F5 power amplifier

[tinitus]

just to make it clear, I do not know this stuff
but from what I have been told, I doubt very much they will be that accurate

have you tried to measure them again, on a different day ?

on the good side, you might have several nice pairs, and not just one


btw, anyone ever tried to blow hot air on them, with a fan or something, to see if they follow each other, and change in unity ?

[alazira]

Originally Posted by tinitus

just to make it clear, I do not know this stuff
but from what I have been told, I doubt very much they will be that accurate

have you tried to measure them again, on a different day ?

on the good side, you might have several nice pairs, and not just one

Hi Tinitus,
Well, I found better accuracy with 100 ohm (then 1V=10mA) and most digital multimeters will give down to hundredths reading. I suppose one could warm the jFets to a specific temp for better matching. It should be good enough for sorting or approximate matching, if not precise Idss.

[Zen Mod]

normal jig with 9V voltage source , DMM as current meter
one , two , three , five
read
next

[tinitus]

well, you still have a lot other components in the curcuit, which probably cause much greater variation
hence I doubt such great accuracy matters that much
but ofcourse it doesnt hurt to try

hell, I need to get on with it and try it myself, and hopefully soon

[tinitus]

Originally Posted by Zen Mod

one , two , three , five
read
next

first I thought, sloppy zenmod

but then this way you might actually take into account the temperature coeffciency, rise time factor, or whatever its called

dont say you didnt think of that

[Zen Mod]

one number for each leg , then some

who needs logic all the time ?

[EUVL]

The methods described in post #11251 and #11254 are both very accurate.

The greatest contributors to inaccuracy are (in random order) :

1) Multimeter accuracy and resolution
2) Thermal drift of the JFET
3) Drift of your multimeter

To observe (2), put a low Idss JFET (say 7mA) on as described in #11251 for 5 minutes, and observe how the reading goes up and down, then breath on it, or just move your body away from the bench. If you have a 4~5 digit multimeter, you can see these effects quite clearly.

Single point methods only give you match at one point. See :

xen-audio
xen-audio


Patrick

[erpiii]

Originally Posted by AndrewT

As a first, test the jFETs for Idss.

Using the method 6L6 supplied, I get 9.15 volts (new 9V batt voltage)... which I'm guessing means they are bad and I need new ones.

erpiii

[EUVL]

If you are using the method described in post #11251, then I do not see a 1k resistor, across which the multimeter should be connected to measure voltage.


Patrick

[Tazzz]

Have anybody looked into using a soundcard to make a simple curve tracer for this small signal stuff?

LR out could be used to sweep Vg and Vd and LR in to monitor currents. (of course with proper signal conditioning in between)

Even tough most soundcards are AC coupled the measurement could be done around a "known" manually set bias point.