Solid State Volume controls

[Bonsai]

I've been doing sopme research for my pre-amp, and considered using a PGA2310/20 integrated volume control and amp.

I came across quite a few remarks on the forum about the aove device, iao comment s about it being garbarge. Then, at the site below, I discovered that AR use a solid state potentiometer on their $5000 LS25 preamp . . . . and its highly rated.

http://www.arcdb.ws/LS25/LS25.html

Here is the digital pot they use . . . .

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

Pavel claims the PGA2310 does not sound good . . . . but CJ Sonograph used it and so does Jeff Rowland on a few of their products. I've also heard that there are PGA2310's ALL over studio mixing consoles (along with loads of op-amps in the signal chain as well).

So where does this leave us?

 

[thespeakerguy]

what's scary about the DS1666 is ...

There is neither a spec or a graph for changes in RDSon as a function of voltage.

They don't tell you if the analog conduction is through PMOS, NMOS, or both in parallel, which is likely the case.

Buffer the output with a FET input amp, is my suggestion if you chjoose to use this.

 

[jan.didden]

I'm using the CS3318 in my 6-channel control for the Behringer DCX. I have had only positive reports about it.
Basically it is a switched resistor network implemented on a chip with a buffer amp. 0.25dB step size, from -96 to +22 dB or so. Distrotion better than -120dB.
Specwise it runs circles around all those other PG's and SS switches.
There is also a 2-channel unit, I believe the CS3308.
Worth to consider.

Jan Didden

 

[analog_sa]

I discovered that AR use a solid state potentiometer on their $5000 LS25 preamp

They also use the same in the LS26 and likely in the REF. I use the same and like it. No, it does not sound as good as a Seiden switch with even simple metal films but i prefer it to any pot i've tried. And it makes balanced/multichannel a breeze. It also sounds much better than the PGA2310, probably because it doesn't include a cheap opamp in the chain.

I would place the DS1666 as the best price/quality compromise volume control i've ever tried; it just doesn't cut it for a statement preamp.

 

[Bonsai]

The point of my post was I hear all these claims about fancy pots, switched resistor netwrks etc, and see a lot of negative 'press' about solid state pots, and yet these same solid state pots get used in 'statement gear'. As Thespeakerguy says, the DS1660 does not even show a distortion plot. do they even measure it in production? How linear is it? etc etc.

What is it that makes a switched resistor network sound 'better' than a PGA2310/20/CS3308 when the measuarable non linearity, cross talk etc for both are below the measurement threshold of all but the very best test gear?

Lets not forget the point about the PGA23xx's in many studio consoles as well.

I think a lot of claims are being made for and against without any decent AB testing to prove the case either way.

 

[Juergen Knoop]

see my comment at the elektor pga2311 circuit.
http://www.diyaudio.com/forums/showthread.php?postid=1582192#post1582192
How much of the bad rep is due to questionable implementions? 😕
regards

 

[analog_sa]

Lets not forget the point about the PGA23xx's in many studio consoles as well.

The main problem of the 2310 is the horrible output opamp. If you can swallow an OPA2132 you'll have no issues with the 2310. And yes, many studio consoles use worse.

A lot of the so called "statement" gear is simply mid-fi inside stunning cases using a couple of boutique components. A badge with history and recognition. Whatever keeps the average audiophile happy.

 

[Bonsai]

hve you tried and compared it to other volume control methods? What were your findings?

 

[Bonsai]

hve you tried and compared it to other volume control methods? What were your findings?

 

[analog_sa]

As mentioned above i have tried DS1666 and PGA2310 against ordinary Alps pots, 6-bit relay attenuators (Nais DS2Y-S-5v) and switched attenuators using Seiden, Elma and Daven switches with both high quality and generic metal film resistors. I don't do any blind tests (life is too short, mine in particular) and any A-B comparisons as that is not the way i listen to audio. So, yes, highly subjective but i really have no point to prove.

The mechanical switched attenuators are simply in a class above the rest, even when using cheap fixed resistors. Followed by the relays and the Ds1666. The relays i tried are probably not of particularly high quality and i find something difficult to express but not very pleasant in their sound.

The Alps pot and the 2310 are at the bottom. They are not similar at all but equally objectionable.

Until recently i also used a transformer volume control. While it had some positive sonic qualities, transparency wise i would rate it below the DS1666 or the relays.

 

[myhrrhleine]

I've been doing sopme research for my pre-amp, and considered using a PGA2310/20 integrated volume control and amp.

I came across quite a few remarks on the forum about the aove device, iao comment s about it being garbarge. Then, at the site below, I discovered that AR use a solid state potentiometer on their $5000 LS25 preamp . . . . and its highly rated.

http://www.arcdb.ws/LS25/LS25.html

Here is the digital pot they use . . . .

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

Pavel claims the PGA2310 does not sound good . . . . but CJ Sonograph used it and so does Jeff Rowland on a few of their products. I've also heard that there are PGA2310's ALL over studio mixing consoles (along with loads of op-amps in the signal chain as well).

So where does this leave us?

>Pavel claims the PGA2310 does not sound good
it's a relative claim.
What are your personal quality standards?
What will you compare this to?

It may work fine for your needs.
If you like the extremes, like the CTC blowtorch, you may find it lacking.

 

[jcx]

digital pots generally use cmos switches to select taps

http://www.analog.com/static/imported-files/data_sheets/AD5290.pdf

may have a little more info than the Dallas/Maxim parts

higher analog supply V gives lower Ron and parasitic C modulation of the switches vs a given signal level

 

[jan.didden]

digital pots generally use cmos switches to select taps

http://www.analog.com/static/imported-files/data_sheets/AD5290.pdf

may have a little more info than the Dallas/Maxim parts

higher analog supply V gives lower Ron and parasitic C modulation of the switches vs a given signal level

If one uses these switches to say, switch taps on an attenuator, the voltage across a closed switch is millivolts. There is no modulation at those levels.

Jan Didden

 

[tvi]

LM1972
"The output of a μPot varies typically between 25kΩ & 35kΩ and changes nonlinearly with step changes. Since a μPot is made up of a resistor ladder network with a logarithmic attenuation, the output impedance is nonlinear. Due to this configuration, a μPot cannot be considered as a linear potentiometer, but can be considered only as a logarithmic attenuator.

It should be noted that the linearity of a μPot cannot be measured directly without a buffer because the input impedance of most measurement systems is not high enough to provide the required accuracy. Due to the low impedance of the measurement system, the output of the μPot would be loaded down and an incorrect reading will result. To prevent loading from occurring, a JFET input op amp should be used as the buffer/amplifier. The performance of a μPot is limited only by the performance of the external buffer/amplifier."

Doesn't seem to have a any internal opamps?

Regards
James

 

[jcx]

If one uses these switches to say, switch taps on an attenuator, the voltage across a closed switch is millivolts. There is no modulation at those levels.

Jan Didden

true jan, V across the sw may be mV but when you want ppm added distortion you need to consider the effects

body leakage and parasitic C modulation give dynamic signal related distortion current/voltage drops across the sw resistance

with Hi-Z load direct r_on modulation may not be very important, but "law faked" linear pot + shunt R will have substantial current in the switches

the on resistance is still a function of gate-source V for each mosfet of the cmos switch, with up to 2 Vrms input (consumer digital sources, phono preamp or pro source may be hotter) the switches at the top of the ladder see their Vgs vary by 5.6 Vpp, this can give r_on modulation large enough to give measurable distortion - especially with +/- 5V supplies

 

[Bonsai]

Good points jcx.


If I go for one of these devices (seems everyone is telling me they 'sound' bad' though - who am I to judge on this specific issue) then th e bes t way to implement the log law is to do it in software - i.e. when the pot is set to a low level, you have to rotate the encoder a lot to get a small increment in volume, while at the top end, the adjustment is much more course. Of course, the pot needs to have a a fine enough resoltion as well to make this work at the user expereince level.

 

[Bonsai]

Here is a review of the coherence pre-amp tha t uses the Crystal 3310.

Amazingly, this preamp has

- Jensen input transformers
- Uses IC's all over the place


http://stereophile.com/solidpreamps/999rowland/index.html

And here is JRDG's take on 'digital volume pots'

http://www.jeffrowland.com/Technology/DigitalVolumeControls.htm

He's also got an interesting article on 'measurement' - check it out.

 

[analog_sa]

Here is a review of the coherence pre-amp tha t uses the Crystal 3310.

Amazingly, this preamp has

- Jensen input transformers
- Uses IC's all over the place

The use of ICs is certainly not amazing. And writing (or reading) reviews is neither easy nor straightforward.

The reviewer only mentions using his active speakers in his assessment. And he's even smart enough to do a "bypass" test 🙂 Only his amazing speakers use a whole lot of OPA2604...pretty much what the 3310 has inside. Could it be that all competent reviewers were unavailable?

 

[jan.didden]

true jan, V across the sw may be mV but when you want ppm added distortion you need to consider the effects

body leakage and parasitic C modulation give dynamic signal related distortion current/voltage drops across the sw resistance

with Hi-Z load direct r_on modulation may not be very important, but "law faked" linear pot + shunt R will have substantial current in the switches

the on resistance is still a function of gate-source V for each mosfet of the cmos switch, with up to 2 Vrms input (consumer digital sources, phono preamp or pro source may be hotter) the switches at the top of the ladder see their Vgs vary by 5.6 Vpp, this can give r_on modulation large enough to give measurable distortion - especially with +/- 5V supplies

True. But wouldn't you use the switch between ground and a leg of the ladder? So they are either closed (with only mV across a few ohms on resistance) or open, with no current true the switch.

What I remember from my old Wireless World project is that I needed to compensate only for the open switch capacitance at the -20dB and the -40dB legs, but that was with CD4007's used as switches. Modern switches are so low cap that probably it isn't necessary now.

Jan Didden

 

[jaycee]

noone's mentioned WM8816 - admittedly its discontinued by Wolfson, but there is a "second source" part - MAS9116 available.