Krell KSA 50 PCB

TO247 or TO264...TO-3 Daddio!

AndrewT said:
Hi,
the MJ15003/4 are To3 devices. They will not fit the output boards.
You need To247 or To264 devices. ONsemi show an L to indicate this. eg MJL4281

I don't have the output boards and I wouldn't be able to use them on the tunnel if I did. It all about TO-3's, see previous post.

The 15003/4 are very robust SLOW devices that were originally used by Krell. The pink selection are 15times faster. Does that make them 15times better? No!! but there may be some sound quality improvement if the speed increase is cascaded all the way back through the schematic.
The 301/281 devices and similar have a high gain at high output current This may be useful in extracting good performance into low impedance loads.
The fast MJLs are also less robust and have only a 150degC Tcmax limit. You need many more of them to equate to the abilities of the 15003/4, That why Pink uses 3pair, but 3pair of 150W devices still leaves the output lacking, That's why I decided to go 5 or 6pair.

I'll run 3 pair of 15003/15004 per channel. I don't think it will be lacking, though I would like to stuff in more but the sink holds 12 TO-3's in total.

I have two 700VA, 40-0-40 VAC toroidals, one for each channel and I'm just deciding on the enclosure and the caps.

Cheers,

Shawn.
 
Input Stage devices

Mark A. Gulbrandsen said:
I prefer the Japaneese devices at the input stage, they sound better and are alot quieter. They are not that hard to find!! Also consider the MJL21194/94 pairs for the output stage. These are also available in TO-3 as well... MJ21193/94 in case you decide on that route.

Mark

Sounds interesting Mark. Could you list the devices you prefer on the input? I think the input stage is more critical than the output, provided the output devices are appropriate.

I have some MJ21193/94's but I want to try other devices too.

Cheers,

Shawn.
 
Re: TO247 or TO264...TO-3 Daddio!

TomWaits said:


I don't have the output boards and I wouldn't be able to use them on the tunnel if I did. It all about TO-3's, see previous post.



I'll run 3 pair of 15003/15004 per channel. I don't think it will be lacking, though I would like to stuff in more but the sink holds 12 TO-3's in total.

I have two 700VA, 40-0-40 VAC toroidals, one for each channel and I'm just deciding on the enclosure and the caps.

Cheers,

Shawn.

Shawn, I've got a similar but much bigger sink than yours with 4 pairs of TO-3 per channel and 37VDC out of two 700VA toroids (note you will have higher DC and higher heat with 40-0-40 toroids).

It does get very hot with 50WPC class A, the sinks approach 60C over 3 hours of use at somewhat lower speeds to reduce noise; so I expect yours to be good at cooking up the eggs as well. I did get 15003/04 devices but went with the MJ21193/94 becuase they were newer (no idea if they are better or not).

Anyway, looks like you are on your way, good luck, I think your design will work out for you.
 
Re: Re: TO247 or TO264...TO-3 Daddio!

lgreen said:


Shawn, I've got a similar but much bigger sink than yours with 4 pairs of TO-3 per channel and 37VDC out of two 700VA toroids (note you will have higher DC and higher heat with 40-0-40 toroids).

It does get very hot with 50WPC class A, the sinks approach 60C over 3 hours of use at somewhat lower speeds to reduce noise; so I expect yours to be good at cooking up the eggs as well. I did get 15003/04 devices but went with the MJ21193/94 becuase they were newer (no idea if they are better or not).

Anyway, looks like you are on your way, good luck, I think your design will work out for you.

Thanks for the good words. :) I think I'll use two fans, one pulling while the other pushes. Yes the tunnel is short; it measures 10.8 cm long, it is extremely well made though. I guess I won't be able to reduce fan speed much with such a short block to cool but I agree it will work well. I may have to juggle fan speed and bias to get some kind of "in between balance" but I would like to run it as hot as I can. :hot:

I guess I'm in the market for a variable fan speed controller, with enough juice to rock two fans. As a side note, the PC world has created a giant market of high CFM/low noise fans and I will be tapping that market when I select them. I thought I saw a fan controller in this thread way back...PinkMouse made it, I think. I'll need to do some more digging on that.

40-0-40 will yeild the "magic" maximum rails I have read about of +-55VDC. Still I thought some one measured 57VDC rails on an orriginal KSA50? I swear I read it in this thread too but I could be mistaken.

Cheers,

Shawn.
 
Pabst fans are probably one the best fans you might consider

Especially the 3414/3412 NGL, it's one of the queitest with the highest CFM/DBA ratings,... from memory I think it's a 92mm
You might check there site for other sizes

These are quality fans that smoke most of the computer cheapies

Also, fan noise goes way up when it's (input) is close to anything.
It will be way quieter if you push/pull air from the top unimpeaded and exit it down and out the case. (one fan only)

I built mine to blow upward and even though I'm an 1 1/2" away, it gets noticably quieter when I put them up on blocks...

Regards
David
 
Re: Input Stage devices

i wasn't asked but that never stopped me before ...
:D

for bipolar input stages, i like 2SA872/2SC1775 when i can find them and 2SA970/2SC2240 when i can't. i know other people have used 2SA1015/2SC1815, but i have not.

it helps if you match like polarity devices and at least choose complementary pairs from the same hFE group (if you decide not to try matching them, too)

mlloyd1


TomWaits said:


Sounds interesting Mark. Could you list the devices you prefer on the input? I think the input stage is more critical than the output, provided the output devices are appropriate.

I have some MJ21193/94's but I want to try other devices too.

Cheers,

Shawn.
 
grimberg said:
Shawn,

The voltage on the original Krell KSA-50 is about 37.5V. This information is available on the Wiki along with some other valuable tips.

http://www.diyaudio.com/wiki/index.php?page=Building+guide

You can also consider water cooled ceramic blocks, as used for CPUs and video chips.

Sorry, my bad, I have two 30-0-30 400 VA's for the Krell KSA50!
~40VDC rails.

And I have 40-0-40 800VA's for Krell KSA100!
~55VDC rails.

Gots a lot on my project table right now, getting hard to keep organized? ;)

OK back to normal here...somewhat. I hope I can knock off both amps this winter? :cool:

Cheers,

Shawn.
 

Attachments

  • toroidalssmall.jpg
    toroidalssmall.jpg
    79.5 KB · Views: 570
Re: Re: Input Stage devices

AVWERK said:
Pabst fans are probably one the best fans you might consider

Especially the 3414/3412 NGL, it's one of the queitest with the highest CFM/DBA ratings,... from memory I think it's a 92mm
You might check there site for other sizes

These are quality fans that smoke most of the computer cheapies

Also, fan noise goes way up when it's (input) is close to anything.
It will be way quieter if you push/pull air from the top unimpeaded and exit it down and out the case. (one fan only)

I built mine to blow upward and even though I'm an 1 1/2" away, it gets noticably quieter when I put them up on blocks...

Regards
David

I'll check out the Pabst but your advice on fan proximity and noise is dead on! I was not even factoring that in but it even changes how I want to lay the amp out inside the chassis. Thanks man!


mlloyd1 said:
i wasn't asked but that never stopped me before ...
:D

Nor should it, bring it on. :up:

for bipolar input stages, i like 2SA872/2SC1775 when i can find them and 2SA970/2SC2240 when i can't. i know other people have used 2SA1015/2SC1815, but i have not.

it helps if you match like polarity devices and at least choose complementary pairs from the same hFE group (if you decide not to try matching them, too)
mlloyd1

I guess I'll be looking at some more data sheets tonight. I'm even looking into the Zetex devices EchoWars is keen on.

Now how about that variable fan control circuit?

Cheers,

Shawn.
 
R-Theta

grimberg said:
Back in December of 2002 I participated in a group buy of large heatsinks made by r-theta in Toronto.

http://www.diyaudio.com/forums/showthread.php?s=&threadid=8542&highlight=

I used mine to build an Aleph 2 clone. It would be nice if we could get enough people to organize a new one and use them to build the Krell clones.


R-Theta is a 40 minute drive from my home. I almost dropped $1,000 on a single extrusion with them just this past summer. They were having some hard times with their union workers on strike at the time and I found a pile of large sinks in a surplus warehouse. I never gave them a PO.

What did you use your 8 pcs of Peter Daniel R-theta on? Pass?

Cheers,

Shawn.
 
well, i have given up trying to find a decent passive sink for cheap, and now i've also given up trying to find a TO-3 fan tunnel similar to the original. i guess i will use plastic devices and the output board after all. I am now eyeing the aluminum extrusions sold by barredboss on eBay. i wanted to make a rear-exiting tunnel out of two sections and now i am just wondering how much of it i will need per channel if i use 3 pairs of MJL21193/94 at 39 volt rails. i am aiming for about 55 watts per channel output, my calculations show that depending on which power supply i go with(not yet decided) the most i will need to disappate per channel is 160 watts so i would like the heatsink to be able to disappate maybe 180 watts to be safe(will that be safe?). i would like to use a quiet 120mm fan hopefully moving about 40 cfm...possibly more if necessary but it depends on finding a quiet enough fan.. can i get away with a 6 inch section per channel or would 9 inches be smarter?

what devices are recommeded for drivers?

lastly, i've asked about this before but not really gotten an answer, maybe better if i post in the PSU forum but i wanted to use a capacitance multiplier supply instead of the "2 big cans" approach. has anyone used one of these before, and is there any reason its performance might not be as good as a large bank of caps or single pair of large cans? the details are here

http://sound.westhost.com/project15.htm

thanks again
 
TomWaits said:
two 30-0-30 400 VA's for the Krell KSA50! ~40VDC rails.
I have 40-0-40 800VA's for Krell KSA100! ~55VDC rails.
both these transformer/amp proposals will give higher voltage than the Krell originals.

The 30-0-30 will give about 4Vdc extra and generate more heat and consume more power.
Your 3pair tunnel will cope with more heat, OK.
Your 400VA will not cope well with more power.
Either reduce the voltage or reduce the Iq. Neither are optimum.
You may get reasonable performance by increasing the 56mF substantially try 75m or 100m.

The 40- 0 - 40 will give about 6Vdc extra and generate a lot more heat and need a lot more power. I think this one is going too far to extrapolate the 100mk2 design.
 
AndrewT said:

both these transformer/amp proposals will give higher voltage than the Krell originals.

The 30-0-30 will give about 4Vdc extra and generate more heat and consume more power.
Your 3pair tunnel will cope with more heat, OK.
Your 400VA will not cope well with more power.
Either reduce the voltage or reduce the Iq. Neither are optimum.
You may get reasonable performance by increasing the 56mF substantially try 75m or 100m.

The 40- 0 - 40 will give about 6Vdc extra and generate a lot more heat and need a lot more power. I think this one is going too far to extrapolate the 100mk2 design.
I agree but I disagree, is it fair to say that? I think in everyday use these values will work well and that it should not be that hard to find some balance. I will proceed regardless and I most certainly will share the outcome. T, sometimes you just have to do it and try not to worry too much. I am not carrying forward on reckless abandon here, nor do I suggest it but I have to experiment in real life scenarios. This is how I learn best, by doing and experimenting. I am willing to take some small chances too but I'm not a gambler unless I'm in Vegas. ;)

Don't hit me with the ruler when I say this, but I think I'm going to try running the 100 off 60VDC rails with 1kVA per channel. I already have the toroids. :eek:

Cheers Man,

Shawn.

P.S. I own a variac! They dial up nice and slow if you want them to. :)
 
Official Court Jester
Joined 2003
Paid Member
AndrewT said:

both these transformer/amp proposals will give higher voltage than the Krell originals.

The 30-0-30 will give about 4Vdc extra and generate more heat and consume more power.
Your 3pair tunnel will cope with more heat, OK.
Your 400VA will not cope well with more power.
Either reduce the voltage or reduce the Iq. Neither are optimum.
You may get reasonable performance by increasing the 56mF substantially try 75m or 100m.

The 40- 0 - 40 will give about 6Vdc extra and generate a lot more heat and need a lot more power. I think this one is going too far to extrapolate the 100mk2 design.

in both cases difference is slightly above 10% rise of mains voltage.....
for me that looks as nothing to care , if heatsinks are good enough

regarding power-that really depends on load (spks he use),nothing more ....and nothing less ;)