1. Things you need to know about the new ‘Conversations’ PM system:

    a) DO NOT REPLY TO THE NOTIFICATION EMAIL! I get them, not the intended recipient. I get a lot of them and I do not want them! It is just a notification, log into the site and reply from there.

    b) To delete old conversations use the ‘Leave conversation’ option. This is just delete by another name.
    Dismiss Notice

ALPS potentiometers, gain, and distortion...

Discussion in 'd.i.y.' started by teddy_pardo, Sep 5, 2009.

  1. teddy_pardo

    teddy_pardo Trade: Teddy Pardo

    While doing distortion measurements on some preamplifiers I discovered incidentally that ALPS (blue velvet, those used by Naim) introduce distortions. The distortion level depends very much on the potentiometer position. At low volume levels 7-12 (o'clock) distortion level is acceptable, but at high volume levels between 12-4 distortion can be very high, it can grow to around 0.5% !!!

    Obviously at full volume the distortion level introduced by the potentiometer is near zero.

    The results were consistent with several potentiometers that I tried (10K and 20K logarithmic). The explanation may be related to the fact that at high volume levels the surface on logarithmic potentiometer is different (changes more rapidly per mm).

    Anyhow, those who reduce the peramp gain, think (or listen) again...

    Teddy
     
  2. trancera

    trancera pfm Member

    0.5% thats totally unacceptable. Must be something else going on somewhere surely ?!
     
  3. mikesnowdon

    mikesnowdon resU deretsigeR

    Any ways to counter this Teddy? RC filter perhaps?

    Did you have any experience with the RK16 type?
     
  4. PigletsDad

    PigletsDad pfm Member

    Teddy, can you test if the distortion depends on the load impedance?

    My mental picture about what is happening is that the sliding contact and the conductive plastic track act a bit like a diode. If this were correct, the effect should be worse with a low load impedance, but a sufficiently high impedance load would greatly reduce it.

    The track is a semiconductor, the contact metal, so it is a bit like a metal/semiconductor junction, as used in say a schottky diode.
     
  5. martin clark

    martin clark pinko bodger

    Teddy, what exactly are you measuring?

    If you are running a high-level signal into the preamp input (say 1vrms) and measurig the output of the preamp, then (given the gain of 13x) much past about '10.30' you'll be driving the preamp into clipping, so yes distortion will increase markedly. And if you are attenuating the test dsignal to get to full volume, do check its not partly teh fault of your signal source...

    Also - what's the spectrum of the distortion you're seeing?

    Just interested...
     
  6. teddy_pardo

    teddy_pardo Trade: Teddy Pardo

    Input signal was 2V RMS 1KHz played from a CD, I measured the distortion right before the potentiometer (was very low), and right after the potentiometer. There was no clipping, and as I said, when the potentiometer was turned clockwise to maximum, the distortion disappeared. I have used an HP/Agilant audio analyzer.

    I have tested it both on a Naim preamp where the volume potentiometer is followed by the balance potentiometer (I think so...) and in any case there is another resistor in the gain stage, and in another case there was an opamp with high input impedance. IIRC, the results were similar.

    PD - your idea makes sense, but still there is a big difference that depends on the position. I didn't make tests with a linear potentiometer (I don't have one), but I think that the results may be different. That is, I think it is related to the fact that the potentiometer is logarithmic.
     
  7. martin clark

    martin clark pinko bodger

    Cool. Watching this with some interest.
     
  8. mikesnowdon

    mikesnowdon resU deretsigeR

    Curious to know at what frequency does this distortion appear worst?
     
  9. binnie

    binnie Binnie

    Interesting Teddy !

    Is the distortion level equal on both channels ?

    Yep, at least in my 72 (729 -> poti -> balance -> 321s -> 325).

    Will shunting the potentiometer accentuate the distortion ?
    Olivier
     
  10. markt

    markt hello

    A little experiment, try 100k or 200k (or whatever is 10x the pot value) metal films across in>out and in>ground both channels, your ears might be surprised but I'm not sure if it would measure any different. Be interesting to see if someone could measure it, anyway I found it made a nice difference and it's a helluva lot cheaper than a fancy pot.
     
  11. martin clark

    martin clark pinko bodger

    Good idea Markt, the in-out R is worth a try.

    My preamp came with a Penny&Giles fader (10k). It puzzled the heck out of me that the end-to-end R varied a little as I mapped the output z balance between channels, untl I realised it must be set up like that inside. Sounds great (and actual channel balance is also extraordinarily good)

    Rememeber a 'log' pot isn't logarithmic! All are effectively linear pots with two different resistances joind in the middle, IYSWIM: a two-slope approximation to a log curve. Which might be enough to explain your result, given PD's suggestion and/or some non-linear capacitance in the load.

    We still have to square your observation with the (rather lower) observed total distortion measured / reported in tests though. Keep digging!
     
  12. markt

    markt hello

    Anything slightly wacky and I'm there...
     
  13. mikesnowdon

    mikesnowdon resU deretsigeR

    As in this kind of thing?
     
  14. martin clark

    martin clark pinko bodger

    ;)

    Anyway, the value my P&G unit must be around 50K - the 10K nominal value drops to 8k at maximum attenuation. That's possibly enough to swamp the effects of nonlinearity at the wiper, and also nonlinearity in the notional 'log' tracking. Of course it demands that sources are happy with such a load.
     
  15. markt

    markt hello

    It's a great way to superfine tune channel balance with R value as well should you have the resistors to play with. Something that bugs me a bit with pot values not always (rarely?) being matched left/right.
     
  16. binnie

    binnie Binnie

    mikesnowdon, yes exactly !
    Now I have the link again :)
    Olivier
     
  17. cromodora

    cromodora foshfishfie

    Regular shunting is passe.

    Optivol anyone?
     
  18. martin clark

    martin clark pinko bodger

    IIRC Pigletsdad already pointed out the ways in which LDRs can lead to quite high distortion.
     
  19. PigletsDad

    PigletsDad pfm Member

    The log attenuation versus position curve is done by varying the width of the track. On a linear pot, the track is constant width. In a log pot, it gets narrower towards the high volume end, IIRC. Maybe this makes the effects of dissimilar materials worse, by giving a smaller contact spot, hence a higher contact resistance.

    Yes, LDRs are highly nonlinear.
     
  20. fran

    fran pfm Member

    They may be, but to these ears LDRs sound much better than either alps or an admittedly cheap but well built stepped pot off ebay.


    Fran
    EDIT: LDRs are often trotted out as being so much better etc etc. Believe me, I would much prefer if that €20 stepped pot off ebay was better. No messng with a power supply for the LDR and above all - no bloody matching!! What a PITA. However, my ears tell me differently, but naturally respect what others desires are.

    Interestingly all the airtight amps ($$ amps) use mono blue velvet pots.
     

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice