Many people are spending a lot of time by matching semiconductors!
My question, is there a real need for a testing/matching board?
If yes, then supply the schematics, I will spend time for layouting, then a GB.
My question, is there a real need for a testing/matching board?
If yes, then supply the schematics, I will spend time for layouting, then a GB.
Just use a standard DVM with a transistor tester on it. Most of the matching is already done by the manufacturer putting Hfe codes on the transistors. In the front end of an amplifier, I would bother to match the differential pair for Hfe and Vbe but otherwise as long as there is enough Hfe in the drivers and output stages, the negative feedback takes care of the rest.
Paralleled transistors need some selection/matching, if you need them to operate at similar temperatures.
Most people in Pass Labs forum are interested in mosfets not bipolar transistors
@2 picoDumbs
I know this article. I want to catch all existing to go eggs laying full-cream milking sow!
I know this article. I want to catch all existing to go eggs laying full-cream milking sow!
A matching device or curve tracer is of no use if you don't have enough devices to match from.
And since this is the biggest problem since Toshiba stopped essentially all analogue discrete devices, your best bet is to buy them pre-matched.
For decent pairing you need a pool of minimum 25~50.
And you will still have a high reject rate (our recent experience with FQP3N30/3P20).
We consider 200~1000 devices as normal.
So if you can solve the device sourcing problem, you can think about equipment.
Until then.
Cheers,
Patrick
PS
As have been discussed elsewhere there is a difference between matching for equal current sharing or distortion cancellation.
You only need a curve tracer for the latter.
And since this is the biggest problem since Toshiba stopped essentially all analogue discrete devices, your best bet is to buy them pre-matched.
For decent pairing you need a pool of minimum 25~50.
And you will still have a high reject rate (our recent experience with FQP3N30/3P20).
We consider 200~1000 devices as normal.
So if you can solve the device sourcing problem, you can think about equipment.
Until then.
Cheers,
Patrick
PS
As have been discussed elsewhere there is a difference between matching for equal current sharing or distortion cancellation.
You only need a curve tracer for the latter.
Last edited:
Who says we don't have enough devices.
I have 300 each of irfp240/9240
200 each of exicon laterals
Besides that it's not just about matching it's about choosing devices with desirable properties.
- Status
- Not open for further replies.