Transistor arrays

[xXBrunoXx]

Hi, I was thinking about transistors arrays for LTP and current mirrors, etc. ULN2003
They are darlington and most likely not what it should be for that use.
But I found these :
https://eu.mouser.com/new/that-corp...T*MTcxNDgzMTczMi4zLjEuMTcxNDgzMTc5Ni41OS4wLjA.

https://pdf1.alldatasheet.com/datasheet-pdf/view/561341/ETC2/300S14-U.html

Anyone used something similar before?.
- Bruno.

 

[MarcelvdG]

CA3046 (out of production, suffered from popcorn noise), SSM2210, BCM53DS, BCM56DS.

 

[Mark Tillotson]

They are expensive and low voltage only, so not really for power amps. Designed for microphone preamps basically with low noise at low voltage.

 

[MarcelvdG]

You can always cascode them for higher voltage applications.

 

[Mark Tillotson]

If you need extremely good matching it might be worth the expense, but do you need extremely good matching? For the same money you could buy ~30 transistors and match them yourself! And there's always the opamp approach if you need the matching, again cheaper.

 

[MarcelvdG]

If 2 mV of V_BE mismatch will do, the last two alternatives in post #2 are a lot cheaper.

 

[as_audio]

In the MPQ series originally made by Motorola you can find something suitable.
I remember the quad MPQ6700 :
"The CENTRAL SEMICONDUCTOR MPQ6700 is comprised of two 2N3904 (NPN)
chips and two 2N3906 (PNP) chips to be used as dual complementary pairs for
general purpose amplifier and switching applications."
Arrays like this may be hard to get today.

 

[dkfan9]

Ssm2212, mat14 from Analog Devices are both NPN, low noise, reasonably high beta. Check out typical beta on the LFCSP 2212. Matched low noise superbetas. THAT300 series has PNPs as well, much lower capacitance and beta.

 

[xXBrunoXx]

If you need extremely good matching it might be worth the expense, but do you need extremely good matching? For the same money you could buy ~30 transistors and match them yourself! And there's always the opamp approach if you need the matching, again cheaper.

So not a " better" ideea, in my case ( audio amplifiers ) , to use that instead of selecting ( matching ) two bjt's from a bag to say so.

 

[xXBrunoXx]

CA3046 (out of production, suffered from popcorn noise), SSM2210, BCM53DS, BCM56DS.

The BCM's are really cheap.

 

[xXBrunoXx]

First tought I had was using a / 6 / 7 , transistor array ic ". I saw a few schematics that have parallel transistors in the LTP, usually 2 or 3. read about it improves noise and what not. I was interested into trying myself. But if they are not " perfectly " ( as it could get ) the "whole purpose " might end up being worse then 2 as I understood ?!. Thank you for your comments !.
- Bruno.

 

[EdGr]

LTP mismatch should not be a problem with the usual large capacitor in the feedback network. You can add a trimpot if you want zero offset.

Emitter degeneration makes current mirrors tolerant of transistor mismatch.
Ed

 

[xXBrunoXx]

@EdGr I have around 5mV dc offset at the amp's ourput in the current design, with matched, glued togheter 2sa1815 as LTP.
My concern for mismatch was if I used more in parallel, let's say 6 transistors in the LTP. 3 - 3 .
#11
Thank you.
- Bruno.

 

[benb]

What's the application, input stage of a power amp? Or for a low-level signal preamp such as mic, phono (MM or especially MC), or tape head? As I think others were saying, you only need matched transistors for the latter applications.

The LM394 was one of the best things for this (see data sheet, it's like two sets of 50 transistors in parallel), but it was discontinued, I think even before TI bought NS. There's an equivalent being made, AS394, but it's several dollars each, and not available everywhere. Google:
AS394 transistor
to find datasheet and retailers.

Here's an old app note on it:
https://www.ti.com/lit/an/snoa626b/snoa626b.pdf

 

[EdGr]

Transistor noise is not a problem in a power amplifier. It is a problem in phono pre-amps.
Ed

 

[basreflex]

space grade components are still available. for example 2N2920A. to be found in nasa's preferred parts list.

 

[MarcelvdG]

LTP mismatch should not be a problem with the usual large capacitor in the feedback network. You can add a trimpot if you want zero offset.

Conversely, a well-matched input stage can be quite useful when you want to avoid that large capacitor and keep the number of trimpots to a minimum.

 

[gerhard]

Renesas HFA3096 & friends.
They also contain some of the very last RF PNPs that have survived, dual or not.


< https://www.renesas.com/us/en/produ...tra-high-frequency-transistor-arrays#overview >

 

[xXBrunoXx]

What's the application, input stage of a power amp? Or for a low-level signal preamp such as mic, phono (MM or especially MC), or tape head? As I think others were saying, you only need matched transistors for the latter applications.

The LM394 was one of the best things for this (see data sheet, it's like two sets of 50 transistors in parallel), but it was discontinued, I think even before TI bought NS. There's an equivalent being made, AS394, but it's several dollars each, and not available everywhere. Google:
AS394 transistor
to find datasheet and retailers.

Here's an old app note on it:
https://www.ti.com/lit/an/snoa626b/snoa626b.pdf

yes, input LTP of an audio amp.

 

[xXBrunoXx]

Conversely, a well-matched input stage can be quite useful when you want to avoid that large capacitor and keep the number of trimpots to a minimum.

What cap you mean ?. input or the one in the feedback?