Leach type input for ne5534

[donovas]

Recently ive been messing with discrete inputs for ne5534 as per circiut below. It works well, sounds good but i wanna get it to work with a leach type input seen on the other pic.

Would it be as simple as using the collectors of q1 and q3 to the opamp's pins?

Please let me know, thank you.

 

[Mark Tillotson]

Interestingly I found conflicting equivalent schematics for the 5534, some with PNP inputs and some with NPN inputs. Tying 2 & 3 to -15V only works for the NPN input schematics, I wonder if there are multiple versions or if your experiments simply invalidate the PNP schematic as nonsense (it has several differences, not just in the input transistors).

Checkout the schematics in these references:

NE5534 Integrated Circuits under Repository-circuits -54440- : Next.gr

http://www.ti.com/lit/ds/slos070d/slos070d.pdf

 

[sgrossklass]

There always some way of doing something but I don't think interfacing a complementary differential input with a single differential pair is going to be easy.

If you need this to be supporting higher input levels, look into adding a cascode with some common-mode bootstrapping to the existing LTP.

 

[donovas]

Thinking back it doesnt seem practical since for that part count one might as well build a DOA.

Singleton input to ne5534 makes more sense.

Any ideas on this?

 

[sgrossklass]

Here's the most simple bootstrapped folded cascode topology I could come up with. This is just meant for illustrative purposes, it might work absolutely terribly for all I know. I haven't checked the currents or anything.

EDIT: OK, now somewhat sane. At least it would probably work like this.

 

[donovas]

Here's the most simple bootstrapped folded cascode topology I could come up with. This is just meant for illustrative purposes, it might work absolutely terribly for all I know. I haven't checked the currents or anything.

EDIT: OK, now somewhat sane. At least it would probably work like this.

How does it simulate?

I would not mind building this

And how about a singleton ver?

 

[sgrossklass]

Latest versions. I haven't actually done much in term of simulating, just op points for now. Looks like you want a floating reference current source for things to remain largely supply-independent.

Can't say I'm too happy with the folded cascode yet, aside from the parts count I also don't have much headroom for the whole common-mode bootstrapping thing (the one advantage of this scheme becomes a disadvantage here), and for that to work well I think I'd also have to replace the resistors down there by current sources again (more parts).

I think a regular cascode it is then. I had a singleton cascode in a version of the 3-transistor preamp, so the normal cascode would be sort of a cross between that and the circuit above.

Do you have any particular application and supply voltage range in mind?

 

[PRR]

Interestingly I found conflicting equivalent schematics for the 5534, some with PNP inputs and some with NPN inputs. ...
Checkout the schematics in these references:
NE5534 Integrated Circuits under Repository-circuits -54440- : Next.gr

You mean this? That's just wrong. Mis-copied. 5534/5532 is NPN inputs. It comes from a period where IC PNP was too crappy to use in good inputs.

 

[aparatusonitus]

Grandpa and dad...

 

[PRR]

All the factory sources show NPN....

 

[Mark Tillotson]

A big pinch of salt is needed with the phrase "equivalent circuit" !!


It also goes to further show just how far ahead of the curve the 5534 was, since PNP is in theory lower noise than NPN, yet few opamps competed with the 5534A for noise for many decades.

 

[PRR]

...PNP is in theory lower noise than NPN, yet few opamps competed with the 5534A for noise for many decades.

The theory is not unbeatable even for "ideal" parts.

The PNPs in older IC technology were crap. A parasite side-bay of a NPN tub. Hans' book gives some ugly details, though he was not too much concerned with hiss.

Are you saying that with 2020 processing, somebody is selling PNP-input '5532s? That would be odd.

 

[MarcelvdG]

The NE5534 schematic with PNP input transistors makes no sense at all. Even if you would want to use PNP emitter followers as the input stage, which would be a bad idea if you want low voltage noise, tying together their collectors and connecting them to a current source would not make any sense. Besides, like PRR wrote, 1970's lateral PNPs are indeed quite poor in all respects.