[runeight]

Perhaps you mean 100R??

 

[wiatrob]

Quote:

Originally Posted by runeight /img/forum/go_quote.gif And since you will need offboard heatsinks we should look at the Hitachi 2sk1058 and 2sj162 lateral fets or the irfp140n and irfp9140n.

And what for us BJT's?

I think I'll gibe the Hitachi's a shot for my three channel MOSFET build. Would there be any other part changes?

 

[wiatrob]

Quote:

Originally Posted by runeight /img/forum/go_quote.gif Perhaps you mean 100R??

DOH! that's what got me last time, yes 100 R!

 

[runeight]

Quote:

Originally Posted by wiatrob /img/forum/go_quote.gif And what for us BJT's? I think I'll gibe the Hitachi's a shot for my three channel MOSFET build. Would there be any other part changes?

Or, if I had a bit more time, we could start prototyping the EHHA II. It's already designed to use the Hitachi and IRF BIG fets. Then you guys wouldn't have to work so hard to heatsink the output devices. And these fets will drive a truck.

 

[wiatrob]

[size=small]MJ15004/[/size][size=small]MJ15003? Or I just resign myself to Fets

[/size]

 

[runeight]

Well ...

You'll notice in the normal build that the BJTs have higher emitter resistors than the mosfets. This to ensure that they don't thermally run away.

As you guys know BJTs have a positive temperature coefficient. This means that they hotter they get the more then conduct. Mosfets have a negative coefficient so they conduct less as they warm up. The fets don't have thermal runaway.

For headphone amps the 10R resistors on the BJTs are no big deal, but for a power amp 10R makes the amp impractical. We can make these smaller, like 0.47R but then we REALLY must ensure that the BJTs stay cool and never get into a runaway condition.

Which means that it will just be easier to use the fets ...

 

[runeight]

Quote:

Originally Posted by tennisets /img/forum/go_quote.gif I searched through this thread for a while, but I still have a couple of questions. I was intending to build a balanced M^3, but this project is tempting. My questions: is there anything else that needs to be done to drive speakers other than use 2W resistors for R32 and R33 and a Zobel network across the speaker terminals? I saw mention of paralleling the output devices. Any more info on this? Headwize being down kind of sucks. In addition, would two sigma22's be recommended for speaker use (as in the b22), or could I get by with one (with appropriately sized transformer)? Thanks!

tennisets, I'm going to repeat my reply because I just wrote it out too fast.

To make an EHHA power amp you should.

1. Change R32, R33 to 0.47R at 2W.
2. Change R13 to 1k.
3. Supply bigger heatsinking for the fets.
4. Remove R22, R23

I think it's right this time.

 

[tennisets]

Quote:

Originally Posted by runeight /img/forum/go_quote.gif tennisets, I'm going to repeat my reply because I just wrote it out too fast. To make an EHHA power amp you should. 1. Change R32, R33 to 0.47R at 2W. 2. Change R13 to 1k. 3. Supply bigger heatsinking for the fets. 4. Remove R22, R23 I think it's right this time.

Thanks for the detailed replies! I don't know if I want to have to deal with offboard heatsinks, so perhaps I will stick with just a headphone amp for now. I think I am going to give this amp a go as opposed to the M^3 though. Sounds like a fun build.

 

[runeight]

Don't let this necessarily discourage you on the speaker amp. It's necessary to give the worst case conditions so that there are no surprises.

However, as we know, music is not a sine wave and most listening is done a low power levels, often only 1W or so. The extra power in the amp is used to get dynamic range in the transient peaks.

With the big heatsinks that are on this board, if you use the tallest ones, you can probably get pretty loud without overheating the fets. It will be a very rare circumstance where you will hammer out a 40WRMS sine wave, unless you're trying out test tones.

If you socket a few resistors you could have both a headphone amp and a speaker amp.

 

[tennisets]

Quote:

Originally Posted by runeight /img/forum/go_quote.gif Don't let this necessarily discourage you on the speaker amp. It's necessary to give the worst case conditions so that there are no surprises. However, as we know, music is not a sine wave and most listening is done a low power levels, often only 1W or so. The extra power in the amp is used to get dynamic range in the transient peaks. With the big heatsinks that are on this board, if you use the tallest ones, you can probably get pretty loud without overheating the fets. It will be a very rare circumstance where you will hammer out a 40WRMS sine wave, unless you're trying out test tones. If you socket a few resistors you could have both a headphone amp and a speaker amp.

Thanks for the response. So you think I would be ok with something like this? I might just go with offboard sinks anyway, since my case will have room inside. I saw mentioned earlier in the thread that the hole locations for the heatsink pins are off. Is that still the case or was it corrected?

Also, why do you recommend no negative feedback for the power amp version? I would probably implement variable NFB with a rotary switch anyway, but I am curious.

Is a second power supply strictly necessary, assuming I would use a large enough transformer for a single supply? Thanks!

Edward

 

[tennisets]

Oops, removed.

 

[runeight]

The heatsink that you want is this one:

Aavid TO220

Except that this is the 2" tall version and Mouser doesn't seem to have the 2.5" tall version. The 2.5" is the one you want if you can find it somewhere.

The holes seem to be slightly off on the boards. We don't know why. However, I have a pair of board (unpopulated) and I have the correct heatsinks and they fit with no problems at all.

If you're using s22 supplies, I think that they are spec'd at 1A continuous current and much higher peaks. You can try one and see what happens. You can always build another one if you really need it.

 

[tennisets]

Quote:

Originally Posted by runeight /img/forum/go_quote.gif The heatsink that you want is this one: Aavid TO220 Except that this is the 2" tall version and Mouser doesn't seem to have the 2.5" tall version. The 2.5" is the one you want if you can find it somewhere. The holes seem to be slightly off on the boards. We don't know why. However, I have a pair of board (unpopulated) and I have the correct heatsinks and they fit with no problems at all. If you're using s22 supplies, I think that they are spec'd at 1A continuous current and much higher peaks. You can try one and see what happens. You can always build another one if you really need it.

Thanks, I actually just saw the heat sink specs on your site (not sure how I missed it before). The heatsink supplied with the Glass Jar kits is a smaller one, I guess.

As far as the supply, I guess I'll try one first. If it doesn't work out I can always just use the amp as a headphone amplifier until I get another s22.

 

[smeggy]

finally got around to doing the shunt mod on my RK27 pot tonight using 51k resistors. Sounds very nice. Dunno if It's my imagination but it seems clearer than standard. My audio memory is notoriously unreliable but I've been wanting to try it since I first read about it. So, mission accomplished!

 

[wiatrob]

Well, the EHHA has broken in nicely since I fixed my stupid resistor mistakes. Full, defined bass and a LOT of power and speed.

Now treats my Balanced RS-1 just as well as the HD600's - in fact, it evens out some of the differences in both cans, if you can believe that.

I still have to finalize the AC wiring (switches and removable line cords). But this amp is really nice. Wishing I had the K1Ks again (thanks for the loaner Smeggy!)