Aion Fx Ares (Clinch EP Pre Clone) JFET Testing

[Big Monk]

Building the Ares according to this:

https://aionfx.com/app/files/docs/ares_documentation.pdf

Only issue is none of the JFETs are matching the values in the document using my TC1 component tester.

Maybe the test voltage? It uses 4.5 vDC if I'm not mistaken. I have 10 of each to choose from. They are all consistent but all low with respect to the build doc.

[Rob Strand]

Like the Aion doc mentions, we don't know the range of allowed values.
Also I could not see the expected drain voltages in the document.
You could work those out using calculations or with spice (you need to make a custom JFET model with
the Aion JFET parametrers).

One issue is those testers (DCA-75 and TC-1) test Vgs_off at a finite current and can produce low Vgs_off values.  I don't think the testers attempt to calculate the true Vgs_off from the measured "Vgs_off" value and Idss.   From the test currents I've seen that should only be upto about 5% error.

Keep in mind the Aion values have such an Vgs_off error built in since they are measured on a DCA-75.

If you want a second opinion for the FET parameters you could try using the ROG test method to compare the values.

However, that should show you the two testers produce reasonably close values.   That leaves you with the problems I mentioned in the first paragraph.

Unfortunately JFET parameters are inherently variable and the only way to narrow down the range is to have a known supplier, buy a lot of parts, and sort them.   Not a practical scheme for a one off build.   Note also there is correlation between parts in the batch.

If your JFET parameters are off you need a plan B.  You need to tweak the circuit to get the JFET drain voltage in the desired range. It is likely you could adjust the drain resistors - provided the adjustments aren't radical.    Radical adjustments will work but then you don't know how close it sounds to the original.

[Big Monk]

I set up an R.G. tester and found devices that match the targets from the Aion build document.

I think I'm good!

[Rob Strand]

I set up an R.G. tester and found devices that match the targets from the Aion build document.

I think I'm good!

With the standard 10k test resistor (Rset not the 2x10k bias resistors), RG's tester produces Vgs_off values that are low in magnitude.   It has the same problem as the commercial testers only more pronouced.  If you change the test resistor to 10M you should see a Vgs_off values closer to reality.

http://www.geofex.com/Article_Folders/fetmatch/fetmatch.htm


FWIW, if you need it here's what I was talking about before.

[Big Monk]

I set up an R.G. tester and found devices that match the targets from the Aion build document.

I think I'm good!

With the standard 10k test resistor (not the 2x10k bias resistors), RG's tester produces Vgs_off values that are low in magnitude.   It has the same problem as the commercial testers only more pronouced.  If you change the test resistor to 10M you should see a Vgs_off values closer to reality.


FWIW, if you need it here's what I was talking about before.

My initial concern was how far off the Aion doc the values were. In the end, I'm not terribly concerned about actuals, but more concerned that I use a device close to what Aion measured.

I as unaware my TC1 would give such values. I'm glad I reached out and then put the tester together.

[Rob Strand]

I as unaware my TC1 would give such values. I'm glad I reached out and then put the tester together.

I guess there's a difference between "testing" a part is OK and "measuring" parameters.   The problems come about
when the displayed values on the tester give the impression they are measuring parameters.  (Having said that
they shouldn't be far off.)

It's certainly possible to take the values from the tester and turn them into better estimates for the parameters.

If the tester shows a measured Vgs_off at some test current IT  (maybe 5uA), then you can tweak the results.    Technically the measured Vgs_off displayed by the tester isn't Vgs_off at all but really only a Vgs measurement at IT.

     Vgs_off_estimate = Vgs_off_measured / (1 - sqrt( IT/IDSS) )

For example if IDSS = 2.76mA and IT = 5uA  then

   Vgs_off_estimate = Vgs_off_measured  * 1.044

And for RG's tester with Rset=10k,   IT = 4.5V / 10k = 450uA then

  Vgs_off_estimate = Vgs_off_measured  * 1.68
 

[Big Monk]

I as unaware my TC1 would give such values. I'm glad I reached out and then put the tester together.

I guess there's a difference between "testing" a part is OK and "measuring" parameters.   The problems come about
when the displayed values on the tester give the impression they are measuring parameters.  (Having said that
they shouldn't be far off.)

It's certainly possible to take the values from the tester and turn them into better estimates for the parameters.

If the tester shows a measured Vgs_off at some test current IT  (maybe 5uA), then you can tweak the results.    Technically the measured Vgs_off displayed by the tester isn't Vgs_off at all but really only a Vgs measurement at IT.

     Vgs_off_estimate = Vgs_off_measured / (1 - sqrt( IT/IDSS) )

For example if IDSS = 2.76mA and IT = 5uA  then

   Vgs_off_estimate = Vgs_off_measured  * 1.044

And for RG's tester with Rset=10k,   IT = 4.5V / 10k = 450uA then

  Vgs_off_estimate = Vgs_off_measured  * 1.68

This is good to know and I'll bookmark it. I don't often work with JFETs. Frankly, I've always been a bit behind on microelectronics. I trained as an electrical engineer but with a focus on "big wires", as I knew I was going into nuclear as a design engineer or eventually working in the utilities (I ended up doing both and am currently a protection engineer in the utilities).

Frankly, I'm always a little perplexed about the hard, academic science behind electronics, and more pointedly, how certain devices work and are modeled, etc.

I'm cool with estimates because with transistors in particular, I'm not too worried about exact values. I simply want to sort them into buckets and use my ears from there.

In this case, I don't think the Aion values in the build docs were hard and fast either for the required JFETs, but rather, a "mile marker" so that quick sorting would yield a suitable device for each position.

I'm sorry if I gave the impression that I was being flippant or ungrateful for your more detailed explanation.

[Rob Strand]

I trained as an electrical engineer but with a focus on "big wires", as I knew I was going into nuclear as a design engineer or eventually working in the utilities (I ended up doing both and am currently a protection engineer in the utilities).

Yes, power engineering has a language of its own.   I remember it well.  We had an old time lecturer who knew everything from the effects of dust on power lines to nuclear power.  He was a truly great lecturer who freely shared his knowledge.  (I still have his lecture notes.)

Frankly, I'm always a little perplexed about the hard, academic science behind electronics, and more pointedly, how certain devices work and are modeled, etc.

It's a tough subject to get traction on.  There's a lot of background knowledge you only get from semiconductor physics books.   Not many of those leave you with any feeling you actually know the subject.    It's a major undertaking to understand to the details of each model.

I'm cool with estimates because with transistors in particular, I'm not too worried about exact values. I simply want to sort them into buckets and use my ears from there.

It's handy to know where you stand and sorting the parts certainly helps.

In this case, I don't think the Aion values in the build docs were hard and fast either for the required JFETs, but rather, a "mile marker" so that quick sorting would yield a suitable device for each position.

Aion's pretty good at pinning down important points. 

In the section on JFET selection he mentions what one modern builder does.   It's very useful to have a marker like that on JFET circuits.  Even more so today where JFETs parameters seem deviate a lot more from the typical values than days of old.

[Eb7+9]

none of the JFETs are matching the values in the document using my TC1 component tester.

Try replacing the 22k and 15k drain loads by 50k trimpots, tweak those and swap-in your jFET's (and their relative position) until you find that magical EP-3 "it" ...