sawtooth-like ground noise from new modular system.

I've been piecing together a system and finally got a two-tier case today with a Mean Well RT-125B PSU that can supply more than enough power for the modules connected. I've quickly noticed that I'm getting a low-volume continuous sawtooth-like base noise coming through the Doepfer output mixer at all times, even if the levels are zeroed out on all modules. It quickly amplifies to interfere whenever I start patching things together. The out is a Doepfer Mini Stereo Mixer A-138s. This feeds into a Yamaha MG10x external mixer, which goes from there to both a stereo receiver and a PC/headphones. I didn't have this issue before when I was working in a 1-tier Synthrotek case, so ... I don't know, I'm too new to know where to start in terms of how to resolve it. I tried removing a couple modules; tried outputting to different lines on the external mixer; tried different cables to mix out with ... it seems that horrible low-volume sawtooth always seems to be faintly there, coming from within the modular build itself somehow. Is it some kind of feedback from the PSU? And if so what would I do about that? Other possibilities?

Thanks for any help anybody here can provide. Happy to provide any info if anyone has ideas but need to know more about the exact setup.

The out is a Doepfer Mini Stereo Mixer A-138s. This feeds into a Yamaha MG10x external mixer, which...
-- eexee

OK, let's stop there for a bit. For starters, there's a big level mismatch going on, since the A-138s outputs signals at modular voltages (up to 10 Vrms) and the MG10x has a 0 dB input reference at .775 Vrms (consumer level, which means that the input pres are NOT up to pro standards). That's a pretty wide difference, and probably plays a part in what's going on.

Secondly, since you mention that the sound appeared after shifting to the Meanwell-powered cab, are you 100% sure that the grounding for the internal P/S is set up correctly? Given that patching more modules yields more noise, I suspect that you've got a grounding issue, since the patch cables will tend to pick up more crud as they increase in number. Mind you, this isn't an issue unless you've got your AC supply going directly to the Meanwell. Also, by "grounding" I also mean that the patch panel itself cannot "float" relative to the internal DC ground, so double-checking the internal DC ground connections might yield a fault.

Third, if these don't work, the next suspicion would be a ground loop between the modular and some other device, particularly something digital given the character of the noise. Or it's also possible that what you're hearing is good ol' 60 Hz with a considerable amount of distortion from the level mismatch. You can test this by using a "ground lift adapter" if the Meanwell has a 3-prong AC input. Put that on the power cord, plug it back in, and if the buzz disappears, you DEFINITELY have a ground loop issue.

Another way to fix ground loops...and one I strongly urge if you're not living/working where balanced AC power is the norm (Europe, for the most part, has balanced power)...would be an isolation module that would go between the mixer's outputs and the inputs for the mixer. This also has the plus of these having level controls; something like Happy Nerding's transformer-isolated Isolator also gives you a ganged stereo level control to bring the level mismatch under control.

Lastly...and also a ground loop issue consideration...would be to "star-ground" ALL of your gear to a common ground point with some basic hookup wire. And even if you don't have an issue with this, it's still a good idea as it'll help lower your overall noise floor by draining any e-crud to that common ground, where it can then go right onto the ground leg of the AC. I should note that I do this in my studio even if there's no apparent noise/ground issues, as it has some very beneficial results that can wring out a bit more S/N, cleaning up the lowest digital levels.

That’s a lot of great information, thanks very much for taking the time. That’s very good to know about the mixer to mixer levels especially. I don’t really understand grounding very well, or what a groud lift adapter is or what star-grounding would mean. Before I left for work I removed a couple modules — a Doepfer BBD and a couple 2HP modules — and the noise was no longer present when I briefly tested it. But I didn’t have much time so, I’m going to test it out a bit more later tonight. ( It was a custom build, though the guy who built it just got an electrical engineering degree... )

So in terms of further general testing I put the modular through my Yamaha AX-497 Stereo Receiver and directly out to headphones from there ... I don't know if this is "professional" level equipment, but it's the only other real output I have for the mix besides the MG10Xu and my impression is it's a generally HQ receiver. Certainly cost way more than the MG10xu anyway! But anyway this in and of itself made a difference in terms of how easily noticeable the sound was, it was basically inaudible until about the 3 o'clock point on the rotary volume knob, whereas audio levels generally are about as loud as you can comfortably listen to by the 12 o'clock mark. However, the sound was still there, and as I was trying some patches and such, it seemed like it became an interference after a bit, like suddenly there was this faint background hum to the patch that wasn't there in the beginning just out of the blue.

So I began changing some modules in and out, double-checking my connections and whatnot, and eventually found a configuration in which the sawtooth-like noise was not there, and at the 9 0'clock and above level all I would get was a bit of white-noise/hiss from the receiver itself (which I'd expect with the volume at that level). Then I started experimentally adding modules back in ... and eventually would hit a point where the sawtooth-like noise came back. I switched a few things out wondering if it was possible a particular module could be causing it, and if so trying to pinpoint which -- but it seemed to vary and right now my impression is that it's once the power draw is past a certain threshold, that the sound becomes present.

I have quite a few modules, but I have an excel sheet calculating the power draw of everything against the specs of the PSU, and in theory none of the rails should be at more than 50% of what that PSU can supply power for even with every single module installed, soo ... in a way I'm back to square one, but until I learn a bit more about how to check or correct the PSU itself, at least now I have a way to mitigate the issue temporarily (which is basically to only use about half of my case's available space).

OK...to check the P/S, you're going to need the following:

1) A multimeter.
2) An oscilloscope.

Fact is, if you're doing electronic music, you should have these anyway. They're essential troubleshooting tools. Anyway, what you would do is to put the multimeter across the different 12V legs while putting the modules back in. Since you're apparently not close to the current limits, what you'll want to look for is a voltage sag. When the "offending module" gets connected and you see a drop in the voltage across (more than likely the +12V) the leg in question, you've found the issue. You might also consider swapping the ribbon on that module out, as there might be a connection and/or continuity issue there. Also, check the pins on both the busboard and the module to make sure one isn't bent, causing a potential shortcircuit.

As for the o-scope, again, connect this across the DC legs one at a time while adding modules. You'll set the scope for the proper voltage coming off of the DC leg, and start with NOTHING connected to that DC leg (except the scope) so that you can check for any indications of noise, ripple, or other crap while the P/S isn't under load. If you see anything other than a dead-flat line once the o-scope is connected and set, then the Meanwell itself is defective. But if you don't see any anomalies with no modules connected to the DC leg, start adding modules one at a time...and at the point where the dead-flat line starts to show crud, you will have found the problem module.

And yeah, it's tedious AF...but this is also part of the "bargain" with modular (or most any other electronic music) gear.

One other thing to check would be the busboards, particularly if you have filtered busboards in the build. If something in the filtering (like a capacitor) has failed, that could also cause this. For this, go back to the "nothing else connected" step above, then disconnect the Meanwell supply from the cab's DC busses altogether, and check the Meanwell again. If the P/S output shows as flatline-clean without the busboards, then a filter component failure is the culprit here. And like the modules, connect the busboards one at a time until you find the crud.

So, the overall sequence of events is:

1) Disconnect the busses from the P/S
2) Check the P/S with the multimeter to determine if there's an undervoltage or overvoltage. If there is, check and see if the Meanwell has a voltage adjustment. If this is set wrong, use the multimeter to reset the DC voltage to a dead-on 12V. But if there's no voltage adjustment, then the Meanwell is probably defective. If not, then --
3) Check the P/S with the oscilloscope to check the DC purity. What you should see once the o-scope is set properly for the P/S leg voltage is a ruler-flat line on screen. If you see ripple, noise spikes, etc, then the Meanwell is...yep...probably defective. But if you get a flatline response, then this is good. So next, then --
4) Connect each busboard one at a time while monitoring the oscilloscope. If there are no anomalies in the DC, then --
5) Connect modules to busboards one at a time while monitoring the oscilloscope. This time, when you encounter anomalous behavior, note the offending device, remove it and its ribbon, then continue reconnecting modules. Each time you see crud on the DC leg, set that module and ribbon aside as suspect. Continue connecting modules to busboards until you have the build populated with the "clean" modules. Then --
6) Swap out the ribbons on the "offenders", then connect them to the busboards again. If the P/S crud goes away when each one is reconnected, the problem was in the ribbons. But if the crud comes back, the module you'd just reconnected has a fault. At that point, you can either contact the module's manufacturer, or give the module a good look-over to see if there are any solder joints that appear lumpy, grainy, slopping over other traces, etc. If you feel confident about your soldering skills, you can try and correct these...but otherwise, take note of which solder joints look bad and include info about these when contacting the manufacturer.

Now, if the Meanwell IS defective, then you've obviously got to replace it. But if this is necessary, one thing I would strongly suggest would be to replace it with a LINEAR power supply. Yes, they're more expensive. Yes, they weigh a ton. But because of their circuit topology, they CANNOT impose noise signals onto the DC busses because linear supplies divide DC via a transformer (while still in the AC domain) and NOT via switching circuit schemes which tend to have more noise, usually in the ultrasonic range. Both of my modular systems (Digisound 80, AE) use linear supplies...and in fact, the Digisound's was specced by the man, myth, and synth legend Kevin Lightner, who definitely knew his power supplies. Plus, my power solution for the AE system was to utilize a lab-grade linear Tektronix supply to feed its 9VDC inputs...again, these do amazing things for synth audio quality, stability, and getting rid of noise. If the Meanwell has to go, this is the better choice anyway.

Thanks again for taking the time to walk thru this... it’ll take me a minute to absorb and try any of this. But the mindset I’m taking is as you say — it’s frustrating, but a part of the decision to get into modular to begin with was to learn more about this stuff and not be intimidated by working w electrical components directly. So it’ll be edifying in that respect if nothing else. I have a couple multimeters, but no oscilloscope. Any you recommend? Also, do you have a link to a commercially available PSU you’d recommend for modular? Can’t afford either at the moment, but even if things work out w the Meanwell, once I move into my new house, I plan to get some woodworking tools and build my own custom case later in the year.

As far as linear P/Ss that'll do a good job with modular in general, keep an eye out for some used Acopian ones. You might also try getting one new, but keep in mind that that'll cost more.

O-scope? Check this: puAAAOSwXzdgXmcI" target="_blank">https://www.ebay.com/itm/174707999737?hash=item28ad6883f9puAAAOSwXzdgXmcI Notice that this also has X-Y capabilities, which is a must; you can then not only use the 'scope for working on gear, but also to check mix phasing via the X-Y display. Check this article out: https://transom.org/2016/its-a-phase-youre-going-through/ So even if that GW Instek isn't what you're looking for, keep in mind that you want a DUAL-trace scope that can do that X-Y display for stereo phase checking; if you do any sort of banger tracks, knowing what your low end is doing is essential, especially if you're planning to press a track to vinyl.

Going thru some of these steps after checking the flying busboard connectors themselves ... all the ones I checked seemed to register +12, -12, and +5 appropriately. Now checking independent modules but bear in mind that the buzzing noise would become present regardless of which modules I chose. It seems to be more related to whenever more than about 10% power is being drawn on the 12v rail or 20-ish% on the negative rail, so far I’ve been unable to link to it a particular module but been able to replicate the buzzing w a variety of different modules, once there are enough of them. Nothing is drawing on 5 at all which someone suggested to me could be part of the issue since it’s a Meanwell psu and they have some issues if nothing is drawing on 5 i guess. But i dunno what i’d do about that —- I have no plans for any modules that would draw on the 5 at present.

Yeah, just to confirm -- after further testing it seems very much to me that it isn't linked to a specific module. At one point I thought I'd isolated it to the 4MS PEG, because when I turned off the "cycle" buttons, the buzzing stopped. But further experimenting and swapping of modules made it clear that the reason it stopped was simply because the power draw of having the LED buttons in the "lit up" state on that module were pushing the overall load above whatever exact threshold it is where the buzzing starts. I was able to take out a module or two and then operate those buttons w/ no noise issue, or remove that module and still replicate the buzzing if enough other modules were added. But again the issue is that the power draw of my modules is nowhere near the specs of what the PSU should be supplying, I'm below 25% load on both +12 and -12.

Out of curiosity, is it possible to repurpose something like this as a PSU for modular?

https://www.amazon.com/Kungber-Adjustable-Switching-Regulated-Adjustments/dp/B08DHZ2X2Q/

Well, I use a Tektronix that's rather like that (but beefier, plus it's a linear supply) for my AE system, but you'd need something that has discrete +12 and -12 outputs, and I'm going to venture a guess that that Kungber probably doesn't, despite the panel markings.

THIS, however...https://www.mouser.com/ProductDetail/Bel-Power-Solutions/HCBB-75W-AG?qs=gCHJnwMCk7Sa%2FhsO6vvSjg%3D%3D Well, it's a linear supply, outputs +12 and -12 at 1700 mA and 5 at 6A. More expensive than a switcher, too. But it's also pretty much the last power supply you'll need for this cab. And if those current specs are still too low, there's a couple of other models that up the +12 and -12V rail current.

It's also worth noting that these Bel units look an awful lot like the OEM ones that Synthesizers.com uses...

What makes you say that the Kungber doesn't have those features "despite the panel markings"? It says it's a linear PSU in the description. Just curious.

Anything available via Amazon PSU-wise that you'd recommend? Or Sweetwater or wherever? I mean the lead time on the one you linked is 38 weeks (!).

Thanks!

Also as a general update -- still no solution. The connections all seem to yield the appropriate power via multimeter, (I have no oscilloscope). The guy who made the case suggested removing the 5V connection from the PSU and taping it off -- none of my modules use it anyway, and the only way he was able to replicate similar interference on his end with same PSU, was with a slight negative power draw on the 5V rail. So I tried that, but it made no difference. Trying to isolate it to a particular module hasn't been fruitful either -- the sound becomes present regardless of which modules I put in, just once there is "enough" of them to (I presume) create a certain power draw level at which point this buzzing noise starts up. I got a recording of the sound today. It's amplified a bit of course, it's subtle at normal listening levels, but problematic if you try to build patches over it.

https://www.dropbox.com/s/gu3fjwnf91vlgfj/modular_sound.mp3?dl=0

Using the same modules in my Synthrotek case creates no such sound, though it is a smaller case with only 1 row (it's the case this new one was intended to replace).

If you've got a second power supply, even if it's underspecced for the time being, try connecting its DC legs to the busses in the new cab after you disconnect the Meanwell that's already there. Then just put in a few modules that appear to be the "usual suspects" here and see if the buzzing continues.

The more this goes on, the more suspicious I'm getting that it might actually be an issue with the filtering on a busboard. By checking these using the older (and obviously functional) P/S, you might be able to eliminate the Meanwell as a source of the noise if the buzzing happens on the other supply.

Well, there are two flying busboards attached to the Meanwell. But thank you for directing my attention thataway, because this morning I finally made some progress towards isolating the problem and solving it.

I attached a row of relatively power-hungry modules to just the top busboard and tested it out. Def a bit more ground noise than I'd like, leading me to think a PSU and overall busboard upgrade/reconfiguration may be a priority regardless, but -- no buzzing/interference sound, and generally work-with-able.

When I attached the same set of modules to just the bottom busboard, I did get the interference sound. Strangely it didn't sound quite the same as what I recorded, but was still hugely noticeable in contrast to the first busboard.

So, my thought is to find I fresh flying bus (I think I have one somewhere) and replace that flying bus. Hopefully that will work and if so this setup will be usable enough until I can afford a more complete overhaul.

Eh ... I spoke too soon.

I'm really stumped now. So after that last post, I replaced the 2nd flying bus with a fresh spare one I had, thinking I had it in the bag now -- but when I added my 2nd row of modules back in ... once again, I got the buzzing.

This leads me back to wondering if it's the PSU somehow, and related to a certain amount of power draw. But if so, why on earth does it have no issue running the same exactly modules on one flying busboard, but not another? Unless I have the bad luck of the spare one also being bad somehow? I might have one other, maybe I'll have to try that. It just makes no sense to me.

Actually, you've managed to rule out the busboards AND found the culprit. It IS the power supply, and there's something specifically wrong with the second flying bus header!

Next step, therefore: pull the power supply and inspect the "buzzy" bus header. For starters, do you see anything inside the header, such as dust, maybe a tiny bit of metal, bent pin, etc etc? If you can get the supply housing open, do you see anything on the soldered connections to the header that resembles a cold joint, or maybe there's some other soldering fault that's having the same effect as a crossed wire from (best guess here) the Eurorack bus line to a power leg?

Yeah, it's confusing...but this is troubleshooting, that process of elimination of causes of screwy behavior. It's tedious, it's annoying...and it's invaluable, because this is exactly how you learn your rig. Look at it as the 21st century equivalent of being an oboe player and knowing that if you want to play at your best, you need to cut and tie your own reeds. What you're learning as you poke around in this thing WILL translate to being able to deal with other equipment faults in an effective manner, over time.

Sorry to sound stupid here, but I'm not sure what you mean by "inspect the buzzy bus header" or what to look for. There are a set of screws that both flying bus cables were manually wired to. Beyond that I'm not exactly sure what I'm looking for? This is the PSU:

http://download.siliconexpert.com/pdfs/2017/7/21/11/32/19/509/mwe_/manual/209402834386489rt-125-spec.pdf

Oooo...part of the answer appears to possibly be hiding on the first data sheet! Check "Note 7" at the bottom, then go to Meanwell's site and see what they say about RFI and noise filtering.

No idea what any of that means to be honest ... I ended up getting a 4MS Row Power and a pair of Synthrotek noise filtering busboards. Wasn't the cheapest solution, but everything's working like a charm now. Now I just need to replace that mixer......