Tektronix Technical Forums are maintained by community involvement. Feel free to post questions or respond to questions by other members. Should you require a time-sensitive answer, please contact your local Tektronix support center here.

237 Output Voltage not as commanded

236, 237, 238 etc
jozereko
Posts: 10
Joined: February 13th, 2015, 11:00 am
Country: United States

237 Output Voltage not as commanded

Post by jozereko » February 13th, 2015, 11:14 am

I have been using the 237 to conduct current measurement sweeps and I began to notice anomalous readings - with positive voltage commanded the 237 reports a negative voltage (and the current sign does not flip when the voltage command does). I have it connected with a triax cable out of the OUTPUT HI port which is split out at the DUT end into three leads: the central core and two others (I inherited this setup and can't tell from inspection which shield goes to which lead).

I replaced the DUT with a ~22 K-ohm resistor bank and ran some diagnostic trials. What I found is that regardless of the commanded voltage or of the method of commanding (locally or remotely via a LabView GUI) the voltage measured across the DUT with a Fluke is -180 or so and the recorded current approximately corresponds to that using Ohm's Law. I have tried multiple combinations of where to put the DUT leads - the results I get suggest no voltage at all is being applied across those combinations. I have attempted to use the OUTPUT LOW port because of the low (+/- 5 V) magnitude of my diagnostic commands, but this only causes even less current to be recorded - so small in magnitude that the noise actually has it crossing zero randomly.

At this point I think it's an equipment problem but it might still be user error. Can anyone help?
Attachments
red2black.jpg
Results from diagnostic run
red2black.jpg (47.42 KiB) Viewed 61030 times

Dale C
Keithley Applications
Keithley Applications
Posts: 2849
Joined: June 10th, 2010, 6:22 am
Country: United States
Contact:

Re: 237 Output Voltage not as commanded

Post by Dale C » February 13th, 2015, 12:51 pm

The output should be from the center conductor of the OUTPUT HI triax connection to output lo. Output LO is on the banana jack or the inner shield of the Sense LO triax connector.
I would suggest to use the OUTPUT LO to GROUND Link.
Then the output would be from the HI triax connection, inner conductor to outer shield.
If you have the triax to alligator clip cable it would be the red and green alligator clips.
Give that a try.

jozereko
Posts: 10
Joined: February 13th, 2015, 11:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by jozereko » February 13th, 2015, 1:37 pm

Thank you for the response.

The latter configuration (with the lo to ground short in place) is actually how we have been doing our tests. I tried the both configurations just now and got largely the same results: a command of +5 V DC on the box front produced a measured current (on the display) of about -5 mA and a measured voltage (with a Fluke) of about -130.

brian.d.smith
Keithley Applications
Keithley Applications
Posts: 402
Joined: June 10th, 2013, 6:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by brian.d.smith » February 16th, 2015, 6:02 pm

It seems that you are unsure of the integrity of your fixture. I would suggest dropping back to something that you absolutely know is giving you HI and LO to your device that is guarded all the way from the 237 to your device. Guard should be dropped just before the device.

I would use the triax to alligator cables. Do you have one or more of the triax to aligator cables available?

brian.d.smith
Keithley Applications
Keithley Applications
Posts: 402
Joined: June 10th, 2013, 6:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by brian.d.smith » February 16th, 2015, 6:22 pm

One additional question. How long since your 237 was calibrated?

jozereko
Posts: 10
Joined: February 13th, 2015, 11:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by jozereko » February 17th, 2015, 8:07 am

Not sure about the calibration question. I'll have to look into that. The test bench was set up in 2011, so the cal might be at many as four or five years old.

I believe my current setup does drop the guard right before the DUT. It uses a triax to bare lead cable - one end has a triax lug and the other end is split off but instead of alligator clips there are bare wires that are inserted into a connection box. (I am enclosing a - of the connection I have set up for my troubleshooting. The DUT here are the resistors connected in series using the red alligator clips and on the right is the end of the triax cable. The inner core is the lead that plugs in to the "top-right" port. Based on checks with the Fluke I am confident that I have isolated the circuit from the metal table despite the primitive appearance of the setup.)

For the time being, this is the only cable I have that is split in this way. I have other single-ended triax cables available but they would have to be split out.

One thing I haven't done is to apply the Fluke directly across the OUTPUT HI terminal (with nothing plugged in) to check the voltage there.

Thanks for all the help with this.
Attachments
20150205_133019.jpg
Connection and DUT
20150205_133019.jpg (401.68 KiB) Viewed 61006 times

brian.d.smith
Keithley Applications
Keithley Applications
Posts: 402
Joined: June 10th, 2013, 6:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by brian.d.smith » February 17th, 2015, 11:01 am

Is black the outer shield and green the inner shield?

brian.d.smith
Keithley Applications
Keithley Applications
Posts: 402
Joined: June 10th, 2013, 6:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by brian.d.smith » February 17th, 2015, 11:03 am

I would drop back from the Labview environment and use the front panel. Make sure that you are on a range that engages as many significant digits as possible.

jozereko
Posts: 10
Joined: February 13th, 2015, 11:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by jozereko » February 17th, 2015, 11:12 am

You have the colors correct.

Also, I have been using both the LabView script and the front panel to control the test with similar results. (I don't have plotted results from the front panel trials, though.)

brian.d.smith
Keithley Applications
Keithley Applications
Posts: 402
Joined: June 10th, 2013, 6:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by brian.d.smith » February 19th, 2015, 8:33 am

I see some issues unrelated to the 237

1. 180V on your bench with an alligator clip is a shock hazard. I recommend using no more than 10V in such an unsecured area
2. 180 v into 22k ohms will create ~ 1.5 watts. Those resistors do not look like power resistors.
3. The alligator clip looks fairly well oxidized. I recommend placing the Hi and LO closer together on your terminal strip, discarding the alligator test lead, using one resistor connected between Hi and LO on the terminal strip.

Have you connected The output LO to Ground strap on the back of the 237?

Are you sourcing V and measuring I or Sourcing I and measuring V?

jozereko
Posts: 10
Joined: February 13th, 2015, 11:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by jozereko » February 19th, 2015, 12:06 pm

I have tried two different configurations per prior advice. In the first, the center conductor of OUTPUT HI is supplying the resistor and current is returning via the inner shield of a cable hooked up to SENSE LO. In this setup the link is disconnected. In the other, which matches the arrangement I was using when I discovered the problem, the center conductor of OUTPUT HI supplies the DUT and the return is via the outer shield of the same cable. In this arrangement the link is made.

In all cases I am sourcing voltage / measuring current.

I don't disagree with the test setup issues, but safety aside for the moment I don't think they're the root cause since this setup is replicating the mismatch between commanded and measured voltage almost exactly. The actual DUT is hooked up using wires that connect to the junction box the same way as the triax leads - short bare leads tucked completely in and securely fastened with the screw.

To address the question of calibration, the last evidence of cal that I have is what appears to be a factory certificate from 15 March 2011.

brian.d.smith
Keithley Applications
Keithley Applications
Posts: 402
Joined: June 10th, 2013, 6:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by brian.d.smith » February 20th, 2015, 4:31 am

The reason I mention 1.5 watts is that if you over heat a resistor, it is not going to behave and you may damage the resistor.

What voltage are you sourcing?

Please drop back to using the front panel. Take an snapshot of the front panel while output is on. Post it here.

brian.d.smith
Keithley Applications
Keithley Applications
Posts: 402
Joined: June 10th, 2013, 6:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by brian.d.smith » February 20th, 2015, 4:56 am

Here is my 237. I am using a triax-to-alligator cable. The load is a 4.97K resistor. I am sourcing voltage and measuring current. At 1mA compliance, the instrument goes into compliance so I changed the current compliance to 10 mA.

Integration time is line cycle. The filter is set to 2 readings. Local sense is enabled.

The least significant digit is rock solid.
SourceV_MesureI 4_97K resistor.jpg
SourceV_MesureI 4_97K resistor.jpg (86.31 KiB) Viewed 60981 times

jozereko
Posts: 10
Joined: February 13th, 2015, 11:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by jozereko » February 20th, 2015, 9:54 am

Apart from the resistance, I ran my test again identically to yours. I will include lots of -. I confirmed the settings you mentioned to be the same except I needed a 10 mA compliance (even though I should not have).

As can be seen, the resistance is 22.88 k-ohm. Unless I severely affected the setup when re-attaching the cable, I assume the resistance remained the same.
resistance.jpg
Resistance check
resistance.jpg (460.91 KiB) Viewed 60975 times
I restored the same triax as before:
wiring.jpg
DUT with triax
wiring.jpg (83.17 KiB) Viewed 60975 times
The back panel:
backpanel.jpg
Back of the box as connected for the test
backpanel.jpg (118.04 KiB) Viewed 60975 times
(Continued next post)

jozereko
Posts: 10
Joined: February 13th, 2015, 11:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by jozereko » February 20th, 2015, 10:04 am

But the current remains the same as in previous tests, with not even the polarity corresponding to the command:
current1.jpg
current1.jpg (124.3 KiB) Viewed 60974 times
Nor was it nearly as consistent as yours:
current1.jpg
current1.jpg (124.3 KiB) Viewed 60974 times
Ohm's Law tells me to expect 0.2 mA for this test. The measured current suggests about -130 V across the resistors. I did not measure with the Fluke because I wanted to minimize the time under load.
Attachments
current2.jpg
current2.jpg (116.74 KiB) Viewed 60974 times

brian.d.smith
Keithley Applications
Keithley Applications
Posts: 402
Joined: June 10th, 2013, 6:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by brian.d.smith » February 20th, 2015, 12:29 pm

It appears that your current has a negative sign. Is it really a negative sign? If so, you are sinking current rather than sourcing current. Look at the snapshot I posted. Sourced voltage into a resistor has a + sign in front of measured current.

Turn off the output. Disconnect the cable from the back of the 237. Leave the voltage at 5V. Turn on the output. You should have 0A measured. If not then please post the snapshot. Turn off the output.

If the current is 0A (or very close to 0), re-connect your triax cable. Disconnect the resistor so that there is no load. Leave the voltage at 5V. Turn on the output. The current reading should be 0A. Post the results here.

jozereko
Posts: 10
Joined: February 13th, 2015, 11:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by jozereko » February 20th, 2015, 1:35 pm

It certainly seems like a negative sign to me, and it appears that way when I run the test using LabView, too.

As to whether it is sourcing or sinking current, this is the original problem: the 237 is being commanded to source +5V and by all measurements it is sourcing something completely different. The response is identical to commands given via the LabView script that had been working perfectly well in months past.

I will try the test you have suggested but probably won't be able to get to it right away.

brian.d.smith
Keithley Applications
Keithley Applications
Posts: 402
Joined: June 10th, 2013, 6:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by brian.d.smith » February 20th, 2015, 2:19 pm

jozereko wrote:It certainly seems like a negative sign to me, and it appears that way when I run the test using LabView, too.

As to whether it is sourcing or sinking current, this is the original problem: the 237 is being commanded to source +5V and by all measurements it is sourcing something completely different. The response is identical to commands given via the LabView script that had been working perfectly well in months past.

I will try the test you have suggested but probably won't be able to get to it right away.
The 237 is a 4-quadrant instrument. I can source and sink positive and negative current. In order to sink current, there has to be something external to the 237 forcing current into the 237. With your setup it should be impossible to obtain a negative current reading. We shall see what your next steps show.

jozereko
Posts: 10
Joined: February 13th, 2015, 11:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by jozereko » February 21st, 2015, 12:19 pm

I ran the test as you described it. To summarize the results:

First I re-ran the +5 V command with the resistors in place. The measured current was -5.8734 mA in my - but fluctuated pretty widely.
With the triax unplugged, the measured current was not exactly zero but it was on the order of +0.1 pA (10^-12). I have assumed in the past that sub-nanoamp current was below the noise floor, but this may have just been a bit of tribal knowledge.
With the triax plugged in but the resistors completely removed, the measured current rose to the order of +2 pA. In both cases the measurement was not steady.

- as requested:

Current across resistor bank:
resistors.jpg
Result with resistors in place
resistors.jpg (119.56 KiB) Viewed 60953 times
Current across bare triax lug:
no triax.jpg
Result with the triax unplugged
no triax.jpg (128.29 KiB) Viewed 60953 times
Current across leads of the triax cable:
triax no resistor.jpg
Result with triax plugged in but no resistors in place
triax no resistor.jpg (129.54 KiB) Viewed 60953 times
(Continued in next post)

jozereko
Posts: 10
Joined: February 13th, 2015, 11:00 am
Country: United States

Re: 237 Output Voltage not as commanded

Post by jozereko » February 21st, 2015, 12:31 pm

I did some further checks that may be of use (or they might be red herrings). I removed the resistor bank and put the leads of my Fluke in its place. With the same +5 V commanded and the OPERATE button pressed, the Fluke measured -185.5 V. When returned to STBY, the voltage remained the same across the Fluke leads. Turning the 237's power off completely cleared the voltage (0.00 V), and it remained at zero when the power was restored - until pressing OPERATE again, at which point measured voltage went to -187.8 V.

This effect of the voltage remaining in place may be an artifact of the Fluke, but it was concerning.
Attachments
residual voltage.jpg
Measured voltage after returning to STBY
residual voltage.jpg (342.33 KiB) Viewed 60953 times

Post Reply

Return to “Other Source Measurement Units (SMU)”

Who is online

Users browsing this forum: No registered users and 1 guest