Re: Curve Tracers

On Thu, 21 Nov 2024 00:00:10 -0000 (UTC), Cursitor Doom
<cd999666@notformail.com> wrote:

On Wed, 20 Nov 2024 12:27:04 -0800, john larkin wrote:
>

On Wed, 20 Nov 2024 18:00:34 -0000 (UTC), Cursitor Doom
<cd999666@notformail.com> wrote:
 

On Wed, 20 Nov 2024 07:40:08 -0800, john larkin wrote:
>

On Wed, 20 Nov 2024 11:32:32 -0000 (UTC), Cursitor Doom
<cd999666@notformail.com> wrote:
 

Gentlemen,
>
Curve tracers reveal useful info about the dynamic characteristics of
semiconductors and make designing for same much more predictable and
dependable than relying on spice models and simulation alone. But
they're typically rare beasts and expensive to come by and boat anchor
varieties are seriously heavy and bulky.
I think therefore that a curve tracer would make an excellent
project,
using the X&Y inputs of a scope as the display. Has anyone here
attempted this? I'd be interested to know what the main challenges are
likely to be.
>
-CD

 
I've considered it. It would be much more valuable if it traced
capacitances too.

>
A VNA is best for that. I've been using my big HP one this afternoon to
label some random bunches of caps and inductors whose values I couldn't
read. Trouble is, there doesn't seem to be much agreement on what
frequency to test these devices at. I've heard 100khz, 1Mhz, 10Mhz and
100Mhz mentioned from different sources. I'm pretty sure the 100Mhz
testing is for RF specified devices only, though. From what I've seen,
there's no reliable alternative to testing each batch for oneself,
because (certainly with ancient NOS leaded (as in non-SMD) stock) you
have no idea what frequency the factory tested them at and different
manufacturers in different countries at different times used different
methods!

 
What I want is C-V curves. I guess a VNA can do that with a some bias
tees and various power supplies.

>
If you only want C/V curves, what's wrong with a plain 'scope and a pulse
generator?

How would you do that?

>
>

The display should be on a computer and the data archived.

>
For people such as yourself, certainly. For me as a hobbyist, a CRT is
fine.

 
You have to read the analog screen and write down numbers. Or take a
picture.

>
It would only be a (slight) issue if I needed to share an image with a
third party. Aside from that, there's nothing I couldn't live with. You're
obviously requiring more than that, though.
>

I often just set up a breadboard and test parts. Last week I blew out
some power mosfet gates. Some have internal zeners; this one doesn't.

>
I can kill a MOSFET at 50' just by looking at it (certainly in winter).
In fact I shock myself from all the static I build up and it ain't
funny. Having dry skin is great for when I touch HV by accident (happens
quite a lot) but the flip side is I accumulate and hold static charge
like no one else I know.

 
Some fets have protective gate zeners. They typically clamp at +-40
volts.
 
Here in San Francisco, we never get static zapped. High humidity.

>
So you never need to wear a wrist/earth strap? That's a big plus. I find
them *so* restrictive and irritating when they (invariably) catch on
something.
>

Most mosfets drain avalanche, but the voltage is never specified, in
fact deliberately hidden. I have to measure that.  A good curve tracer
should measure that too.
 
Currents should go down to picoamps.
 
The fact that there aren't many curve tracers for sale suggests a
small market.

>
But a market, though small perhaps, there must be.

 
Sure. It would be a fun project.

>
And with thanks to Trump's tariffs, you might even make a few buxx out of
it, too.

Is there going to be tariffs on USB capacitance-measuring curve
tracers?