Beta Matching

Thread Starter

Joster

Joined Jun 12, 2013
95
Hi All,

I have a situation where I have 7 mje13009's that must be beta matched. My question is why does the beta vary so much from transistor to transistor. Does it get closer if they are all from the same batch? If so, how would I get them all from the same batch through Digikey or Mouser? How many would I realistically have to buy to get 7 matches?

Thanks,

Joster
 

GopherT

Joined Nov 23, 2012
8,012
Normally, they come off of a reel or in sleeves and you get a set of brothers all made the same day or even consecutively - since it is easier from a materials handling point of view, to keep everything oriented and together.
 

ErnieM

Joined Apr 24, 2011
8,046
I've worked with matched devices in die form and you would get a waffle pack with devices literally in the same order as they came off the wafer as adjacent devices have the best match. There are just too many minor variables with large effects on Beta for anything else to work.

The best way to get matched devices is to go to the source, the manufacturer. However, and this is an obsolete device...

How close do you need the match? Why are you trying to beta match power transistors?
 

Thread Starter

Joster

Joined Jun 12, 2013
95
I've worked with matched devices in die form and you would get a waffle pack with devices literally in the same order as they came off the wafer as adjacent devices have the best match. There are just too many minor variables with large effects on Beta for anything else to work.

The best way to get matched devices is to go to the source, the manufacturer. However, and this is an obsolete device...

How close do you need the match? Why are you trying to beta match power transistors?
does the manufacturer offer matched devices? I suppose they would cost alot more

that are for a type pf battery charger that pulses a 7 winding coil and sends the high voltage spikes to the battery to charge it.
 

crutschow

Joined Mar 14, 2008
25,132
.......................
that are for a type pf battery charger that pulses a 7 winding coil and sends the high voltage spikes to the battery to charge it.
You don't want to go with matched transistors as that would be difficult to do and expensive. You should be able to design the circuit with some emitter negative feedback so the value of the beta has only a minor effect on circuit operation. That's the normal way to design a circuit. If you post more details about what you need (frequency response, current values, coil inductance, voltage values, etc.) someone on this forum can likely help you with the design.
 

Thread Starter

Joster

Joined Jun 12, 2013
95
You don't want to go with matched transistors as that would be difficult to do and expensive. You should be able to design the circuit with some emitter negative feedback so the value of the beta has only a minor effect on circuit operation. That's the normal way to design a circuit. If you post more details about what you need (frequency response, current values, coil inductance, voltage values, etc.) someone on this forum can likely help you with the design.
Thanks that's good to know...I'll post a circuit asap.

Wondering...lets say I find an equivalent device. Does anyone know roughly how many I would need to buy in order to get 4 beta matched devices?
 

ramancini8

Joined Jul 18, 2012
473
Transistors from the same wafer seldom match unless the wafer has been designed to yield matching transistors. Even with matched die from the same wafer different mountings alter the matching, thus matched transistors are often mounted together. A design that depends on matched transistors is a guaranteed failure because so many secondary effects ruin the matching. Take Crutschow's advice and use a circuit configuration that is beta independent; this usually entails the use of degenerative feedback. I acted as fireman for Texas Instruments where I would fix customer circuits, and non-specified parameters (like matching) were the biggest problem.
 

ErnieM

Joined Apr 24, 2011
8,046
Ideally they would be exact or pretty close
So you will run these devices in a chamber to maintain the exact same temperature on all devices, and pass the exact same current thru each device?

Let either (and several other parameters) vary and even the best matched devices will wander.

That is why circuits are designed such that the current gain may vary (sometimes by more then an order of magnitude) and still operate property.
 
Last edited:

Sensacell

Joined Jun 19, 2012
2,568
Matching Beta should only be done as an absolute last-resort solution.

There should always be a better solution obtainable through good design practice.
 
Top