PSU schematic

[Diabolical Artificer]

Working on sorting out a noisy PSU for an Alesis Quadraverb which involves tightening up the laminations and a few other bits. in the past it caused EMI with recording software so I put the tfmr in a metal box ( old PC PSU) and used a screened lead as the power lead. i did this in the past before I knew about split supplys, so followed the layout as was.

The tfmr has two secondary's - 0 - 9V 0 - 9V, I'd wired them in parallel. Checking over the PCB yesterday I saw it uses op amps which usually use split rails, sure enough it does, but there is only a two wire lead, so no centre tapped AC supply. I found a schematic for a later version of this FX box which uses two caps in series connected in parallel with the AC in, so that's how they're getting the minus DC rail, surely this is daft as the transformer secs could be wired in series to give 9v 0v 9V AC. It uses a four pin din plug for power too, so no worries there.

So what are the disadvantages of doing things this way? Why bother with a centre tapped secondary to get a split rail if all we have to do is use a capacitive or resistive divider to get split rails? Secondly, would there less power available from the same tmfr if the secs were connected in series?

Andy.

PS, they've also made things tricky by making the chassis positive so I can't use a metal DIN plug as a shield.

[Nick]

I suspect they want the extra current on the +5V rail - more than one winding of that specific transformer would carry.

The +/-12V rails will draw tiny current for the opamps - the doublers will cover that fine.

Who knows?

[ppppenguin]

Same power available from a transformer regardless of whether secondaries are in series or parallel. If I've got a centre tap available then a bridge rectifier, +ve and -ve regulators and job's done.

The arrangement on your circuit only makes sense if the bulk of the current is from the 6V rail and the +/-12V rails only supply a small current.

[Terry]

I would expect the ±12V rails to be lightly loaded in comparison with the +5V supply and it is probably the current drawn by this that dictates that both secondaries in parallel are required to source it.

I find you earthing comment puzzling, Andy, as the drawing clearly shows the chassis as being the negative side of the +5V and +12V rails as normal. It is, of course, also the positive side of the -12V supply, as you would expect.

We can't see the entire circuit so, if the metalwork is bonded to one of the positive rails - which one? You don't say - at some point, all those earth symbols are very misleading!

[Refugee]

It also does not make all that much sense to me either.
The 12 volt rails are driven by half wave voltage doublers and only have little 78/79 regulators.
The big problem is the high power low voltage rail with only a half wave rectifier. The transformer would have to be de-rated to allow for this.

[Mark Hennessy]

This sort of thing is very common. It makes the outboard PSU very simple (and cheap!). The same PSU "brick" will be used for many models in the range, which saves cost (design and approval). It means a simple 2-pin connector (and single switch) can be used for power. Perhaps not the most obvious way to do it, but good production engineering.

Would I do it this way myself? Perhaps. Indeed, I have.

In this case, it means I can spec the transformer to deal with the 5V load, and not have to have extra windings for the +/-12V rails.

Alternatively, if I did use a circa 12-0-12 transformer in the usual way to make +/-12V DC (post-regulators), I'd have a problem with heat dissipation in the 5V regulator. A switched-mode DC-DC converter might not be welcome, and a custom transformer with tappings to do the 5V rail might not be desirable either.

Vive la difference

[ppppenguin]

I've used a different trick to split rails from an AC wall wart with just one winding. I've used a 12V regulator to make the "outer" rails, then a -5V regulator to split the supply into +5 and -7. When doing this sort of trick you need to be very clear about the relative loads on the 2 rails. Regulators can source current, they can't sink current. If necessary you preload one rail with a resistor to ensure this is true. The attachment is from another design I did that went into production. The 22R resistor was necessary and got quite hot. Might have been able to avoid it by using a pair of 7905 -ve regs but I'd need to think about that.

[Mark Hennessy]

The attachment is from another design I did that went into production. The 22R resistor was necessary and got quite hot.

Presumably the -5V rail was quite lightly loaded in that case? Might've been better to use a lower AC voltage suitable to the +5V load and a 2-diode/2-capacitor circuit to get the -5V? Done that a few times...

Yes, changing to negative regulators would be more elegant.

[ppppenguin]

From what I remember the -5V rail was quite lightly loaded. Unfortuantely we'd already done the PCB before I realised that a pair of 7905 would have been better. Glad I remembered the site for the resistor. I think we were already using the wallwart for another design so it made sense to use the same one again.

The problem with a 2 diode 2 cap circuit for -5V is that I think I'd have to have used a half wave rectifier for +5V, something I would rather not do.

As is often the case, there's more than one way to do a design. Some ways are better than others. I'd be a liar if I said all my professional designs were as good as they could be - I've made my share of mistakes and some of them have gone into production. Nobody died.

Once there was very nearly a lethal problem on one of my deiigns. Not my fault at all but still scary. I'd specified a mains transformer (something like a 12+12 30VA toroid) and the client found they could get a good deal from a smaller compnay whose main business was custom wound parts. One day their technician got a tingle from the metal case. It SHOULD have been earthed but for odd reasons it wasn't. A whole batch of these transformers had faulty insulation. So much for the transformer manufacturer having proper test procedures in place. Fortuantely none of the kit had been shipped to customers yet. While my client would have been first in line for the public liability claim if there had been a nasty accident, some flak could easliy have come in my direction too.

[Refugee]

I had that problem and blasted out a warning about being told to use an unsuitable transformer.
The job was to charge up 8 AH deep discharged 6 volt dryfit batteries and my spec was a 317 regulator with the thing thermally regulated by the heat sink size all powered by a 9 volt 20VA transformer.
I was told to use a transformer that was left over from another job where it was removed and the power taken from a type approved transformer that ran the rest of the kit avoiding running mains outside the internally live part of the housing. That was OK and it was correctly fused.
Using it on this charger was not on but my non technical boss objected and left me throwing a warning that the job would bounce.
When it bounced it came back accompanied by an insurance claim and the transformer I chose was fitted once the boss had learned the hard way and the job never bounced again.
I still have one of the "wrong" transformers and will de-rate it a bit if I do use it with only a 230 volt primary.