The diyAudio First Watt M2x

Member
Joined 2011
Paid Member
Make the conservative assumption. Or, if you don't like assumptions, find someone who has intentionally and deliberately magnetized this particular interstage transformer made by Edcor, while measuring the current. Then ask him. I have not tried the experiment myself.

If you feel energetic and want to build a test fixture perhaps this page offers a few good ideas

Power Inductor Checker
 
I will try not to perform destructive tests.
May play Mozart instead of Rammstein. Rammstein may magnetize Edcor?

Picture is 1000 Hz in 4 ohm load (5V / Div). It is 9,1 V RMS (measured with true RMS DMM). So that is more than 2A RMS current. Bias is about 1.3 A so M2X has left class A I assume?
That is only about 20W but amp should be able to deliver 40W in 4 ohm....but not in class A mode? ….the 40W in 4 ohm is the clipping point?
 

Attachments

  • DSC_4014_00001.jpg
    DSC_4014_00001.jpg
    502.4 KB · Views: 444
It is interesting the heatsinks are just above ambient running this test. So power is delivered in load and not much in mosfets. The load gets hot. That is the way to make big class A amps with small heatsinks…...just use the power? ….so maybe Rammstein is back again! …...but I know that is not the idea…..it is more like use less than 1W with efficient speakers and have low distortion…...and big heatsinks.
 
Any interest?

Do folks have any interest in these?
They are dirt cheap to have made, so I could order some extras for folks if they want 'em. The same board works for either end - just put the proper connector on. 14 boards are needed for the full set of daughter boards including Melbourne. I've included the Gerbers for folks who live in places that don't have a flag on the moon yet since shipping there is silly expensive.

They use AVX 9159 board to board connectors.
https://www.mouser.com/ProductDetai...ffgRu4KC1R7KDMLYmRnj1blbnaGtSRlDuVb6hMf9Bxg==
https://www.mouser.com/ProductDetai...ffgRu4KC1R7KDMLYmRnj1MfV4ng0egik2C1mWSrZe5g==

and will take either a standard 5mm Eurostyle screw block, or 16awg wire for connection to the main amp board.
They should also accept the Melbourne board - it is designed so the overhanging capacitor is over the connector rather than the mounting screw.

The connectors are SMB, but look pretty easy to solder. I added an exra .2mm to the footprint to make it easier.

I'm not trying to make anything off these, so keep headache to a minimum I will only ship 15 at a time (gives you 1 spare). PM for cost (the connectors are ~4x the cost of a board)

-Josh
 

Attachments

  • B2B.jpg
    B2B.jpg
    92.6 KB · Views: 463
  • BtoB_Connector.zip
    17.9 KB · Views: 51
The other M2X mono block has same nice behaviour. Scope picture shows a 20 Hz sine in 4 ohm load (5V / Div). Then a couple of pictures of build. After the DiyAudio Store standard CRC PSU (using 33 mF caps) I added LC filters (2.5 mH / 10 A DC filter chokes) and 4-pole 33 mF Jensen caps to remove last bit of 100 Hz noise (I hope). Will see when I have made an interface from power amp to Focusrite so I can measure distortion/noise. I think my amps are very silent. Both toroid and Edcor has magnetic shields. First I have to finish the build completely. But it is very close. Now I know that the basic functionality of amp works and both mono blocks have had some burn-in tíme putting some watts out into a load. No failures so far.
 

Attachments

  • DSC_4018_00001.jpg
    DSC_4018_00001.jpg
    677.6 KB · Views: 611
  • DSC_4017_00001.jpg
    DSC_4017_00001.jpg
    340.4 KB · Views: 681
  • DSC_4016_00001.jpg
    DSC_4016_00001.jpg
    444.3 KB · Views: 681
  • Like
Reactions: 1 user
Here a couple pictures of my almost complete M2X amplifier. Coming from Class AB (updated Hafler DH-500) and Class D (Pioneer SC-35) I now see what all the Class A fuss is about . It really sounds wonderful and I can listen for hours without tiring. Thanks to everyone who has made this DIY project possible, especially Nelson!
 

Attachments

  • CD877F98-F5AC-4AB9-B926-4310B52A58BA.jpeg
    CD877F98-F5AC-4AB9-B926-4310B52A58BA.jpeg
    714.2 KB · Views: 1,344
  • 007B4FB0-0476-4E8F-BB50-95DE8BD138BB.jpeg
    007B4FB0-0476-4E8F-BB50-95DE8BD138BB.jpeg
    557 KB · Views: 1,347
Nice build, SLK23! Very neat wiring. Glad the SLB is working out on the M2X. That chassis looks nice, can you give link to it? Looks like AliExpress CNC type all aluminum.

I have the same Antek 3223 and SLB with ground from RCA to star hub on SLB. That’s the way to do it - no ground loop hum.

Thanks for the feedback! And thanks also for the SLB PSU; it performs beautifully in this amp.

I bought the chassis on eBay from a domestic USA seller who apparently doesn’t stock it regularly. But it’s available from other sellers and on AliExpress for around $250 including shipping.

Class A Chassis on AliExpress

Class A Chassis on eBay

I posted pictures of my M2X on Reddit too. One commenter suggested running the signal input ground to the amp board rather than the PSU star ground. I replied with a link to your post in the SLB thread about reducing hum. He/she also recommended taking the speaker/output ground from the PSU instead of the M2X board. Comments?
 
Member
Joined 2011
Paid Member
It might be a good idea to turn on your M2x and wait 60-90 minutes for it come up to thermal equilibrium. Then measure the case temperature of the two power devices which use the chassis floor as a makeshift heatsink. The floor has no "fins" for improved thermal transfer to the ambient air, so it might not perform as well as the extruded aluminum heatsinks on the left and right sides of the chassis, where the amplifier main boards are mounted.

Each amp transistor on the main heatsink, dissipates (Vsupply x Ibias) watts.

Each power supply transistor on the floor, dissipates (Vdropout x 2 x Ibias) watts since they supply both amp channels.

Maybe the case temperatures of the devices bolted to the floor are as low or lower than the case temperatures of the devices bolted to the main heatsinks. If so, congratulations. Or, maybe the case temperatures of the floor devices are higher. After measuring, you'll know.
 
Maybe the case temperatures of the devices bolted to the floor are as low or lower than the case temperatures of the devices bolted to the main heatsinks. If so, congratulations. Or, maybe the case temperatures of the floor devices are higher. After measuring, you'll know.

Mark, is your comment directed at me or MEPER?

I've 'measured' the temperatures by touching the area around the transistors on the heatsink and chassis bottom. The PSU transistors a quite cool; as far as I can tell the area around them is the same temperature as the whole chassis after the amp has been powered up for an hour. The big heatsinks are warm but it's no problem resting my hand on them for an extended period.
 
Founder of XSA-Labs
Joined 2012
Paid Member
Since your floorpan is probably 3mm thick aluminum - I would imagine it can dissipate 7.5w each BJT no problem (2.5A x 3v).

Regarding the comments about where to ground stuff:
I posted pictures of my M2X on Reddit too. One commenter suggested running the signal input ground to the amp board rather than the PSU star ground. I replied with a link to your post in the SLB thread about reducing hum. He/she also recommended taking the speaker/output ground from the PSU instead of the M2X board. Comments?

Try it and if it is quieter you can use it - but I doubt that it will be. I found that I had basically no ground loop hum with RCA ground connected to the star hub. Having the speaker returns going to star hub is also best. If one is going to use star hub topology, might as well make all ground connections there. If you have it going back to the amp board and the amp board ground to the PSU is a nice thick low impedance wire, it’s probably fine. But that’s why I had so many gnd FASTON tabs put on the board.

Thanks for the link for the case. I guess the trade wars with China are starting to take their toll as prices are now closer to $300 including shipping- still a good price for an all aluminum CNC’d Class A chassis.

attachment.php
 

Attachments

  • 9E1AFD23-B532-4B90-AB76-93A36CDB141B.jpeg
    9E1AFD23-B532-4B90-AB76-93A36CDB141B.jpeg
    56.4 KB · Views: 776
Last edited:
Since your floorpan is probably 3mm thick aluminum - I would imagine it can dissipate 7.5w each BJT no problem (2.5A x 3v).

Regarding the comments about where to ground stuff:

Try it and if it is quieter you can use it - but I doubt that it will be. I found that I had basically no ground loop hum with RCA ground connected to the star hub. Having the speaker returns going to star hub is also best. If one is going to use star hub topology, might as well make all ground connections there. If you have it going back to the amp board and the amp board ground to the PSU is a nice thick low impedance wire, it’s probably fine. But that’s why I had so many gnd FASTON tabs put on the board.

Thanks for the link for the case. I guess the trade wars with China are starting to take their toll as prices are now closer to $300 including shipping- still a good price for an all aluminum CNC’d Class A chassis.

Thanks, I'll try using the PSU star ground for the speaker return. However, my amp is already pretty quiet with one channel silent and the other with a hum I can only hear with my ear 8" from the speaker.

BTW I still need to hook up the front panel power indicator LED. Perhaps the next PSU you design can have an extra Vout connector for that?

The eBay seller has the chassis for $246 with shipping. There may be other, cheaper sellers; that was just the first I found. Still, a Deluxe 4U Aluminum Chassis from the diyaudio store at $339 is quite reasonable considering that you don't have to measure, drill, and tap. That is, unless you enjoy that sort of work and value the character your mistakes will give the project :)
 
Founder of XSA-Labs
Joined 2012
Paid Member
BTW I still need to hook up the front panel power indicator LED. Perhaps the next PSU you design can have an extra Vout connector for that?

You could put a set of PCB header pins where the D1/D2 LEDs are on the SLB and use them to connect external LEDs on the front panel. They are in series with 10k resistor as power indicators and to slowly discharge the bulk caps.
 
You could put a set of PCB header pins where the D1/D2 LEDs are on the SLB and use them to connect external LEDs on the front panel. They are in series with 10k resistor as power indicators and to slowly discharge the bulk caps.

Or run a fiber optic from the led to the front panel...

Hopefully you dumped the orange thing he called for and put a blue one in there :)