Make it (the buzz) stop... please

[idle_chatterbox]

QUOTE:
" If a bad grounding is not the problem then the easiest way to find the hum source would be to probe the signal output of the various preamp stages with a capacitor coupled wire feeding a high impedance amplifier plugged to small speaker."

hmmm.... sounds like a good idea. I'll try it. Wish I had a schematic, but from what I can gather the company that made the amp was in business about 6 months, so they're not online anywhere.

Thanks for the suggestion and the help.

[teemuk]

OK. I guess you can pretty much ignore the theory of sharing the mains ground return I presented in the previous post then. What you describe sounds like an adequate switching configuration - something found from many amps. IMO it's just safer to break the connection of both L and N. This way you can be sure that only the switch (and probably the fuse are in mains potential). I assume the sleeve of the switch has no electrical connection to any part of the switch either...

I edited my post while you wrote your's but my main points were to say:

"If the problem would be the bad capacitors the amp would hum all the time - even with the Pandora plugged in. ...Same thing with the "mixer" section: if the hum would be induced there you should hear it even without having to plug someting into the amp"

and...

"you do not need a separate mixing stage to induce hum problems if the jacks share the same ground path (chassis) at the input."

I hope this helped to see thing in a better light. Anyway, just to make sure I have not completely misunderstood your description let me ask this once again: Does the amplifier hum with nothing plugged in?

If hum is present only when something is plugged in it means that disconnecting the input breaks the signal path of the hum. To visualize the thing better examine the attached picture and consider how it affects your amps input signal in different situations. This schematic most probably shows what you're dealing with. Now, you should find out what causes the signal loop of the hum and how you could get rid of it.

BTW, What's your amp model?

[idle_chatterbox]

Ok, thanks again. If that's the safest way to run the switch (with both of the mains instead of just the one going to fuse) I'll do that. The switch can definitely handle it since it's got the lugs on both sides.

As to your question: no, the amp does not hum when nothing is plugged into an input jack. Even if I crank all the volume knobs, I only get a hissing (maybe better to describe it as "seashell against your ear" sound?). It's a loud hissing, but it's pretty much the same sound that you get if you crank up an ordinary stereo that's switched to "Aux" or something like that.

The humming only arises if I plug in a guitar. At that point, I can't even break past the "1" on the volume of the master. The hum is just too loud and overpowering. Increasing the bass/treble, etc. has the expected effect of making it worse.

Like I said earlier, running the same guitar through the pandora doesn't just quiet things down a bit, it completely eliminates the hum. I've tried to make it hum, and even on some of the pandora settings that are meant to model a high-gain amp with distortion, it just won't. Interestingly, if I just run the Pandora's drum fill (with no guitar attached), there's no hum either. Since I started trouble-shooting this, I've switched the guitars to amps that I know are good, because I really want to believe that this is just a guitar issue. But on other amps they are silent, especially the guitar with the EMGs.

RE: the model of the amp, it's called "Electric Spirit 4" by Holmes. Fairly simple, in fact most of the amp head is empty space inside. There's a small reverb pan, but no other on-board effects.

With the hum gone, the amp sounds wonderful. I'll never be able to use anything but a few watts, but I don't care, it still sounds great. Just for the heck of it, this morning I ran the guitar through my Art SGX (which is usually noisy) and into the amp. No hum. But then again, that is grounded through its own 3-prong plug. Either way, it sounds good. I'll keep trying to figure out the mystery, but it's nice to at least have a way to avoid the hum.

Thanks. I'll look at the schematic you posted. 

[teemuk]

Ok. I could not locate a schematic either.

Here's a simple and quick test you can try: If the inputs have non-insulated jacks and if they are connected to ground trace on the PCB then lift one of the jacks off from the chassis, insulate the jacks's ground pin so it won't touch anything by accident (i.e. wrap it it in plastic film) and then plug in a guitar (without Pandora) to the lifted jack, See whether you still have the hum problem. If you can do this you have either located the source of hum or eliminated one possible source.

[idle_chatterbox]

ok, I'll try it and let you know. I take it that the point is to "unground" the jack from the chassis? Or disconnect the ground wire coming off the back of the jack as well?

[teemuk]

You can try both. Start with "ungrounding" the jack from the chassis.

Another point: Do the input jacks have a switching mechanism and if they do how is it connected?

If you can post a detailed picture showing one input section and it's connection to PCB that would be helpful. I see the amp photos in a size of a postcard. I guess haven't figured out how to view your photos in higher resolution.

[teemuk]

This is what I highly suspect: The chassis acts as a high current ground return path which causes it to appear as an alternate voltage source between input stage's input and reference ("ground") nodes. The signal amplitude of this voltage source doesn't have to be greater than few millivolts in order to cause a notable hum. Note my deliberate use of quotation marks: There can exist only one true ground point in your system.

Now consider what happens if you would unplug the guitar cable, thus removing the "voltage source" from the place of the jack symbol: The same hum voltage would now affect at both input and reference nodes. Therefore the voltage difference (V input) reading would be zero. If there is no voltage difference then there will not be amplification either.

Note: Your guitar pickups do not have a ground but if you were to really ground the input "ground" - and at the same time the chassis - to another system, i.e. through Pandora or any other device then the current return path on the chassis would become grounded as well. If both ends of the hum voltage source are connected to ground it means the hum can not be in series with the input jack voltage source. The result: No hum.

If you lift the input jack "ground" from the chassis you can test whether my theory is right or not. In worst case, you may have to unmount all input jacks from the chassis to check this out. The PCB looked like it contains a ground trace where it hooks all the ground wires from the input jacks, therefore the input can make a connection to chassis through any jack still left mounted.

I need some sleep now. Good luck in troubleshooting.

[idle_chatterbox]

ok... I've now disassembled and cleaned all 22 pots again, re-checked all solder connections (most of them are large, so that's easy), and tried plugging the guitar into a jack that's not secured to the front panel. No luck... I have to admit I'm starting to get disillusioned. I need to order some parts for other projects, so will order some caps and replace those.

Still the same issues. With nothing plugged in, fairly loud hiss. Guitar plugged in, loud hum as as soon as the volume is increased. Run guitar through pandora, hum is pretty much gone. Gone enough to not worry about. Just to see what happened, I ran it through a Yamaha digi-stomp. That didnt' take away as much hum, but it did make a difference. Even with the gain turned up all the way, the hum didn't really get noticably worse.

Now that I've been messing with the thing so much, I do notice that the hum doesn't really increase proportionally with adjustments to the master volume. I.e., sometimes it will become 100% louder with a tweak from 1-2, and other times only 50%. Could all of my problems be just a bad pot, or is the pot just "reporting" a problem elsewhere in the circuit? I guess it wouldn't be that hard to swap in a new one and see what effect that has, given that I'm now intimately familiar with all 22 of those stupid thing.

Thanks for all the suggestions. I tried to upload a bigger photo of the front panel, but there's a size limit for this site.

[teemuk]

Ok, you should try my suggestion and lift all jack grounds from the chassis. The trace circling around the PCB is the "ground" and it seems to make chassis connection through several screw mounting points. Something like this can be an open invitation for ground loops. Then again, I repaired an amp which used a similar grounding method few months ago and it didn't have any problems with hum whatsoever. Sometimes grounding issues can be very tricky...

Another easy method to test my theory of ground return currents in the chassis is the following: Plug the guitar in normally but make an electrical connection from the amp's chassis to a known good ground point i.e. a chassis of another amplifier. Before doing this check if a notable voltage difference exists between these points.

Here's a result of intuitive thinking:
1) I doubt that you have no mechanical component fault in the amplifier (i.e. faulty / dirty potentiometer etc.) If your amp works fine with Pandora or any external effect this option should be overruled. If your controls would have a fault it would appear in every situation regardless of the signal source.
2) The problem is not induced ripple in the supply line from aged capacitors: You would hear this as well. You could still control the volume of preamp induced hum but it would appear regardless whether the guitar is plugged in or not.
3) If you can control the amplitude of the hum with master volume potentiometer then the source of hum must be located before this potentiometer. You can apply the same rule for every control that adjusts the sound and amplitude of the hum. This should help you narrow down the problem.
3) But I consider this point as the all narrowing you need: If your amp works alright but hums when you plug in a guitar then you have found the source. The way I see it there's mainly two options:
a) the guitar is faulty or picking up external interference: You have overruled this option by testing the amp with other guitars and signal sources.
b) the ground path induces hum. Have you tested the amplifier with another mains outlet? The one you use may not neccessarily have a safety ground return which means your chassis is most likely floating in other potential than zero. Have you checked the continuity of the safety ground wire in the power coord?

Even though all points would prove the chassis is connected to ground adequately it doesn't imply the potential of all points in the chassis is zero. Remeber, there can exist only one good ground point in your system. Ideally all grounds should connect it directly (star ground topology). The only problem causing the symptoms you describe I can think of is overlapping ground return paths. The current in this return path must likely be fairly high since it can induce a notable voltage over a (assumably) small resistance.

Anyway, I assume your amp did not induce this hum from the beginning: The injection of hum is not a normal condition and definitely something has gone faulty in your amp. I still suspect a bad solder joint or mechanical connection somewhere. Now, you should check what high current stages overlap their ground return with the inputs. In order to do this you must locate the secondary ground (hint: three most common places to look for are secondary center tap, the main filter caps or rectifier). Follow the ground returns of inputs and high current stages (mainly PA and speaker jack) to this point and see whether they overlap. The chassis is tricky but most likely the electrons find the shortest/least resistive path). Check the path for for loose connections, cracked traces, corrosion, broken parts and anything abnormal.

[joecool85]

Thanks for all the suggestions. I tried to upload a bigger photo of the front panel, but there's a size limit for this site.

If you need something really big you can always use photobucket or another image hosting site.

[teemuk]

Exactly. I've had good experience with http://imageshack.us/ so far...

[idle_chatterbox]

Thanks for the suggestions on the photo sites. I tried both. ImageShack seems to require more intelligence than I have to upload photos, and with photobucket, I can upload, but if there's a way to keep the photos their original size, it's apparently beyond me. In any case, I tried.

http://s96.photobucket.com/albums/l164/idle_chatterbox

I know when I'm beaten....

Anyway, back to the amp. Took it apart again, checked with eyes and meter for bad connections. Recleaned the 22 pots. No change.

grounded the amp to another grounded unit. No change in noise. Tried other guitars, cables, speakers. Unconnected one of the input jacks, plugged in, no change.

I've got no ideas.... Replacing the caps at this stage seems like a shot in the dark, and I'm wondering if it is cost-effective to solder good caps into a noisemaker.

:grr

[teemuk]

If you disconnect only one input jack from the chassis it will still remain connected through the wiring of other jacks. Grounding the chassis to another unit should however indicate that the chassis is not the hum source... at least I think so. You could try probing various ground points with a wire that has the other end connected to central ground point. (Where is it located by the way?) This should show you if another ground route would stop the humming - what it most likely should do if the hum is indeed ground induced. Start with the input jack's ground lug. Remember to keep the amplifier's master volume low as possible - you only need to hear if the hum goes away or decreases. Be careful not to short anything by an accident.

And another thing to check out: Do the input jacks have switches. (Check if the inputs are grounded when no plug is connected.) It looks like they don't have a switching mechanism but then again your amp pics are postcard size and only show an overview. At this point it has become impossible for me to help you anymore unless you provide more details (i.e. schematic, sketch of the grounding paths or more detailed and larger photos - preferebaly with the important sections labeled (central ground point, power amplifier, power supply, etc... you got the idea.)

Replacing the capacitors as a cure for the problem seems like a far fetched idea indeed. I sincerely believe that an increased amount of ripple voltage (= hum) due to bad capacitors would be heard in all situations. It might work in case the capacitors have began to leak current to ground path. Curing the hum problem with an external effect unit is a very clear sign of some problems existing in the grounding path and topology.

....However, I advice you to change the filter capacitors anyway (both main and preamp filters of course). Actually, I advice you to change all electrolytic capacitors while you're at it. Your amp doesn't seem to have a massive amount of them - actually it has amazingly few - so doing that is not a big deal. Capacitors cost what, few bucks - not a big deal either. Decoupling capacitors can start leaking DC and a bad bootstrap capacitor in the power amp doesn't sound good either. Sometimes aging capacitors do make strange problems: My 20 years old Yamaha multitracker makes sudden loud pops, hums now and then and the "lowest" VU meter leds sometimes burn constantly (DC in signal path). I don't want to touch it: this thing houses at least 100 electrolytic caps. Compared to work needed for monster circuits of 80's (where virtually every capacitor was an electrolytic) your "cap job" seems like a child play. If you replece those capacitors you have both increased the life of your amplifier by, say, 20 years and eliminated the capacitor's as being any (even strange) source of problems.

[idle_chatterbox]

Yeah, I wish I had a schematic. But at least last time I searched online for one, I came up with nothing.

I guess you're probably right that it won't make it sound any WORSE if I replace the caps, and like you say, there's not that many. The only good part about it is they all seem to be out in the open and not under or tangled up with other components.

As far as where the ultimate ground is, I haven't been able to tell (yet). The main cord ground wire (green, the "3rd prong") of course grounds to the chassis right next to the transformer, but there is as far as I can see only a few wires that connect the back panel (where the trans is) to the front panel (where the inputs and pots are). There's no obvious wire leading from either PCB to the chassis, for example, or metal screws/mounting points that would do that. Or at least I haven't seen them yet, and I've been staring at this thing a lot lately. Two of the wires connecting the back and front panels probably don't count, since I think they are only the "On" light circuit.

Thanks again and I'll keep tinkering with it and let ya know. 

[teemuk]

For starters, it looks like the PCB on the front panel makes ground connection to the chassis through the mounting screws and input jacks. ...Should I even use a word "chassis", your amp seems to be built on two separate metal sheets? Does the casing have a metal sheet or anything tying these two panels together?

The interesting point is indeed how these two panels connect each other, It's quite fair to assume that the only ground - not just a ground return is located somewhere in the rear panel and near the power supply. Therefore you should have at least three wires running from rear panel to front: Signal, ground return and Vcc. If the supply is dual (which one is it?) then you should also have Vee. (Unless the preamp uses single supply). If the whole amp is single supply then the ground counts as Vee.

The above configuration seems reasonable if the front panel houses only small current stages - ideally, if you have to bundle small current stage ground returns together the signal ground returns should use a return path separated from the supply ground returns (ie. filter caps). Shared supply and signal ground returns can cause hum so it seems like you should be concerned about this at some level. I think that grounding the front panel to a separate system could actually cure hum caused by this. However, a quiet amp without a guitar plugged in doesn't fit this theory. There's a heatsink looking thing attached to front panel also but I assume that it is not a section of power amplifier but the case for spring reverb? Note: The input impedance of most spring reverbs is quite small so this should be - at some level - considered as a stage using high currents. After all, it is a small power amplifier driving the unit.

Front panel:

The section on the left (by left I mean the side housing that "spring reverb case" ) interests me and a better picture of it (or a set of them) would definitely help. I assume that this is the section which ties the circuits together. Looks like three wires run out from it - can you identify them? The section at the far left looks like the switch you installed but there's something under it as well. Now, I saw some pictures of this amp while looking for the schematic and recall that this place holded direct inputs to power amp - is this true? I can't seem to locate the (ebay) site with those pics anymore. Looks like you lifted one jack off to house the switch.... make sure the jack can not short anything or that removing it has not ruined any important continuity. If this indeed is a direct input to power amp the next thing is important: Do these jacks hum as well. If they do is the hum amplitude equal, smaller or higher. This might tell you a lot about where the hum source is located.

A mains wiring near small signal stages looks like trouble but if it induces hum it should - again - do it all the time. You have also removed your switching arrangement once and it did not make any difference. I guess you can pretty much ignore any effect of this wiring.

The rear panel:

Three wires run from the transformer so I strongly suspect you have a center tapped configuration. Now, I see only two diodes - might be a single supply configuration or dual with half wave rectification. If it's single I see no output capacitor - unless it's the other one of those huge caps. Enough assumptions, the picture does not show all the details and forces to guess. You can easily measure with a DMM against ground to find out. If the configuration is dual then the number of wires (three?) does not sum to equation of 2 supply wires, and at least 1 ground and 1 signal wire. (Note: Unless the preamp uses single supply). Identify these wires, it will help you - if not now, then at least in the future if you ever have to work on this amplifier again.

The two jacks between PCB and transformer - are they speaker jacks? If so, the point where they connect the PCB is assumably the main ground point - at least reasonably near it anyway. The PCB spot would fit perfectly.

Those two green wires, are they for the lamp? Interesting sections here (I need more pics) are the bundle of wires near the place where the power switch is located and the three wires leaving the PCB and connecting the other one at front panel. I assume one of these wires must be the ground return of the preamp, trace out where it connects. (Ideally it should run straight to ground without overlapping any high current return path.

The safety ground: Only in rear panel? Use a very thick (grn or yel/grn depending of the country code) wire to hook it up to the front panel as well. If you use "fork end" in the wire you can mount the other end, for example, under any of those screws holding the PCB on place. If there indeed is not any other ground return between those two panels except a thin wire - which also runs through PCBs - I do not consider it as adequate "safety grounding". All conductive places that the user can touch should be safety grounded.
Now, adding this wire could cause a ground loop - as well as prevent one - so finding an optimum hookup place is essential. For safety point of view, optimum is near that switch you put to the circuit. I guess this is also the shortest place to connect the wire. Even better if that lamp is near this place as well. See, I guess the lamp is hooked to mains potential in parallel with the transformer (a usual configuration when leds were expensive). It might have a resistor for voltage drop and current limiting but "mains potential" is the key word here. Should any of those wires ever come loose and touch the chassis the current would find the least resistive route to ground instead of running through the fragile PCB traces and thin wire connecting the boards together.