GGG bipolar power supply help

[blues123]

Hi,
Just finished building the GGG bipolar power supply (15v version using 7815/7915). The purpose of this build is to power an old EH Memory Man (the old blue one with echo/chorus switch and extra boost input) that has a blown internal transformer and NE5554N regulator. These old regulators are difficult to find, so being as the transformer was faulty as well, the best option appeared to be a 24V AC wallwart and the GGG PSU.

I identified the ground, -15V & +15V points on the circuit board and cut the traces to the old PSU circuit on the board. I connected my bench power supply to the 0,-15 & +15 points on the board and bingo, the old girl fired up and works a treat with warm chorus and echo.

Just a case of substituting my bench supply for my new wallwart/GGG PSU combo, or so I thought. Before connecting the GGG circuit, which I built on vero, I tested the voltages and all seemed well with -14.8 and +14.8. However, when I connected it to the board all is not well. I do get chorus and echo, but also some strange artifacts as well, in some positions in both modes I get strange oscillations and noises and when I hit the footswitch, I can still hear the echo slightly. Non of these problems are apparent with the bench supply.

I thought it was my build so I have completely rebuilt the cct on perfboard using the board layout on the GGG sight, I checked all the capacitors on a tester (electro 1000u and film .1u) along with the two diodes.

For the record, the layout shows 0.01u film caps when the schematic shows 0.1u. I have used 0.1u. Not sure which is correct or if it would make difference anyway.

http://www.generalguitargadgets.com/pdf/ggg_bipolar_ps.pdf

If anyone can see where I am going wrong or if I need to add more circuitry, please advise, as this is driving me insane

Thanks for listening....

[R.G.]

There is a problem, but it's not yours. I think you did the power supply correctly.

There is a compatibility problem with that power supply and some ICs. Being as it is two half-wave rectified supplies, with one half necessarily before the other, the two power supplies +V and -V rise from zero volts non-simultaneously. Some ICs don't like this at all. Back when memory chips had three power supply voltages, they had to be applied in the correct sequence or the chips would self destruct. Even today, some ICs will refuse to operate correctly if power is applied non-simultaneously. Not all, but some. So I think you have a power supply sequencing problem. One voltage rises appreciably before the other polarity even starts up. By the time the other side comes up 16ms later, the ICs are already in some mode that's not good and that they don't get themselves out of.

When I ran into this, I concocted a circuit that had two transistors in series with the input voltges to the regulators that were left off initially. There was a third transistor that only turned on when one of the voltages was over half of its full voltage. That ensured that the other voltage was at least partially up when both raw DC voltages were allowed into the regulators. This seemed to work well, and with the delayed turn on setup, my opamps never got funny. Without the delayed turn on, they didn't operate right about half the time.

There are other solutions, I'm sure. But I do know that when the circuit does not work with a circuit like the double halfwave, you need to change the power supply.

[blues123]

RG, you are the man.

Thank you so much for the speedy reply - I thought I was going mad!!

If I understand you correctly, a full wave rectified cct is the way to go then to avoid my problems. Do you think this one will be more suitable?

http://www.nutsvolts.com/~downloads/January%201998%20Thomas%20Henry%20-%20Power%20Supplies%20for%20Electronic%20Music.pdf

Thanks again

[R.G.]

If I understand you correctly, a full wave rectified cct is the way to go then to avoid my problems. Do you think this one will be more suitable?

Not exactly. A full wave centertapped power supply is ONE way to go. But you can also use what you have, given only that you do something to let the power be up on both positive and negative sides of the power supply when it's applied to the circuit.

That could be one of the following:
(1) the automated transistor thingy I mentioned.
(2) a 555 running as a delay that turns on either two transistors, one for + and one for -
(3) as in (2) but with a small DPDT relay
(4) A manual switch. Have one switch for AC power, and the second one actually applies power to the effect board, No matter how fast you are, if you have the "power to the circuit" switch off when you turn on the AC power switch, both + and - will be good enough when you turn on the DC switch. Or have a manual switch that's always off when you plug in the AC power, and no AC power switch.

Anything that does a little delay works.

[blues123]

Thanks again R.G.

I'll give it a go and see what happens.