Voltage multiplier cap values

[nashvillebill]

Okay, I figured I'd try making a voltage multiplier for B+ on a 12AX7 preamp. I've got a transformer with 60VAC (RMS) on the secondaries.


So I whipped up a 3 diode voltage multiplier. Only thing is...evidently I need a whole lot more capacitance. The voltage multiplier examples on the 'net are using a 5V p-p AC with 1000pf caps. I tried that...no go. Tried 0.33 mfd, helped a little but still nowhere close.

Exactly how much capacitance does one need (based on your experience), even for a modest preamp load like maybe 6 or 8 mA? Did not see a formula anywhere and I don't have spice loaded on my computer here.

p.s. I've asked for Merlin's power supply book for my birthday next week...

[loudthud]

At 50 or 60 Hz you need to use about 100uF @100V.

[nashvillebill]

Well, immediately after posting my original post, I rummaged through my parts drawers, and after experimenting with everything else, and guess what...I wound up with 100 ufd. With 60 VAC RMS, and the 3 diode/cap arrangement, this yielded around 250 VDC.

The only problem I see is the ripple, I've got almost 60 volts ripple on there.

[tubeswell]

Well, immediately after posting my original post, I rummaged through my parts drawers, and after experimenting with everything else, and guess what...I wound up with 100 ufd. With 60 VAC RMS, and the 3 diode/cap arrangement, this yielded around 250 VDC.

The only problem I see is the ripple, I've got almost 60 volts ripple on there.

Then you need still more capacitance.

[Chuck H]

With 60 VAC RMS, and the 3 diode/cap arrangement, this yielded around 250 VDC... I've got almost 60 volts ripple on there.

Noticing that the input voltage and the ripple voltage are about the same, is it possible that you've wired something wrong and all the ripple from the input is getting through unfiltered?

[nashvillebill]

I suspect something is wrong, perhaps I miswired it (though this is pretty simple) or maybe one of my diodes is dead. Here's the scheme I'm using.

[paggerman]

My first build used 2 30v transformers in parallel to give 60v. This was then fed into a voltage quadrupler to give a B+ of 300oddVDC. The schematic and values are available at Grant Willis' site Valve Heaven - Home
Look for the Lamington DIY valve amp link. Mine works just great. Good luck.

[loudthud]

With an odd number multiplier, the output is taken from the "driven" side and the 60VAC is coupled right to the output. With an even number multiplier, the output is taken from the "grounded" side and the ripple is much smaller.

Check this thread: http://music-electronics-forum.com/t4086/

[nashvillebill]

That makes sense! It did seem more than a coincidence that the ripple was exactly equal to my input voltage.

The whole purpose of this exercise was to piddle around with a power supply and preamp made from parts I've got lying around--until I get a paycheck rolling in again I'm not going to buy any more caps or whatnot, so maybe I'll try the next crazy notion. I've got two transformers that are both 115 VAC primaries and 60VAC secondaries, I could hook them up back to back (secondary to secondary) to yield 115 VAC at my new "secondaries" and then use a conventional half wave voltage doubler. Again, this is more of a "experimentation and learn" exercise. When I get a job and decide to build something more permanent, then I'll just order a suitable PT.

Thanks!

[nashvillebill]

Okay, on a whim I added another diode and cap to make it a voltage quadrupler instead of a tripler. Ripple stayed 60 Hz but resulting voltage DC actually dropped!

This tells me C1 in my circuit is still WAY too small. Makes sense in a weird way, with the inductance and series resistance of the transformer.

So I backed up and made a 6-diode voltage doubler, according to post 4 in this thread http://music-electronics-forum.com/t15181/ (thanks guitardude). The result, with a 60VAC input, was over 170 VDC with only 150 mV or less of ripple, when loaded with a 33K resistor. Quite an improvement!

So...my next step...maybe I'll put my other power transformer to use, and wire the secondaries in series, to get approx 120VAC into the voltage doubler.

Quite an educational day.

[J M Fahey]

This is one of those things that I understand but can't find the right words to explain, buy I'll try.
In a multiple stage voltage multiplier, there are "multiplying" caps and "filtering" caps.
Don't know if this makes sense to you, but in this case I call C2 a "filtering" one and all others "multiplying". In odd circuits you end necessarily with a multiplying one, you do not hqave "ripple" per se but the full transformer winding voltage which by definition is "unfilterable", no matter the capacitance.
In even ones the situation is much better, but still "ripple multiplies while capacitance divides", *much* better than before but not enough.
You need an extra "non multiplying" stage, which is simply an extra diode and a last cap, which stands full output voltage and is directly grounded.
Try , say, a 47uF or 100uF x 300 or 350V.
Post results.

[nashvillebill]

I tried that last night, I think, but either I had too much ripple for the cap or else I miswired it, the cap I was using at the time started to overheat.

Now I've got two transformers with the secondaries wired in series for approx 120 VAC, and the voltage doubler is yielding somewhere around 350 VDC with very little ripple, a little over 100 mV ripple with a 33k resistor for a load. This isn't perfect, the two transformers aren't perfectly matched (I fired it up on a 40 watt dim bulb tester on the primaries, and the bulb was glowing--but not too bright, a 100 watt bulb did not glow at all) but it'll do for a while. This will let me play with making a preamp or two to experiment with, then I'll buy a real transformer later--by then I'll have Merlins power supply book.

When playing with power supplies, I have quickly discovered the usefulness of the dim bulb tester!