LM301 substitution

[zamproject]

Hi all

I have consoles that use the 301 which is now obsolete in dip8 package (soic seem to stay alive?)
The LM301 is used in a specific way I don't remember to see elsewhere, where the input stage is bypassed and a discrete differential pair is used in place and sent to the output stage via comp/bal pin.

I'm not expert at all in opamp topology/design and like to know what other of opamp is able to be used like that.

Here the idea of discussion is not to replace for "better" opamp, I have no complain about the sound of my 169 and 289, but anticipate a possible substitution when my spare spent out or when NOS stock get completely overpriced...

Best
Zam

 

[JohnRoberts]

you mean like this

for more discussion  https://www.proaudiodesignforum.com/forum/php/viewtopic.php?f=12&t=1135&start=10

The old 301 also had a trick that only worked in inverting topology to use a capacitor to bypass HF around the input stage for faster slew rate.

You can probably just buy a modern high performance op amp to use in place of the entire old kluge.

JR

 

[zamproject]

Hello JR

Yes, more or less what you show with a 5534.
Tks for the link, I'll have a read.

(one configuration attached, sum amp, but also used at the line out amp)

Best
Zam

 

[tommia]

:-XMouser 0,5 euro and use an adaptor pcb socket.
Never seen this schematic, interesting
Can you post other parts of the console where the lm 301 is used in such an odd way? What is the brand? I guess Dynacord or so, schematic feels German.

 

[JohnRoberts]

Hello JR

Yes, more or less what you show with a 5534.
Tks for the link, I'll have a read.

(one configuration attached, sum amp, but also used at the line out amp)

Best
Zam

I would personally still be tempted to use a modern op amp for the whole shebang (technical term), but the 5534 looks close to a drop in.

I might also reconsider the transistors used in the long tail pair if sticking with that approach, there are lower noise modern devices (probably). If you want to be closer to the original (why?) use the old transistor part numbers.

JR

 

[zamproject]

:-XMouser 0,5 euro and use an adaptor pcb socket.
Never seen this schematic, interesting
Can you post other parts of the console where the lm 301 is used in such an odd way? What is the brand? I guess Dynacord or so, schematic feels German.

This is swiss...you'll find this in Studer x69 and x89 series (service manual and schemo all around)
Yes, soic to dip is an option, might stock soic until they become obsolete too.

I would personally still be tempted to use a modern op amp for the whole shebang (technical term), but the 5534 looks close to a drop in.

I might also reconsider the transistors used in the long tail pair if sticking with that approach, there are lower noise modern devices (probably). If you want to be closer to the original (why?) use the old transistor part numbers.

JR

Ok you say 5534 should work, now I wondering why they don't use it by that time then.
Is any single non compensated opamp able to do this (via bal/com pin) or some internal topologies can't ?

I was told and read that only 301 is a candidate for that, and I repeat this myself, but never search deeper on this subject.

Again I'm not here to discuss the improvement of this design, otherwise I will not use those desks. I know your position about this and you'r right mostly I guess, there is for sure better design today for noise figure, THD etc... but that's what I have and enjoy to use.
But don't make me wrong, I'm not that conservative...When bc109/179 died in the console I have no scruples to use 550/560  ;D

Best
Zam

 

[JohnRoberts]

This is swiss...you'll find this in Studer x69 and x89 series (service manual and schemo all around)
Yes, soic to dip is an option, might stock soic until they become obsolete too.

Ok you say 5534 should work, now I wondering why they don't use it by that time then.
Is any single non compensated opamp able to do this (via bal/com pin) or some internal topologies can't ?

Most op amps do not break out multiple internal nodes like that. Modern companies also pay attention to reliability so invest in extra protection circuitry (like clamps) on every pin exposed to the outside world. This is an extra cost in silicon real estate and packaging pins for a feature nobody needs when they can just buy a quiet modern op amp. 

JR

I was told and read that only 301 is a candidate for that, and I repeat this myself, but never search deeper on this subject.

Again I'm not here to discuss the improvement of this design, otherwise I will not use those desks. I know your position about this and you'r right mostly I guess, there is for sure better design today for noise figure, THD etc... but that's what I have and enjoy to use.
But don't make me wrong, I'm not that conservative...When bc109/179 died in the console I have no scruples to use 550/560  ;D

Best
Zam

 

[user 37518]

Ok you say 5534 should work, now I wondering why they don't use it by that time then.

You have to consider that the ubiquitous 5534 wasn't the cheapest opamp back then as it is today, it was an expensive opamp, so many manufacturers used 5534s only in the most crucial stages. Engineering is not just about making things sound good, cost is also an issue.

 

[ruffrecords]

The first time I saw that circuit configuration, before the  5534 came out, it used an LM748 op amp.

Cheers

Ian

 

[zamproject]

The first time I saw that circuit configuration, before the  5534 came out, it used an LM748 op amp.

Cheers

Ian

Your right...early version of the schemo I post use 748
there is also an intermediate version with a TBA931, without discrete differential pair, as JR suggest in a previous post to modernize the circuit.
But the later rev is the one with 301 and dif pair.

You have to consider that the ubiquitous 5534 wasn't the cheapest opamp back then as it is today, it was an expensive opamp, so many manufacturers used 5534s only in the most crucial stages. Engineering is not just about making things sound good, cost is also an issue.

Yes I know that, but don't think Studer will mess (by that time) with this, this is the summing section of the top mixer they have then...and there is TDA1034 in less critical modules in the console.

Best
Zam

 

[JohnRoberts]

The first time I saw that circuit configuration, before the  5534 came out, it used an LM748 op amp.

Cheers

Ian

I never used a 748 back in the day but think they were pretty similar to 301. I just looked at a 748 data sheet and they show the same inverting mode feedforward trick to use a cap to bypass around the input LTP at HF. 

The ua748 appears to be Fairchild numbering and LM301 National Semi, both companies are now gone... looks like TI inherited the LM301 from National, Fairchild SGS Thompson (?).

JR 

 

[user 37518]

Yes I know that, but don't think Studer will mess (by that time) with this, this is the summing section of the top mixer they have then...and there is TDA1034 in less critical modules in the console.

Best
Zam

I wouldn't underestimate the Studer engineers, if they went the 301 route its because they realized it was ok, a 5534 may not make the difference you imagine. I always give the same example, around 4-5 years ago I decided I wanted to replace the 5534s in my Tektronix SG505 with LME49710 to further reduce its THD, I did and in fact the THD went higher, perhaps it was mainly due to noise since the 49710 has higher noise current than the 5534, in any case, the better opamp wasn't an improvement and the 5534s went back in. I was reminded that a good opamp is not the end of the story, good design is more important than a good opamp, one shouldn't be so quick to question the original designer's decisions.

BTW the TDA1034 is not bad at all, it has very similar specs to the 5534.

 

[Gold]

I remember an engineer telling a story about how they went in to improve their Studer mixer. They replaced everything with higher spec lower noise stuff. It lowered the noise but changed the noise spectrum for the worse. They went back to the original chips.

 

[zamproject]

I wouldn't underestimate the Studer engineers,

I don't !!! I have 3 Studer desk...
I just don't think they chose 301 for cheaper BOM as you suggested. As I say previously I'm not expert in opamp internal topology, I just search compatible one for this particular design.
Not for better (neither for worst...) "sound" just for guardrail as 301 is now obsolete and already overpriced

(To my knowledge TDA1034 is NE5534)

I remember an engineer telling a story about how they went in to improve their Studer mixer. They replaced everything with higher spec lower noise stuff. It lowered the noise but changed the noise spectrum for the worse. They went back to the original chips.

I can believe this 

Best
Zam

 

[JohnRoberts]

I remember an engineer telling a story about how they went in to improve their Studer mixer. They replaced everything with higher spec lower noise stuff. It lowered the noise but changed the noise spectrum for the worse. They went back to the original chips.

I've told this story several times but at Peavey we had a house number dual op amp (iirc MRPS # 70403678) that was basically a ne5532 but graded for noise floor specta... The testing sorted out op amps with high 1/f noise relative to the rest of the noise floor.

This was related to the WFO test that consumers would do, cranking all the pots full up and listening to the noise floor. 1/F noise sounded like something broken or about to break. 

JR

 

[tommia]

I always give the same example, around 4-5 years ago I decided I wanted to replace the 5534s in my Tektronix SG505 with LME49710 to further reduce its THD, I did and in fact the THD went higher, perhaps it was mainly due to noise since the 49710 has higher noise current than the 5534, in any case, the better opamp wasn't an improvement and the 5534s went back in.

I had the same experience trying to upgrade an orban EQ. Even the 741 went back in!

For the mixer I would have suggested to change the differential transistor pair to a real transistorpair in one case. But after knowing it is about Studer, I guess they matched the transistors, I would not touch it and leave all like it is. People love Studers like they are.

 

[zamproject]

Hello

For the mixer I would have suggested to change the differential transistor pair to a real transistorpair in one case. But after knowing it is about Studer, I guess they matched the transistors, I would not touch it and leave all like it is. People love Studers like they are.

For the shemo I show I'm not 100% sure about that, there is no mark on them.
I have various studer tools at the studio and when a pair is matched, or a selected transistor in a circuit, you usually have paint marking on them, an in house colour code.
For example the FET (2N5639) used in the PWM compressor in that desk have 3 colours lines at them.
Unfortunately I don't know what the code refer to.

So far about the topic, this seem complicated... I'll try a 5532 when back to the studio, just to see how it goes, and probably put a load of SOIC lm301 in my next RS order.

Best
Zam

 

[JohnRoberts]

Hello

For the shemo I show I'm not 100% sure about that, there is no mark on them.
I have various studer tools at the studio and when a pair is matched, or a selected transistor in a circuit, you usually have paint marking on them, an in house colour code.
For example the FET (2N5639) used in the PWM compressor in that desk have 3 colours lines at them.
Unfortunately I don't know what the code refer to.

So far about the topic, this seem complicated... I'll try a 5532 when back to the studio, just to see how it goes, and probably put a load of SOIC lm301 in my next RS order.

Best
Zam

5532 is a dual opamp and does not provide access to internal nodes.

try a 5534 (single) or source a lm301, note pin outs may not be identical between 5534 and 301 so confirm connections.

JR

 

[zamproject]

5532 is a dual opamp and does not provide access to internal nodes.

For sure...I think 5534 and wrote 32...this is the whole point here, having access to output stage of the opamp

(having called the single one 5531 would have saved me the quasi systematic confusion I have when talking about 553x  : )

Best
Zam

 

[user 41968]

The 5534 expects to see about Vcc-2.5V at the compensation pins.

The LM301 may be different. If a 5534 is dropped into the LM301 socket you may need to adjust the DC operating point of the front end transistors to use a 5534.