digital potentiometers

[Andy]

I have a raspberry pi, and I want to use it to build a digitally controlled tube preamp initially, and eventually adding a power amp. the idea is that all parameters would be programmable, including the number and order of gain stages, the position of the tone stack, power amp plate/screen voltage and bias. Analog Devices has the AD5290, which is a digital potentiometer with +30/0 or +/-15 voltage capability. so I have a few questions:

1. will preamp tubes perform acceptably with a 30v plate supply?
2. would the digital potentiometer have a noticeable negative impact on tone if it were in the signal path?
3. are there any other devices with higher resistance element voltage ratings? I've googled, and 30v Vss/Vdd is the highest I've found.

[Steve Conner]

See the Matsumin Valvecaster for an example of a tube preamp running off 9V. However you might struggle to make it sound like a tube preamp running at full voltage, the distortion character is quite different.

Digital pots have distortion and stray capacitance, but you can work round that. You might also like to try their close cousins, multiplying DACs. The TLC7528 is one example.

The 30V limit is imposed by the IC process the things are made on. You may get 36 or 42 if you're lucky, but never 250 or 400: the market isn't there.

[exclamationmark]

I've used digital pots in my amp, but only for the tone control section. The tolerances on them were really quite terrible though. I remember measuring one at the max value and it was almost 30% off! Perhaps they've gotten better now... The tone control section on my amp only ran at something like 18v, and was just a dual state variable filter (sort of like a parametric eq, but without the annoying effects of variable q when you move the frequency around, and other stuff) made with op amps. I had the bright idea to try and get my MCU to calculate the bode plot of the filter and display to an LCD, but back when I built it I didn't realize I didn't have hardware support for floating point operations! I should probably revisit that now, since we have things like raspberry PI's and ARM's with FPU's.

I've also dabbled with a programmable guitar amp quite recently and used around 80 reed relays (Hamlin HE751 type relays - rated to around 300v and hermetically sealed) for a 3 stage guitar amp. I set it up so each gain stage was in order but you could optionally bypass each stage to make a however many stage pre-amp you wanted. I found the linear approach of bypassing stages to work better than routing signals all over the place, which induced feedback and made it much, much more complicated. I had the B+ at something like 250V, and this caused some problems I didn't initially anticipate. I had several coupling caps in parallel with each other, all relay switched, initially without a pre-charging circuit in place (I made a thread on that subject before). Well of course when I tried to parallel these capacitors together by activating the appropriate relay, a non charged cap + a charged cap naturally spelt disaster. I didn't realize even at 0.022uf and something like 150v, it would weld my relay contacts together due to the capacitors trying to balance voltages (and hence the huge inrush current). So uhh... lesson learnt. Another problem was the EMF put out by the switching relays, which wreaked havoc with my shift registers used to drive them. I solved the problem by sticking a low-pass filter on the latch pins, as the emf must have been causing the shift register to latch with whatever crap happened to make it's way into the shift register at the time.

I don't like the reliability factor of mechanical relays (especially when you're looking at several hundreds of the things for a fully fledged programmable guitar amp), so I'm currently trying some SSR's. I've ordered up some Vishay SSR's with only 22pF capacitance at 0V(!) but they haven't arrived yet. I have a test bench set up in anticipation though, with a dual trace oscilloscope, balanced 12AX7's, identical circuits and one hell of a breadboard According to my SPICE simulations there should be some differences in the 10kHz+ area, but I'll be looking out for things like crossover distortion and distortion in general when I actually test them out (hopefully tomorrow).

[Andy]

You might also like to try their close cousins, multiplying DACs. The TLC7528 is one example.

do you know of any such devices with higher voltage ratings? I can live with 30v, but I'm concerned that a voltage as low as 5.5 would not give me enough headroom.

[Gregg]

You might want to check this old thread where digital pots and some other solutions are discussed:

http://music-electronics-forum.com/t1015/

If you want to build a high voltage digital pot MAX4800/4802 may be suitable.
In my programmable tube preamp I currently have and LDR pot for Gain and digital pots for the rest.

[Enzo]

It isn't what you are doing, but you might find it interesting:

Peavey.com

[Andy]

You might want to check this old thread where digital pots and some other solutions are discussed:

http://music-electronics-forum.com/t1015/

If you want to build a high voltage digital pot MAX4800/4802 may be suitable.
In my programmable tube preamp I currently have and LDR pot for Gain and digital pots for the rest.

I read the thread on LDRs as pots, and I'd consider it as a fallback solution, but it involves many more components, and for the sake of simplicity, I'd love to just stick with the digital pots. the max4800 series is interesting, although they're just switches, so at best, they'd be viable for use in a programmable gain amplifier. I wouldn't mind running my preamp at 30v, and I'd expect it to still sound good. I'll probably have to use smaller plate resistors, and adjust the values of some other components, but I think it will be a fun project no matter how I end up doing it.