[tomb]

Quote:

Well it works! Both the MOSFET max and the grub DAC seem to work just fine. I used elna silmic II 1000uf caps in the signal path, used the wima for C9, and the vitaminQ in the other bypass position. Initial impression was that the low end was not as strong as my minimax with black gates, nor was it as musical, it had an etchy-ness to the top end, and tended to be somewhat simblant. It has opened up and mellowed out a bi with some use, etchy-ness is mostly gone, and the low end is improving, but I still think it is not as fluid or musical as the BJT I built a while ago. I am listening through a set o M-50's. I am hoping it will continue to improve as it burns in, but I also think I might start looking at different bypass caps as well to smooth it out a bit more. Excellent kit and web support though. Thanks.  

You've got to use at least 22R on the output resistors (RB14-L/R) or the MOSFETs will oscillate and contribute to that "etched" sound.  FYI, it's really horrible if you had used KZ's, but try the 22R's - they settle everything out just perfectly.  Unlike the MiniMAX, additional resistance on the output has little effect on detail, but the 22R is enough to tame the MOSFETs for excellent detailed, but smooth highs.  I've built several and the additional resistance on the output is key.  All of mine have Wima's at CA9, VitQ's at CA8, and Elna's at CA2 and CA7.  The Elna's should not require a long break-in period - maybe 24 hrs, that's it. 
 
Note that the 1.15K for R8 and 11.5K for R9 is important, too.  I think I saw in the other thread you used 1.13K for R8.  R8 actually governs the CCS - R9 only needs to be at least 10X more resistance.  However, 1K results in a CCS current of 0.65.  1.13K results in 0.575, but 1.15K results in 0.565.  jdkjake may be correct in that 2% difference is not significant, but if you went with the older 1K/10K, it's a big difference.

 

[jdkJake]

Quote:

 
Note that the 1.15K for R8 and 11.5K for R9 is important, too.  I think I saw in the other thread you used 1.13K for R8.  R8 actually governs the CCS - R9 only needs to be at least 10X more resistance.  However, 1K results in a CCS current of 0.65.  1.13K results in 0.575, but 1.15K results in 0.565.  jdkjake may be correct in that 2% difference is not significant, but if you went with the older 1K/10K, it's a big difference.

TomB,
 
In your opinion, is it more critical to get closer to the design point of 0.565 than say a 1% resistor? Does the design spec for the CCS stay consistent over a wide set of tubes (say a 12FK6 vice 12AE6)?
 
Thanks.

 

[tomb]

Quote:

Quote:

 
Note that the 1.15K for R8 and 11.5K for R9 is important, too.  I think I saw in the other thread you used 1.13K for R8.  R8 actually governs the CCS - R9 only needs to be at least 10X more resistance.  However, 1K results in a CCS current of 0.65.  1.13K results in 0.575, but 1.15K results in 0.565.  jdkjake may be correct in that 2% difference is not significant, but if you went with the older 1K/10K, it's a big difference.

TomB,
 
In your opinion, is it more critical to get closer to the design point of 0.565 than say a 1% resistor? Does the design spec for the CCS stay consistent over a wide set of tubes (say a 12FK6 vice 12AE6)?
 
Thanks.

No ... because 0.560 is actually the original design point picked out by Pete Millett.  I've still picked a point (0.565) where the resistors are available at economical prices.  The real point is that there is a big difference in performance from the original 1K/10K combination that resulted in a 0.650 operating point.
 
No - the CCS operating point is only a compromise among the tubes.  Each tube (12AE6, 12FM6, and 12FK6) will have its own optimum.  I don't know what that is, though - but I did test for the differences and the lower number works better with all three tubes.  There is a great possibility, though, that you might find an even better operating point with a given tube.  Even Pete said, "There's nothing sacred about using 0.56mA as the plate current - looking at the plate curves, I thought it looked like a good point to operate the 12AE6 tube."
 
The original Millett Hybrid used a constant current diode to set the CCS.  However, Colin Toole's design of the MAX/MiniMAX/MOSFET-MAX uses a cascade transistor arrangement that allows us to change the resistors and alter the CCS currents.  Colin had always hoped that people would do some experimentation with the CCS.  Doing so has not been talked about much, but after I saw how much difference it made with Dsavitsk's new Torpedo (coming soon!), I decided to try it out and things worked out great.

 

[H22]

You've got to use at least 22R on the output resistors (RB14-L/R) or the MOSFETs will oscillate and contribute to that "etched" sound.  FYI, it's really horrible if you had used KZ's, but try the 22R's - they settle everything out just perfectly.  Unlike the MiniMAX, additional resistance on the output has little effect on detail, but the 22R is enough to tame the MOSFETs for excellent detailed, but smooth highs.  I've built several and the additional resistance on the output is key.  All of mine have Wima's at CA9, VitQ's at CA8, and Elna's at CA2 and CA7.  The Elna's should not require a long break-in period - maybe 24 hrs, that's it. 
 
Note that the 1.15K for R8 and 11.5K for R9 is important, too.  I think I saw in the other thread you used 1.13K for R8.  R8 actually governs the CCS - R9 only needs to be at least 10X more resistance.  However, 1K results in a CCS current of 0.65.  1.13K results in 0.575, but 1.15K results in 0.565.  jdkjake may be correct in that 2% difference is not significant, but if you went with the older 1K/10K, it's a big difference.

 


I'm pretty sure I used the 22R's for R14, I read that R14 was needed and not optional in the MOSFET version, so I used what was on the BOM. Same with the CCS resistors, I ordered extra 1k and 10k, just to have them, but I also ordered and used the 1.13k & 11.3k from the BOM after I read yor post on he minimax thread. I tried to order the 1.15/11.5 resistors, but mouser does not have them in the line that the others are, my thought was that the diff between 1.13 and 1.15 was just not a big enough difference to worry about.

For an update, it has improved greatly with about another 20 hours of burn in. It also seems to helpmif Iet it " warm up " for 10 min or so before I start listening.
This amp is really starting to make me want to buy a better set of cans though, so I need to start looking at what would be a good can for me, initially the HD650's would seem like a good choice, but I'm a bit of a bass head, so I'm thinking another can might be better.

Again, amp is sounding great, wish I had the mini here to compare them side by side, but thy will have to wait.

 

[tomb]

Quote:

 
I'm pretty sure I used the 22R's for R14, I read that R14 was needed and not optional in the MOSFET version, so I used what was on the BOM. Same with the CCS resistors, I ordered extra 1k and 10k, just to have them, but I also ordered and used the 1.13k & 11.3k from the BOM after I read yor post on he minimax thread. I tried to order the 1.15/11.5 resistors, but mouser does not have them in the line that the others are, my thought was that the diff between 1.13 and 1.15 was just not a big enough difference to worry about.For an update, it has improved greatly with about another 20 hours of burn in. It also seems to helpmif Iet it " warm up " for 10 min or so before I start listening. This amp is really starting to make me want to buy a better set of cans though, so I need to start looking at what would be a good can for me, initially the HD650's would seem like a good choice, but I'm a bit of a bass head, so I'm thinking another can might be better. Again, amp is sounding great, wish I had the mini here to compare them side by side, but thy will have to wait.

OK - that all makes sense.  Yes, the Millett-type tubes take awhile to warm up.  I think tubes perform better when they're hot, but at the low voltage of the space-charge tubes, it takes longer.  I think best listening is definitely available within 15 minutes, perhaps sooner.
 
As for a different pair of headphones, I would recommend the K701's or Grado HF-2's for the MOSFET-MAX.  Just MHO, of course.

 

 

[Tacoboy]

Where currently is the best place(s) to order parts for to build the Millett Hybrid Mosfet Max?

 

 

 

[jdkJake]

The BOM has part numbers for Mouser. I would start there.

 

[Tacoboy]

Quote:

The BOM has part numbers for Mouser. I would start there.

BOM?
Mouser?
 

 

[MrDavis]

BOM is Bill of Material.  A.K.A a parts list.  http://www.diyforums.org/MOSFET-MAX/MOSFET-MAXbom.php
 
Mouser is a parts suppler.  http://www.mouser.com/
 
The Mouser parts list on the BOM is more up to date. 
 
Quote:

BOM?
Mouser?
 

 

 

[Beefy]

Quote:

BOM?
Mouser?
 

Uh-oh......

 

[W-Moran3]

IT WORKS!!!
I’m so pumped!!!!
Hooked it up and queued the inaugural song last night.
 . . .
Everything In It’s Right Place by Radiohead.
1)      I’m a huge Radiohead fan
2)      It’s the perfect title (if it works)
3)      I’m a huge Radiohead fan
4)      IT WORKED!
 
Although I’m posting this after that first night, I’m still excited that it has been up and running without any problems.  My fabrication skills way outstrip my electrical skills and knowledge.  My previous EE project was one of those kits for a blinking Christmas tree ornament many years ago. So there’s hope to those that are on the fence with respect to starting a build.  I would like to give back to the community. Enclosed are some hits and misses during the project.  Thanks to TomB and jdkJake for their help!
 

 

[jdkJake]

Congrats on the build! Looks great!!
 
The lessons-learned is a neat idea. What a great presentation of that material.
 
I love, Love, LOVE the AutoCAD work! Very, very cool.
 
Enjoy your amp. I know you will. 

 

[tomb]

Holy cow!!  Saying, "a MOSFET-MAX lives

" just doesn't seem to do that justice.  Very nice AutoCAD work, by the way ... very!!

 

[HiGHFLYiN9]

Quote:

 
Although I’m posting this after that first night, I’m still excited that it has been up and running without any problems.  My fabrication skills way outstrip my electrical skills and knowledge.  My previous EE project was one of those kits for a blinking Christmas tree ornament many years ago. So there’s hope to those that are on the fence with respect to starting a build.  I would like to give back to the community. Enclosed are some hits and misses during the project.  Thanks to TomB and jdkJake for their help!

Nice work and incredible documentation. You have enough there for a magazine article The preliminary design would actually make a pretty cool looking amp too:
 

 
 

 

[BobSaysHi]