DIY Power Supply Thread - Questions and Help

[analogoutput]

@TimMJN

I noticed the hum emerged whenever I plugged two modules from different PSUs into eachother, so my immediate thought was: ground loops.

Just to check, do you get the hum if you just touch a cable from one jack to the other? That is, if you connect just the ground from one module to the ground from the other?

And it’s only with modules on two different PSUs, right? Any two PSUs?

With no modules plugged in, is there a direct connection between the grounds of the PSUs? If you check continuity between the center pins of one of the 16 pin headers on one PSU and those of a header on another, is it there?

Does the hum change if after plugging in the cable between two modules, you move it around, twist it, etc? Does it depend on the length of the cable?

@TimMJN

Here’s the things I will try in the coming weeks:

• replace the transformer with a centre-tapped one
• replace all wiring with 15 AWG
• create more space between 12VAC and GND wires

One other thing to consider trying is building a Faraday cage around the transformer and mains input.

Added: If it is the case that you get the hum if one module is plugged into PSU1 and the other into PSU2, but not if both are plugged into PSU1 or both are plugged into PSU2, then one possible culprit is the part of the loop formed by the long-ish leads you have running from the transformer to the PSUs. You might try having these leads branch out close to the PSUs instead of at the transformer — maybe just run 2 wires from the transformer to one PSU, then 2 wires from there to the next PSU, then 2 wires from there to the third. (Like this.) Or it might work to keep the separate wires from each PSU to the transformer but put them parallel and close to one another. You might also consider putting the PSUs closer to each other. The point being to minimize the area between the ground wires between one PSU and another. You can try running the 12VAC wire far away or not and see if that matters. My guess is it won’t matter much once you’ve minimized that loop but I could be wrong. Getting rid of hum basically involves trying every (safe) thing you can think of, whether you understand what good it’ll do or not, until it goes away.

[TimMJN]

Thanks for the great reply!

@analogoutput

Just to check, do you get the hum if you just touch a cable from one jack to the other? That is, if you connect just the ground from one module to the ground from the other?

Yes I do. Also when I e.g. connect the ground of one PSU to the scope or mixer ground.

@analogoutput

And it’s only with modules on two different PSUs, right? Any two PSUs?

Yes and yes

@analogoutput

With no modules plugged in, is there a direct connection between the grounds of the PSUs? If you check continuity between the center pins of one of the 16 pin headers on one PSU and those of a header on another, is it there?

All ground are tied together on a terminal block, so there is continuity. But, as you mentioned, the leads are kind-off long, and wires on the thin side. I also made a ribbon cable which connects the bus board grounds together (ie only connecting pins 3-8). When I plugged that in, I could hear the hum level drop. Not disappear, but a small noticable drop.

@analogoutput

Does the hum change if after plugging in the cable between two modules, you move it around, twist it, etc? Does it depend on the length of the cable?

I’ve also suspected bad cables, but couldn’t find any evidence for that. Tried different cables and moved them around. I didn’t check for the influence of patch cord length though.

@analogoutput

One other thing to consider trying is building a Faraday cage around the transformer and mains input.

I thought about this. You’re a physicist, I’m just a measily chemical engineer (), so please correct me if I’m wrong. Hum due to ground loops is a magnetically induced effect, right? If I remember Young and Freedman correctly, a Faraday cage blocks electric fields, but not magnetic fields. Unless I enclose it with something like mumetal, but .

@analogoutput

maybe just run 2 wires from the transformer to one PSU, then 2 wires from there to the next PSU, then 2 wires from there to the third

That’s a good suggestion, I will try this! That’ll basically transform it from 3 separate loops into 1. I will have a nice rewiring session soon.

Thanks again!!

[analogoutput]

@TimMJN

I also made a ribbon cable which connects the bus board grounds together (ie only connecting pins 3-8). When I plugged that in, I could hear the hum level drop. Not disappear, but a small noticable drop.

Seems like that’s roughly equivalent to my wire shortening suggestion, though with the original larger loops still (I assume) present. So there’s hope that might work.

@TimMJN

If I remember Young and Freedman correctly, a Faraday cage blocks electric fields, but not magnetic fields. Unless I enclose it with something like mumetal, but .

Not quite; it won’t block static magnetic fields. (That’s what mu metal’s for.) It will block (sufficiently low frequency) electromagnetic radiation.

[TimMJN]

Well, I spent the better of part of today rewiring and testing my PSUs. There’s no solution (yet), but I seem to be getting closer to the cause.

I started off by taking out the 2/3 PSUs, and properly wiring the remaining one with 1.5mm^2 wire. After hooking up the scope, this is what the 12V supply rail looks like:

As can be seen, there’s a ~30mV 50Hz ripple on the supply. There’s also an opposite ripple on the negative supply. I ran some tests by plugging in a sequencer and looking/listening to the output. What should be a steady DC voltage, has that same 30mV audible ripple on it. From the tests where I have connected only 1 PSU, 1 module, and where I’m listing to the signal through headphones, I have concluded that the problem is NOT one of ground loops. In this situation there are no ground loops present, still the hum persists.

Okay, so clearly the problem is a supply ripple, not a ground loop noise. So far, I have used only 1 of the routemasters AC inputs. Getting two AC inputs with opposite phase will cut the ripple by half, right? So, as my next step, I replaced my single-secondary transformer with a centre-tapped one. After wiring that up, this my supply rail:

Oh god, it got worse. Now the sample ripple is there, it just doubled in frequency. Well this is just very, very odd. Let’s look at the input of the regulators:

The problem lies not in a sagging input of the regulator. Instead, it’s somehow unable to deal with the peak of the input. Is it just too slow to respond to the transient here? I was very surprised to find this.

Just to verify, I wired up my previous PSU, a FC microbus, and found a fairly clean supply with ~2mV ripply. No audible hum on the DC output.

So it seems the problems lies within the regulators, or the surrounding components. I will go check my boards and components, maybe see if I can swap out the regulators. These are the exact ones I used:

LM317T STMicroelectronics - IC: voltage regulator | linear,adjustable;...

LM337TG ONSEMI - IC: voltage regulator | linear,adjustable; -37÷-1.2V; 1.5A;...

From a quick look at the datasheets, I can’t see any reason why these would not be suitable. If I’m overlooking something, please tell me.

So yeah, no solution yet, but I seem to have found the problem.

[fredrik]

What are the input levels? Are you dropping out of regulation?

(317 has better specs than 78xx but iirc it has a slightly higher voltage drop)

[TimMJN]

I’m rectifying 12VAC, so that gives me 12V*sqrt(2) = 17V. Minus a forward rectifier drop is rougly 16V. The 317 datasheet mentions a minimum regulation Vin-Vout=3V. So I should be on the safe side. Also, if that was the problem, shouldn’t I see the ripple at the lowest point of the input? I see it at the highest.

Thanks!

[analogoutput]

Couldn’t the high point be where it’s going into regulation, not out of it?

If you haven’t measured your nominal 12VAC, that might be a good thing to do.

[TimMJN]

@analogoutput

Couldn’t the high point be where it’s going into regulation, not out of it?

Fair point. I just went ahead and measured, my transformer outputs ~14.5VAC, the output of the rectifiers read +/- 19.5VDC. This doesn’t seem to be the problem…

[analogoutput]

Have you done these tests with all three PSUs? Do they all behave the same?

If so it’s presumably something in the design, unless it’s a fault in the PCBs. FC doesn’t show schematics that I can find but it seems the Routemaster and Microbus are broadly similar, for instance they both use 6 x 4700 µF to filter the 12VAC. They differ in that the Microbus uses LM7812/LM7912 fixed regulators while the Routemaster uses LM317/LM337 adjustable regulators. I don’t know why the latter would perform worse… unless the support circuitry is wrong. I’ve previously noted the capacitors used with the regulators in the FC Power are mysteriously orders of magnitude different from what the datasheets recommend, but whether that’s the case with the Routemaster too and whether that has anything to do with your problem, who knows? But if all three have the same problem I’d say the regulators and support circuitry are where to look. (LM317 datasheet)

[TimMJN]

@analogoutput

Have you done these tests with all three PSUs? Do they all behave the same?

I tested four PSUs built with identical parts, all show identical outputs. So a one-off soldering error or faulty part seems unlikely.

@analogoutput

They differ in that the Microbus uses LM7812/LM7912 fixed regulators while the Routemaster uses LM317/LM337 adjustable regulators.

Yes, the difference between the Microbus and the Routemaster is LM7X12 vs LM3X7, and 2 vs 4 rectifier diodes (the Microbus doesn’t support centre-tapped transformers).

I’ve traced out the schematic of the routemaster. I see little qualms with posting it here, as it basically just the circuit from the LM317 datasheet. I don’t see anything out of the ordinary…

I’ve done some more testing, but haven’t been able to make much sense of the results.

Here are the +12V and -12V supply rails. There is quite some higher-frequency noise, but this seems to be mostly caused by my improvised scope probes. There’s a peak in the rails at every peak in the AC waveform, but one AC phase results in higher peaks than the other. Also, the effect is stronger in the positive rail than in the negative.

Here’s the input of the +12V regulator and its output. The same alternating peaks are present at the input, at roughly twice the amplitude.

And here is the output of the regulator and the input of the rectifier. Notice how the peaks in the input are much more narrow than the AC waveform. The AC waveform seems to have an odd flat top, but from zooming in on it, it seems that this is mostly an effect of my scope.

Here’s that zoomed in picture. There is definitely something fishy going on here, notice the distorted AC waveform.

Here’s that same zoomed-in picture without the PSU conected. No distorted waveform here.

It seems to me that I have all the evidence here, but I just can’t piece it together to find the underlying issue. Something related to not actually being an electrical engineer, I guess. My current take is that the root issue lies with the regulators, which appear to be shorting their input to the output at some point. This then in turn distorts the voltages measured at the rectifier input and output. I might, however, be very wrong…

Cheers!

[analogoutput]

@TimMJN

I’ve traced out the schematic of the routemaster. I see little qualms with posting it here, as it basically just the circuit from the LM317 datasheet. I don’t see anything out of the ordinary…

This is similar to the FC Power in that the output capacitor (C6/C13) is 100 times larger than what’s shown in the datasheet. I have no idea why or what the effect would be. Datasheet says “C_O improves transient response, but is not needed for stability.”

The datasheet calls C7/C8 “C_ADJ” and does not supply a value for it. “C_ADJ is recommended to improve ripple rejection. It prevents amplification of the ripple as the output voltage is adjusted higher.”

[TimMJN]

@analogoutput

The datasheet calls C7/C8 “C_ADJ” and does not supply a value for it.

If you check the example in Fig. 13, they indeed suggest 10uF for C_ADJ.

@analogoutput

This is similar to the FC Power in that the output capacitor (C6/C13) is 100 times larger than what’s shown in the datasheet. I have no idea why or what the effect would be.

Yes, though even throughout the datasheet, various values are used. E.g. Fig. 23 has 47uF marked.

[TimMJN]

By the by, I’ve asked Rick Holt (FC) about this in a Facebook post. Because, as much as I hate to admit it, I’m kinda at loss here…

[zorch]

A quick question about heatsinks:

Are these at all adequate? The size of the footprint on the PCB is giving me pause. If I do need to replace them, I would rather do it now rather than later.

[TimMJN]

I think this should generally be fine, depending on the number of modules you’re planning on hooking up. The PCB space seems more of a one-size-fits-all. But is that a rubber ring I spot there?

[TimMJN]

The plot, it thickens…

I decided to use the one remaining empty PCB I had to pinpoint the problem. Build up the different elements of the circuit one by one, and see where the problem emerges.

First up is just the rectifier. Four diodes, nothing more.

I’m sorry if you’re colourblind like me, but this is exactly what I expected. The AC inputs, but rectified.

Okay, next I added the 6 4700uF filter capactors, 3 on each rail. Now I’m measuring just two DC voltages, as there is no path for the capacitors to discharge through. So to see what they would do on a small load, I added a 1k resistor across each rail. This is my circuit now:

I’d expect my output to look something like this:

(courtesy of https://www.capacitorfaks.com/post/the-role-of-capacitors-in-ac-filter-circuits)

with the ripples at 100Hz and a rise time comparable to the AC waveform. The typical amplitude of the ripple I’d expect at around Vripple = Vpp(1-exp(-1/fRC)) = 14mV.

I think you can feel it coming, it’s not what I found. Here’s the two rectifier outputs:

What. The fudge. There are a few weird things here. The peaks are at 100Hz, but they are alternatingly large and small. It’s more of a 50Hz two-stage ripple. The positive and negative ripples differ. The negative ripple is in the right ballpark, but the positive one is much higher than expected. Also, those peaks on the positive rail are múch faster than the rise time of the AC waveform. And they don’t fall exponentially like expected, but have steep drop first.

So… are my filter capacitors just sh*t? These are the units I used: EWH1EM472L30OT AISHI - Condensator: elektrolytisch | THT; 4700uF; 25VDC; Ø16x30mm; ±20%; CE-4700/25PHT-Y | TME - Elektronische Componenten . I don’t see anything out of order on their specs. On the previously mentioned, fully functioning Microbus, I used these: SD1E478M1631MBB SAMWHA - Condensator: elektrolytisch | THT; 4700uF; 25VDC; Ø16x31mm; ±20% | TME - Elektronische Componenten . Apart from the operation temperature, the specs seem exactly the same.

Time to order some new capacitors, it seems…

[update] Swapped out the capacitors, no change whatsoever. Something something wit’s end.

[zorch]

It’s an isolator. The heatsinks came with them as well as a nylon bushing for the screw. I figured I would use them as I didn’t have any spare thermal paste laying around.

[JaggedNZ]

What are you connecting ground to on the transformer?? If it’s connected at all? Otherwise I’d suggest disconnecting the gnd wire from the transformer and seeing what happens.

[TimMJN]

@JaggedNZ

What are you connecting ground to on the transformer?? If it’s connected at all?

The transformer centre tap, PSU GND and protective earth are all tied together. I don’t see how disconnecting the transformer would work? That’ll just leave it floating…

[JaggedNZ]

Sorry I’ve been working on a single tap transformer supply and answered a bit hasty.

So your center tap is connected to mains earth? Try disconnecting that, you might be shorting your transformer to neutral depending on wiring standards in your location.