Why did this fry my bridge rectifier?

[Ripthorn]

So, the continuing adventure of my submini amp brings me to this.  I have a bridge rectifier for the heater voltages and it has been steady as a rock, worked great with no problems.  Then I tap off the transformer secondary (12.6VAC) to do the double half-wave rectification as shown on RG's power supplies page to produce a bipolar power supply.  Hooked it up and tested it with no load, perfect.  Then I hook the circuit up to my device (TL072 running on the bipolar supply functioning as a gain recovery stage) and turn on the power.  Well, the magic smoke got out.  I thought maybe I exceeded the current rating of the rectifier (only 1.5A, not much safety margin, I know), but the thing that doesn't make that make sense to me is that I wasn't pulling any extra current through the bridge rectifier when tapping off the transformer secondary.  Is there something I am missing here?  I've got a new rectifier rated at 4A on the way, but more than anything, I want to understand why this happened.  Any ideas?

[R.G.]

Where is the winding grounded for the heater voltages?
And where is it grounded for the bipolar power supply generation?

[Ripthorn]

Both are grounded to the chassis (which is grounded to the earth prong of an IEC socket), but at different points (I did a quasi star ground).

[R.G.]

Both are grounded to the chassis (which is grounded to the earth prong of an IEC socket), but at different points (I did a quasi star ground).

Where ON THE WINDING are they grounded? The point is that you may be grounding the output of your bridge rectifier inadvertently.

[Ripthorn]

I'm at work right now, but I will check when I get home.  That may well be the issue.  Perhaps I am creating a reverse voltage across the rectifier which would fry it.  I will check and report back.  Thanks RG.

[Ripthorn]

Okay, I looked at this now and at the risk of sounding stupid, could you explain the grounded on the winding thing, RG?  My 12VAC is coming from a non-CT transformer, and I take one end of the winding and it goes to the bridge rectifier and also the the positive side of the voltage doubling circuit.  The other winding also goes to the bridge rectifier and also to ground in the voltage doubling circuit.  Is the problem that the second winding is going to ground?  If so, how do I fix this so that I can use the voltage doubler circuit and the rectifier?

[Ripthorn]

So I was thinking that maybe an isolation transformer would keep my rectifier from blowing up.  Just put my voltage doubler circuit on the other side of the iso transformer from the bridge rectifier.  Would this work?  I am specifically looking at this one http://www.murata-ps.com/data/magnetics/kmp_da100.pdf

[R.G.]

Okay, I looked at this now and at the risk of sounding stupid, could you explain the grounded on the winding thing, RG?  My 12VAC is coming from a non-CT transformer, and I take one end of the winding and it goes to the bridge rectifier and also the the positive side of the voltage doubling circuit.  The other winding also goes to the bridge rectifier and also to ground in the voltage doubling circuit.  Is the problem that the second winding is going to ground?  If so, how do I fix this so that I can use the voltage doubler circuit and the rectifier?

Using an isolated winding to make 12Vdc is fine. So is using an isolated winding to make +/- voltages. But doing them both at the same time from the same winding creates the risk that you will connect the external circuits which these things power. If you are making a bipolar supply for signal circuits, you're going to be grounding the center tap of the bipolar power supply, and that means in turn that one side of the 12Vac winding is tied to ground.

If you full wave rectify the winding as well, everything's still OK - I think! - until you connect either side of the full wave rectified power to ground. When you do that, you have one side of the 12Vac winding connected solidly to ground, and the other side connected to ground through the diode of the full wave rectifier bridge. So the diodes are handling the full shorted current of the 12Vac output. They die. If your full wave rectified output is for filaments, it simply has to be ground referenced somewhere or you'll get intolerable hum, so you're probably grounding one side or the other.

[Ripthorn]

Thanks for the response, RG. That makes sense.  I think what I might do is use the rectified power and do some charge pump conversions.  I posted a schem for an idea I had in this thread http://www.diystompboxes.com/smfforum/index.php?topic=77597.0 though I am not sure the negative voltage multiplication will work (I know the positive multiplication and negative voltage generation will).  If it doesn't work, I think I could just use a second charge pump to multiply -9V to -17V by using -9V as my ground reference.  I am still considering all the options, though.