[WilCox]

Quote:

Originally Posted by Daveze /img/forum/go_quote.gif Wilcox, very tidy build. Pardon my ignorance but how is the Super-E configuration achieved?

It's just two Black Gate non-polars in parallel, making sure that one is reversed relative to the other so that inductance artifacts are cancelled. Here is a paper describing the Super E-Cap:

black gate super e-cap

I'm not sure how much snake-oil we have here, but in any case, you get twice the capacitance which lowers the cut-off frequency.

 

[BoilermakerFan]

Quote:

Originally Posted by WilCox /img/forum/go_quote.gif It's just two Black Gate non-polars in parallel, making sure that one is reversed relative to the other so that inductance artifacts are cancelled. Here is a paper describing the Super E-Cap: black gate super e-cap I'm not sure how much snake-oil we have here, but in any case, you get twice the capacitance which lowers the cut-off frequency.

Which DB combo did you use, the 2SC3422/2SA1359 or the 2SC3421/2SA1358?

 

[tomb]

Quote:

Originally Posted by BoilermakerFan /img/forum/go_quote.gif Well, I was going to try to get this posted up over the weekend, but ah well, all you married with family guys know how that goes!

Yes to that - and it seems not to get any easier as they get older, either. It's a blessed thing to bear, though - although not so conducive to building prototypes lately.

Quote:

So I have my K601s here and no decent amp to play them on but I have (3) MAX boards, several sets of caps and TEN sets of potential diamond buffers. Brian D. at TPA confirmed that I can use one Darwin to switch my source and sole pair of headphones between the 3 different MAXes. So that will allow me to do side by side listening comparisons on the fly. Now for the detailed plan: I am going to temporarily mount all three MAXes on the top of one of my NABU chassis. I am going to layout and drill the top of the NABU so I can access the machine screws that will be holding my heatsinks in place from the bottom of the NABU without having to dismount the boards. I will be using the transistor sockets at all positions on the MAX boards so reverse pin out DBs aren't a problem. I'll use unthreaded spacers for the screws and heatsinks so they will be very stable and solidly mounted. All DB transistors are being mounted in sets to 2-1/2" tall heatsinks so I can crank up the bias on a few key chips to really get them into their sweet spots based on the curves. Each set of DBs will be moved between the three different boards and listened to a few different bias levels to get a feel for the sound of each cap/DB/bias level combo. Yes, 40 heatsinks and forty transistors, this isn't a cheap trial and I only have 12 heatsinks at the moment (but that's enough to get started!)... Volume will initially be controlled by ALPS pots since I have three of them, but after I boil down my favorite combos, I will take the time to wire in the OptiVol for a final side by side comparison. (If I can easily pull the signal from all three boards through one volume attenuator, then it will be the OptiVol, but I haven't checked into that closely yet.) MAX Board ONE will have Cerafine caps with VitQ.18 bypass caps and JFETs for the matched pairs so I can use the MOSFET DBs with it in it's final intended build. MAX Board TWO will have Muse ES bypassed with .18 Vit Qs. MAX Board THREE is up in the air, but I'm leaning towards the Panasonic FC caps since several builders I really respect use them as their standard cap when they don't use BGs. These will also use .18 Vit Q bypass caps. I also have Silmic II and Muse KZ caps on hand, but I don't intend to swap those in unless I just really like the sound of the Cerafine caps over everything else with every DB combo, then the FCs will be swapped out for Silmics and the ESs out for the KZs. BUT I don't see that happening anytime soon.

If it were me, I'd use the KZ's or the Silmic's - assuming you're already going to try the ES's and Black Gates and you have no height issues. FC's are nice caps, and decent sounding for quality "power" caps, but not up to the standards of a true boutique. I've remarked about this before, but FC's benefit from a guitar-amp-recapping reputation at DIYAudio and other places. That's not the same thing as a true signal cap application such as at the MAX's CA7.

I recently built the SSMH with FC's all around, too, and while the sound of that little amp is beyond impressive at the cost, it's easy to tell that the FC's are imparting their own signature to the sound. The difference from something like a Black Gate or Cerafine is quite obvious.

The Millett Hybrid tradition has been to use the film bypasses and this mitigates the effect of even a power-quality electrolytic in a signal path. However, it stands to reason that you'd be better off starting with the KZ's or Silmics, IMHO.

Quote:

So the bypass caps will stay .18 Vit Qs across the board for now since they are low cost and generally the default recommendation. I have .22 K42s, Wimas, and Vit Qs too in case I run into a disagreeable combo. Oh, all tubes will be FK6 initially since I have 3 pairs of RCA tubes that are all similar. Once things are boiled down or I see a trend of very bright highs and no bass across all boards with a DB combo, then I'll try a set of AE6A tubes.

Be sure to keep the Wima's at all positions other than the output bypasses. Try the K42's if you have a mind - I noted lately that they are a still a worthy option and can really enhance the bass slam, while having extended highs. They passed out of favor for awhile after Dsavitsk included them in his excellent Notes on Coupling Capacitors web page - making me wonder where we had gone wrong!? I noticed that he's revised his review upward, though, and recent listening on my part confirms this re-assessment.

Quote:

My sources will be my Pro-Ject 1.2 TT with Grado Green cartridge and an assortment of vinyl from monophonic Beach Boys to the Foo Fighters on 180gram vinyl! I'll be building a very simple OPA627 phono stage that has received great reviews on diyAudio for the initial phono stage, but I have parts and plans to build 3 other phono stages in the future. The other source will be my Sony A816 DAP straight from it's headphone jack via mini-to-RCA cable because I have my music as 320kps or lossless and I listen to the Sony every week for hours on end, so I know how the files sound on my Denon C700s and Shure E2Gs. One good source, one ok source. My 47 Labs Flatfish clone and DAC won't be finished for some time so I won't have a great source for a while, but the Grado Green cartridge might upgraded to a Gold sooner than later, as well as the addition of a set of Ultrasone HFI-780s since my B-Day is approaching in October.

A DAC - even an Alien or the soon-to-come BantamDAC - with FLAC files will be best in the long run, IMHO. One thing about the MAX was that I was able to tell the shortcomings in mp3 files almost immediately.

Quote:

So, on to the initial 3 DB combos: The 2238s/968s, the MOSFETs, and either the Toshiba 2SA1930/2SC5171 or Sanyo 2SB1274/2SD1913 - leaning towards the Toshibas since they are on the same price level as the 2238s if not a hair more expensive and the curves look to great potential with higher bias levels. The remaining contenders: 2SA1011/2SC2344 2SA1358/2SC3421 2SA1359/2SC3422 2SB649A/2SD669A 2SB507/2SD313 2SA699/2SC1226 - Note, these are discontinued and being salvaged off an old Garrard integrated receiver, but they fall into the 5W range like the others and the curves are decent. Why not?

Again, just MHO, but the greatest range in sound and effect would be to build your first three with the following pairs:

2SC2238/2SA968
2SC3422/2SA1359
2SC3421/2SA1358

At least with those, you will get a fairly wide spread of different sound signatures and at the same time be assured that the quality of listening is up to a Millett MAX standard. Roll in the 2SC2344/2SA1011's if you want to save the 2SC2238/2SA968's for something special - the differences are very subtle, at best, between those two pairs.

Also, don't hesitate to roll in a pair of 12AE6's right away. You might want to settle on your "best" pair of tubes, first. They have as much or more of an effect than anything else in the amp. The only reason I've implied otherwise in the past is that rolling the tubes in and out is "easy" and "assumed."

Look for 12FK6's that have punch or 12AE6's that have extended highs - those are your primary differences. You may find that some 12FK6's are very extended in the highs, but somewhat "soft" in dynamics - much like the difference between a BUF634 output buffer vs. a discrete output buffer. 12AE6's on the other hand, will have dynamics that bounce the headphones off of your head, but it may take some searching to find a pair that have highs that are as extended as a 12FK6 - but it is possible - they are out there!

Quote:

If any of you guys want to weigh in on which of the following four sets should be the first to go in 2SA1930/2SC5171, 2SB1274/2SD1913, 2SB649A/2SD669A, or 2SB507/2SD313; post up your opinions! If side by side comparisons have been done to death between the MOSFETs and 2238/968s, then maybe I'll start with two previously untested BJT combos. I'll let you guys have some say in what starts off the competition, I mean comparisons... Phwew! Long post. Now I know how tomb must feel sometimes! - BMF

Good luck!

 

[WilCox]

Quote:

Originally Posted by BoilermakerFan /img/forum/go_quote.gif Which DB combo did you use, the 2SC3422/2SA1359 or the 2SC3421/2SA1358?

This build uses the 2SC3422/2SA1359 combo which seems to work out well. My first build was with the 2SC3421/2SA1358 since I was looking for more bass slam out of my ATH-W5000 and W1000. However, with the 2SC3422/2SA1359 and Black Gate Super E-Cap, I seem to get the added treble clarity while retaining the bass slam.

My next build, likely the Mini, will be with 2SC2238/2SA968.

 

[Daveze]

Wilcox, thanks for that info. I was at work and had seen that website but didn't seem to be able to wade through the marketing...guff. Your distilled summary is exactly what I was after. I'm wondering just how unique their manufacturing is though, I'd be quite surprised if an automated cap-making-machine introduced much variation...must investigate more.

 

[BoilermakerFan]

Quote:

Originally Posted by tomb /img/forum/go_quote.gif If it were me, I'd use the KZ's or the Silmic's - assuming you're already going to try the ES's and Black Gates and you have no height issues. FC's are nice caps, and decent sounding for quality "power" caps, but not up to the standards of a true boutique. I've remarked about this before, but FC's benefit from a guitar-amp-recapping reputation at DIYAudio and other places. That's not the same thing as a true signal cap application such as at the MAX's CA7. I recently built the SSMH with FC's all around, too, and while the sound of that little amp is beyond impressive at the cost, it's easy to tell that the FC's are imparting their own signature to the sound. The difference from something like a Black Gate or Cerafine is quite obvious. The Millett Hybrid tradition has been to use the film bypasses and this mitigates the effect of even a power-quality electrolytic in a signal path. However, it stands to reason that you'd be better off starting with the KZ's or Silmics, IMHO.

OK, thanks Tom, I have the room, hence the 2-1/2" heatsinks, so I'll use the KZs since their described sound signature is different than the ESs and Cerafines. Aren't the Silmic IIs close to the sound signature of the Cerafines, just not quite as nice? Edit: Now I think I need to order Cerafines for my gainclone too!

Quote:

Originally Posted by tomb /img/forum/go_quote.gif Be sure to keep the Wima's at all positions other than the output bypasses. Try the K42's if you have a mind - I noted lately that they are a still a worthy option and can really enhance the bass slam, while having extended highs. They passed out of favor for awhile after Dsavitsk included them in his excellent Notes on Coupling Capacitors web page - making me wonder where we had gone wrong!? I noticed that he's revised his review upward, though, and recent listening on my part confirms this re-assessment.

To clarify, I was planning to keep Wima's at all the other .22uF positions, I would only roll in the others into the bypass positions if I heard a distinct issue with one of the boards with the .18 Vit Qs not getting along with the Cerafines or KZs (we know they play well with the ES caps). I don't expect it to happen, but I have them. Eventually I may roll in the K42s to replace Wimas in the other positions, but that's another set of evaluations and I have other stuff I'd rather build...

Quote:

Originally Posted by tomb /img/forum/go_quote.gif A DAC - even an Alien or the soon-to-come BantamDAC - with FLAC files will be best in the long run, IMHO. One thing about the MAX was that I was able to tell the shortcomings in mp3 files almost immediately.

Oh, I agree, but I want to see if any of the combos is more forgiving of lesser sources while still sounding good since some peeps might want to build a MAX to get into headphones before they start working on upgrading the rest of their sources and/or cans. I'm waiting on a BantamDAC.

I have a CD transport and I'll be building a NOS DAC along the lines of Dsavitsk's design with the ability to bring in USB FLAC files or the SPDIF signal from my transport. I just don't have the time to get it finished before I build the MAX, Darwin, phono stage, and the gainclone amp for the console.

Quote:

Originally Posted by tomb /img/forum/go_quote.gif Again, just MHO, but the greatest range in sound and effect would be to build your first three with the following pairs: 2SC2238/2SA968 2SC3422/2SA1359 2SC3421/2SA1358 At least with those, you will get a fairly wide spread of different sound signatures and at the same time be assured that the quality of listening is up to a Millett MAX standard. Roll in the 2SC2344/2SA1011's if you want to save the 2SC2238/2SA968's for something special - the differences are very subtle, at best, between those two pairs.

All of those will be tested, I just wanted to start with the 2238/968 and either the MOSFETs, or two of the untested BJT DBs since nobody else has built a MAX with them. For my evaluation to have any value, I have to listen to all of them to be able to say what works best and what isn't worth the trouble. I'm now leaning towards the two untested 649A/669A and 1274/1913 combos for the initial testing, but the heatsinks and sockets for the rest of the boards will be here before I get the boards fully populated, so it's not a huge deal. I will probably use hex head cap screws to mount the sinks so I can swap out the DBs even faster and safer.

Quote:

Originally Posted by tomb /img/forum/go_quote.gif Also, don't hesitate to roll in a pair of 12AE6's right away. You might want to settle on your "best" pair of tubes, first. They have as much or more of an effect than anything else in the amp. The only reason I've implied otherwise in the past is that rolling the tubes in and out is "easy" and "assumed." Look for 12FK6's that have punch or 12AE6's that have extended highs - those are your primary differences. You may find that some 12FK6's are very extended in the highs, but somewhat "soft" in dynamics - much like the difference between a BUF634 output buffer vs. a discrete output buffer. 12AE6's on the other hand, will have dynamics that bounce the headphones off of your head, but it may take some searching to find a pair that have highs that are as extended as a 12FK6 - but it is possible - they are out there!

I'll roll in the 12AE6s too, I was just going to start with 12FK6s since I will need to use the FK6s with my gainclone amps for the console in order to keep the gain down a bit to work better with the gainclone's gain of 32. I planned to use AE6s or FM6s for my stand alone desktop MAX. All of the DB combos will be evaluated on all of the boards with all of the tubes. By my calculations with 3 boards, 10 sets of DBs, and at least 11 pairs of tubes I have 330 variations to test and that's before I start to cherry pick my favorite individual tubes on one set of

. I plan on taking very good notes so I can narrow down the combos for further listening with HFI-780s and the other tweaks such as JFET vs. BJT for the small signal caps and K42s instead of Wimas. A lot of listening, but some of it will be done while I'm building other stuff like the phono stages, DACs, and refurbishing my older Magnavox 6V6GT tubed mono console.

Quote:

Originally Posted by tomb /img/forum/go_quote.gif Good luck!

Thanks, and thank you for the feedback.

 

[manaox2]

Quote:

Originally Posted by digger945 /img/forum/go_quote.gif If Tom wont take a beer, this tarheel will!! Good to see you got it goin' Does it sound good or what?

It sounds greater by the second. Changes during warm-up and the tubes settling aside, It'll make me shed a tear any day now from just the shuffle it seems! If I get that ballsie for my Opus I have populated already, I will have a source that at least will compare to the iMod I sold to afford it. Glad that I finally have a real headphone amplifier again for my favorite phones.

So surprise, I also originally bought the parts for three of these beasts also with a NABU case...

As soon as I put the funds aside and get the attenuator for the real project, we will see what I learned from the experience. Pabbi1 should see a brother amp hopefully coming in the new year (if that darn updated Joshua tree attenuator gets released at least!).

BTW, that crazy blackgate super-E with Vit-Q bypass up there really gets the geek in me drooling! Wondering if I can afford to get eight of those sweet blackgates now.

 

[digger945]

manaox,
I wanna know if your eyes water when you get the Opus feedin' the Max.

Yea that's the only reason I didn't go home (Asheville) this year, cuz I bought way too many diy stuffs and goodies

 

[pabbi1]

Quote:

Originally Posted by manaox2 /img/forum/go_quote.gif So surprise, I also originally bought the parts for three of these beasts also with a NABU case... As soon as I put the funds aside and get the attenuator for the real project, we will see what I learned from the experience. Pabbi1 should see a brother amp hopefully coming in the new year (if that darn updated Joshua tree attenuator gets released at least!).

Oh, and Sam, I happen to have an unused Ballsie board..

3 sets of boards? Hmmmm... but, this does bring up a question: Has anyone else built a balanced Maxx? Not necessarily with a DAC, but, certainly, I can't be the only one so far... and the most FUN amp I have ever had.

I have already built a WAY badder brother for Maxxie... they are far different animals, 350v vs 28v. No end to the madness...

 

[BoilermakerFan]

Quote:

Originally Posted by manaox2 /img/forum/go_quote.gif <snip> (if that darn updated Joshua tree attenuator gets released at least!).<snip> BTW, that crazy blackgate super-E with Vit-Q bypass up there really gets the geek in me drooling! Wondering if I can afford to get eight of those sweet blackgates now.

I'm waiting for the JT too for my destop MAX in the NABU. I'll definitely be incorporating the VFD!

There may be a lower cost way to duplicate the concept with other non-polarized caps... Maybe it would work, but worst case you just have more coupling capacity as WilCox said.

Quote:

Originally Posted by pabbi1 /img/forum/go_quote.gif <snip> 3 sets of boards? Hmmmm... but, this does bring up a question: Has anyone else built a balanced Maxx? Not necessarily with a DAC, but, certainly, I can't be the only one so far... and the most FUN amp I have ever had. <snip>

The thought has occurred to me. Greg at SKA even said that a 100K ALPS Blue Velvet could be used to operate two OptiVols for a lower cost way to implement stereo volume control since he could tweak the output resistors on the OptiVols to better match them than trying to match 8 sets of LDRs. They ALPS Blue Velvets two channels track close enough to work. So maybe I have done a little more than think about it. But my TT is SE and I haven't done the homework to find a SE to Balanced phono stage.

I just figured I would build two stereo Octal Aikido boards into a balanced headphone amp if I was going to go balanced, but there is a significant price difference between that and a balanced MAX.

Three MAXes would make a really nice preamp feeding (6) LM3875 based gainclones (one bridged and paralleled for much higher output power to the subs) for home theater applications too! A motorized encoder driven JT would handle the volume attenuation.

- BMF

 

[rockcod]

I am stumbled. I initially installed LM338 as VREG, rectifiers are HFA08TB60 HEXFREDs. Voltage measured across V+ and GND is 17.8V and keeps dropping, MOSFET bias is overloaded (beyond the mV range) and QB9s get hot quickly (even smoke after 10 seconds) and so does VREG. Then I changed VREG to LM317, voltage measured across V+ and GND is now 13.4V and keeps dropping quickly, VREG gets hot (it's dropping too much voltage), but QB9s remain normal. Adjusting RR3 doesn't do anything.

I've replaced DR1 & DR2, CR2, CR4, & CR5, checked RR1 & RR2. I don't think it is the problem with the HEXFREDs as they are used in my other builds, besides, voltage across CR1 is normal at 27.8V.

Any insight?

Edit: Ok, found the problem -- I mistakenly used 10K instead of 1K in RB2s & RB3s. I guess I shouldn't be listening to music while soldering -- it is too distracting

 

[RubberDuk]

Just dropping by to mention I bought a MAX built by a fellow head-fier and I'm in love. That is all. Carry on.

 

[amphead]

Happy to hear that you are enjoying your Max RubberDuk! They were well designed and should last at least 50 years or more! So you should have some listening pleasure for quite awhile. Once they are dialed in they just keep on ticking like a Timex watch, but sound like a Rolex Submariner! Edit: good that you found that Rockcod! Should be an easy fix from there.

 

[tomb]

Quote:

Originally Posted by rockcod /img/forum/go_quote.gif I am stumbled. I initially installed LM338 as VREG, rectifiers are HFA08TB60 HEXFREDs. Voltage measured across V+ and GND is 17.8V and keeps dropping, MOSFET bias is overloaded (beyond the mV range) and QB9s get hot quickly (even smoke after 10 seconds) and so does VREG. Then I changed VREG to LM317, voltage measured across V+ and GND is now 13.4V and keeps dropping quickly, VREG gets hot (it's dropping too much voltage), but QB9s remain normal. Adjusting RR3 doesn't do anything. I've replaced DR1 & DR2, CR2, CR4, & CR5, checked RR1 & RR2. I don't think it is the problem with the HEXFREDs as they are used in my other builds, besides, voltage across CR1 is normal at 27.8V. Any insight? Edit: Ok, found the problem -- I mistakenly used 10K instead of 1K in RB2s & RB3s. I guess I shouldn't be listening to music while soldering -- it is too distracting

Glad to hear you found your problem so quickly!

 

[ruZZ.il]

Quote:

Originally Posted by rockcod /img/forum/go_quote.gif ...... Edit: Ok, found the problem -- ....

its the fumes man! I bet cause you solder so fast that you get a steady stream of them and by then 1, 10, same thing

good stuff on finding it so quickly though! you recover well from fumes