The Multi Amp aka Dynalo Mk2

[spritzer]

This one has been brewing for a while now but since I wanted to build a nice balanced amp it moved up in line.    Basic setup is the same as the old Dynalo but now with Super Symmetry input, higher rail voltages and naturally a bit more power.  Input stage changed to THAT340 which is a quad bipolar transistor array with matched NPN and PNP pairs.  It is cheap and easy to source so no jfet drama.  The other transistors are new as well and the pinout is different to reflect that. 

 

Here is a look at the new PCB design which is pretty much ready:

 

 

It is 138*86mm. 

 

Another thing from the same vein is a fully discrete unbalanced to balanced circuit that we have been working on so no DRV134 garbage.  I'll start with this one though and order some proto boards this week. 

Edited May 6, 2013 by spritzer

[Bidoux]

I remember amb told us about it a while ago, glad to see it is finally here. With Super Symmetry, is it possible to wire two output jacks in order to use it single ended or is it simply not an option ?

[jwzhan]

.

Edited May 7, 2013 by jwzhan

[kevin gilmore]

It's super symmetry!

One channel per board.

Unbalanced or balanced in, balanced out.

[jwzhan]

..

Edited May 7, 2013 by jwzhan

[spritzer]

Sorry about that, I had the thought of modifying the board so the same one could be setup for L+R instead of just balanced while writing the post.      Stock it can't be done but I'll see about changing it.  Might just change the board and offer a single ended one as well as I'm not really a fan of cryptic jumpers etc.  It should just have all the parts on the PCB, stuff it and it works. 

[Emooze]

Color me interested. I was actually pondering a balanced dynamic amp last weekend. But when are we going to see some supercharged, off-board heatsink version?

 

On a serious note, how critical is matching the transistors? And would an o22 be sufficient to power this?

[RudeWolf]

Looking good! I do like my Dynalo, but lately I've found it to be a tad rolled off at the extreme ends of the spectrum. At least compared to other amps I've listened to- Beta22 and simple hybrid amps. After I'm done with my speaker + amp project I might do another headamp. So far the Sjostrom QRV08 has been quite interesting to me.

[spritzer]

To be fair, the B22 is more than just a bit bright. 

 

Color me interested. I was actually pondering a balanced dynamic amp last weekend. But when are we going to see some supercharged, off-board heatsink version?

 

On a serious note, how critical is matching the transistors? And would an o22 be sufficient to power this?

 

There is already a supercharged offboard version... it's the Dynahi.    We might do an on board version of that though given that the LSJ109 will finally be available soon. 

 

Matching isn't all that critical, the amp will function just fine without any matching and 1% resistors.  The servo should have no problem taking care of any extra offset. 

[Pars]

To be fair, the B22 is more than just a bit bright. 

 

 

There is already a supercharged offboard version... it's the Dynahi.    We might do an on board version of that though given that the LSJ109 will finally be available soon. 

 

Matching isn't all that critical, the amp will function just fine without any matching and 1% resistors.  The servo should have no problem taking care of any extra offset. 

Thanks for making me fall out of my chair...

Edited May 8, 2013 by Pars

[spritzer]

No really, the LSJ74's are in production.  Seriously... they are finally seeing daylight. 

[Bidoux]

Matching isn't all that critical, the amp will function just fine without any matching and 1% resistors.  The servo should have no problem taking care of any extra offset. 

When I built my dynalo, the transistor and diodes were too poorly matched for the servo to function. After swapping leds (and taking advices from Justin) it kicked in to reduce the offset to less than 1mV.

[RudeWolf]

I had matched input JFETs and LEDs on my Dynalo and it was possible to dial it in so it had about 1mV offset. The servo has saved my ass on one occasion tho- when my DAC output stage was misbiased and outputted 0.5V of DC.

 

In any case matching the LEDs is trivial. And in the case of Dynalo it was easy to get matched input devices as well.

[spritzer]

...and on this one they are already matched. 

[kevin gilmore]

 Part Number Package 1-99 100-999 1000+
LSJ74A SOT-23 3L $6.93 $4.83 $4.59
LSJ74A TO-92 3L $6.63 $4.62 $4.38
LSJ74B SOT-23 3L $6.57 $4.14 $3.93
LSJ74B TO-92 3L $6.27 $4.02 $3.81
LSJ74C SOT-23 3L $6.21 $3.66 $3.48
LSJ74C TO-92 3L $5.91 $3.51 $3.33
LSJ74D SOT-23 3L $6.21 $3.66 $3.48
LSJ74D TO-92 3L $5.91 $3.51 $3.33
LSJ74E SOT-23 3L $6.21 $3.66 $3.48
LSJ74E TO-92 3L $5.91 $3.51 $3.33

 

The prices are way high, but at least they are available.

Edited May 7, 2013 by kevin gilmore

[Nebby]

That's....amazing.

[padam]

Definitely interested in this one.

[dubiousmike]

Looks awesome! Can't wait to hear more about this project as it comes together.

[c12mech]

Looks nice Birgir.  I like that it can be unbalanced or balanced in.  That makes source choice a lot easier.  The fact that parts are actually available is a plus too.

[fishski13]

nice. calculating TX VA, what's the safe max quiescent current bias and total current draw from each board/channel?

[FraGGleR]

Very interested in this project.  Thank you, Birgir, for pointing it out to me.

[spritzer]

Look what the cat (fedex) dragged in:

 

 

Quite a bit smaller than the old Dynalo boards. 

[Bidoux]

Very nice Spritzer. If I may ask, how much did you pay per dynalo board ? (and where did you have them made ?)

[spritzer]

Some progress aka I'm out of parts...

 

 

I did use the W transistors on this one as it's all I had.  More power handling which is good in my book.

[spritzer]

Some progress on this as well:

 

 

I've moved the 10uf servo caps to make the oversized Wima caps a better fit but tantalum's will be far smaller so not an issue.