DIYHiFi.org

Gang,

I need to look at remote control volume for my preamplifier. I have some motorized pots and they work pretty well but I am using DHT as gain/output tubes and the motor makes them a little microphonic. So I am thinking of either solid state (i.e. Digtial Pot) or relay.

Any comments?

Sorry I have been away from the forum. I was trying to get the 24/96 stuff done for async mode. Now that it's all working, I can move on to other things.

Thanks
Gordon

_________________
Gordon
Chief Scientist
Wavelength Audio

Hello Gordon,

I use Maxim DS1666 in my preamplifier:

http://www.maxim-ic.com/quick_view2.cfm/qv_pk/2747

I control them with a PIC uC and a remote handset. They work very well and at normal line levels the distortion is unmeasurable (but note that I put a buffer before and after them), I strongly suggest them: they're cheap and I can even get them locally. And you could even control them with 2 buttons and some logic gates, so no need for microcontrollers. Some pics:

http://img251.imageshack.us/img251/1055/immag003cv9.jpg

http://img517.imageshack.us/img517/1944/immag006ug4.jpg

Gordon,

If you are going to use MCU's (PIC, Atmel or other), I suggest trying DS1666 as told by Giaime, PGA2310/2320 from TI (my preferred) or WM8816 from Wolfson Microelectronics (discontinued but still available).

The very best is the DS1802 from Dalals-Maxim and it can be controlled only with a couple of push buttons.

_________________
Pedro Martins

Playing with Canon EOS 5D mkII...

Hello Gordon,
I'm using the PGA2320 like this, with good results. Please note Jocko's comments...

I'm not using any buffers (not needed in my system). But I'm switching the grounds togheter with the signal when selecting the active input.

The Elektor uC that I bought does a nice trick: it can 'learn' the remote (any RC5 - Philips) that the user has available, so no need to search for a specific one. Other features that can be set by the user: equalize relative levels for max. 8 inputs, set the min/max level, set the ramp speed, fade-out & fade-in when switching inputs or muted while switching, balance,...etc.

The same article from Elektor was published later in aXp, worth having a look at it.
About the sound: well, after living with it for +3 years I still like it or I can't find any major faults or I'm getting bored. (pick any and you won't be wrong)

Anyway for me it is time to evolve enter switched ladder and...jogging

I think Pedja has made something out of this, years ago.
Found it:
http://www.pedjarogic.com/spot/index.html

_________________
Carlos Filipe

"Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius - and a lot of courage - to move in the opposite direction." Albert Einstein

That's how every preamp should be made.

_________________
Carlos Filipe

"Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius - and a lot of courage - to move in the opposite direction." Albert Einstein

Gang,

Jocko suggest the DS1666 also yesterday when we had a talk. One question on the data sheet is the Vb. If Vb is 0 volts does that mean it will clip at 0? Or to put it another way... to get their specification of -5v to +5v signal i/o do you need to fix Vb to -5v?

Wolfson part looks good but they are getting pretty goofy on their website about buying parts. I would prefer something available and I like the idea of mono units.

I will be using a PIC probably for the control and display of the output as well as the RF remote control input. I have been using the RF over the IR as I get much better control and when it's asleep it is really dead quiet.

I guess discrete is one way too go with either relays or fet switches... any body have favs here they want to share?

One idea I had was a two stage latter idea ... 6 large steps 8 small steps.

I do need a stable input impedance as the input to the preamplifier is isolated via a custom transformer.

Thanks
Gordon

_________________
Gordon
Chief Scientist
Wavelength Audio

Yes mr. Gordon, the "substrate" or whatever it is should be connected to -5Vdc, I did this way and it does not have measurable distortion straight up to the supply rails.

Hi,

I've got a few samples of the TI chips I've yet to undertake implementing, PGA2320 specifically. I'd be interested in knowing how the Maxim chips measure up to those since Maxim doesn't seem to provide any such specs in their datasheet.

Can anyone who's tried the 2320 comment on the sound though, would it have the BB opa flavor or what? Thanks.

Has anybody done any listening comparisons between the diferent chips? I'm assuming (maybe incorrectly) that they sound different.

Yeah, the '3320 sounds like a cheap op-amp. I have been unable to make it sound better.

The '1666 sounds about the same as the typical conductive plastic pot...........you know.......the affordable ones, that are comparably priced.

Jocko

I'm really sceptic about using digital pots to control volume...
Although I've seen an implementation on a Krell integrated amp I've worked on that ended up sounding quite good (after removing a lot of crap from the preamp). It used an AD 12 bits dac for volume.

Gordon, it seems weird to mix candles with chips.
I suppose you have your motorized volume pot in the main PCB, near the active stages.
What about replacing it for a normal pot, a belt and an offboard motor, decoupled from the chassis?
Mmm...

_________________
Carlos Filipe

"Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius - and a lot of courage - to move in the opposite direction." Albert Einstein

One way to get around worries about digital pots in the signal path (somewhat) would be to use one to provide the control voltage in a variable-gain amplifier stage.

Ultimate inspiration here must be the Ayre KX-R preamp, but I'm sure there's a way to implement this for less than $18,000 (mind you, it's that wonderful preamp's Swiss watch of a non-digital gear-control system that accounts for much of the price). Its designer, your Diyhifi colleague Mr. Hansen, argues that this approach keeps S/N consistent at varying listening levels... It would also serve, if done more simply, to keep the digital pot removed at least one step from the audio path.

I've been meaning to breadboard a trial version using a digital pot and an open-loop AD844 as a transimpedance amp - - someone here suggested that a while back.

Carlos,

The problem is that I am using the 71A directly heated triode for the gain/output stage. With the motor attached to the pot attached to the top plate it will ring the filament and cause audible noise. The top plate is very heavy and has tons of mass (18 pounds of iron) on it but it doesn't matter.

This is an option for the user. I prefer to use the Penny and Giles RF15. But some people are lazy so I am offering as an option the RF remote running into a pic controller and have some visual readout of gain (led string).

I thought about relays but... that's allot of them and tons of resistors to buy even to try it out. I thought about a two stage thingy with 5 resistors into 8 resistors (selected via 74138). So basically 26 relays if I do mono or 13 for stereo fixed balance. It's still and option I am looking at. If I did this I would use rubber isolated standoff's to assure even the relay click would not effect the filaments.

One really silly thing I am thinking about with the DS1666 is too use a joy stick and have both vol and balance in one. Then the remote can do the same. Since the amount of i/o goes down this would be very doable with a 28 pin pic controllers.

One note... if you guys are using pic controllers get the "C" compiler from CCS. I have written assembler for PIC for years. It's kinda of a backwards neumonics compared for most processors. Anyways... I did a SPI thing to initialize the Wolfson dac using a 12F509 and man it took literally more time to install the software than it did too get it too work. Bam... init the controller go to sleep. I did not have to bit bang at all the library functions did everything.

Thanks
Gordon

_________________
Gordon
Chief Scientist
Wavelength Audio

To my ears, two resistor passives sound much more natural and open. The P&Gs are good but are 'coloured' in comparison. 5Rs into 8Rs is not enough and really 0.5 dB steps are needed. I have 125 steps or so in my Placette balanced which sounds better than anything else. Even then some uneven steps still cause problems with ideal playback volume. 34 steps is just rubbish.

If there are no impedance issues, then a fixed R into a grounded Pot sounds good.