There are two electrolytic capacitors that concern me. C20, C7 are good capacitors. But these are electrolytic capacitors and an audio signal passes through them.
Why not film?
An electrolytic capacitor has ionic conductivity and this is not very good for sound.
Why not film?
An electrolytic capacitor has ionic conductivity and this is not very good for sound.
I've never been able to measure any sonic impact of the Nichicon MUSE UES-series. I don't see widespread complaints about the sound of the Modulus-86 either. So it seems like a non-problem, thus requiring no solution.
Any sonic impact of the two UES-series capacitors would be about two decades below the audible spectrum, thus, inaudible.
Most have commented that they prefer the sound signature of the Modulus-86 Rev. 3.0 (which has the UES caps) to Rev. 2.4 and below, which does not have the caps. So maybe there's more to sound signature than a couple of capacitors. 🙂
Tom
Any sonic impact of the two UES-series capacitors would be about two decades below the audible spectrum, thus, inaudible.
Most have commented that they prefer the sound signature of the Modulus-86 Rev. 3.0 (which has the UES caps) to Rev. 2.4 and below, which does not have the caps. So maybe there's more to sound signature than a couple of capacitors. 🙂
Tom
It’s just that a film capacitor is more durable, doesn’t dry out, doesn’t leak, and doesn’t make noise.
The UES capacitors don't make noise either.
A 100 uF, 35 V film capacitors would be the size of a hand grenade and incredibly expensive. It would offer higher ESR and ESL than the electrolytic capacitor, hence, be further from an ideal capacitor than the UES.
But you're absolutely right that a large film capacitor would command attention from some audiophiles. I doubt Duelund has any issues selling their components for example. But I cater to those who value science and engineering higher than marketing babble and urban legends.
It's no secret that the Modulus-86 contains two capacitors in the signal path. That's immediately obvious from the information available on the Modulus-86 Product Page. If that doesn't suit your fancy, you can either omit the capacitors (and get higher DC offset) or you can realize that it doesn't tickle your desires and keep scrolling. I'm fine either way. This is DIY after all and we don't all have to like the same stuff.
Tom
A 100 uF, 35 V film capacitors would be the size of a hand grenade and incredibly expensive. It would offer higher ESR and ESL than the electrolytic capacitor, hence, be further from an ideal capacitor than the UES.
But you're absolutely right that a large film capacitor would command attention from some audiophiles. I doubt Duelund has any issues selling their components for example. But I cater to those who value science and engineering higher than marketing babble and urban legends.
It's no secret that the Modulus-86 contains two capacitors in the signal path. That's immediately obvious from the information available on the Modulus-86 Product Page. If that doesn't suit your fancy, you can either omit the capacitors (and get higher DC offset) or you can realize that it doesn't tickle your desires and keep scrolling. I'm fine either way. This is DIY after all and we don't all have to like the same stuff.
Tom
Great. Now go measure the distortion of a 100 uF, 35 V Nichicon MUSE UES-series when loaded by 2.2 kΩ, ±0.1 % metal film. Let me know what you find.
Here's the distortion of the Modulus-86, including the two Muse UES caps, at just below clipping levels. Let me know which part of this distortion curve you believe is from the capacitors.
Tom
Here's the distortion of the Modulus-86, including the two Muse UES caps, at just below clipping levels. Let me know which part of this distortion curve you believe is from the capacitors.
Tom
An electrolytic capacitor introduces distortion, this is a fact.
If the goal is to minimize distortion of the audio path, then electrical capacitors should be eliminated. This is the logic.
When you make the next version you can take this fact into account.
If the goal is to minimize distortion of the audio path, then electrical capacitors should be eliminated. This is the logic.
When you make the next version you can take this fact into account.
My reading of Self’s book is not “Do not use electrolytics in the signal path” it is “When you use an electrolytic in the signal path, ensure that it is sized such that there is always less than 80mv across the capacitor at the lowest frequency of interest.” Not following that advice means you would see a kick in distortion at low frequencies. There is no such kick in Tom’s distortion graph which means that Tom has done this correctly.
If you’re so hung up about it why don’t you design and market you’re own version?
I do not argue. Of course that's right. Tom made a good sound amplifier. But it’s better when there are no capacitors in the sound path.
Except that the evidence suggests otherwise. Done right it isn’t a problem.
Anyway, I don’t want to derail Tom’s thread. I’m pretty sure this has been discussed elsewhere and I think Nelson has a capacitor thread going somewhere here. That might be a better place to discuss the merits of what type of capacitor to use, if any, for signal coupling.
Suffice to say that if I had the budget to build a few more amps, Tom’s would be at the top of my list. (Not that I think they are unduly expensive, it’s just that I’ve already spent as much as I can reasonably justify on amplifiers already).
Bruce Hofer of Audio Precision fame has made similar observations.
It's like many other things. A tool in skilled hands can produce great results. The same tool in unskilled hands can produce disaster.
Yep. Measurements don't lie.
Tom
Hey Tom,
I am eyeing your modulus 86 boards, and I am about to pull the trigger on a pair of them. I have a question though. I have a very special amplifier chassis that I want to use for this, and I have very limited space. It means that I would need to minimize the height of the board. To me it seems that the two elecrolytics are the tallest part of the assembly. Can I use shorter caps there? I would also like to design a power supply board that has the same footprint as the mod86, and place it above the amp board with standoffs. How big clearance do you think I should have? What would be the smallest possible distance between the two boards? Thanks in advance.
torzsok
I am eyeing your modulus 86 boards, and I am about to pull the trigger on a pair of them. I have a question though. I have a very special amplifier chassis that I want to use for this, and I have very limited space. It means that I would need to minimize the height of the board. To me it seems that the two elecrolytics are the tallest part of the assembly. Can I use shorter caps there? I would also like to design a power supply board that has the same footprint as the mod86, and place it above the amp board with standoffs. How big clearance do you think I should have? What would be the smallest possible distance between the two boards? Thanks in advance.
torzsok
The capacitors I include in the Modulus-86 Stereo Kit stand 25 mm tall (nominally). According to the data sheet they might reach 27 mm.
The LM3886 will reach at least 0.810" (20.6 mm) above the board and its pins will protrude through the board by 0.104" (2.64 mm).
The connectors for the power supply and output are 21.5 mm in height.
Those are the tall boys on the board. You could switch to the Nichicon UBY1H102MHL1TN for the capacitors. That'll get you down to 21.5 mm max height.
I would allow at least 4-5 mm between boards to make sure that the pins of one board don't make contact with any of the components on the other board. If you make sure to trim the leads you could use 25 mm long threaded standoffs to space out the boards. It would be tight, but it would be doable if you trim the pins after soldering. Personally I'd rather go with 30 mm standoffs and have a little more wiggle room.
Tom
The LM3886 will reach at least 0.810" (20.6 mm) above the board and its pins will protrude through the board by 0.104" (2.64 mm).
The connectors for the power supply and output are 21.5 mm in height.
Those are the tall boys on the board. You could switch to the Nichicon UBY1H102MHL1TN for the capacitors. That'll get you down to 21.5 mm max height.
I would allow at least 4-5 mm between boards to make sure that the pins of one board don't make contact with any of the components on the other board. If you make sure to trim the leads you could use 25 mm long threaded standoffs to space out the boards. It would be tight, but it would be doable if you trim the pins after soldering. Personally I'd rather go with 30 mm standoffs and have a little more wiggle room.
Tom
Thanks, I think it will try to make it work with the 25mm standoffs. I understand that I need to do a very clean job on the underside of the PS board.
As for the PS board: I will use separate boards for each of the two channels. I am going with an ideal rectifier (surplus from an earlier project). How big capacitors do you think I should use?
As for the PS board: I will use separate boards for each of the two channels. I am going with an ideal rectifier (surplus from an earlier project). How big capacitors do you think I should use?
I hope you'll take the time to post some pictures along the way. Your build could serve as inspiration for others.
2x 10000 uF to 2x 22000 uF would be about right.
Tom
2x 10000 uF to 2x 22000 uF would be about right.
Tom
Oh, crap, I can't fit that much in, or I need some radically new idea for the placement. Also my voltage would be almost ±34V. Maybe I should go for another project for this chassis. Thanks for the help though, I appreciate it.
I suppose you could squeeze it down to 2x 4700 uF, but I doubt that saves you much space. I also think that's a bit skimpy for a stereo amp.
What chassis do you have in mind here?
Tom
What chassis do you have in mind here?
Tom