HFE matched tolerance

[oran.outan]

Hi,

What will be the tolerance or the maximum HFE interval for matched transistor ?
I'd like to match (pair) some NOS Philips BC549c but with 10 of them, no one have the same HFE value.
I only have one close pair = x1: 553 / x2: 551 . Will that be acceptable ? or does it need to be exactly the same values ?

thanks for your feedback.
Best
0-0

 

[gyraf]

question does not compute - entirely depends on what your intended circuit needs, what you're trying to do

 

[JohnRoberts]

What is the particular circuit's sensitivity to beta (HFE)? If the circuit is very sensitive to beta, test the devices in a similar circuit.

JR

 

[Walrus]

Surely 0.4% difference will be good enough??

 

[oran.outan]

question does not compute - entirely depends on what your intended circuit needs, what you're trying to do

so, will that work for a 2520 style DOA like a Gar or similar ?
I also have a schoeps style LdC whose I will start soon… I already have 2x onsemi bc550 matched but maybe give a try with Them !?

Thanks
0-0

 

[Walrus]

The 2520 i was making I matched the input pair to within 10 of each other.
They work perfectly.

 

[Voyager10]

so, will that work for a 2520 style DOA like a Gar or similar ?
I also have a schoeps style LdC whose I will start soon… I already have 2x onsemi bc550 matched but maybe give a try with Them !?

Thanks
0-0

Yes, 1% or so matching is good enough for both applications.

For a DOA, matching of transistors can affect the input offset voltage and current (and therefore the DC offset at the output), and the common-mode distortion (i.e. how that offset changes as the inputs swing up and down).

For the Schoeps circuit, matching will affect the DC offset between XLR pins 2 and 3. This is only a worry if feeding a transformer-coupled preamp, where it causes a DC current to flow in the primary. Too much DC will cause distortion (although I can't find figures for 'how much' for typical transformers).

 

[oran.outan]

@Voyager10 , thanks for the explanations ! good to know...

best
0-0

 

[Whoops]

What will be the tolerance or the maximum HFE interval for matched transistor ?

As close as possible, but it doesn't need to be 0%.

Probably most of the times less than 5% is enough.
I guess less than 1% could be considered a "perfect match"

Please correct me if I'm wrong

I'd like to match (pair) some NOS Philips BC549c but with 10 of them, no one have the same HFE value.
I only have one close pair = x1: 553 / x2: 551 . Will that be acceptable ?

Yes it will be acceptable, but there's more to it, the match only happens at the same conditions the transistors will be used in the circuit.

So if you measure the HFE with a tester, it will give an HFE value for the specific current draw that tester uses.
But the circuit might have other conditions, and the match you made with the Tester might not be match at the circuit conditions.

Jakob and JR will be able to phrase this better and explain it better than me.

or does it need to be exactly the same values ?

No

 

[JohnRoberts]

As close as possible, but it doesn't need to be 0%.

Probably most of the times less than 5% is enough.
I guess less than 1% could be considered a "perfect match"

Please correct me if I'm wrong



Yes it will be acceptable, but there's more to it, the match only happens at the same conditions the transistors will be used in the circuit.

So if you measure the HFE with a tester, it will give an HFE value for the specific current draw that tester uses.
But the circuit might have other conditions, and the match you made with the Tester might not be match at the circuit conditions.

Jakob and JR will be able to phrase this better and explain it better than me.



No

Generally you can only match as closely as the parts you have available to select from. Back in the 80s I sold a preamp kit that benefited from using matched JFETs (lower distortion) in the front end. I bought several hundred parts and tested each one in the exact circuit I was matching them for. I put the parts in paper envelopes and wrote the value of the parameter I was matching for. After that I just paired together the closest matches for my kit customers.

Since then I avoided making designs with such sensitivity to component matches. Sometimes you can purchase arrays with matched components inside a single array.

JR

 

[Cqwet Dbdfte]

Legacy-ware is expensive. Glass to metal hermetic packaging is expensive. Not a problem for DoD or NASA.

https://www.digikey.com/en/products/detail/linear-integrated-systems-inc/LS351-TO-78-6L/13688005

https://www.digikey.com/en/products/filter/bipolar-transistor-arrays/277

If you can up your game to SMD assembly, there are choices. Otherwise getting boards made and populated by JLC is not expensive.

https://www.digikey.com/en/products/detail/nexperia-usa-inc/PMP5501Y-115/1157748 maybe?

https://www.mouser.com/ProductDetail/THAT/320S14-U?qs=9Udfh7QmL4u%2BQY2LzA4HXA==