15v dual power supply problem. need help

[poldus]

I built two units of the bass equalizer kit from elliott sound products that includes an on-board dual power supply.
They are working well except for the following glitch: The units will not work upon power-on. Ground to negative rail reads 15v on the voltimeter but ground to positive rail reads 0. Now, when I briefly shortcircuit + and - they suddenly power up and give a correct +15v. Each board has its own power supply but they share a common AC transformer.
Any hints? I sure don´t want to be jump-starting the units like this everytime.
Thanks

Here´s a link to the project that includes a circuit diagram
Eight Band Sub-Woofer Equaliser

[AndrewT]

check the AC voltage and check the three DC voltages at the PSU. Agh, half wave rectified.

[diy_audio_fo]

Originally Posted by poldus

I built two units of the bass equalizer kit from elliott sound products that includes an on-board dual power supply.
They are working well except for the following glitch: The units will not work upon power-on. Ground to negative rail reads 15v on the voltimeter but ground to positive rail reads 0. Now, when I briefly shortcircuit + and - they suddenly power up and give a correct +15v. Each board has its own power supply but they share a common AC transformer.
Any hints? I sure don´t want to be jump-starting the units like this everytime.
Thanks

Here´s a link to the project that includes a circuit diagram
Eight Band Sub-Woofer Equaliser

Try to put two diodes in backward direction on the regulator outputs (catode to +15V on positive output and catode to GND on negative output).

According to National data sheet for IC voltage regulator it is beneficial expecially when you have a transverse load (not ground referred).

In the power supply you are referring to the problem is exacerbate due the half wave rectifier: the postive and negative rails "rise" with half mains period delay between them

[Scott594]

My guess is that due the 1/2 wave design and smallish input filters, the IC input voltage is too low to give the required minimum 2V input to output differential. When you short the outputs (not a great thing to do repeatedly) The ICs go into protection, and allow the input caps (C2 - C3 470uf) to charge up beyond the dropout voltage. I'd guess there's a bit more load on the positive leg, causing it to be the one effected (once one regulator shuts down, the opposite one gets enough input voltage to avoid dropout). One thing that may help is to increase C2 and C3 as much as practical (the author states he used 2200uf in the prototype). This will help keep the ripple valleys from dropping low enough to cause trouble. You can check this easily if you have a scope. The raw DC ripple valleys have to be higher then +-17v - higher would be better. I didn't see a Xfmr secondary rating anywhere, but it's probably a tad low. A better design can be found here LM7800 pdf, LM7800 description, LM7800 datasheets, LM7800 view ::: ALLDATASHEET ::: on the last page of the spec sheet, but that will require new Xfmr(s). Hope it helps.

[AndrewT]

with a little bit of work you can add the two extra diodes to make a bridge rectifier and thus a full wave rectifier.
But, you need a centre tapped secondary or dual secondary transformer to get your dual polarity supplies.

[Mooly]

Originally Posted by poldus

I built two units of the bass equalizer kit from elliott sound products that includes an on-board dual power supply.
They are working well except for the following glitch: The units will not work upon power-on. Ground to negative rail reads 15v on the voltimeter but ground to positive rail reads 0. Now, when I briefly shortcircuit + and - they suddenly power up and give a correct +15v. Each board has its own power supply but they share a common AC transformer.
Any hints? I sure don´t want to be jump-starting the units like this everytime.
Thanks

Here´s a link to the project that includes a circuit diagram
Eight Band Sub-Woofer Equaliser

Don't guess... find the fault and fix it ! It's a very basic circuit that should work perfectly.
"Ground to Negative rail reads 15volt".... are you sure ? do you mean "reads minus 15 volt" ?

With it in the faulty state, and with the meter negative lead left connected to the "ground" rail measure to each end of C2 and C3 (The inputs to the regs) and tell us what you measure.
If shorting the plus and minus rails makes it work it could be a poor connection somewhere and the "shock" of that causes a small arc and remaking of the connection... or perhaps D1 is intermitent... it happens.

Two minutes with a DVM should isolate the fault for sure. Have you tried it with no load connected to the outputs ? Does the regulator get hot when there is no output ?

[poldus]

I added a 2,200uf capacitor in parallel to both C2 and c3 and it the problem is gone. Thanks to all those who posted for your help!!!

[Mooly]

Ok then you have fiddled the fault

Why not try replacing C2 and C3 one at a time... it's so basic a circuit (if it were me ) I would want to know why it didn't work originally. Maybe one of the caps is open circuit etc.

[poldus]

Mooly: So it happens I thought it was fixed but it isn´t.I powered it up again today and the problem remains.

The transformer output is 14.6v while connected to the board. 14.75 on the positive and -14,90 on the negative rail (relative to ground).

Later,I will measure C2 and C3 when in faulty condition and post the readings. If I turn it off and on again now it works fine, probably because the capacitors remain charged.

I may try adding two extra diodes (same as D1 and D2 but after the rectifier if I understand correctly) as diy-audio-fo suggested.

I do not have a scope, only a voltemeter

[Mooly]

A meter is fine... try and find the fault first...
If you have +15 and -15 coming out of the regulators then there isn't a problem with the PSU... or do you mean they read that when working ?

Connect meter to ground line and measure the input to each reg first and then the outputs. In your first post you said there was no +15 volts... so you need to recheck that.