How to Calculate Transistor Leakage?

[mac]

Somebody MUST create a leakage current function of temperature and phase of the moon coordinates..

Ileakage(ic, T, a) = IRG(ic, T)*[R + r*cos(a/2)]

where
ic is the collector current
T the temparature
IRG is the RG leakage current
R is the average Sun to Earth distance
r is how much the Earth drifts from the average distance as it circles around the Sun
a is the angle of the orbit in radians

This Eq should be corrected to include moon phase and Mojo.

mac

[mac]

would I still use the standard (V-Vc*)/2472 to calculate leakage or would I also altar that slightly because of the 9.4 volts being fed into the tester?

Just measure the voltage drop across the 2472 and divide the result by 2472.

mac

[duck_arse]

mac, don't forget the minus sign - that lot are in the northern hemisphere.

[zbt]

@GibsonGM, OK 

Due to different VBE even my lovely OC44 has 0.25V, also Si and Darlington,
this draft come to mind!



combine with Ian Fritz method for match vbe, maybe we can also match hfe,
using current constant source.

[GibsonGM]

Thanks, I just 'approximated' the R.G. method, and managed to turn up 2 BJTs out of a bunch I have that seem like they'll work in a FF!     I need good germanium transistors approximately once every 20 years, ha ha   

[zbt]

Yup for germanium, just measure real value of resistor, and voltage, a little counting is enough for precision.

I need a versatile one, also for superfuzz match.

[Phend]

Just bb the keen tester using a series of resistors to get 2.2m and 2.472k, of course the dmm has an inaccuracy, but good enough. I'm gonna think like a dog, "my ears" will be the ultimate judge, I won't "chase my tail" with irrelevant minute differences, else I will never get a "muff fuzz".
But I hope to sort the good, bad and ugly.

From Freestompboxes, a table:
2.472v = 1mA
1.236v = 500 uA
0.989v = 400 uA
0.741v = 300 uA
0.494v = 200 uA
0.247v = 100 uA
0.124v = 50uA
    ^ Voltage across 2.472K resistor with 2.2M lifted.

And RG says:
"How much leakage is too much? 100uA is common, 200 happens pretty often. More than 300uA means the device is suspicious, and more than 500uA I would say is bad."

[Phend]

Interesting thread,
How many decimal places should I calculate the leakage to ?
(At my age leakage is observed in drops per hour)

[mac]

How many decimal places should I calculate the leakage to ?

If you want that Hendrix tone as many as Pi! 

mac

[PRR]

(At my age leakage is observed in drops per hour)

https://lunderg.com/confidence-clamp/

[Phend]

Humm..
I am finding that my Elcheapo Tester and Cheapo DMM and the my state of the art Keen BB differ.

Case 1 1T308V
Tester 72
DMM  75
Keen   42

Case 2 2N518A
Tester 153
Keen   154


Case 3 (unknown)
Tester 102
DMM  166
Keen   89

I will use RG's test , since he has shown the math's and explanations. 
The Chinese units I have are good for sorting resistors.
Those don't account for than darn leakage.
Apparently to make the holy grail FF these things need to be known.

[Rob Strand]

"How much leakage is too much? 100uA is common, 200 happens pretty often. More than 300uA means the device is suspicious, and more than 500uA I would say is bad."

If you are vetting devices then it depends on what you expect.    You might not be overly concerned with an AC128 at 300uA but you would certainly be thinking something is wrong if it was an OC44.   Most devices are around the values RG quotes.

As for using the correct leakage for a given circuit, that's a different story.   Maybe some AC128 work and some AC128's don't because the leakage is too high.