electrocompaniet amp

[mikep]

I just stumbled across this schematic:

http://users.otenet.gr/~athsam/electrocompaniet.htm

a sucessuful pro mixer friend of mine went thru an electrocompaniet "phase" last year. I think he's back on bryston now but he assures me their amps sound very nice. im interested again after seeing the schematic. I like the VAS. with a different output stage and lower rails this topology would make an interesting (low signal) opamp.

mike p

EDIT: I can't spell

 

[bcarso]

It's all greek to me...

Seriously, I'll give a more considered review in a bit :grin:

 

[pstamler]

My admittedly dim recollection of Electrocompaniet reviews from the past is that they were notoriously unstable.

Peace,
PAul

 

[bcarso]

Paul, did you check out the schematic? It looks at first sight as if it would be a pretty good oscillator.

 

[Samuel Groner]

Looks like a variation on a design presented at IEEE 20 or 30 years back. The positive rail surely looks like it might better be higher than 19 V.

It looks at first sight as if it would be a pretty good oscillator.

I smell positive feedback as well--or did I miss something?

I like the VAS. with a different output stage and lower rails this topology would make an interesting (low signal) opamp.

Why do you like the VAS? This three-stage architecture is very prone to be difficult to compensate, while not providing serious audio advantages over more conventional approaches. Not surprised about Paul's statement.

Samuel

 

[thermionic]

This appears to be pretty similar to the notorious ETI P104 amp: http://novagraph.com/examples/mosfet.html

Both amps are takes on the famous Hitachi Lateral MOSFET apps data topology, albeit with less stability. From what I've gathered, many a DIY-er has come unstuck building the 104.

If you're contemplating building such an amp, my advice would be to go for a proven design such as Rod Elliot's Hitachi-based amp: http://sound.westhost.com/project101.htm

Make sure to take the advice regarding the PCB layout - I've been there and these amps can be very 'twitchy'.

Justin

 

[bcarso]

[quote author="Samuel Groner"]The positive rail surely looks like it might better be higher than 19 V.
I smell positive feedback as well--or did I miss something?
Samuel[/quote]

I'll bet the schematic is just wrong. Surely the feedback network should connect to the right-hand base of the input pair.

And indeed, +19V has got to be a typo. OTOH the -44V seems a trifle excessive to drive an output stage swinging towards -26V.

There appear to be a number of other implausible values. I would be very scared about building this as it is shown, even after hazarding a guess at the correct topology.

The cascade of diff stages with all that gain reminds me of a high-voltage-compliance current source I did in early days, when I knew a whole lot less. The requirement was actually not that stringent, but I saw an opportunity to play with more loop gain than was prudent. I was also relatively unsupervised and full of youthful enthusiasm.

It took forever to compensate, and to make matters more fun, the error amp sat at about 500V above the chassis potential. So trying to poke in there with a few hundred picofarads somewhere inevitably led to either circuit destruction or a massive jolt.

Years later someone wanted to take the thing on a road trip, which would have disabled another spectroscopic instrument. So I threw another power supply together in a matter of a few days, which worked perfectly on first power-up. How times had changed.

 

[Guest]

[quote author="mikep"]I just stumbled across this schematic:

http://users.otenet.gr/~athsam/electrocompaniet.htm

a sucessuful pro mixer friend of mine went thru an electrocompaniet "phase" last year. I think he's back on bryston now but he assures me their amps sound very nice. im interested again after seeing the schematic. I like the VAS. with a different output stage and lower rails this topology would make an interesting (low signal) opamp.

[/quote]

3 stages + 2 followers... Nice oscillator! :thumb: :green:

[quote author="Samuel Groner"]
I smell positive feedback as well--or did I miss something?
[/quote]

It is positive on DC and possibly some infrasound frequencies only.

 

[mikep]

indeed it is probably wrong in several regards. it probably says so in the text, for all I know. Bablefish wont even work on this page.

This was the only EC schematic I could find in 5 minutes of searching. word on the street is their amps are "inverting". anyone have a reliable schematic?

what I liked about the VAS was the mirror. hadnt seen that before. I guess i'm easily impressed after a 14 hour work day + 1 glass of Riscal.

regarding 2 stages, the first thing I thought was this looks a little 2520ish.

EDIT: if you reduce it to one stage degenerated diff amp and then the mirror loaded VAS it seems easy to compensate in my simulator. low open loop gain ~65dB but good open loop BW ~9kHz. As I see it, there are two advantages to this. relativly constant loop gain v/s frequency compared to a conventional opamp which presumably will help flatten distortion v/s freq, a very important property, IMO. the other thing happening is the error amp generates the most harmonic distortion. IMO this is preferable to a loop in which the VAS makes significant nonlinearity that is reduced by feedback in a high gain, high precision, current-mirror equiped error amp, as this situation leads to high order harmonics.

mike p

 

[bcarso]

One thing that can be said favorably for the design intent, is that it probably would have pretty good power supply rejection, given the amount of differential gain up to the third composite stage.

Actually, Rich May did an amp design that looked a bit like this but with one less composite stage. He liked putting the dominant pole in the first diff stage's collectors, as does Borbely. Different strokes...

 

[mikep]

I found some more info on EC. apparently their design decended directly from Otala's work. At first built as-is, they eventually added some minor modifications. the company is now owned by someone else. here are some more schematics to ponder (first one closly resembles the opamp I am fiddling with):

http://www.geocities.com/ResearchTriangle/Node/2356/newscheme.html

Im hoping that the title: "The influence of schematics on power amplifier parameters" is bad translation, it is too funny taken literally. too bad the link is dead.

After trying a couple of different approaches in the simulator, it does seem most effective (to me) to stabilize with RC between input stage collectors, FWIW. Im going to build a few opamps based on this topology and see how they sound subjectivly. low closed-loop gain inverter is the desired application I had in mind.

Edit: here is another link of interest regarding EC:
http://home.online.no/~tsandstr/OtalaStory.htm

mike p

 

[Gus]

A warning
at the last link in the above post I clicked a link and java started to load and the cancel did not work, it got by the firewall without having a warning pop up.

 

[Consul]

Sorry for the off-topic.

Gus, are you running Firefox with the NoScript extension? http://noscript.net/

It's been a lifesaver for me.