• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

12AU7 / 12AX7 push-pull

I'm taking a shot at a mini stereo amp. Per-channel it uses 2 dual triodes to make the PI (12AX7) and output (12AU7) stages. I'm using a cheap $5 PA matching transformer for the output transformer. A couple cheap 12vac transformers will make my power transformer. I'm stepping 120 down to 12 to grab heater power then stepping 12 back up to 120 with the second transformer. A voltage doubler provides me about 300vdc. My output transformer has an 8-ohm secondary and quite a few taps on the primary.

0
62.5
125
250
500
1000
2000
4000
8000

I'll use the 4000 for my center tap and 0 and 8000 for the 2 phases. B+ will be around 225~275v to the 12AU7 and 200v to the PI.

Here's a pic of the single-ended monstrosity I built just to try it. The power stage is a parallel 12AX7 and one side of another 12AX7 for a preamp. No Criticism! I'm still on my learning curve. :D

I've seen a design like this somewhere online but I can't find it. Anyone have it?
 

Attachments

  • DSCF2416 (Medium).JPG
    DSCF2416 (Medium).JPG
    161.3 KB · Views: 1,275
I'd suggest changing to "floating" paraphase inverter for better balance.

If I knew what it was.

Also you're throwing away 2/3 of your voltage swing at the 12AX7 (220K plate load and 100K grid resistor).

As far as I know the 100k on the output tube's grid is to match the PI's voltage divider (100k+4.7k) on the other output tube.

And 2K is the center-tap...

So 2k center tap, 0 and 8k for phases? Why is that? Shouldn't they be equal impedance?

but 8K is pretty low for 12AU7s

It's a cheap easy to find transformer.
 
Last edited:
Much info on line transformers as output transformers:
Cheap Output Transformers

Thanks. Lots of usefull data there. It still doesn't answer why I'd want a difference impedance for the 'push' then I would for the 'pull'. :confused:

I built the thing. Sounds like @#$%. More for fun then anything. I have a feeling I wired something wrong though. I had some time to kill and figured what the heck why not but not enough time to debug it. It sounded much better single ended on the 8k tap. A battery powered version of this, as a portable mp3 amp, would make a nice conversation piece. I'd make it look just like a mini DIY hifi amp.

I might fiddle around with it after I get my stereo amp finished.
 
Last edited:
Thanks. Lots of usefull data there. It still doesn't answer why I'd want a difference impedance for the 'push' then I would for the 'pull'. :confused:
The impedances are the same (symmetrical), but by using the centre tap you end up with 4ka-a rather than the 8k you expected.

For small-signal valves you really want one of those really small line OPTs, the ones that offer you several primary taps up to about 1W (usually 0.1W, 0.25W, 0.5W, 1W). That way you can get impedances up to 10k, which are more appropriate.
 
The impedances are the same (symmetrical), but by using the centre tap you end up with 4ka-a rather than the 8k you expected.

For small-signal valves you really want one of those really small line OPTs, the ones that offer you several primary taps up to about 1W (usually 0.1W, 0.25W, 0.5W, 1W). That way you can get impedances up to 10k, which are more appropriate.

Right. So wouldnt I want 0-4000-8000 with 4000 as the center tap? That leaves 4000 in each direction. Tom said 2000 is the center tap. How can that be? That makes 0-2000-8000. I'll have 2000 on one phase and 6000 on the other.
 
Right. So wouldnt I want 0-4000-8000 with 4000 as the center tap? That leaves 4000 in each direction. Tom said 2000 is the center tap. How can that be? That makes 0-2000-8000. I'll have 2000 on one phase and 6000 on the other.

OK this is how I'd explain it..

What you really want is an equal number of TURNS on both sides of the center tap. Think volts not impedance.

So the number on turns from one end to center is 1/2 the number of turns from end to end. Doubling the number of turns increases the impedance by a factor of four. Remember impedance is proportional to square root of the number of turns. So the center tap to end is always 1/4 the total impedance in any transformer with a center tap.

Try this experiment: Connect a signal generator to 1/2 of you output transformer and look at the volts on the secondary. (Use a 12V heater transformer or wall wort for the signal generator if you have to.) Then switch the signal so it is across the other half of the OPT. You should see the same volts on the secondary (give or take a few percent) if you have the "correct" center tape. This way you can pick the correct tap even without having too understand this. Just pick whichever one give symmetric results.
 
Last edited:
Interesting. I have a boat load of 12AU7s from and organ pull. Might be fun to make a little computer amp or headphone amp using this approach. I assume that a 70V line tranny could be used in a similar way, just different turn ratios at the same power level taps.

That's what I was thinking. It seemed to work much better single ended then it did push-pull. If single ended, you could even grab a UL tap and use a tetrode. Maybe use a 12AU7/6V6 or EL84. Check this thread. http://www.diyaudio.com/forums/tubes-valves/64604-el84-se-advicenschematics-needed.html

Remember impedance is proportional to square root of the number of turns. So the center tap to end is always 1/4 the total impedance in any transformer with a center tap.

So 2 2000ohm coils in series is 8000ohms because they share the same core right? Otherwise 2 inductors in series equal the sum of the inductors. So a center tap IS the center tap of the coil but the measured series inductance is 2x the sum of each seperate coil due to the coils sharing the same core.
 
Last edited:
i wanted to build a el84 triode PP amp, but meanwhile accidentally saw that 5 sections of 12AU7 ~ EL84 triode quite near, regarding µ, rp, transconductance, down to heater current, so i will give that a try. they will be able to dissipate slightly more than EL84s,
because 5 units are inside 2.5 glass bottles, while a EL84 sits in a tube about 4/5 the size. (still, the AU7s are rated a bit cooler and not dissipate as much as that factor relatively, but little less).
first complete ;p
 
Last edited: