2-channel preamp (power amp added)

[Joe]

Thought I would share this, it's basically a 2-channel Dr. Boogey I just got done building into an amp. Does a pretty good job of simulating the original, though a bit more harshly. A clean switch on the orange channel would be nice, but I ran out of room/didn't want to bother. Have a great day.

[PRR]

[SimonS]

Just thinking here: won't the disengaged channel still load down the active signal when you switch like this?

[defaced]

The relay contacts are arranged to select on or the other, so there is a complete break in the circuit when a channel is disengaged.  This should prevent the unused channel from loading the used channel.  It looks like a cool design, any sound clips? 

[SimonS]

I'm sorry but that's not correct: e.g. the red gain potentiometer + circuit is still connected to circuit around orange gain when orange gain is selected (and vice versa).

[Joe]

There might be very slight interaction between the channels, but there's a lot of resistance there and I couldn't notice any. Ideally there should be two DPDT relays, to switch both the input/output of the gain/tone sections.

PRR: Nice idea for the LED's. I'll change the schematic when I have time to test that.

In the meantime, I drew the LED's in the wrong place, so here's the updated schematic:

[Joe]

Here is the power amp, please note that it must be biased correctly:

1. Adjust balance to get 0 volts at output no load.
2. Adjust bias to get 50-75mV between the output emitters. (Or connect speaker and turn just until crossover distortion is gone.)
3. Re-check balance, as it may have shifted slightly.

The design is a little unusual, but it soft-clips, good for guitar. Also, it's inherently stable and not prone to oscillation unless the feedback resistor value is too low. Sounds really good cranked to the max. Original design here (circuit #3): http://depalma.pair.com/Analog/analog.html.

Edit: Fixed trimpot per PRR's suggestion. (My pcb was actually wired this way, but didn't think to check.)

[defaced]

I'm sorry but that's not correct: e.g. the red gain potentiometer + circuit is still connected to circuit around orange gain when orange gain is selected (and vice versa).

I stand corrected.  Given the behavior of gain controls though, I wouldn't expect much bleed.  Not that way I'd do it, but if it works it works.

[mth5044]

Love the poweramp! What's the wattage?

[PRR]

> the red gain potentiometer + circuit is still connected to circuit around orange gain

Yes, but these are Meg-ohm pots driven from the ~~30K drain resistor.

It would take a while to estimate "all" the possible interactions, but I'll bet it is hardly audible. To make a "zero interaction" switching would need more buffers or more relay contacts.

> What's the wattage?

Has +/-35V rails, 70V total.

Might sag under load.

The peak-to-peak output must be less than 70V or whatever it sags to.

Being over-optimistic: 70V p-p is 25V RMS. 24V RMS in 8 ohms is 3 Amps and so 24V*3A= 72 Watts in 8 ohms. Half that in double impedance, 36W in 16 ohms.

Depending on supply, a little more or considerable less. I've seen amps with +/-35V at idle barely crack 40 Watts.

In any case, this is a serious amp. It needs big iron, big heatsinks, and careful attention to details like thermal coupling. Find a good how-to for a similar 70W amp.

I "object" that _when_ the bias trim pot wiper loses contact, bias goes way-high. Smoke. I used to earn daily-bread fixing such things. Kept me busy. But on my own stuff, I like the trim organized so a bad-contact pot goes to minimum bias. You can finish the gig with a little crossover. Meantime, use a high-quality bias trimmer.

[Joe]

I "object" that _when_ the bias trim pot wiper loses contact, bias goes way-high. Smoke. I used to earn daily-bread fixing such things. Kept me busy. But on my own stuff, I like the trim organized so a bad-contact pot goes to minimum bias. You can finish the gig with a little crossover. Meantime, use a high-quality bias trimmer.

Fixed, thank you. Although you're probably screwed anyway if the balance control fails in either direction 

[PRR]

> if the balance control fails

Ah, yup. Seen a lot of bias-pots let go. Few amps had balance; those which did often failed unspectacularly (a few volts DC to the speaker, or even just reduced low-THD power). This one will probably slam to a rail? No, not that bad since the full-pot resistance would be there. You get a 33K:37K unbalance. Even with the hardly-any NFB.... actually there's significant NFB below 160Hz, so the balance pot may never tip as far as a volt of DC to the speaker? I have not seen an amp work quite this way so I'm hazy on failure modes which are old-hat on "conventional" designs.

[Joe]

37K on the spice sim put 2.2VDC on output.

Had an idea a while back to avoid the trimpots: use fixed resistors and switches, for example a 500ohm pot would use these resistors in series: 250 125 62 31 16 8 4 2. To program it short out the correct combination of resistors (sort of a binary thing).