I was recently putzing about with simple stereo volume controls using dual-gang pots.
It struck me that the variance between the single shaft's rotation against two separate wafers would lead to situations where that simple volume control becomes an unintentional balance control. In other words I couldn't really guarantee that the volume pot was accurately reproducing the levels coming out of the source.
I did some tests on various cheap dual gang pots I had laying around and was pretty surprised to see big variances not only each wafer's total resistance (i.e. measure lug 1 against lug 3) but also measuring each gang;s wiper with a given shaft position.
Not very accurate devices I'd say.
So it seems my choices (if I want to be anal about it) are:
- High precision dual-gang pots (where to find?)
- Stepped attenuators (gah, I can't imagine assembling those)
- Dual vactrols controlled by one single gang pot (but would I just be trading variances in pot wafers for variances in vactrols?)
- Digital
I did some searching here and found various tidbits, but should I assume there aren't any other reasonable solutions others than those listed above?
Thanks!
It struck me that the variance between the single shaft's rotation against two separate wafers would lead to situations where that simple volume control becomes an unintentional balance control. In other words I couldn't really guarantee that the volume pot was accurately reproducing the levels coming out of the source.
I did some tests on various cheap dual gang pots I had laying around and was pretty surprised to see big variances not only each wafer's total resistance (i.e. measure lug 1 against lug 3) but also measuring each gang;s wiper with a given shaft position.
Not very accurate devices I'd say.
So it seems my choices (if I want to be anal about it) are:
- High precision dual-gang pots (where to find?)
- Stepped attenuators (gah, I can't imagine assembling those)
- Dual vactrols controlled by one single gang pot (but would I just be trading variances in pot wafers for variances in vactrols?)
- Digital
I did some searching here and found various tidbits, but should I assume there aren't any other reasonable solutions others than those listed above?
Thanks!
Hi,
dual gang linear pots match far better than dual gang logarithmic law pots.
Dual pots always match better at low attenuation rates than at high attenuation rates.
At high attenuation rates log pots are atrociously matched. The balance swings off and can sweep from side to side as the pot approaches the off position.
If you want to avoid this then you must go to stepped attenuator but this requires quite accurately matched resistor values if you want good channel to channel balance at all attenuation settings.
If you can settle for less precision, then a pair of log faking resistors added to a dual linear pot will give some semblance of channel balance that can be far better than dual log.
dual gang linear pots match far better than dual gang logarithmic law pots.
Dual pots always match better at low attenuation rates than at high attenuation rates.
At high attenuation rates log pots are atrociously matched. The balance swings off and can sweep from side to side as the pot approaches the off position.
If you want to avoid this then you must go to stepped attenuator but this requires quite accurately matched resistor values if you want good channel to channel balance at all attenuation settings.
If you can settle for less precision, then a pair of log faking resistors added to a dual linear pot will give some semblance of channel balance that can be far better than dual log.
dano12 said:
I've read that linear pots are more consistent than log/audio pots, but don't know how much of an improvement they would be, or how adaptable to your proposed use.
See http://myweb.tiscali.co.uk/nuukspot/decdun/gainclonepre.html#gcprevc
Various stepped attenuators are available fairly cheaply from Chinese/HK sellers on eBay. I have a few in the parts box though have not bothered testing them. Certainly not the sort of thing I would solder up myself for fun...dano12 said:
Interesting alternative, in spite of the dogmatic response to 'all things digital' that they evoke. I'd be interested to read what you find in this area.dano12 said:
I think my next build will probably incorporate a balance control of some sort- to compensate for room effects as much as volume pot variances.
Interesting thread- thanks for starting it up.
Cheers
John
Here is a clever scheme using a single potentiometer to control the gain of two channels at the same time:
http://electronicdesign.com/Articles/Index.cfm?AD=1&ArticleID=4247
Adapting it to make the variation law logarithmic would be an interesting exercise.
Anyone up to the challenge?
http://electronicdesign.com/Articles/Index.cfm?AD=1&ArticleID=4247
Adapting it to make the variation law logarithmic would be an interesting exercise.
Anyone up to the challenge?
Digital is a good way to go - PGA23xx series from TI.
ADI also have many devices you could try.
I can confirm AndrewT's comments. I have a Marantz CD player which feeds a Marantz amp. The output of th e CD playe r is generallyh too high for th e amp so th e volume control is normally turned way down and the balanc e is all over th e place - even over small rotational angles.
ADI also have many devices you could try.
I can confirm AndrewT's comments. I have a Marantz CD player which feeds a Marantz amp. The output of th e CD playe r is generallyh too high for th e amp so th e volume control is normally turned way down and the balanc e is all over th e place - even over small rotational angles.
I was petrified of building a stepped attenuator myself, but I took the plunge and it came out pretty good. Got it from a DIY site in Australia. I didn't follow the layouts recommended, did my own thing and it turned out just fine, and my soldering skills back in the day were really less than desirable. Uses 1% Dale resistors, and is decent enough tolerance for me.
I don't use a mechanical device at all now, preferring the soundcard's driver panel to control the volume into the amp. Doesn't get better matching than that. I don't have a problem with digital, IMHO as long as your speakers/ears/room are less than perfect (they all are, like it or not) anomalies in channel balance are far more detrimental to the listening experience than the 11th harmonic that is -130dB. At least, to my ears. The way I have things connected are straight from the DAC's I/V converter into the amplifier input, no coupling cap, no pot, nothing. The DAC is fed off a digital output from a soundcard.
Finally the channel imbalance in dual ganged pots is totally real and audible. We have these tiny little Philips pots that are used in literally all of DIY gear all over my country, they claim worst case 3dB tracking between the halves of the pot. 3dB is a *lot*, and is clearly audible as the singer seems to shift positions as the volume control changes. Highly irritating. Worse, the insertion loss and the way they affect the impedance of the amplifier means a buffer is pretty much essential - this is not corrected with even a stepped attenuator. Lawfaking a linear pot may exacerbate the impedance / noise issue, though it does indeed offer better tracking due to the simpler manufacturing process.
Edit: Didn't someone here post a volume control using a LDR? Was it NP?
I don't use a mechanical device at all now, preferring the soundcard's driver panel to control the volume into the amp. Doesn't get better matching than that. I don't have a problem with digital, IMHO as long as your speakers/ears/room are less than perfect (they all are, like it or not) anomalies in channel balance are far more detrimental to the listening experience than the 11th harmonic that is -130dB. At least, to my ears. The way I have things connected are straight from the DAC's I/V converter into the amplifier input, no coupling cap, no pot, nothing. The DAC is fed off a digital output from a soundcard.
Finally the channel imbalance in dual ganged pots is totally real and audible. We have these tiny little Philips pots that are used in literally all of DIY gear all over my country, they claim worst case 3dB tracking between the halves of the pot. 3dB is a *lot*, and is clearly audible as the singer seems to shift positions as the volume control changes. Highly irritating. Worse, the insertion loss and the way they affect the impedance of the amplifier means a buffer is pretty much essential - this is not corrected with even a stepped attenuator. Lawfaking a linear pot may exacerbate the impedance / noise issue, though it does indeed offer better tracking due to the simpler manufacturing process.
Edit: Didn't someone here post a volume control using a LDR? Was it NP?
As Andrew stated - log pots are the worst culprits and yes - high attenuation settings make things worse for both type of laws.
So one should use linear types if appropriate or use law-faking in the first place if possible.
The variance at high attenuation settings can be lessened by two methods:
1.)
Simply don't use those pots at high attenuation settings ! While this sounds strange at first it is quite easily implemented actually. Simply use two series resistors at the "hot" end of the pot which can be shorted by a double switch. I have indeed seen some equipment with such switches quite a few times. The most usual ration was ten (i.e. 20 dB).
2.)
The variance can be made smaller by lowering the possible attenuation ratio. I did this by inserting small resistors at the "cold" end of the pot. This way the variance is not eliminated but reduced. This is of course at the cost of how low you can turn your volume down.
Regards
Charles
So one should use linear types if appropriate or use law-faking in the first place if possible.
The variance at high attenuation settings can be lessened by two methods:
1.)
Simply don't use those pots at high attenuation settings ! While this sounds strange at first it is quite easily implemented actually. Simply use two series resistors at the "hot" end of the pot which can be shorted by a double switch. I have indeed seen some equipment with such switches quite a few times. The most usual ration was ten (i.e. 20 dB).
2.)
The variance can be made smaller by lowering the possible attenuation ratio. I did this by inserting small resistors at the "cold" end of the pot. This way the variance is not eliminated but reduced. This is of course at the cost of how low you can turn your volume down.
Regards
Charles
Thanks for all the info and ideas.
I pulled apart a few cheap dual gang pots last night. One unmarked type actually had a different type of metal on each gang's wiper. LOL I've also only found that almost no datasheets rate the tolerance between gangs.
As for the LDR route, I sketched out a few ideas last night using a VTCL5* part, but can't make it into a voltage divider without adding more components.
I am in the process of building a stepped attenuator just for the misery of all that soldering.
I'll be spelunking through some of the Maxim parts for digital volume controls now.
I pulled apart a few cheap dual gang pots last night. One unmarked type actually had a different type of metal on each gang's wiper. LOL I've also only found that almost no datasheets rate the tolerance between gangs.
As for the LDR route, I sketched out a few ideas last night using a VTCL5* part, but can't make it into a voltage divider without adding more components.
I am in the process of building a stepped attenuator just for the misery of all that soldering.
I'll be spelunking through some of the Maxim parts for digital volume controls now.
sangram said:
All I could find was George's lightspeed attenuator which seems to get rave reviews. However, I didn't understand the need for a dual gang pot and two LDRs per channel. Maybe I'm just a bit thick, but it seems a workable solution would be something like this the eliminates the need for the dual gang pot:
An externally hosted image should be here but it was not working when we last tested it.
(assuming of course I could match LDR1 and LDR2....)
One of the best advices is indeed to avoid the higher attenuation ratios as nearly all rotary faders tend to loose tracking in this region; it´s quite common that one channel has already reached the cutoff region while the other is still playing.
Possible solutions were rotary faders from TKD or penny&giles, back a couple of years Noble pots were quite good in this regard too.
The TKD are matched to ~0.5-1dB afair; the (quite expensive) Penny & Giles rf15 is spec´ed a bit worse, but at least back in the good old days, pg was willing to match down to 0.1 dB for an extra fee.
Not to mention that the pg will not have any problems with dust over the years.
Possible solutions were rotary faders from TKD or penny&giles, back a couple of years Noble pots were quite good in this regard too.
The TKD are matched to ~0.5-1dB afair; the (quite expensive) Penny & Giles rf15 is spec´ed a bit worse, but at least back in the good old days, pg was willing to match down to 0.1 dB for an extra fee.
Not to mention that the pg will not have any problems with dust over the years.
dano12 said:
The problem is that this circuit has a varying output impedance. The answer you seek is on post 4 made by George - using 2 LDRs means a more or less constant impedance of the total circuit, as one LDR increases resistance while the other reduces. In principle, it is exactly the same as a stepped attenuator, where the values of both the resistors change in tandem (or, for that matter of fact, a pot itself).
Any dual pot I can afford is totally unsuitable for audio service. It just isn't that hard to wire up a stepped attenuator and even if the resistor series isn't perfect (based on what's in my junk box), you're unlikely to notice it. You can get the matching almost perfect with an ordinary DVM. The kits are reasonable and make it far easier. IMHO, the worst stepped attenuator is better than the best pot.
Here's an interesting little article that I just came across via Google:
http://www.nutsvolts.com/index.php?/blog/post/what_price_for_perfection/
The interesting part was the graph:
I realize that stepped attenuators are the solution, but if I was to attempt to do small-scale manufacturing of custom audio gear, that would jack the price up pretty high. Of course, as I'm sourcing parts, I'm finding the price of most components is pretty high in the DIY audio world
http://www.nutsvolts.com/index.php?/blog/post/what_price_for_perfection/
The interesting part was the graph:
I realize that stepped attenuators are the solution, but if I was to attempt to do small-scale manufacturing of custom audio gear, that would jack the price up pretty high. Of course, as I'm sourcing parts, I'm finding the price of most components is pretty high in the DIY audio world
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.