Help with DMM: Swapping Probes Gives Different Measurements

[luvdunhill]

so, here's my circuit. LED CCS with a pot (pot + series resistance, with another resistor in parallel with the above) to set the current through a JFET. This times 5. Yay, it's working. But, depending on which probe (black / red) is on the gate and which is on the source, I get different Vgs readings My meter is a B+K 2706A. The amounts are off quite a bit, ~200mV to ~600mV for example. I tried using a Cen-Tech cheapie and it did the same thing. Any ideas, besides get a bench-top DMM?

[jackinnj]

the best probes are no probes at all and cheapo probes purchased on EBay (of which I have, mea culpa, purchased many) are worthless for precision work.

Stick with Pomona.

On the hp/agilent Yahoo group a couple folks were discussing probes, solder, thermal gradients, etc. Real serious probes come from Keithley -- no solder, crimp or screw terminals, so no thermals. if you have to measure the nanovolts this is the way to go.

[luvdunhill]

Originally posted by jackinnj
the best probes are no probes at all and cheapo probes purchased on EBay (of which I have, mea culpa, purchased many) are worthless for precision work.

Stick with Pomona.

On the hp/agilent Yahoo group a couple folks were discussing probes, solder, thermal gradients, etc. Real serious probes come from Keithley -- no solder, crimp or screw terminals, so no thermals. if you have to measure the nanovolts this is the way to go.

so, the issue could be the probes themselves, not the meter? Seems strange to me. This really isn't precision measurement, I mean at 6mA through a 2SK170 with Vds being around 9v, I should get 170mV or so for Vgs, not exactly nanovolts. Could something else be the issue?

[Mooly]

Is your CCS stable -- not doing things it shouldn't 'scope it !
Meter probes can be a problem. The probes can look clean and yet have a "tarnish" that prevents proper contact.
If in doubt rig two resistors up across a battery to give say 600 mv and see if it reliably reads this OK.
The fact you have tried another meter points to instability. DVM's sometimes have quite a bit of low level "hash" across the open circuit probes, is this tipping your CCS over the edge ?

[luvdunhill]

Originally posted by Mooly
Is your CCS stable -- not doing things it shouldn't 'scope it !
Meter probes can be a problem. The probes can look clean and yet have a "tarnish" that prevents proper contact.
If in doubt rig two resistors up across a battery to give say 600 mv and see if it reliably reads this OK.
The fact you have tried another meter points to instability. DVM's sometimes have quite a bit of low level "hash" across the open circuit probes, is this tipping your CCS over the edge ?

ok, I can try the battery test to 100mV or so and see what happens. I don't see why this would be a problem, but I'll give it a go. If I have issues with this, perhaps I'll look at the Pomona 6213.

As for the CCS being stable, I just assumed so how would I go about looking into this with a 'scope? is there a common fix that I could also try? I'll add some small 10uF capacitors on each of the battery rails as well and see if that does anything.

[Mooly]

Hi,
If you have a scope, measure the point on the CCS you are trying to measure the voltage at. The 'scope will show if it's oscillating. Leaving the 'scope still connected now measure again with your DVM as you were trying to do before. If it's unstable the 'scope will show it when you couple the meter leads up. If it now seems OK for some reason it could still be unstable and the scope input capacitance that is helping to kill it.

[luvdunhill]

Originally posted by Mooly
Hi,
If you have a scope, measure the point on the CCS you are trying to measure the voltage at. The 'scope will show if it's oscillating. Leaving the 'scope still connected now measure again with your DVM as you were trying to do before. If it's unstable the 'scope will show it when you couple the meter leads up. If it now seems OK for some reason it could still be unstable and the scope input capacitance that is helping to kill it.

ok, I gave it a whirl. I setup the scope using the calibration and got nice 500mV 1kHz square waves. (though the tops were a bit thicker than the rest, and I couldn't narrow them) I set triggering to auto and touched the gate and source of the JFET and got a nice thick band on the screen (I used to have a probe ground clip, but couldn't find it, so I didn't use it). Same when I touched the battery rails. I tried to zoom in, but I didn't' really get much detail. I upped the probes to 10x and got a bit more resolution, but nothing really clear. it seems that the pattern does change a bit when i touch the DMM probes across the GS point.

What next?

[Mooly]

Are you sure you have a license to drive that thing
You have to have the 'scope ground connected to the ground in your circuit at all times.
When you say a " a nice thick band " on the screen, well that doesn't sound right at all. That's just what oscillation would look like at a low sweep speed
How about posting the circuit of the CCS

[jackinnj]

Originally posted by luvdunhill


I used to have a probe ground clip, but couldn't find it, so I didn't use it

Ground clips themselves can be a source of problems.

Use the manual trigger, widen the trace out with the time base such that you have a top and manually adjust the trigger control -- that fat band at the top of the square wave is actually a bunch of dancing asparas.

My TEK 2465 does this wonderfully.