The diyAudio First Watt M2x

When member 6L6 powered up his M2x the first time, he and I watched the ammeter on the 115VAC mains (USA). Starting from zero milliamps, it slooooooowly crept up over the span of about 20 seconds or so. Would anyone care to guess why? Hint: the answer is found on the amplifier board, not the power supply board.

677253d1524762146-diyaudio-watt-m2x-m2x_v1b_mainamp_schematic_image-png


_

I’ll play to learn.

Does it have to do with optocoupler 4N35 and the large 3300uf capacitor?

I know optocouplers can provide electrical isolation between an input source and an output load using just light. Current from the source signal passes through the input LED which emits an infra-red light whose intensity is proportional to the electrical signal. The emitted light falls upon the base of the photo-transistor, causing it to switch on and conduct in a similar way to a normal bipolar transistor.

Perhaps that takes time and hence the slow increase in current for the output stage bias.

Best,
Anand.
 
That was 1,176 Jfets per channel, and recall that I kept the leads on
them nice and long for easy re-use by future generations.

And matched!

:p



Ok.....only 1176......think I found it here:
http://www.firstwatt.com/pdf/art_beast.pdf


One big advantage........you don't need any heatsink for this Class A amp. People which sells matched pairs on ebay for 50 USD pr. pair seems to have got boxes with 1000 or 2000 in each box.
I wonder what causes the demand for them went down since they stopped production......or maybe there was never a huge demand for these jfets.
 
Member
Joined 2011
Paid Member
Even with the CL60 inrush current limiter(s) in the standard First Watt power supply, the DC voltage rails reach their final value (plus and minus 25 volts) quickly. However, over on the M2x amplifier PCB: it takes a very long time for resistors R6 and R7 to charge up the enormous capacitor C3.

Because the output transistors Q1 and Q2 have very little VGS, they are both OFF. Thus no current flows into the LED side of the optoisolator Q5, and no current flows in the output transistor side of Q5 either. R6 and R7 are charging C3 while the optoisolator is completely shut off. Eventually the voltage across C3 grows large enough to turn on both output transistors, and then the negative feedback control system wakes up and sets the bias current. Just as Nelson Pass drew it up.

_
 

Attachments

  • psu_fw.png
    psu_fw.png
    226.5 KB · Views: 3,167
Last edited:
Member
Joined 2011
Paid Member
PCBs for a skunkworks project (see 6L6's remarks in post #2 above) had a very busy weekend. According to DHL, these boards flew from China to Ohio, changed planes, flew to Utah, changed planes again, and then flew to San Francisco. They must be exhausted by now. When they arrive at my lab in Silicon Valley this afternoon, I'll give them a good long rest before we start working.

_
 

Attachments

  • DHL.png
    DHL.png
    50.3 KB · Views: 2,193
Mark - and that's when everything goes according to plan. Last year I ordered a chassis through the DIY Audio Store, and FedEx lost it somewhere in the New Jersey area - it took over 2 months for them to find it and finally deliver.

For the record, everyone at DIY Store fell over themselves assisting with tracking / following up on the order - to the point of preparing to duplicate the order and make the loss claim against the shipper.

As frustrating as this was for all parties, I couldn't have asked for better service from folks here.
 
Member
Joined 2011
Paid Member
Make a comparison table in Excel (or on paper), of the nine or ten most important characteristics of these JFETs: Idss, gm, maxVds, etc. Post it here. Members will suggest which line items you probably want to focus upon, when deciding which JFET to use on your MountainView input stage PCB.
 
Member
Joined 2011
Paid Member
I'm pleased to announce that input stage daughter card #5, named "Austin" for Motorola's fab site, has arrived. I've got it on the lab bench and it's testing out very well. Details in 45 minutes.

This means that, finally, I'm able to sell full sets of M2x PCBoards. As of today I have thirty (30) "sets" of boards, where each set is what you need to build a 2-channel M2x amplifier. A "set" consists of
  • 1 Channel-A amplifier main board ("mother board")
  • 1 Channel-B amplifier main board (mirror image of Channel A main board)
  • 2 Ishikawa input stage daughter boards (Toshiba JFET)
  • 2 Tucson daughter boards (Burr Brown Opamp IC)
  • 2 Mountain_View daughter boards (Fairchild JFET Class A single ended)
  • 2 Norwood daughter boards (surface mount opamp + video buffer)
  • 2 Austin daughter boards (discrete BJT "Diamond Buffer")

Price per set of 12 PCBs, including shipping, will be $40/set for USA and $50/set for the rest of the world.

If you are interested in buying a set of 12 PCBs, please send me a Private Message here on diyAudio, expressing interest. I don't need anything else at the moment, just a count of the number of sets desired. When demand flattens out, I'll send payment info and collect your ship-to address, et cetera. Today, just a "M2x for me please!" message is all that's needed.

SEND ME A P.M. IF YOU WANT TO PURCHASE M2X AMPLIFIER BOARDS Sold out
 
Last edited by a moderator:
Member
Joined 2011
Paid Member
Here is the fifth M2x input stage, named "Austin" for Motorola's fab.

Austin turns out to be kind of a misnomer because, at the eleventh hour, I yanked out the Motorola input transistors and replaced them with (gasp!) Fairchild devices. But we've already got a Fairchild input stage board, and besides, Austin is a pretty cool city. So the name stays.

The circuit is usually called a "Diamond Buffer" since Q1, Q2, Q5, and Q6 roughly form a diamond shape on the schematic. Diamond buffers are frequently used in audio circuits and there are quite a few threads here on diyAudio about them. Bob Cordell's textbook discusses Diamond circuits in section 10.4, for example.

Transistors Q1 and Q2 operate as emitter followers, with Q3 and Q4 providing current source loads. Q5 and Q6 are another pair of emitter followers, which drive the output.

I've used relatively high precision "C-grade" (0.5%) IC voltage references, to get precise control of the emitter currents of Q1 and Q2. Don't worry, they cost less than a dollar each. Precise emitter currents establish precise base-overdrive voltages thanks to 1% resistors R4 and R5. The precise base-overdrive, plus 1% resistors R9 and R10, establishes a solid and fixed bias current in the output stage.

I'll see if I can upload the Austin schematic, and its Bill Of Materials, into the .zip archive attached to post #1. However, don't expect immediate miracles!

BTW-- If you've got access to some New Old Stock of Motorola MPS8099 and MPS8599 transistors, they'd work great in this board. But right now they're only available in surface mount packages, not TO-92s.

~
 

Attachments

  • M2X_Austin_V1_2018_05_02.pdf
    26.2 KB · Views: 391
Member
Joined 2011
Paid Member
Hi Graeme,

I bought a few dozen different Zetex part numbers a couple years ago and measured them. Now I refer to my table of measurements when choosing devices. You guessed it, the ZTX795 is in my table but the ZTX796 is not. I didn't even consider it. Looks like it would work fine though.

Or as Wayne Colburn of Pass Labs says, when asked about a parts choice, "We had lots of em in the stock room."

MJ


SEND ME A P.M. IF YOU WANT TO PURCHASE M2X AMPLIFIER BOARDS Sold out
 
Last edited by a moderator:
Thank you Mark. I made my choice by looking through the data in the Zetex catalog. I was searching for an upscale replacement pair for the a42/a92 and wanted to stick with the to-92 package.

Great work on M2X by the way. I’m subscribed here. Carry on.

Graeme
 
Member
Joined 2011
Paid Member
For those who hate, hate, hate Adobe Acrobat, here's the Austin schematic as an embedded image file. Click once to get rid of the obvious foreshortening distortion, and click again on the white "X" at bottom left, to see it full size.

attachment.php


~
 

Attachments

  • image_file_not_acrobat.png
    image_file_not_acrobat.png
    84.8 KB · Views: 2,413

6L6

Moderator
Joined 2010
Paid Member
Absolutely fantastic!!!!!


I'm really looking forward to trying the diamond buffer. All the rest have been superb so far, they all work perfectly and honestly all sound wonderful. The various input stages are a happy abundance - we have so much to choose from! :D :D :D
 
I once tried a diamond buffer at the front of an F6 and rather liked it! (I was waiting for the FETs to arrive) Nothing exotic, just with BD139s and 140s. But it sounded excellent. I could not do a direct comparison to a FET version, so I cannot tell if there were any subtle differences. When the FETs arrived I put them in and it sounded just as good to me, so I left it that way (and it looked better as it was not jury-rigged above the PCB).