Need help engineering a replacement for an NLA part in an EML 101 Synth.

[mkiijam]

This is an EML 101 synthesizer.

Exponential Amplifier 3 has failed. The uA726 is no longer available in any form that I could find or feel comfortable with.

I don't see why I need the matched pair aspect of the original uA726, nor the heated aspect(?)

Could it be just as simple as popping in an off-the-shelf exp. amp?

[Swainster]

Disclaimer: No relevant experience in this area - just interested in old semiconductors.

OK, that out of the way, hopefully, this will at least help to bump your thread...

Obviously the matching is not required for this design, however the heating is critical for stable and repeatable performance of the exponential amplifier. This is probably quite important for a musical instrument. I would think that you could test the surrounding circuit by replacing the uA726 with a single transistor, something with high gain like BC109C. If the synthesizer comes back to life then it confirms that uA726 is the culprit. In order to get the stability back then you could then consider adding a crystal heater to the BC109C.

Actually, come to think of it,  if you are adding a heater then perhaps you could choose a transistor with a bigger TO-220 package, then tape the crystal heater to the heatsink tab (with suitable electrical and thermal insulation)

[Audiorepair]

It is also worth considering that the original circuit with u726 may have distortion etc characteristics that are highly desirable, sometimes what makes "vintage" synths sound, well, vintage.

So you could change/modify the VCA using any modern circuits/IC's, but then you might not have a EML 101 any more, maybe a modern modular instead.

[mkiijam]

I seem to remember that exponential amps have like two transistors in them and more stuff?

[mkiijam]

"...the pA726 works in circuits where the original µA726 has pins 3 and 10 (the transistors emitters) connected only, which is the most widely used design. It won’t work in a few rare exceptions such as the EML 101 or Moog Sonic Six"

[james_s]

Are both halves of the UA726 faulty? If not, since only one half is used here you could bodge it so the other half is used instead.

[Audiorepair]

The CA3046 has often been cited as an alternative workaround for this device.

However, if you buy them on Ebay, or otherwise from China, they are most likely not going to work as most are fake.
As I found out.


You want the proper Harris ones.

[mkiijam]

Are both halves of the UA726 faulty? If not, since only one half is used here you could bodge it so the other half is used instead.

Both sides read "Diode junction" with my atlas DCA55 transistor checker. So, to be honest I'm not sure. It for sure isn't heating.

[fzabkar]

https://tonatelierlauter.de/ua726-replace/

uA726 Replacement Chip

The first real drop in replacement for the famous uA726 heated transistor array. This chip comes pre-built, pre-calibrated, is the most compact in size and doesnt need any circuit modifications, as it has the same characteristics as the original uA726. Suitable for repairs and new builds.

[mkiijam]

https://reverb.com/shop/tonatelier-lauter

Nothing in his store.

[fzabkar]

Maybe it's because he wants 100 euros for the real (?) thing?

https://reverb.com/item/67392578-ua726hc

Still, it's not quite as rich as this guy:

https://www.aliexpress.com/item/32789581800.html?spm=a2g0o.productlist.main.1.48924b409WKQQy

[james_s]

Are both halves of the UA726 faulty? If not, since only one half is used here you could bodge it so the other half is used instead.

Both sides read "Diode junction" with my atlas DCA55 transistor checker. So, to be honest I'm not sure. It for sure isn't heating.

Well if it's not heating then you're pretty much screwed. I think I would try replacing it with discrete transistors just as a proof of concept.

[Swainster]

Haha, d'oh! Of course it is worth trying the unused transistor of the pair first.

If it isn't heating but the transistors are ok then it wont have a stable gain, however it should still work as an exponential amp so you should see some output. If you are getting nothing then there is probably another problem. Can you share some voltage measurements from around the circuit - the voltage at the IC pins shown in your schematic excerpt would be useful to know.

[james_s]

It doesn't seem like the circuit is doing anything *that* special. Surely there's a modern amplifier that could be configured to operate similarly?

[Swainster]

I guess that the modern way would be doing everything in software, making use of a high dynamic range ADC and DSP techniques - no temperature dependencies to worry about.

[james_s]

Is an analog exponential amplifier not a thing anymore? It's not something I've ever encountered that I can recall. Isn't the UA726 just a pair of matched NPN transistors with a temperature controller? What am I missing?

[Swainster]

No, it's still a 'thing' - I was considering using one in a design just recently, but in the end decided that the temperature dependance and general parameter variation was not worth the hassle in a production environment. Adding compensation didn't make sense for the low cost device that I was working on - luckily in my case a linear response was a reasonable compromise in cost vs functionality. As it happens, after the first prototypes came out, the product manager came back and cut so much functionality that the whole question became moot.

[mkiijam]

I'll take some measurements and post more in a bit, but wanted to say I tried a NPN transistor and it didn't seem to work or solve anything. Looking at the internal diagram in the spec sheet of the uA726 seems like a lot more going on, but is all that stuff just for the heater?

[james_s]

I looked at the datasheet yesterday and it looked like it was just two transistors, and then an analog circuit to control the temperature.

[Swainster]

Agreed, exactly that. A monolithic matched transistor pair with temperature stabilisation. In this particular circuit it is just used as a single temperature stabilised transistor. If you ignore the matched part, the only thing that I noticed is that the transistor has fairly high gain.

[fzabkar]

AISI, either the transistor is faulty, or the temperature regulation is faulty. The two components appear to be isolated, so if the circuit isn't working at all, this suggests that the transistor must be the culprit. If this is the case, then switching to the unused transistor should have restored the function of the circuit. Or am I missing something?

BTW, this is the datasheet:

http://www.synthdiy.com/files/2004/ua726.pdf

[james_s]

It's entirely possible for both parts to be faulty, even though they're isolated I think it's very possible, probably likely that they are all part of the same die.

[fzabkar]

What about disconnecting the power to the heater circuit, ie open pins 5 and 8?

[mkiijam]

Both sides read "Diode junction" with my atlas DCA55 transistor checker. So, to be honest I'm not sure. It for sure isn't heating.

[james_s]

Don't you have a multimeter? You can check a transistor with the diode test feature, B-E and B-C junctions should appear as diodes. I suspect it is faulty though, have you tried swapping in an ordinary transistor as a test? Assuming the gain is high enough it should work at least to some degree. If you can find a transistor that works then you ought to be able to build an oven to stabilize it, using a circuit similar to that of the original, or one that has been reverse engineered from an ovenized oscillator.