Amek M2500 transistors

[marcus4audio]

Hello guys. I would like to build this pre, but wonder if i substitute the transistors (Ztx214) for some other PNP small signal -low noise, do i have to re-bias them? I tried to simulate the circuit and figured out that currents and voltages are changing a lot with, for example BC650c. Now, if I have to re-bias them, should I look for the same voltages-currents like ztx214's or I have to target some other values to achieve best from transistor substitute?

 

[Harpo]

... with, for example BC650c...

the BC650c being a NPN Transistor.
On a sidenote, the ZTX214 has a different pinout compared to the BC214  ...

 

[Chris_V]

I think he means BC560C, which should be ok and as the same pinout (but a slighly bigger package).

 

[marcus4audio]

... with, for example BC650c...

the BC650c being a NPN Transistor.
On a sidenote, the ZTX214 has a different pinout compared to the BC214  ...

sorry, typo. Bc560c. I dont care about the pinout. I wonder should i re-bias the voltages to be like ztx214 or i have to go with other q points based on the choosen substitute? I know that direct drop works but i wonder do I have optimum performance withouth chanhing the resistors values around diff group and current sources.

 

[abbey road d enfer]

I tried to simulate the circuit and figured out that currents and voltages are changing a lot with, for example BC650c.

Then your sim must be wrong, because the operating point of the transistors is defined by the surrounding elements, principally by R5 and R20, thet in turn define the collector current of Q7 and Q2, which feed the input quads of transistors.

 

[marcus4audio]

I tried to simulate the circuit and figured out that currents and voltages are changing a lot with, for example BC650c.

Then your sim must be wrong, because the operating point of the transistors is defined by the surrounding elements, principally by R5 and R20, thet in turn define the collector current of Q7 and Q2, which feed the input quads of transistors.

Thanks, yes, right. I made a mistake.  Actually I used Tina software but when I switched to good old Circuit maker everything settled down.  diffences are negligible. I should change the topic name to "whats wrong with Tina" or why it's so easy to make non accurate simulations

 

[joaquins]

... or why it's so easy to make non accurate simulations

That's it!

Now, I don't know about those two transistor options, IIRC someone here came out with a rule of thumb for biasing BJT for the best noise performance, I guess he did some measurements which aren't for anyone to do since are quite tricky. I don't remember who or the number but I think it was a proportion of the nominal current.

In your case that's 100mA, let's say it was 1/4th of that. So 25mA but you don't want to get close to have thermal issues, in fact the cooler the better for the noise, so keep your voltage down. That's another preamp. In this case ~3mA should do, or you are looking for something other than the A2k5. Also I didn't take into account the quad transistors here and putting them in parallel will definitely have an effect on the optimum bias point, going towards lower bias current I guess.

JS

 

[Chris_V]

I did not look at your schematic before answering but BC560C is not the best choice here.
For that kind of topology, 2N4403 is the classical choice (and easy to find/cheap)

 

[Henke]

bc560 isn't "low noise" at low source impedances like this, you're looking for a low rbb' part such as 2sa1316 or 2sa1084. If you have difficulty sourcing these exotic transistors, try the 2n4403 from the jelly bean department.

 

[synthiaks]

hmm...

This is an interesting circuit, what's the transistors Q3-Q10 doing?

S

 

[Henke]

hmm...

This is an interesting circuit, what's the transistors Q3-Q10 doing?

S

It's actually a very straightforward circuit but it's drawn in a rather confusing way.
Q3-Q6 (and Q7-Q10) are just paralleled devices (lowering noise) in a simple transconductance stage

 

[synthiaks]

It's actually a very straightforward circuit but it's drawn in a rather confusing way.
Q3-Q6 (and Q7-Q10) are just paralleled devices (lowering noise) in a simple transconductance stage

Quite confusing! so Q11/12 is a longtail?

 

[joaquins]

It's actually a very straightforward circuit but it's drawn in a rather confusing way.
Q3-Q6 (and Q7-Q10) are just paralleled devices (lowering noise) in a simple transconductance stage

Quite confusing! so Q11/12 is a longtail?

No, current sources for Q3-Q10 polarization. Q3 to Q10 could be just two, and the circuit would work just fine, depending on the transistors specs the noise improvement could be significant.

JS

 

[marcus4audio]

Hey guys, thanks for the comments. I mentioned Bc560c like the example. For sure i can find better tr for the low noise application. My question was more like general rule of thumb for possible re-biasing the transistor in this kind of application. But as abbey road d enfer pointed there is no need for that. Most of them are acting very similar and I decided to experiment by ear So far I'm happy with the 2sa970

 

[synthiaks]

It's actually a very straightforward circuit but it's drawn in a rather confusing way.
Q3-Q6 (and Q7-Q10) are just paralleled devices (lowering noise) in a simple transconductance stage

Quite confusing! so Q11/12 is a longtail?

No, current sources for Q3-Q10 polarization. Q3 to Q10 could be just two, and the circuit would work just fine, depending on the transistors specs the noise improvement could be significant.

JS

Sorry for going a bit off topic here but i would like to understand this a bit better. I do understand the basic concepts of Transistor input stages Bias, longtails and push/pulls but this is still strange to me. Should i see it as Paralleled common collector (voltage followers) with a constant current source? Or am i lost here?

S

 

[joaquins]

The parallel transistors are just for noise, they work exactly the same as it if where just one (one for Q3 to Q6 and another for Q7 to Q10, I'll call them Q3 and Q7 from now on)

Q11 is a constant current source at the emitter of Q3 for polarization. Same for Q12 and Q7.

Q3 and the upper half of the TL072 (U2a from now, U2b is the half on the bottom) is configured as a transamp, in other terms, Q3 and it's polarization components, (C6, R5, R6, R7, R13 and Q1, sharing CR1 CR2 and R12) make a new input stage for U2a working effectively as a whole opamp, with better input noise performance (and other improved specs)

Everything at the left of Q3 and Q7 is the input protections and bias for the bases of them. R25 is to provide a DC path to U2 non inverting inputs, C9 is to reduce the noise at that resistor, which going to a rail doesn't make much sense compared to ground which should be cleaner.

Once you picture all that as an opamp, there is a instrumentation amplifier configuration. R8 and R11 are the feedback resistors, R3+R26 is the gain resistor, everything from U2 to the right form the subtraction amplifier.

JS

 

[synthiaks]

Thank you! 

 

[metalb00b00]

What about Fairchild KSA992FB? Lower noise than BC560C, although with a lesser gain bandwidth. Can it be used as a direct drop in replacement? If so, it'll be cheaper to have 4 of them instead of 10 BC560C ($0.20/piece from Mouser)

 

[PRR]

> KSA992FB? Lower noise than BC560C

Where do you see that? Fairchild quotes voltage noise in a strange way at a much higher impedance than a microphone.

I bet the main thing is: '992 is sold as "low noise", '550 is sold as a Power device where hiss is not the main thing. The '560's spec-sheet hiss number may be very far from typical for current production. Certainly experience shows it is reasonably low-hiss at mike interface. And it is available at far more places.

But sure, '992 will drop-in on all '560 preamp applications. Let us know.

 

[metalb00b00]

Thanks for the explanation PRR.

Truth is, I don't really know if it's really really low noise. Just a thing I skimmed read from somewhere about KSA992FB more suitable than BC560C for a preamp application. Line preamp probably :

Sadly I don't have any so I can't try it out myself. The ones I saw at a few local store scream out as "counterfeit", because they were marked A992FC. Datasheet only mention FA/FB hFE classes.