forgot to mention this earlier, but I did test it's temp for about 10min. I finger tested it and it felt too hot, so I rigged up the meter with some heatsink goop and tested it's temp at half volume. It went up to about 50C fairly quickly, and stabilized at about 58C. If I maxed the volume with no signal (just the hum) it made it to 60C. This seems quite warm, but manageable. I imagine that if I hooked it up and made it do something it could easily melt down.
 

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18085
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Could a consumer product really be designed this badly?
« Reply #4 on: May 14, 2010, 06:46:38 am »
well consumer grade parts work up to 70 C so your getting quite near thas unless it is a 85C chip (check the datasheet)
 

Offline rossmoffett

  • Regular Contributor
  • *
  • Posts: 234
  • Country: us
Re: Could a consumer product really be designed this badly?
« Reply #5 on: May 14, 2010, 07:12:22 am »
Cheap piece of crap practice guitar amps all have that problem.  For starters, probably the top is unshielded.  If your buddy has a cheap guitar amp, probably he has a cheap guitar too and it's unshielded on the inside.  If he has single coil pick-ups, rather than "humbuckers," then there will always be hum without special noise-gate pedals.

Overdriving that op-amp could be done for distortion.  It's not ideal distortion, mind you, but they probably don't care about that in such a budget application.  Often high-performance silicon is sold with lower ratings in order to create multiple product lines.. maybe this is another case like the Rigol ADC chips.  That's just two ideas.

I think you're barking up the wrong tree.  Shield the guitar, shield the amp chassis, then use shielded wire from the amp's guitar input jack to the circuit board (unless it's mounted on a PCB, which is likely).  Look for ground loops, but again, there's not much you can do about this because likely it's on the PCB.  

Is it low-frequency hum?  Then it's from the power supply - something is feeding out from there.  Is it high frequency whine?  Then there's probably a switched-mode power supply with crappy filtering.

Bottom line, I don't think you can do much about it, and I think your buddy needs to buy or build a better amp instead of worrying about this, but good luck.

Troubleshooting for tube guitar amps - some good tips
« Last Edit: May 14, 2010, 07:16:15 am by rossmoffett »
ArcAttack - A group of musical Tesla coil performers with semi-regular blog updates.
 

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18085
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Could a consumer product really be designed this badly?
« Reply #6 on: May 14, 2010, 12:24:13 pm »
power supply or picking up mains hum due do it's sensitivity could be a possibility
 

Offline Ferroto

  • Frequent Contributor
  • **
  • Posts: 289
  • Country: ca
Re: Could a consumer product really be designed this badly?
« Reply #7 on: May 14, 2010, 05:50:04 pm »
Put a voltage divider between VCC and the op-amp.
 

Offline dengorius

  • Regular Contributor
  • *
  • Posts: 96
Re: Could a consumer product really be designed this badly?
« Reply #8 on: May 14, 2010, 06:03:12 pm »
This is probably not the case, but also fluorescent lights create lots of hum
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Could a consumer product really be designed this badly?
« Reply #9 on: May 14, 2010, 08:41:10 pm »
Some things to try:


If J1 is not grounded when unused it could pick up 50-60Hz hum

Check if CD input jack is not shorted, if it is it will make IC2B a voltage follower instead of a differential amp and will easily amplify ripple

Some of power supply ripple could be fed into the amp stages, from other means, if you have a large cap handy, carefully place it across the + and - lead of the full wave bridge, it should reduce the ripple, and possibly the hum if this is the culprit

Usually powering amps above their rated power should cause distortion and overheating, but not hum pickup




Best Wishes,

 Saturation
 

Offline DJPhilTopic starter

  • Frequent Contributor
  • **
  • Posts: 511
  • Country: 00
Re: Could a consumer product really be designed this badly?
« Reply #10 on: May 16, 2010, 04:37:30 am »
Thanks for all the feedback guys, I appreciate the sanity check. :D
I'm sorry I didn't get back a bit sooner, but I was getting over some weird illness I picked up. I think it was a cold, but I'm not sure. I know I didn't have a serious fever thanks to my multimeter's temperature function (no other thermometer in the house). Thought you guys might get a chuckle out of that one.

More responses and information I forgot to mention earlier (that was likely important) follows. I had to put together this response over several hours, so I apologize if it seems a bit disjointed.

well consumer grade parts work up to 70 C so your getting quite near thas unless it is a 85C chip (check the datasheet)
You've got it right, the chip on the board isn't marked as their industrial temp version so it's good to 70C ambient. They list the junction max as 150C, and a quick skim through the thermal data implies that at max power it'd be borderline overheating. It's interesting to see that such a large supply rail overdrive isn't producing proportional thermal problems. I bet the answer's in the equations somewhere, I'll have to go back and look through the thermal design sections of the big grey book again.

Cheap piece of crap practice guitar amps all have that problem.  For starters, probably the top is unshielded.  If your buddy has a cheap guitar amp, probably he has a cheap guitar too and it's unshielded on the inside.  If he has single coil pick-ups, rather than "humbuckers," then there will always be hum without special noise-gate pedals.
Yeah, the top and short sides are unshielded. The guitar's actually fairly good, it's got multiple pickups (the arrangement of which I don't completely understand) and the cable's decent. Between his experience using other people's amplifiers and the testing I did on his guitar/cable pair with a slapped together, shielded, single transistor grounded emitter amp I'm confident that the problem (or the largest part of it) is in the guitar amp itself. It is the cheapest one ever made by Peavey, so it definitely fits into the 'crap' category.

Overdriving that op-amp could be done for distortion.  It's not ideal distortion, mind you, but they probably don't care about that in such a budget application.  Often high-performance silicon is sold with lower ratings in order to create multiple product lines.. maybe this is another case like the Rigol ADC chips.  That's just two ideas.
True. I think the opamp is just used for summing and low-ish gain buffering in this circuit though. There's a large discreet block dedicated to Peavey's tube simulation distortion, but I'm just beginning to understand the subtleties of deliberate distortion. It's a bit of an odd concept at first.

I think you're barking up the wrong tree.  Shield the guitar, shield the amp chassis, then use shielded wire from the amp's guitar input jack to the circuit board (unless it's mounted on a PCB, which is likely).  Look for ground loops, but again, there's not much you can do about this because likely it's on the PCB.
The amp is symptomatic with nothing attached, and plugging appropriate things into the jacks has had no effect for better or worse on the hum. It's amplitude tracks with the volume knob, so I'm guessing it's leaking in somewhere between the power stage and the input jack.
As an aside, I've saved up for the last four months or so, and I'm halfway to a Rigol scope. I'll definitely pick up an analog scope later on, but I think the value of the extra features (like rudimentary FFT with wider bandwidth than a sound card) is worthwhile enough to keep saving. It's things like this that really burn when there's no scope handy.

Is it low-frequency hum?  Then it's from the power supply - something is feeding out from there.  Is it high frequency whine?  Then there's probably a switched-mode power supply with crappy filtering.
It's definitely power supply hum, it's measurable at 120Hz steady on the multimeter. The power supply is definitely linear and not particularly well filtered, I'm just not sure where it's leaking into the signal yet.

Bottom line, I don't think you can do much about it, and I think your buddy needs to buy or build a better amp instead of worrying about this, but good luck.Troubleshooting for tube guitar amps - some good tips
The plan is for me to develop enough understanding to work up a good amp for future use, this endeavor is just a start. I felt it would be good to know why this one's misbehaving, even if it was a wild goose chase. I know this flies in the face of common sense a bit, but I've got the free time and it's become a bit of a vendetta. :)

Put a voltage divider between VCC and the op-amp.
That was my first thought too, and I will try that when I get clearance from the owner to perform permanent modifications. I'd probably do both rails, as it's running from a split supply. All I've done so far is desolder twenty or so key parts to check them for function. I've got some extra 317s and 337s lying around, so I might redo the whole power system, but (as Ross was saying) that's really polishing a turd.

This is probably not the case, but also fluorescent lights create lots of hum
I mentioned that to my friend before I had a chance to go out there. He tried the amp in the garage with nothing else on within 20ft and nothing changed, so I got to rule that out early. When I first started tinkering with the amp I thought of the bit I'd heard somewhere about the photoelectric effect in glass cased diodes causing weird problems. We did manage to rule out these problems in the end. I did notice that he never batted an eye while I was trying (or asking him to try) all this stuff. It's gotta look like made up voodoo to someone who doesn't know why it might or might not help. I'd have been saying, "Uh, ok, but how's that going to help anything?" He probably didn't want me to go on another electrorant.

Some things to try:
If J1 is not grounded when unused it could pick up 50-60Hz hum
Check if CD input jack is not shorted, if it is it will make IC2B a voltage follower instead of a differential amp and will easily amplify ripple
The jacks themselves are kinda crappy, the spring terminals rest lightly on their seats to short the connection when unused. They're easily bent out of shape and stand open, causing what you're describing with J1. I noticed this when I hooked up his guitar to see if the noise changed, and noted it to be worse that it started after the guitar was unplugged.
I see what you mean about IC2B (I think). I'll check it at the trace level on the pcb, and maybe pull out R40 and R41 just to disable it altogether for testing.

Some of power supply ripple could be fed into the amp stages, from other means, if you have a large cap handy, carefully place it across the + and - lead of the full wave bridge, it should reduce the ripple, and possibly the hum if this is the culprit
Usually powering amps above their rated power should cause distortion and overheating, but not hum pickup
Excellent idea, I should have thought of this earlier! I've got to dig for something with a high enough voltage rating. I think most of my larger caps are 50V rated, which might be cutting it a bit close. I think I'll wire it in temporarily just so I don't have to hold it in my fingers.
The last bit matches my intuition, but it's good to know it's true.

Summary
With your help, I think I've ruled out most of the easy stuff. If it's not something mentioned above that I've yet to test then I'm inclined to think it's either one or a few bad discreet parts or board layout. Either way, I'm fairly confident that it's not a silly simple problem that I've overlooked, and that makes me feel much better.

Thanks again everyone for your input. :)
 

Offline rossmoffett

  • Regular Contributor
  • *
  • Posts: 234
  • Country: us
Re: Could a consumer product really be designed this badly?
« Reply #11 on: May 16, 2010, 06:42:40 am »
Quote
Is it low-frequency hum?  Then it's from the power supply - something is feeding out from there.  Is it high frequency whine?  Then there's probably a switched-mode power supply with crappy filtering.
It's definitely power supply hum, it's measurable at 120Hz steady on the multimeter. The power supply is definitely linear and not particularly well filtered, I'm just not sure where it's leaking into the signal yet.

It's leaking everywhere, if that makes any sense.  Unless the PCB is very well designed it has antennas all over it.  Look at the power supply section and try to determine whether it's isolated enough from the pre-amplifier and output sections.  If the AC power is running close to those sections then that's where the hum pick-up is and you can't do anything about it because it's a PCB.  Filtering the power lines may help, but I bet they're filtered enough already on a hunch - because it doesn't take much filtering really.  In the amp I built, for example, there were three 20uF caps in parallel with some small resistors between them.  That was all the filtering it needed!  Bypass caps are common in mixed-signal applications, but not always necessary in audio.

The only advice I have for you now is to perhaps twist the AC input lines if you can - this will have the effect of cancelling out the magnetic fields coming off of them.  Maybe you can put a passive 60 Hz filter on the speaker?

If you want to know what they did wrong, here are some ideas.

Layout - As I said above, the Output, preamp, and power sections should be partitioned from one another.  They're on different boards entirely in expensive amps.  The leads should all be as SHORT as possible - especially in the pre-amp!  Those aren't just signal traces, they're antennas.  If there is a resistor going to an input to something, put the resistor directly on the input.  If there is a resistor connected to an output of something - put it as far away as possible from the output (except in cases of feedback).  That's because the stronger the signal is, the less susceptible it is to noise.  If you let the resistor take away voltage right away, then you leave the weaker signal open to noise pick-up.  Same for series capacitors.

Ground Loops - These are a complicated subject and audio engineers have mixed opinions about them.  There are names for different grounding schemes, like "Star" grounding.  Some people believe in one master ground point, some believe in grounding the output, power and preamp stages at different parts of the chassis.  There's lots of reading material about this available.

Amplification - Because the purpose of a guitar amp is to introduce distortion, some get carried away and do it right from the start.  If the first stage introduces distortion you're in trouble though, because the signal is not yet totally separate from the noise and the output will show this vividly.

My first guitar amps were of the same variety and I couldn't get rid of their noise.  That's why I learned how to make my own, actually!
ArcAttack - A group of musical Tesla coil performers with semi-regular blog updates.
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Could a consumer product really be designed this badly?
« Reply #12 on: May 16, 2010, 12:54:26 pm »
You're welcome.  My bad.  A clearer head this morning, just to clarify, I've attachments.

On the guitar jack input, if the jack doesn't ground R16, then the input will appear has high impedance, and inject whatever it picks up into the amps.

On the CD input, the circuit is appears more like a non-inverting low pass filter.  Its inverting input forms an RC circuit  C21, R39 and 38.  



A voltage divider output is used as a volume control to the next stage.

If this jack is faulty, then R40, 41, 22kohms  either alone or in parallel now is in parallel with R36 changing the low pass filter specs.  Its not a voltage follower.


Some things to try:
If J1 is not grounded when unused it could pick up 50-60Hz hum
Check if CD input jack is not shorted, if it is it will make IC2B a voltage follower instead of a differential amp and will easily amplify ripple
The jacks themselves are kinda crappy, the spring terminals rest lightly on their seats to short the connection when unused. They're easily bent out of shape and stand open, causing what you're describing with J1. I noticed this when I hooked up his guitar to see if the noise changed, and noted it to be worse that it started after the guitar was unplugged.
I see what you mean about IC2B (I think). I'll check it at the trace level on the pcb, and maybe pull out R40 and R41 just to disable it altogether for testing.

Some of power supply ripple could be fed into the amp stages, from other means, if you have a large cap handy, carefully place it across the + and - lead of the full wave bridge, it should reduce the ripple, and possibly the hum if this is the culprit
Usually powering amps above their rated power should cause distortion and overheating, but not hum pickup
Excellent idea, I should have thought of this earlier! I've got to dig for something with a high enough voltage rating. I think most of my larger caps are 50V rated, which might be cutting it a bit close. I think I'll wire it in temporarily just so I don't have to hold it in my fingers.
The last bit matches my intuition, but it's good to know it's true.


Thanks again everyone for your input. :)
« Last Edit: May 17, 2010, 07:49:14 pm by saturation »
Best Wishes,

 Saturation
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf