How to measure the gain of MPSA13?

[gigimarga]

Hello,

I have a cheap (but not the cheapest ) DMM...for my electronic skills it's ok...but when i select the hFE section on the display appears only 3 digits  ...so i can't measure gains over 999...so i  can't measure the hFE of a MPSA13 (it displays only a big 1 - overload)...what can i do (without buying a more expensive one )?

Thx a lot!

[R.G.]

MPSA is a darlington in a single package. The typical gains will be on the order of 3000 - 20,000. The simple answer is - you probably can't. I doubt any packaged multimeter will measure it. Some special purpose transistor testers might. And on top of that, a darlington's gain changes even more with operating conditions and temperature than a normal transistor's, so any measurement is only good for that one test condition.

The whole point of a darlington is that the gain is so big you don't need to know it exactly.

[Dragonfly]

The whole point of a darlington is that the gain is so big you don't need to know it exactly.

Summed up into one nice little sentence !

[gigimarga]

Thx all for your answer R.G....very  cool!

I want to know exactly the gain of a MPSA13 because i want to build a Fuzz Face with 2 of them and i want to use the calculator from Analog Alchemy (i used this calculator for another Fuzz Faces and i am very happy with the results - i saw that the value of R1 is very important, too).

It is possible to use a technology like that is described in the amazing article "Technology of Fuzz Faces"?

Thx again!

[R.G.]

A darlington like the MPSA13 is a prepackaged device with a small-current transistor emitter internally connected to a medium current transistor base. It's normal to shoot for a minimum gain of 100 in making any signal transistor these days; 300-400 is very common. Both the small-current first transistor and the medium current second device will be optimized to make this true, in the nominal sense.

However, the current gain of a bipolar transistor is a function of its collector current. Low current => low gain, high current => high gain. But at even higher currents (relative to the device's size, of course) gain falls off again. Current gain variation of 3:1 are easy to get. In a multiplier cell, you deliberately change the current to change the gain, and that works over a range of typically 1000:1.

In a darlington the gain is the product of the two transistor gains. If the first one varies 3:1, and the second one varies 3:1, you get a 9:1 variation in the total gain. That's a pretty big window, big enough to make putting a single number on it very difficult.

As you probably know, I'm not a fan of special purpose calculators. They prevent you from learning what's underneath the thing being calculated. That's fine if you never want to know, but my mind has trouble understanding how any human could never want to know what's happening; just my quirk.

As a possible suggestion, go to the calculator and stuff in a whole range of values. Try 1000, 2000, 5000, 10,000, etc for the first transistor and the same for the second transistor for each of the first transistor values - sixteen trials there. See what the calculator does with it. What you hope for is that there is little change in the component values (I'm guessing that this thing calculates component values.) for some ranges of gain. If they're all 'way different, then you are stuck with tuning to fit the devices you actually have.

It is possible to use the technology of the tester in the TOFF. But you need to feed the tester a much, much smaller current - about 100x less. If you change the base resistor to 22M from 2.2M, and change the measuring resistor to 3000 ohms, then the result on the meter is 1V per gain of 1000. If you use a 300 ohm resistor, the result on the meter is 1V per gain of 10,000. Note that this is only valid at the measuring current, which you have no control of at all, because you're putting in a fixed base current and letting the collector current go whereever it wants to.

A better tester would force the emitter current to be some fixed value - 1.000ma, 100uA, etc. - and then measure the base current necessary to make that true. That can be done, but it's more complex than the setup I did, and I already get a steady ration of complaints that the tester can't be made to work.

[gigimarga]

Thx a lot R.G. for your answer...you're very kind!!

As i saw in your answer there is no chance to measure, in a simple way, the gain of MPSA13...so i reformulate my question, based upon the cause, not the effect : how could be biased Q1 in a Fuzz Face using a trimpot instead the collector resistor?
As i know (and saw...) the Q2's collector voltage must be near 4.5V...is something like that for Q1?

Thx a lot again!

[mac]

If you use a darlington, Vbe is about 1.0v to 1.2v, and Vc1 could be at 1.8v or so, so you may need about 4k or less total resistor at Q2 collector if a 33k is used at Q1 collector.
Another way is to use a 33k, the 470 and 8.2k, but a 2k or so at Q2 emiter. Or a 2k fuzz pot and a little tweak of the 8.2k.

Since I'm not a fan of hyper gain distortion, I would add a small resistor at Q1 emiter to soften things a little.

mac

[gigimarga]

Thx for your answer, but it's not clear yet for me how to bias Q1!
If i assume that Q1 and Q2 has a gain of 10000, the the Analog Alchemy's calculator said that R1 must to be 280 ohms, not 33K
Why?

[Dragonfly]

Heres a simple answer for you....

1 ) Choose a value for the resistor for Q1's collector....anywhere from 10k to 47k should work fine...you may decide to change this later, but for now it's fine.

2 ) use a 1k resistor in series with a 10k trim pot for Q2's collector.

3 ) use bias trimmer to set Q2 to about 1/2 the supply voltage

What I've found with FF style circuits is that, provided you make all the connections correctly and you choose values that are at least "in the neighborhood", then biasing Q2 generally pulls Q1 into a workable range. The two transistors work together....

[gigimarga]

Very good news for me, Dragonfly...thx a lot!