...and oddly enough this thing actually works.    There are still issues to sort out but it sounds pretty ballsy even in the crap test setup. 

 

 

The output tubes are run too conservatively so that needs to be changed and I need to figure out why the left channel sometimes cuts out. 

 

Does look pretty though...

 

Edited June 29, 201313 yr by spritzer

and since the krell ksa5 klone is wrapping up...

 

this should be final version

 

http://gilmore.chem.northwestern.edu/squarewave.jpg

 

as usual with all singlepower stuff, the sound was very right, and the implementation was very wrong.

well this fixes all of that, adjustable bias, heatsinks for the transistors, and a dc servo, and otherwise

just about as simple as possible.   Just kickass for hd800

 

3.1 x 3.1 inches

 

 

Arizona desert ironwood knobs coming out absolutely gorgeous. harder than ebony.

and a lot less trouble.

Edited June 29, 201313 yr by kevin gilmore

How readily available are the transistors for that Squarewave clone? I'm guessing the power supply would be the usual Sigma 22 or the like?

How readily available are the transistors for that Squarewave clone? I'm guessing the power supply would be the usual Sigma 22 or the like?

 

All are easy to find and in current production.  PSU can be pretty much anything bipolar.  When Kevin was working on Jacob's amp I found some Elpac units what would work and are quite cheap.  A 317/337 unit like the ones Twisted Pear sell would also work.  Ebay is filled with suitable PSU's too. 

All are current production mouser parts.

 

mje210 863-mje210g    1 0.49 0.49
mje200 863-mj3200g    1 0.54 0.54
2n5210 512-2n5210ta    3 0.21 0.63
2n5087 863-2n5087g    2 0.13 0.26
2n4393 610-2n4393    2 2.69 5.38
opa134 595-opa134pa    1 2.94 2.94
1n914 512-1n914    2 0.1 0.2
heatsink 513202B02500G  (newark)  2 1.67 3.34
dip8 socket 571-2-1571552-2   1 0.81 0.81
2 pin terminal 651-1729128   2 0.82 1.64
3 pin terminal 651-1729131   1 1.21 1.21
.1uf 100v ls=5mm 80-R82EC3100DQ70J 4 0.32 1.28
4.7uf 63v wima 505-MKS24.7/63/5  2 4.33 8.66
22pf  ls 3mm 598-CD5EC220JO3F  1 1.79 1.79
10 ohm 71-rn60d-f-10    2 0.11 0.22
49.9 ohm 71-rn60d-f-49.9   1 0.11 0.11
100 ohm 71-rn60d-f-100    2 0.11 0.22
200 ohm 71-rn60d-f-200    1 0.11 0.11
221 ohm 71-rn60d-f-221    1 0.11 0.11
332 ohm 71-rn60d-f-332    2 0.11 0.22
10k  71-rn60d-f-1.0k    4 0.11 0.44
34.8k 71-rn60d-f-34.8k   1 0.11 0.11
75k 71-rn60d-f-75k    1 0.11 0.11
1M 71-rn60d-f-1M    2 0.11 0.22
50k pot 652-3299w-1-503lf   1 3.25 3.25
    
    
    34.29  per channel not including the circuit board (should be about $5 each if we make 100)

plus power supply, chassis, pot... cheapest thing the stax mafia have done to date.
 

Edited June 30, 201313 yr by kevin gilmore

Next weekend when I got the transformer and make sure the new kgsshv works fine, I will organize a small groupbuy of every board available. Hopefully it will go smoothly like last time...

What is the power supply for the Squarewave?  Nothing complicated I suppose.  Looks like a nice little "cheap" amp.

Any bipolar that can give enough current will do just fine.  Sigma22 being the obvious choice but there are dozens and dozens of PSU's out there. 

The 50mm high heat sinks (532-513202B25G) are out of stock at Mouser (at least for me). Will the 25mm be ok or will they get too hot?

I'd go for the 2.5" instead. 

I found these instead, which appear only to be slightly different:

 

http://jp.mouser.com/Search/ProductDetail.aspx?R=531202B00000Gvirtualkey53210000virtualkey532-531202B00

Those need to be tapped for screws which is why I prefer not to use them.  Just extra work and a bit annoying to use. 

I know this thread's well past its prime but I finally put mine together and I think there's an error on the board. You need to mount the 2N4393 from the bottom of the board or twist the leads for the pinout to match the TO-18 part. It's running fine now though.  This is on the boards with the onboard heat sinks.

 

Haven't had much luck with the servo though, and I notice it doesn't match the schematic, which uses a second op amp for inversion. The ouput doesn't drift more than maybe 10mV in the one I have running, any reason why I shouldn't skip the servo and connect the gain resistor (R3 or R17) directly to ground without the caps?

 

E: the schematic, for reference.

Edited September 12, 201411 yr by PretentiousFood

on the first board you have to flip the leads on the 2n4393.

 

if you well match the two fets, you can delete the servo.

Got it, thanks! It's running without the servo now, just need to finish up the power supply and box it up.

finally putting the single ended version in a chassis

 

http://gilmore.chem.northwestern.edu/squarewave4.jpg

I have some Squarewave boards on the way, and I was unable to find instructions on how to measure bias and what values are recommended. Can anyone help?

 

//UFN

about 120ma, measure across the 10 ohm resistors

Thanks Kevin! 

 

//UFN

Finally managed to put the boards together and test them. They seem to work fine, but I am unable to get the voltage across the 10R resistor below app. 1.45V and I have app. 2V offset on the output. Any ideas on what the problem is??

 

 

//UFN

lots of variation in the fets. i had to hand pick, and throw out a few.

 

the 1.45 volt is not that out of range, i can actually get mine to 160ma.

 

you measuring the 2V without the servo? the servo takes about 30

seconds to bring it to about .1v or less.

Think I may have found the problem (or at least part of it). The trimpot is 25K in the schematic and I used a 20K (because that's what I had) but it is 50K in the BoM above. That might be why I haven't got enough adjustment range?

 

The 2.2V offset is with the servo in place BTW.

 

 

//UFN

yep.  if you match the fets with the transistor tester you can get that 2.2V down to about 100mv

but you need a bunch of them, over a 4:1 range on idss

 

done

Oh well, I replaced the 20K trimpot with a 50K and the result is - nothing! Exactly the same values as before, meaning 1.45-1.5V across the 10R resistor and 2.15-2.2V offset on the output. Will have to keep looking I guess...

 

 

//UFN

Try changing 75K resistor to 62K range,combined with 20K pot I can reach 125mV without problem

 

 

Oh well, I replaced the 20K trimpot with a 50K and the result is - nothing! Exactly the same values as before, meaning 1.45-1.5V across the 10R resistor and 2.15-2.2V offset on the output. Will have to keep looking I guess...

 

 

//UFN