Joe Davisson's new Guitar amp

[joecool85]

Look what the cat dragged in

Sweet 

Did you order those or etch them yourself?

[Joe]

I ordered the boards. Rebuilt the amp using the new design and it works, except I think the balancing control is too touchy (easy to fix). The rest seems okay, so I'll run it full blast tomorrow and see what happens.

[joecool85]

I ordered the boards. Rebuilt the amp using the new design and it works, except I think the balancing control is too touchy (easy to fix). The rest seems okay, so I'll run it full blast tomorrow and see what happens.

Sweet.  And we want pics of the build - as usual 

[Joe]

Checked this out some more.

With the output shorted, and no input signal, one of the output transistors gets warm, and one of the emitter resistors reads .650v across it. It's either one transistor or the other, but never both, depending on how the balance/bias controls are set. I can hold it for 20-30 second, and it's fine except for the warming. Does a VI limiter always throw to one side or the other like that?

The amp is pretty quiet if the balance/bias is tweaked slightly after plugging in the speaker. (plugging in the speaker throws the balance off very slightly). Raising the quiescent from 20mA to 75mA gets rid of most of the hum.

So I guess it's working alright. Sounds nice cranked up, which is what counts. Some build pics:

[rowdy_riemer]

Are you going to post any sound clips?

[joecool85]

It looks excellent, Joe.  Are you going to be selling those PCBs?

[Joe]

I'd get the boards out there if people wanted them, but there are some caveats: Because the output drifts slight with temperature, exact output centering is not guaranteed. I'll explain the biasing procedure to see what you all think:

1. With speaker/input sources removed, adjust dc balance and vbe multipler back-and-forth until the output is zero and voltage between output emitters is around 75mV.

2. Connect speaker (or load resistor) and adjust controls again. Check again after a warming period. In the future, the amplifier will begin with a very slight offset (and accompanying hum) which disappears after warming up.

3. Glue trimmers into place.

So the balancing/biasing is a little tricky. I've made changes to the balancing network, which now limits the possible output deviation to around +/-2V, making adjustment easier and somewhat safer.

The amp also requires choosing positive-feedback values appropriate for the application. (This can also program the level where power amp distortion begins.

I have a nice preamp going, using some leftover MPSA06's running off the +35V supply. Gets about a "Marshall" amount of gain, and seems to be a perfect match. Will try to post a sample.

[teemuk]

This is an interesting design but I have to say that on my part I can't really verify this "diode compression" effect taking place. Unfortunately, my experience of this circuit only limits to few SPICE simulations so I don't doubt if the effect actually happens on a real circuit and with the VAS/output transistors in question.

Anyway, here's what the SPICE sim I ran seems to show. Hopefully it provides some alternative food for thought and some further insight to this design.

- The output indeed compresses rather softly but it does so without the diodes as well. With the diodes bypassed, the 470K resistor between the VAS bases can be removed as well.
Where does the compression come from then: The DC bias injected to emitters (instead of a more typical base-bias scheme) is also a resistive divider, that is an alternative shunt path for the feedback. Because of it, the corrective feedback is reduced and this provides the more gradual saturation characteristics. The symmetric topology is also much more immune to "rail sticking" than the generic bootstrapped/CCS-loaded circuits.

- The diode's capacitance introduces less resistance at HF, which in turn gives the amp more gain at high frequencies. If the diodes are bypassed the gain becomes more linear and consequently the high frequency response of the amplifier gets narrower. The effect is only few dBs though and really starts to show above 7 kHz. Basically, aside phase shifting, it's insignificant. Nevertheless I found it interesting.

[Joe]

You're right, the diodes really don't do a whole lot. They round out the waveform a bit more, especially if the 470K is increased (but too high and the transistors won't bias). The hi-fi and headphone versions I've made don't use diodes and sound great.

On the preamp circuits the diode helps a lot, because it keeps the transistors out of saturation, but I doubt it's more than a subtle effect here, except perhaps at full blast.

Still need to do an extended full-power test through a load resistor, but other then that seems to be a nice little amp that gets plenty of volume.

Thanks for the comments and help guys!

[Joe]

I took the diodes/resistor out and couldn't tell the difference, so I guess I'll forget about that idea.

The higher voltage preamp didn't pan out either, went back to a 4-stage vulcan circuit at 9V which sounded much better.

So at this point everything is sounding good and even got some tone controls in there. When I get it all boxed up I'll try to record some samples.

[Joe]

Made some minor changes to the balance control and lowered the two 10K emitter resistors to 1K:
http://www.diystompboxes.com/analogalchemy/sch/diodepwr.html

Power supply schematic (using the Radio Shack transformers):
http://www.diystompboxes.com/analogalchemy/sch/rs_pwr.html

I think it does sound a little better to use the diodes, so I went back to that on my build.

[Brymus]

Thanx for the update Joe.
Wouldnt be OK to use 4amp diodes X4 instead of those 25amp bridges?
I'm curious cause thats what I was planning on using with those trannies along with snubbers or is that too low a rating?
And are snubbers unnecessary ? I noticed you didnt use any.

Edit OK I went in the garage and looked and those are only 2amp trannies so I think 4amp diodes will be fine.
But what do you think about adding snubbers?

[Joe]

I don't know anything about snubbers, so I can't help you there. I'm sure diodes of proper rating are fine.

Edit: sorry I yanked the preamp circuit, needs more work.

Edit: Preamp circuit is back, seems to work pretty well this way:
http://www.diystompboxes.com/analogalchemy/sch/v4preamp.html

[Joe]

Working on a transplant. Sort of a Recto knock-off, but with an effects loop for each channel, and less features. (Nothing is really mounted yet, just doing a test-fit.) Still working on the channel-switching and fx-loop circuits.

[Brymus]

Wow again it looks really nice Joe!
Cant wait to see more.