I have a question for those of you who may have done this sort of thing, or know more about it - the volume control potentiometers on the xDuoo headphone amps are 50k ohms. They all don't have enough taper - they turn up loud too quickly. They start at about 6:30 o'clock position, and by 8 o'clock they are more than loud enough.
I want to replace the 50k ohm pots with 100k ohm, and I have found such a dual ganged pot on Mouser:
https://www.mouser.com/ProductDetail/Bourns/PDB182-K430K-104A?qs=h2IHEVivlqD8cnChHAVPeg%3D%3D
I think this is the physical match for the unit in these headphone amps. They are wicked cheap at $1.91 each.
My questions are:
1) What pot can I use that will have better channel balance?
2) What pot will have high quality wipers for better sound quality?
3) What is the difference(s) between so-called audio taper, reverse audio, and linear taper?
The MT-601/602 have the dual ganged unit, and the MT-604 has separate volume controls for each channel, so they are not ganged.
I want to replace the 50k ohm pots with 100k ohm, and I have found such a dual ganged pot on Mouser:
https://www.mouser.com/ProductDetail/Bourns/PDB182-K430K-104A?qs=h2IHEVivlqD8cnChHAVPeg%3D%3D
I think this is the physical match for the unit in these headphone amps. They are wicked cheap at $1.91 each.
My questions are:
1) What pot can I use that will have better channel balance?
2) What pot will have high quality wipers for better sound quality?
3) What is the difference(s) between so-called audio taper, reverse audio, and linear taper?
The MT-601/602 have the dual ganged unit, and the MT-604 has separate volume controls for each channel, so they are not ganged.
Nothing less than the key to your problem, quite possibly:
Beginners' Guide to Potentiometers
I suspect the original pots may be linear taper, what are they labeled as? Should be easy enough to check with a multimeter. Lin pots tend to give quickly rising volume far down but very little more happening by the 12 o'clock point.
I would not go higher than the original 50k unless your source absolutely requires it (very few things do these days). Even going down to 10k may be worth considering in the interest of lower noise, although channel tracking on average seems to be better for the higher values.
If the original pot is logarithmic and you're still confined to the very lowest ranges, you have too much overall gain in the playback chain and learning about gain staging would be in order. You should normally be at 10-11 o'clock at normal volume. 6:30 to 8:00 max is way low, no wonder you've got tracking issues. This may be a combination of multiple factors, e.g. sensitive IEMs on an amp of highish gain combined with a 2 Vrms source and no playback volume normalization (e.g. ReplayGain) active.
The MT-601, for example, seems to have a gain of +18 dB, which is definitely higher than average. A Schiit Magni Heresy gives you a choice of +15.5 dB and 0 dB, with the IEMagni even adding a setting with overall negative gain for yet greater flexibility. Given the bandwidth of headphone sensitivities (easily 30+ dB) this feature is direly necessary unless your headphone arsenal is restricted to medium to low sensitivity models.
If you have no issues with amplifier noise, the easiest solution aside from the playback software side may be adding a line-level attenuator in front of the amplifier. An L-pad with R1 = 10k, R2 = 3k3 (for ~-12 dB and change) should be a decent start.
Okay, I appreciate your reply - I will try to check with a multimeter. The only distinguishing mark on it is "IDL" and "A50K".
With the multimeter probes on the first and second pot leads, and the pot turned all the way down (~6 o'clock) it reads ~1.5 ohm
~7:30 o'clock = ~1k ohm
~9:00 o'clock = ~2.76k ohm
~10:30 = ~6.46k ohm
~12:00 = ~16.4k ohm
~1:30 = ~32.9k ohm
~3:00 = ~44.6k ohm
~4:00 o'clock = all the way up = 45.08k ohm The same for reading the first and third pot leads.
I am using Monolith M570 headphones: 32 ohm, 96dB
These amps would be waaaaaaay too loud for any IEM.
Thanks for the link - the guide was very helpful. If I put the know at the approximate halfway point - at about 11 o'clock - it reads about 7.5k ohm. So, this seems to be a log/audio taper? This would seem to confirm the A50K marking; most likely being the "new code".
For the L-pad you're suggesting: 10k ohm on the series resistor and 3k or 33k (or something else?) for the shunt?
With the multimeter probes on the first and second pot leads, and the pot turned all the way down (~6 o'clock) it reads ~1.5 ohm
~7:30 o'clock = ~1k ohm
~9:00 o'clock = ~2.76k ohm
~10:30 = ~6.46k ohm
~12:00 = ~16.4k ohm
~1:30 = ~32.9k ohm
~3:00 = ~44.6k ohm
~4:00 o'clock = all the way up = 45.08k ohm The same for reading the first and third pot leads.
I am using Monolith M570 headphones: 32 ohm, 96dB
These amps would be waaaaaaay too loud for any IEM.
Thanks for the link - the guide was very helpful. If I put the know at the approximate halfway point - at about 11 o'clock - it reads about 7.5k ohm. So, this seems to be a log/audio taper? This would seem to confirm the A50K marking; most likely being the "new code".
For the L-pad you're suggesting: 10k ohm on the series resistor and 3k or 33k (or something else?) for the shunt?
Attachments
Last edited:
I am trying to think how the L-pad could be built - it would involve an L-pad for each channel, and then connecting it would be 4 RCA jacks, and another pair of interconnects. And the MT-604 would be XLR's.
That would be harder than swapping out the pots.
I got this idea from Gishelli Labs - they offer a 10k and 100k pot as options. The added noise is a concern that I didn't realize was possible.
On the pot I linked to in my first post - is the "carbon" element something that might be lower noise and/or better channel tracking? Or is there an element that would be likely to be better quality?
That would be harder than swapping out the pots.
I got this idea from Gishelli Labs - they offer a 10k and 100k pot as options. The added noise is a concern that I didn't realize was possible.
On the pot I linked to in my first post - is the "carbon" element something that might be lower noise and/or better channel tracking? Or is there an element that would be likely to be better quality?
Sorry again, sometimes I get into rapid-scroll mode and don't read carefully before replying.
I often see (not so much here, but on another site that claims to be "scientific") postings that are confused about volume controls. Typical mistakes are claiming one amplifier is more powerful than another because it plays louder at the same volume setting. Or hand-wringing because the rated power of the amplifier (typically with a switchable input attenuator set to "low") cannot be reached with normal input voltage levels.
It's all about gain distribution and overload margin. You want to be able to drive the headphones to a comfortable level with usable volume adjustment range, plus or minus, and enough headroom to avoid clipping. If that's the case, then nothing else much matters.
Barring a defect in the amplifier, your issue (as sgrossklass nicely summarized) is almost certainly too much amp gain for your source signal level. I usually use stepped attenuators and design them to ramp up quickly over the first few steps. The reason is if they followed the log curve all the way down, there still be music playing at the lowest step. So, typically, the last few steps are steep to smooth the transition to zero volume. This lets you use finer steps over the middle of the range where you would normally set the control. I design my headphone amps for a gain of three (about 10 dB), set the volume at about two o'clock, and then use digital control (Roon) for listening adjustment.
Adding a voltage divider (L-Pad) at the input lets you knock down the excess gain. Because resistors generate thermal noise and throw away signal, this will reduce your signal-to-noise ratio. That's unlikely to be a problem because the signal levels are high and the gain of the headphone is relatively low, meaning the added noise from the L-Pad shouldn't be amplified enough to be audible over the music. You also don't want to make the input impedance seen by the source too low. But with a 10K "R1", it should be fine for any solid-state source component.
Another consideration is that the L-Pad will increase the source impedance seen by the amp looking backwards from the input. This may reduce the HF response a little, but with the suggested resistor values it's unlikely to be a problem.
Modifying the amplifiers internally to have less gain (not just attenuating the signal) is a theoretical possibility but not at all practical except for gurus who are willing to void the warranty.
I often see (not so much here, but on another site that claims to be "scientific") postings that are confused about volume controls. Typical mistakes are claiming one amplifier is more powerful than another because it plays louder at the same volume setting. Or hand-wringing because the rated power of the amplifier (typically with a switchable input attenuator set to "low") cannot be reached with normal input voltage levels.
It's all about gain distribution and overload margin. You want to be able to drive the headphones to a comfortable level with usable volume adjustment range, plus or minus, and enough headroom to avoid clipping. If that's the case, then nothing else much matters.
Barring a defect in the amplifier, your issue (as sgrossklass nicely summarized) is almost certainly too much amp gain for your source signal level. I usually use stepped attenuators and design them to ramp up quickly over the first few steps. The reason is if they followed the log curve all the way down, there still be music playing at the lowest step. So, typically, the last few steps are steep to smooth the transition to zero volume. This lets you use finer steps over the middle of the range where you would normally set the control. I design my headphone amps for a gain of three (about 10 dB), set the volume at about two o'clock, and then use digital control (Roon) for listening adjustment.
Adding a voltage divider (L-Pad) at the input lets you knock down the excess gain. Because resistors generate thermal noise and throw away signal, this will reduce your signal-to-noise ratio. That's unlikely to be a problem because the signal levels are high and the gain of the headphone is relatively low, meaning the added noise from the L-Pad shouldn't be amplified enough to be audible over the music. You also don't want to make the input impedance seen by the source too low. But with a 10K "R1", it should be fine for any solid-state source component.
Another consideration is that the L-Pad will increase the source impedance seen by the amp looking backwards from the input. This may reduce the HF response a little, but with the suggested resistor values it's unlikely to be a problem.
Modifying the amplifiers internally to have less gain (not just attenuating the signal) is a theoretical possibility but not at all practical except for gurus who are willing to void the warranty.
Thanks - I appreciate the follow up. Every reviewer I have seen, who has used these xDuoo MT-60x headphone amps has said they have too much gain. Sadly, they lack a gain switch.
And every single reviewer who has used these, say that despite the awkward volume control slope, that they sound very good. They are tube buffered (obviously) and they state the discrete solid state output stage is Class A. The odd thing is the MT-601 is the only one that has heat sinks on those output transistors.
The MT-604 has separate volume pots for each channel, so balancing things is both easier and not convenient.
I may try replacing the pot on the MT-602, just for grins. If it gets too noisy, then I can go back to the stock A50K pot. But I now know for sure, that the log/audio slope is the one to use.
Thank you for educating me on this!
And every single reviewer who has used these, say that despite the awkward volume control slope, that they sound very good. They are tube buffered (obviously) and they state the discrete solid state output stage is Class A. The odd thing is the MT-601 is the only one that has heat sinks on those output transistors.
The MT-604 has separate volume pots for each channel, so balancing things is both easier and not convenient.
I may try replacing the pot on the MT-602, just for grins. If it gets too noisy, then I can go back to the stock A50K pot. But I now know for sure, that the log/audio slope is the one to use.
Thank you for educating me on this!
3.3k. Sorry, using the prefix for the decimal point is Euro nomenclature.
You'd typically be building one into a cable extension (female to male), towards the amplifier input side. 1/4 W axial MF resistors and shrinkwrap galore (or tape galore if you must) so things can be kept both slim and isolated, i.e. the L is basically folded up and squished flat... with the requisite ground connection it turns into more of a Z.
Then you'd need a double L-pad as shown here. (In these days of inexpensive precision resistors, this configuration is actually preferred over the 'n' pad, which does not reduce the amplitude of the common-mode signal component. As a DIYer you can easily buy a bunch of cheap 1% Yageos and hand-match them with a multimeter.)
Yeah, but swapping pots will do nothing about excessive gain. A volume pot is wired as a rheostat, so two pots of the same law but different values will give pretty much the exact same output for a given position.
Any run of the mill pot has carbon tracks, but quality may still vary widely from dodgy cheapies to ALPS Blue Velvet (RK27).
Last edited:
Thanks for all the additional information! I wondered about the 3k3 - it makes sense now that you mention this convention. We are often doing a bit of "cultural" translations, here on the Intertubes.
The best part of the L-pad method is - it doesn't affect the amp. And with your mention of the wattage of the resistors, I will do this. I only recently did my first analog crossover, that used an L-pad. XSim did the calculations, including the wattage required. I needed two in parallel on the series resistor, but the shunt takes less power, so was fine with one.
On the single ended amps, I can add the L-pad onto the end of the interconnects, that I put together. I have shrink wrap sleeves and a good heat gun, and a new soldering iron, for that matter.
EDIT: on the article you linked to, it mentions using the N-pad for both a single ended or balanced - in your example of 10k and 3k3 for the L-pad - would the 10K be split into two 5k for the + and - lines? Also, what is an MR resistor?
On the two different value pots having similar values at the same position - I would imagine this is essentially true at the beginning of the slope, but by definition, the difference would have ramp up as the pots are turned up? So, merely doubling the max value would have a minimal effect at the low settings - that are required by the high gain of these amps.
Again, thank you for taking the time and having the patience to guide me through this. I don't know if I mentioned this earlier in the thread, but I know the pots used in my Audible Illusions Modulus 2C are better than the ones used in the 2B - they apparently have some sort of "plastic" tracks, that apparently are better sounding, and they have lasted and are completely noise-free with about 34 years of hard use.
I now also have a Modulus 3A, which has a Swiss made stepped attenuator. It sounds great, but with my newest MLTL speakers, which have 92dB sensitivity woofers, the steps at the bottom are noticeably farther apart than what I can get from the M2C volume pots. For more typical sensitivity speakers, like the first pair of MLTL speakers I built, that take about 3 clicks more gain to get similar SPL, the steps are *fine*.
The best part of the L-pad method is - it doesn't affect the amp. And with your mention of the wattage of the resistors, I will do this. I only recently did my first analog crossover, that used an L-pad. XSim did the calculations, including the wattage required. I needed two in parallel on the series resistor, but the shunt takes less power, so was fine with one.
On the single ended amps, I can add the L-pad onto the end of the interconnects, that I put together. I have shrink wrap sleeves and a good heat gun, and a new soldering iron, for that matter.
EDIT: on the article you linked to, it mentions using the N-pad for both a single ended or balanced - in your example of 10k and 3k3 for the L-pad - would the 10K be split into two 5k for the + and - lines? Also, what is an MR resistor?
On the two different value pots having similar values at the same position - I would imagine this is essentially true at the beginning of the slope, but by definition, the difference would have ramp up as the pots are turned up? So, merely doubling the max value would have a minimal effect at the low settings - that are required by the high gain of these amps.
Again, thank you for taking the time and having the patience to guide me through this. I don't know if I mentioned this earlier in the thread, but I know the pots used in my Audible Illusions Modulus 2C are better than the ones used in the 2B - they apparently have some sort of "plastic" tracks, that apparently are better sounding, and they have lasted and are completely noise-free with about 34 years of hard use.
I now also have a Modulus 3A, which has a Swiss made stepped attenuator. It sounds great, but with my newest MLTL speakers, which have 92dB sensitivity woofers, the steps at the bottom are noticeably farther apart than what I can get from the M2C volume pots. For more typical sensitivity speakers, like the first pair of MLTL speakers I built, that take about 3 clicks more gain to get similar SPL, the steps are *fine*.
Last edited:
On Mouser, I figured out how to see the axial metal film resistors, and after I select precision metal film resistors, and 3.6k and 8.2k, I get these two Vishay resistors:
RN60D3601FB14 Vishay / Dale | Mouser
RN65E8201FB14 Vishay / Dale | Mouser
I can buy six of the 3.6k, and twelve of the 8.2k and have them shipped for just shy of $25.
But these are not small - the 8.2k are 4.57mm in diameter X 14.27mm long. And the 3.6k are 3.68mm in diameter X 8.74mm long.
I will try without the precision ... selection.
Okay, their data search is pretty powerful - these are the smallest diameter 1% 1/4 watt units:
MFR-25FBF52-3K6 YAGEO | Mouser
MFR25SFTF52-8K2 YAGEO | Mouser
I can buy twelve of the 3.6k, and twenty five of the 8.2k, shipped for less than $11. Just $2.46 for all 37 of the resistors, and $7.99 for the least expensive shipping - they offer at least 3 ways to ship at this same price.
These are much smaller - the 8.2k are 1.9mm diameter X 3.4mm long. And the 3.6k are ironically larger - 2.4mm in diameter X 6.3mm long.
RN60D3601FB14 Vishay / Dale | Mouser
RN65E8201FB14 Vishay / Dale | Mouser
I can buy six of the 3.6k, and twelve of the 8.2k and have them shipped for just shy of $25.
But these are not small - the 8.2k are 4.57mm in diameter X 14.27mm long. And the 3.6k are 3.68mm in diameter X 8.74mm long.
I will try without the precision ... selection.
Okay, their data search is pretty powerful - these are the smallest diameter 1% 1/4 watt units:
MFR-25FBF52-3K6 YAGEO | Mouser
MFR25SFTF52-8K2 YAGEO | Mouser
I can buy twelve of the 3.6k, and twenty five of the 8.2k, shipped for less than $11. Just $2.46 for all 37 of the resistors, and $7.99 for the least expensive shipping - they offer at least 3 ways to ship at this same price.
These are much smaller - the 8.2k are 1.9mm diameter X 3.4mm long. And the 3.6k are ironically larger - 2.4mm in diameter X 6.3mm long.
Last edited: