Allen&Heath S02 Op-Amp question

Hello, i am not too deep into electronics.. enough to do repairs.. but far from beeing able to do circuit design. I ve a question about the discrete build Allen&Heath S02 Op-Amp. Does the voltages in the picture suggest that this opamp is biased into class a operation? 

And has anybody some other info about this opamp? Or what do you think about the design?

I just got one old Syncon A console and are about renovating it. Many functions are down and maybe i run into broken S02 opamps.

Of cause i like to restore original condition but i also need to go into production soon.. 

The transistor types used on this op amp boards are not so easy top get.. How critical would it be to just use some complementary types?  Is it likely that Allen&Heath used selected transitors in the outputstage or will the circuit take care of slight missmatchings?

Sorry for the stupid sounding questions..but..living is learning ;-)

> class a operation?

Is that good or bad?

> biased into class a operation?

You want to know if the power (last) stage idle current is large or small compared to the load current.

Do some math. Sadly there's too many odd voltages on that drawing. I've picked the Q6 Q7 Emitter resistor voltages. From Q6 E to Q7 E is (1.55V-1.1V) or 0.45V. From Q6 E to Q7 E is (180r+180r) or 360 ohms. 0.45V across 360 ohms is 0.00125 Amps or 1.25mA.

What is the load current? What is the load? What is the maximum output swing?

You don't run +/-24V supplies to get a few volt signal. By eye the output can swing over 20 Volts peak.

Load.... well, you would know more about the architecture and context than most folks. Is it driving 600 ohm loads? Just a few 10K pots? Just 100K pots?

But as a wild guess, those output emitter resistors would normally be much-less than the load. So if they are 180r, you expect loads over 2K.

20V peak in 2K is 10mA peak current.

10mA peak current is much greater than the estimated 1.2mA idle current.

Class AB.

> selected transitors in the outputstage

Not fussy at all. Use anything big enough. If stalled worst-case DC load a transistor and a resistor will dissipate 0.8W each. Worst-case AC/Audio, 0.4W each. That's in line with modern TO-92 jellybeans.

Q1 Q2 match should not be way-way-out, but there's no point in paying for "Matched" because the input pair runs a 2:1 unbalance already.

thank you...  i will find out soon enough..but after closer investigation of the real thing in front of me ..there are no indicators of transistor selection ..just all out of the same batch.

The loads these op amps are driving are  similar to actual op amp mixing desk circuits..  Theese discrete s02 opamps sit everywhere where you find in later consoles TL071 or 5534 stages..

THey arealso  the output stages and the minimal load these can take is 200 ohms.


When i understood you right that would make the peak current even more greater than the idle current..therefore they are in AB operation.. is that correct?

thnx anyway.

I don't understand how these voltages are measured. Surely, the opamp should be surrounded by at least a DC FB path and a galvanic path to ground for the non-inverting input, and these would influence the measurements. The differences between analog voltmeter and DVM seem to be more a reading/precision error than an actual loading effect.

@PRR: What makes you say that "the input pair runs a 2:1 unbalance already"?

abbey road d enfer said:

I don't understand how these voltages are measured. Surely, the opamp should be surrounded by at least a DC FB path and a galvanic path to ground for the non-inverting input, and these would influence the measurements. The differences between analog voltmeter and DVM seem to be more a reading/precision error than an actual loading effect.

@PRR: What makes you say that "the input pair runs a 2:1 unbalance already"?

Click to expand...

The schematic indicates around 130mV DC voltage between the two input LTP bases. IIRC it's something like 1dB current difference for every 3+mV of Vbe difference... My back of the envelope math suggests a lot more than 6 dB difference, while I don't know how trustworthy those voltage numbers are.

That said neither the input or output voltages are 0V so there is a "rest of the story" circuitry not visible.

It doesn't look like a hot rock, with values shown so maybe build one and see... I wouldn't expect any magic, looks pretty bare bones. 

JR

As I wrote earlier, these values, in particular the input offsets are so dubious they can't be used as the basis of any conclusion.
Meas at point 5 suggests 2-20uA base current (assuming measurement done with 10k-100k res at + input). Considering the input pair operates at about 300uA, it's more or less adequate.
But nothing justifies negative base current for Q2 (Vout being 1.3V negative respective to Vin-).

> @PRR: What makes you say that "the input pair runs a 2:1 unbalance already"?

I was tired, the slipstick slipped.

The ratio is nearer 1800 against 3300, the bias and load resistors, which is indeed pretty near 2.

The output bias can also be figgured by R4 R5 and junction drops. This leads to like 0.6V across (180+180) or 1.7mA.

So loads over 10K are in A, and 600 ohm loads must cross into AB for most signal peaks.

I'm not at ALL convinced it matters, with those monster 180r ballast resistors. The active emitter impedance varies from ~~20 ohms to less than 1 ohm, but the total resistance change is only 200 to 181 ohms, nowhere near what we usually call "crossover distortion".

I don't think this design is worth deep debate. It's adequate but not extravagant. Gosh, it's only 3 stages of current gain! If you need a VERY-clean sound, the best fix is to torch off the parts and find a good +/-24V chip

PRR said:

> @PRR: What makes you say that "the input pair runs a 2:1 unbalance already"?
The ratio is nearer 1800 against 3300, the bias and load resistors, which is indeed pretty near 2.

Click to expand...

The plot thickens... I read the current in Q2 is about 180uA into R1 + about 10uA Q4 base current, versus the total 330uA LTP current, so Q2 runs at 190uA and Q1 at 140uA. It could be better, but it's not really bad. Or is there something else that I don't see?

Hi again,

i was able to try them now.. they do sound good..overall performance sounds better than the typical 072/ 5534 middleclass desks..

I still wonder a bit what causes this subjeltive better sound.. the supply voltages are higher...  And while i look at them..

Why do they have a separate power rails for the output stage of the op amp? The voltage is distributed to the channel strips only one time..splits from that point and really runs extra traces for all 4 power rails of the op amps..  why that?


sorry when this is a stupid question.. but you dont see this kind of things with integrated opamps.

thanks again..
Sven

3phase said:

Why do they have a separate power rails for the output stage of the op amp? The voltage is distributed to the channel strips only one time..splits from that point and really runs extra traces for all 4 power rails of the op amps..  why that?

Click to expand...

These rails are probably decoupled differently. Since the output stage operates in class AB, it can produce rectified currents, which, injected in the ground path can produce nasty distortion. One can minimize that type of pollution by decoupling these rails to an ad hoc dirty ground. Now generally, this is accompanied by some kind of series rail decoupling, like low value resistors or inductors.

interesting.. thanks