A*P*I DOA question...

[mikefatom]

Hello,

Just would like to know if the original version of the 2520 (or the currently in production version) has a current source at the differential input pair tail or just a resistor. I have seen basically just 1 schematic of the 2520 which I'm not sure is original (the one on fabio's site) and it doesn't have the current source. However, a lot of DOA clones that are supposed to be "improved" 2520s have the current source implemented.

Comments?

mike

 

[Walrus]

As I understand it, the 2520 that has Studio Systems Division on the bottom of the label, are pre current source versions. The ones that have Melville or Huntingdon on the bottom will have the current source circuit.

 

[mikefatom]

It wouldn't seem so.

I just partly answered my own question.

From the huntington 2520 spec sheet, it mentions the fact that it's a "9 transistor" design which would mean that it doesn't have the current source.
This assumption is simply based on the fact that the basic 2520 building blocks require 9 transistors, which would mean another one (the tenth) would be a current source transistor.

Also, if I understand correctly, the huntington version came out after the melville version which would suggest that the melville don't have the current source either (as the current source is an 'upgrade').

Correct me if I'm wrong here guys!

cheers,

mike

 

[Walrus]

You're wrong! 
The Huntingdon data sheet, and also the one with the Melville address on, is showing a picture of a Studio Systems Division 2520. This has 9 transistors. The 2520's with Melville or Huntingdon on are 10 transistor current source types.

 

[mikefatom]

Hey Walrus,

Thanks for the reply!

Are you talking about this data sheet? :
http://www.waltzingbear.com/Schematics/API/API_2520.htm

I'm really confused  ??? because it clearly says it's a 'nine transistor amplifier'.
Could that mean that they are not accounting for the tenth transistor which they deem not 'amplifying'?

Would you happen to have the Melville data sheet by any chance?

cheers,

mike

 

[burdij]

Maybe a little timeline would help.

Company founded in 1968 by Lou Lindauer and Saul Walker

Studio Products Division in Farmington - up until late 1971 seems to be using Melcor products in their custom consoles such as the one they produce in association with Fedco Audio Labs in 1969. Their main product, the 550A equalizer also uses Melcor op amps. Their other main product is conductive plastic linear faders.

Company moves to Melville NY in late 1971 and remains there until around 1974 when they move to Huntington NY.

They are out of business by 1978.

In 1981 we see the first mention of a 10 transistor 2520 in a datasheet during the Datatronix era.

 

[drpat]

...

 

[PRR]

> Are you talking about this data sheet? :
http://www.waltzingbear.com/Schematics/API/API_2520.htm
I'm really confused   because it clearly says it's a 'nine transistor amplifier'.

That's what it says. So does the schematic:
http://www.waltzingbear.com/Schematics/API/API_2520_schematic.pdf
Note that this schematic has flaws!!!

BUT: look at the common-mode input range: there is no way they can do +/-12V with +/-15V rails using a simple resistor. If inputs are at +12V then we have 26.4V across 150K or 0.176mA. Half to each side, 0.088mA. In 82K or 90K load resistor, that's 7.2V drop down from the +15V rail.. but the bases are 3.6V down from the rail so the collectors can't be much lower than 4.1V down. Not 7V down.

The schematic does say that R1 is replaced with current source, but gives no details of "when".

You can determine if a specific module is R1 or current source. Wire as unity gain follower. Wire 10K pot from V+ to V-. Run 100K from pot wiper to opamp input. Power with +/-15V. If you can swing a full +/-12V output, it is current-source. If you can only swing +/-7V, it is R1 type.

If you run the amplifier at closed-loop gains over about 10, the difference is quite small. The current source has advantage mostly at low closed-loop gain.

 

[burdij]

Where does "Studio Systems Division" fit into that timeline?

Sorry, that's a case of not being able to read my own handwriting. There always was a "Studio Systems Division", from the start of the company. All the conductive faders manufactured during that early period are marked with that name, so that really isn't a good indicator of an historical period. I still haven't found a good reference as to when the so-called "Melcor engineers" show up on the scene and develop the 2520 but I will continue searching the AES archives. It apparently didn't happen right at the start of API.

 

[sliebers]

I've found a current source transistor on all the Huntingtons I've dissected.  10 transistors total.


Scott

 

[Walrus]

I've found a current source transistor on all the Huntingtons I've dissected.  10 transistors total.


Scott

Ditto the Melvilles I've dissected. However, all the Studio Systems Division ones I've done have been 9 Transistors.

 

[mikefatom]

Very cool piece of history there guys. Thanks a lot for all the replies!

PRR: Thanks for the lil technical note! I will have to do some pspice simulation of both versions to really see the difference for myself.

cheers,

mike

 

[mikefatom]

PRR: can you elaborate on the r1 limitations? I tried doing a DC sweep at the input with a r1 opamp but got a full output swing...

 

[Bo Hansen]

Mike,

The 2520 which I have dissolved and drawn a diagram from, have a pc-card mark with "A5A" have a CCS on the first NPN diff. pair, and the second diff. pair are PNP with emitters up towards +v.
So there shall be total 10 transistors on the "A5A" version.

You can find the "A5A" mark on the solder side of the PC board if you manage to get the top off.
It is possible to carefully file around the edges to release the thin epoxy top layer, and then also lift away a thin PCB that acts as a screen.
(of course just to investigate already broken 2520)

What I know, the "A5A" version was the "Datatronix/Reston V.A" version under the 1980´s, maybee they start with this earlier in Huntington before API convert to Datatronix.

BTW, under the 1980´s it was a lot of problem with the 2520 op-amps, but the best version in my opinion was the "Melville N.Y. version", because we purchase 6 of 312 cards in the middle of 1970´s, and they still work perfect with out any trouble, but on the later API product I have change a lot of faulty 2520 op-amps, even on the new products to day.

--Bo

 

[opacheco]

You can determine if a specific module is R1 or current source. Wire as unity gain follower. Wire 10K pot from V+ to V-. Run 100K from pot wiper to opamp input. Power with +/-15V. If you can swing a full +/-12V output, it is current-source. If you can only swing +/-7V, it is R1 type.

Dear PRR, Could you exlplain us this procedure in more in detail way please?...If you add a drawing for will be great!!

Thanks a lot anticipated
opacheco

 

[FIX]

The 2520s all had current sources, it's the only way you can really stabilize a high gain stage. Most, if not all, of the changes in the different models/versions were because of component unavailability. They weren't focused on "making them the same" as much as making them work. That obsession was never their problem... The schematics I keep seeing posted all over the interwebs are all pretty much wrong. CAPI makes a good one, API make theirs as close as they can to the original (but which original? LOL), and the rest is history. The reality of the older op-amps that everyone made was first, to make them work and not explode, and to fit them into the 1" square standard instrumentation package. Slew rates and things like that weren't the focus, as if you got past 2.5v/usec, you were clear of the audio band, and transformers didn't go much above that anyway.

 

[Walrus]

The 2520s all had current sources,

The Studio Systems Division labled ones didn't, the input LTP just has a resistor to the negative rail.

 

[sr1200]

So what kind of crap are they using now. I have 2 friends that have bought authentic API products in the past decade, and the 2520's in them (one just a 500 series preamp, the other one of their small consoles) the pre amp has had to have the DOA changed out at least twice. The console has had, i can't even count how many have been replaced now, especially in the master section.

 

[Walrus]

Modern 2520's use a lot of surface mount components, whether that is a factor I don't know.....

 

[JohnRoberts]

I am not the API expert but it seems like the 2520 DOA was been widely copied/knocked off. There are perhaps less than authentic units floating around. Component substitutions and even things like rail voltage can impact performance/reliability.

Who exactly are the "they" you refer to?

Hopefully Paul Wolff will chime in, now that this thread has been "cleaned" .

JR