Loading one side of a bipolar supply?

[therecordingart]

So I'm trying to adapt a stomp box to the API 500 series. The power supply is a bipolar supply ( 16+ | 0 | 16-). I don't need the bipolar supply and only want to use the 16v+ side of the supply. In my mind, I feel like the 16v- side needs or wants to see a similar load to keep the power distribution at equilibrium. Otherwise, an API rack full of modules will have a larger load on the 16v+ side than the 16v- side if my module was loaded into it. I'm ignorant to this topic and would like a little help.

[tubegeek]

If the PS can handle differing numbers of modules loaded into the lunchbox, it can handle differing loads on one polarity vs. the other, is my guess. The answer would depend on the regulation of the power supply (that's a spec, % difference between loaded and unloaded output.) The PS probably uses active regulation so the output will probably not be an issue. If it is an unregulated linear power supply I'd be VERY VERY surprised, but that is possible, and the unloaded side would be slightly higher into the other modules that do use that polarity.

But individual modules may contain their own regulation and if so the PS output will not be a problem even if it were unregulated.

Modules that are looking to use the highest rail voltage available might not re-regulate because they depend on the higher rail voltage for full signal headroom and don't want to drop any of that voltage to provide the headroom for the regulator to operate.

So, in short: it depends. But you're probably quite safe.

[PRR]

> a larger load on the 16v+ side than the 16v- side

So?

I drive to the store and buy a six of beer. Do I sit exactly in the center of the car? No, and obviously Honda didn't think I should, I have to sit on one side to reach the controls and strap myself. Do I put the 6-pack right on the center hump? Maybe, but probably not.

I am in the US and my house-power comes as split-240V, 120V on each side. When I turn-on a 120V lamp, must I also turn another lamp on the other side of the power line? No; in most houses you wouldn't even know one side from the other.

I have a plank on a rock, teeter-totter. If I put the 80-pound dog on one end, do I have to put another dog on the other end? Well, maybe yes; this dog would spook on a tippy-plank. However if I want a stable porch, I use three rocks under the plank, and then I can throw dogs on one side or the other (or both).

Most audio bipolar supplies are built as two separate supplies tied together. They are, in effect, "three rocks". There is a maximum load (don't put a moose on my plank), but any smaller load can be put on the left or on the right, on the + or on the -, without concern for "equilibrium". Adding dummy loads just makes useless heat.

[anotherjim]

It depends on the power supply. If it's got a transformer with 2 separate secondaries or a single with centre tap and full wave rectification, the power rails are almost as good as independent. If it's the old trick of only half wave rectification with a centre tap ground, it's going to be unbalanced. That's unlikely with any recent designs - it was only done in the days rectification was a significant expense. That said, you can find half wave designs out there, but I wouldn't expect it in pro-audio (hah!)

Modular synth racks use balanced power too, but a lot of modules have circuits that mostly load the pos rail.

A stomp circuit's unlikely to take enough current to cause any issue.

With high currents, it could be a problem, if the chassis ground distribution is poor and a common 0volt rail exists (you don't see separate 0volt in most rack distributions). A balanced supply with equal loading has zero net common return current. In the unlikely event of it actually having exactly equal loading, the 0v common return to the PSU could break and you'd never know unless it also broke the screening ground and everything starts picking up noise. Then if the common PSU ground return has a fault causing some resistance, say 1ohm, and the load was unbalanced, you get a "ground lift" fault. If only the pos supply was loaded, say 1A, then the 0volt at the load will be raised 1volt.

I've seen this cause a real problem. Production manager wants a "goods inward" test of a balanced PSU. Just a simple go-no go indication that the rails are in spec under the real loading. Test equipment engineer comes up with a test box containing dummy loads and comparators driving LED's. Loads are unequal, the product takes 4A from the pos and 2A from the neg, although the the PSU is rated 6A on both rails. A cable is made to connect the PSU to the test rig, but this is far longer than in the product because the PSU will be under a perspex safety cover; out of the way in case it blows electrolyte from it's caps. Same gauge cable is used as in the product.
Result - no PSU will pass! It's a stupid ask! Unbalanced return current caused ground lift and no way would the comparators ever see in-spec voltages. Solution - equal the dummy loads to both at 4A. Of course, in the assembled product, the PSU cable was quite short and it caused no significant ground lift with unequal loading.