diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Pass Labs (http://www.diyaudio.com/forums/pass-labs/)
-   -   "Le Mutant" Class A (http://www.diyaudio.com/forums/pass-labs/191986-le-mutant-class.html)

wrenchone 3rd July 2011 11:11 PM

"Le Mutant" Class A
 
1 Attachment(s)
Attached is a pic of a new Class A amp prototype inspired by my work with the "Half Nelson" amp described elsewhere in the Pass Labs forum, and also as a result of a little email back-and-forth between me and Papa after last year's Burning Amp. It solves a few of the problems I encountered with the Half Nelson, at the cost of a tiny bit of extra complexity (actually, the parts count is probably very similar between the two - I'd have to open up both schematics and do a head count).The new design also eliminated a lot of nasty coupling caps.Schematic will follow when I get the thing powered up and nail down the bias values. I'm also working on a version of the Half Nelson using the Semisouth R085 jfet, but that one is on hold for a while, as this amp may be easier to spit out. I'm also working on a tube-type design using an R085, but that will be the subject of another thread. Burning Amp should be interesting this year...

wrenchone 5th July 2011 06:13 AM

1 Attachment(s)
Attached is a tentative schematic for the Le Mutant. I have both channels running thermally stable (one at a time) with a bias current of ~1.5A. Even with such a large heat sink, it gets real hot, so there will most likely be a toned down DC fan to augment cooling. I've made no attempt to pre-select or match output/bias FETs, so each channel needs individual tweaking for both the bias current and so that the input differential stage is in balance with no signal. Right now, I'm getting about 18mV of output offset for each channel, not too bad for the first attempt. The resistors to tweak are R4 and R5. A 2.2K resistor was added in series with the collector of Q9 to reduce its power dissipation. It still runs considerably hotter than Q8, unbalancing the mirror somewhat. A subsequent version of this amp will use some nice Fairchild TO-126 devices coupled to a common heat sink. I will also investigate what can be accomplished by selecting/matching Q4-Q7. I may also add a pot to help tweak in the bias.

The 2200uF capacitors on each rail are only the local bypassing. The main filter caps will be much more massive.

I still have some work to do before I unleash this thing on some speakers, but it won't be too long.

wrenchone 29th August 2011 04:54 AM

I took the mutie off the big heat sink. As of tonight, both channels are on an aluminum heat spreader plate with a pair of microprocessor coolers on the back side to pull heat out. It's a much more compact setup. Next is to complete the new amp modules incorporating the changes I mentioned in the last post, as well as matching up some output devices. I should be powering this baby up in the next week or so.

knutn 29th August 2011 01:34 PM

Nice work. Are Q4 and Q7 mounted on the same heat sink as the power transistors?

Nelson Pass 29th August 2011 03:51 PM

I think that's very clever.

:cool:

jacco vermeulen 29th August 2011 04:36 PM

Not sure what i like more, the schematic or the name.

(first time i read it a number of days ago, it made me toss coffee on my keyboard :clown: )

wrenchone 29th August 2011 06:20 PM

Jacco, you're easily amused...

Q4 and Q7 are indeed mounted right next to their respective output devices on the heat sink. Bias was stable when I fired up the first revision of this circuit. In the next revision, I have added a pot to tweak in the bias/balance, and have changed the mirror transistors to TO-126 devices on a common heat sink, in addition to a resistor to help equalize power dissipation. I need to sit down and sort some mosfets before I can finish populating the next revision boards. I'm expecting that the better matching will improve DC offset a bit. The next schematic will follow when I have the boards populated and tested.

This one should be finished in time to play at Burning Amp.

wrenchone 30th August 2011 04:22 AM

Passing-for-Pass designs for this year's Burning Amp will probably include the one posted here, the "SiC Puppy" (once I do the heat spreader/microprocessor cooler routine with it as well), and a revved-up version of the "Half Nelson" with a higher VCC supply and a fan on top to prevent melt-down. I have tube amp designs in the works as well, but that's another thread, another forum. I actually had another inspiration for a Pass-type design last night before I dropped off to sleep, but I'm putting that one on hold so I can get something finished.

wrenchone 30th August 2011 05:38 AM

The stuff I bring to Burning Amp may consist of engineering sculptures lashed up on a piece of plywood for expediency's sake (Papa does it, why can't I?). In which case, god/dog help the inquisitive fingers that attempt to palpitate the Sic Puppy - 300+ volts (with some authority behind it) is no joke...

wrenchone 31st August 2011 05:46 AM

This afternoon, I sorted a bunch of IXYS 200V, 0.17 ohm P-channel fets to serve in the output stage. I'm hoping they will be a better match to the IRFP240 than the usual 9240 types. Once I go through my IRFP240s the same way, I can sort the driver fets. The output fets are strapped gate-to-drain, and labeled with the Vgs for a bias current of 1.5A (bench supply in constant current mode). The power devices are bolted to a fan-cooled microprocessor heat sink during characterization so that thermal drift is held to a minimum, and I can get a stable value for Vgs relatively quickly. The driver transistors will be sorted and labeled for 10ma drain current using my usual fet sorting jig. Once I have the numbers on everyone, I'll start pairing up driver and output fets.


All times are GMT. The time now is 02:31 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2