Digitally driven tube pre-amp

[eranrund]

Hello all,

Lately I've been thinking about constructing a digitally controlled tube-amplifier.
What I have in mind is a tube-driven line-stage, having source selection and volume attenuation controlled by a microcontroller, most likely an Atmel AT91SAM7S.

Source selection is done using the usual relays method - seems to provide more than good results for everyone, and the only way around all in all. The pre-amp does not need to accept any balanced connections at this stage.

Volume attenuation requires more attention, as there are a few options:
1) dedicated iC (for example PGA2320) - probably the easiest and "suckiest" of all
2) Motorized ALPS pot - Expensive, might have other disadvantages I am not aware of (such as motor control - will an h-bridge be necessary?)
3) Resistor ladder controlled by relays (e.g. 6 relays, providing 64 attenuation steps)
- From what I understood, this would yield the best performance. However, I'm having trouble regarding the load variations that happens when I switch resistor values.
Is this any matter of concern? What would be good ways to overcome it in a tube pre-amplifier?
In a solid-state I could put a buffer before and after the attenuation stage. Would this be desired? What about a tube pre-amp?

Extra control features include the mandatory IR remote control and a small LCD display.

The pre-amplifier needs to have both un-balanced and balanced outputs - what would be a suggested way to achieve a balanced output without using two op-amps in an inverting and non-inverting configuration?

A little background about myself - I am pretty much new to analog designs. I work as an embedded programmer and am familiar with the world of digital electronics. I have also built an LM3875 'gainclone' amplifier, using my own PCBs/schematics and it worked perfectly until I moved an apartment

A general question about pre-amps - where would you place the attenuation stage? Seems most if not all designs place a voltage divider before the buffer. What would be the effects of placing it after the buffer? What would happen if I used variant resistance on the gain loop of a typical op-amp to alter its gain?


Lots of questions, but hey - I'm starting out on this whole tubes/pre-amps thing. Might not be the best newbies project, but I think I will manage to handle it. I hope so, anyway

Many thanks.

 

[eranrund]

Small correction

Small correction... on the topic, I meant to write 'Digitally controlled' and not 'Digitally driven'

Too bad I cant just edit it :/

 

[analog_sa]

1) dedicated iC (for example PGA2320) - probably the easiest and "suckiest" of all
2) Motorized ALPS pot - Expensive, might have other disadvantages I am not aware of (such as motor control - will an h-bridge be necessary?)
3) Resistor ladder controlled by relays (e.g. 6 relays, providing 64 attenuation steps)

1) Certainly the "suckiest".

If you are already using a micro, may i suggest a DS1660 - easy to implement, very nice sounding, ARC use it in their top preamps and it's very easy to do balanced.

2) Not much enthusiasm for any kind of Alps pots, i'm afraid. Other than being cheap and easy to control i see no reason to use them.

3) I've built several relay attenuators and the results, using Nais relays and cheap metal film resistors, were really mediocre compared to a switched ladder/shunt attenuator using same tupe resistors. Finding nice sounding relays and hoping they would go on sounding nice is a long shot.


Tube preamps are not very suitable at driving attenuators at output. You either need a large resistance value (which makes an active pre silly) or a transformer coupled stage capable of driving low resistance (600ohms?). That's the reason 99.9% of tube preamps have the pot at the input. It's certainly possible to sandwich it between stages but this is very seldom done.

 

[eranrund]

Are you talking about the DS1666 (http://www.maxim-ic.com/quick_view2.cfm/qv_pk/2747) ?
I can't find information on a DS1660..


Thanks

 

[mashaffer]

I have seen active volume controls using tubes but I can't remember where. One was some kind of beam deflection tube with two plates that had a pair of grids that directed the beam in varying amounts to one or the other. The other I think used a remote cut off tube of some sort. As I recall one of the issues with these was the total amount of attenuation being limited.

You might do some googling on that general theme as you might be able to come up with some sort of variation or hybrid.

mike

 

[oldeurope]

I have seen active volume controls using tubes but I can't remember where. ...

mike

EF86 as a volume pot

Schematic

 

[analog_sa]

Are you talking about the DS1666

I am. Sorry.

 

[eranrund]

Unfortunately I do not read German. Add to that my (current) inability to properly read and understand tube-containing schematics, this leaves me quite helpless.

I should have a 2nd look at the schematics once I familiarize myself with some tube schematics.

 

[analog_sa]

I have seen active volume controls using tubes but I can't remember where.

I think the coolest implementation is John Swenson's BDT preamp.

 

[eranrund]

Regarding the DS1666, should one be worried about the signal limitation imposed by the +5/-5v rails?

The most interesting thing about this chip seems that its control interface can be very easily opto-coupled, thus completely isolating it from the digital part of the circuit, where some high frequencies run

What about its 20% tolerance? Should that be of any concern?

Many thanks