precisely setting a resistance in stereo with one knob?

Discussion in 'The Projects Forum' started by dextro, Jan 12, 2012.

  1. dextro

    Thread Starter New Member

    Dec 7, 2010
    14
    0
    Dear all,

    I am working on a project building a stereo audio delay. The project is basically two rebote 2.5 delays in an expensive 19" rack enclosure.

    I am nearly finished, I etched two identical boards and everything works as it should. except for the fact that the PT2399 chip was a bad decision for a stereo project. The IC is a very cheap one and the delay-time (length) is adjusted by changing the frequency the chip operates at. this is done by setting a potmeter configured as rheostat which is connected to the internal RC network.

    When i first tested the two board in parallel i noticed how the two chips operated at different speeds while connected to the same stereo-potmeter (obviously). While anyone (including myself) could have told me this at the beginning I did not foresee this issue.

    Since I already invested quite alot in the project I decided not to give up. I solved one part of the issue by ordering 25 PT2399's and found two which are practically identical (<0.1% difference in delay time).

    The other issue still remains, the setting of two identical resistances with one knob. I bought a relatively expensive stereo-potmeter but it still has like a 1% discrepancy between the two outputs.

    So finally my question: is there a good solution (possibly digital?) to set two resistances (up to 47k) nearly identical with one knob, which is preferably cheap and easy?

    I'm looking forward to hear all your brilliant solutions!
     
  2. MrChips

    Moderator

    Oct 2, 2009
    12,449
    3,365
    I do not have any experience or knowledge of sound recording electronics but I would have assumed that by now they would have moved away from noisy analog pots. But I can imagine that a slider pot is still the easiest type of control on a mixer board.

    I would replace the dual ganged pot with a single analog pot or a digital knob.
    (btw we call it a pot or potentiometer. We usually reserve the term "rheostat" for power control instruments.)

    A digital control may take many forms, two up/down buttons, knobs or a multi-turn knob or dial. A single analog pot can also be used.

    I would then take the signal from the control knob (either analog or digital) and use it to control two digital pots. The two pots will track each other very closely, limited only by the non-linearity of the each pot which is usually better that 1%.

    Hope this helps.
     
  3. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    Does the chip have an option of CLOCK IN or similar to replace the RC delay? I didn't look at a datasheet for it.

    If so, you could get high precision by using one pot to control a single oscillator, then split the output of that oscillator to the clock in of both ICs.
     
  4. dextro

    Thread Starter New Member

    Dec 7, 2010
    14
    0
    @MrChips:
    I am not sure if you understand the point i tried to make with the "rheostat configuration", I meant that the (stereo) pot in the rebote design is used as a variable resistance and not as a variable reference voltage. I have never used digital pots so i do not know much about them. Is it possible to use a digital pot as a variable resistor?

    @thatoneguy:
    The PT2399 does not have such a function. other IC do, i think it is the PT2396.
    however i already made the delay circuits so i'm stuck with this chip.
     
  5. crutschow

    Expert

    Mar 14, 2008
    13,052
    3,244
    Unfortunately the absolute resistance of a digital pot when used as a rheostat (as required in this circuit) is much worse than the potentiometer linearity accuracy, which is what you are quoting. So I don't think a digital pot is the answer to the OP's problem.

    Perhaps a small value trimpot in series with each stereo analog pot would allow enough adjustment so the two will track with sufficient accuracy.
     
  6. dextro

    Thread Starter New Member

    Dec 7, 2010
    14
    0
    @crutschow:
    I was able to sample two AD5206 256 bit digital pots just now. but i agree those will probably have the same inaccuracy as analog pots when used as variable resistor.

    Using an added resistance on the smallest of the two stereo pot lugs only halves the maximal error since the error varies linear along with the value. So this still is no real cure since the 9€ pot i have has up to 4% error at some intervals
     
  7. rogs

    Active Member

    Aug 28, 2009
    279
    37
    Some of the information and links from this page may be useful.
    Seems as if the 'VCO' input (pin 6) that sets the delay time is in fact a current source - which is both a bit confusing, and a bit of a nuisance! Why call it a VCO (Voltage Controlled Oscillator) if it's not?

    Anyway, some of those ideas maybe of some help....
     
  8. crutschow

    Expert

    Mar 14, 2008
    13,052
    3,244
    The spec on the Analog Devices pot is 0.25% nominal and 1.5% maximum resistance match between two specific pots on the same chip. I would expect that the correction of that amount of initial deviation with a trimpot would allow adequate tracking over the pot's adjustment range sufficient for the OP's needs.
     
  9. dextro

    Thread Starter New Member

    Dec 7, 2010
    14
    0
    Yes. i also saw that particular spec and it looks promising. i ordered two 6 channel devises and since i only need 2 i assume there will be at least one usable pair.

    the only thing which makes me doubt is this: on my analog pot the total resistance matches quite well (~0.5%) but the error of the values on the wipers when the pot is halfway go up to 4%. meaning the pots are not exactly linear and more importantly non-linear in a different way from each other. I hope the digital pots do not have this issue.
     
  10. dextro

    Thread Starter New Member

    Dec 7, 2010
    14
    0

    Yes that looks very good indeed. I had run in to the page before but did not notice the guy actually succeeded in controlling the delay time with a current source. Also i found this thread with someone else who used the same technique with succes.

    I do not have any experience with current sources But i would assume it is possible to control two currents at the same time with good precision, easier i would guess than two resistances, right?

    My circuit design is suitable for this solution since i have a direct connection to pin 6 because the pot is not on the board.
    Thanks for the tip! and i will look into it.
     
  11. rogs

    Active Member

    Aug 28, 2009
    279
    37
    I would think it would be possible to be considerably more accurate. In the configuration shown, the use of a DC summing junction at the -ve input of the op amp means you should be able to 'offset' anty diffenence between the two devices with the 'delay' pot shown as the calibration, and then simply control both with a common control voltage input. Should track pretty well......
     
  12. crutschow

    Expert

    Mar 14, 2008
    13,052
    3,244
    You can generate two currents with fairly close values by using three identical transistors in a "current mirror" configuration. One transistor is the master and two are the slaves which would connect to the delay chips.
     
  13. dextro

    Thread Starter New Member

    Dec 7, 2010
    14
    0
    thanks! i looked it up and that is probably the easiest solution to a problem i have ever had. Also i think the adjustable current source is not needed since i can regulate one delay with a single pot and just copy the current coming out of pin 6 to the other delay. However putting a NPN somewhere between pin 6 and ground might also influence the PT2399 because of the collector emitter voltage.

    I will try this as soon as possible and report back.
     
    Last edited: Jan 13, 2012
Loading...