[Nerull]

!

Holy crap, I just happened on the website after looking for some information about the M^3 and lo and behold, he's got a new project in the works! The Beta22 amplifier.
I'll let you read the details, but man, it looks amazing!
Hopefully, it'll be out by Christmas or some time when I have money...but wow, I was waiting for the step-up and here it is.

(Link here! -> http://www.amb.org/audio/beta22/)

This actually has me super excited, and I cannot wait to build one. No doubt, it's a beautiful design, and man am I jumping ship to that thing! I've been dying for a discrete, true class A amplifier, that's DIY and superior in design to the M^3. (and more easy to build than the Dynahi...)

Sorry if I rambled and sound like a kid going to DisneyWorld, but man, it looks so cool and I love DIY!

~Tom

 

[jarthel]

how do classify something as high-end?

 

[Garbz]

does not look much easier than a dynahi

 

[bperboy]

That looks pretty sweet. I think I'll still plan on building an M3 though, to not outstrip my Millett. I want them to complement, not compete! I imagine with 3 boards for the active ground version, this Beta22 could get kind of pricey!

 

[Rescue Toaster]

I don't see a huge advantage of a balanced version of this over a dynamid or dynamite, except perhaps for the single pair of cascoded output devices instead of the paralleled non-cascode ones the dynas use.

Loop feedback across the built in cap. multipliers makes me wary of using this without a fully regulated supply as well. Of course loop feedback always makes me wary.

 

[tomb]

Hmm ... some of the first comments seemed to imply the design is self-power-regulating; that anything special in terms of power supply might even be counter-productive.

Amb is still focusing on the MOSFET's (yay!), so comparisons to the Gilmore designs may not be valid.

Jarthel - maybe you can suggest to Amb how to build a flexible PCB if you're unsure of his design's high-end status.

 

[Teerawit]

I'm game

 

[amb]

Quote:

Originally Posted by Garbz does not look much easier than a dynahi

This is all discrete just like the dynahi, which means more parts to solder onto the board, but one of the goals of the β22 project is ease of construction, so I plan to make the mechanical interface to the case less labor-intensive. If you use onboard heatsinks then the assembly shouldn't be much more difficult than the M³. I also plan to make it possible to use larger, case-mounted heatsinks as well; since there are only four power devices per channel vs. dynahi's eight, the β22 will have less precision holes to drill on the sink. Also, due to β22's high PSRR, a much simpler PSU is needed.

 

[MASantos]

SWEEEEEEEET!!!

I'm starting to consider building an "high end" amp in 2007, either a Dynahi or balanced dinalo. This just ruined my thoughts and now I have another amp to consider

 

[kevin gilmore]

Very nice. A good board layout with integral heat sinks should
make this fairly easy to build. You are still going to need a beefy
and high quality power supply as the power supply rejection ratio
on this, similar to the dynahi is poor.

 

[amb]

Quote:

Originally Posted by kevin gilmore Very nice. A good board layout with integral heat sinks should make this fairly easy to build. You are still going to need a beefy and high quality power supply as the power supply rejection ratio on this, similar to the dynahi is poor.

I dunno, even if the pspice simulation is a bit optimistic and we get somewhat less than the 92dB PSRR in real life, I would hardly characterize a figure in that ballpark as "poor"...

 

[Rescue Toaster]

Quote:

Originally Posted by amb I dunno, even if the pspice simulation is a bit optimistic and we get somewhat less than the 92dB PSRR in real life, I would hardly characterize a figure in that ballpark as "poor"...

What's the open-loop PSRR?

If you wouldn't use an electrolytic as an output coupling cap (and I wouldn't either), you can't have one in the power supply current loop. You've got to isolate your current loops from nonlinear elements in the supply. Just counting on feedback to take care of it is bad news, imho.

 

[kevin gilmore]

The power supplies are going to have to be very stable and
matched because since there is no dc servo the output dc voltage
would move with the power supplies.

 

[Andrea]

Quote:

Originally Posted by kevin gilmore The power supplies are going to have to be very stable and matched because since there is no dc servo the output dc voltage would move with the power supplies.

Kevin, if you don't mind I'd have a question to ask, related to this subject.

It's just to be sure -- do op-amps and integrated buffers behave as if they had a DC servo (which perhaps they do have?), or will their output DC voltage also fluctuate like in this case?

 

[kevin gilmore]

In general, the higher the open loop gain, the more the amplifier
is resistant to changes in power supply. However offset voltages
and currents at the input of the amplifier still apply and will effect
the output dc voltage. Temperature changes also effect this.
So opamps with an open loop gain of >100db are going to ignore
power supply variations much more than a discrete amp with
an open loop gain of about 30db.

In many circuits a servo is a necessary evil.