Author Topic: DIY High Resistance bricks  (Read 4696 times)

0 Members and 1 Guest are viewing this topic.

Offline VintageNut

  • Frequent Contributor
  • **
  • Posts: 521
  • Country: 00
DIY High Resistance bricks
« on: January 24, 2017, 06:38:29 pm »
I decided to make two different high resistance bricks that will be similar to the Keithley 5155. I already own the resistors. They are vintage devices used for everyday low current measurements.

Below is the 100G device

Next the enclosure is a Hammond 1590A

There will be triax connectors on each end.

A teflon tube will cover the entire resistor and its leads.

A metal guard tube will cover the teflon tubing. The guard tube will be soldered to the triax guard terminal.

The plan is to build a 10G and a 100G brick to be calibrated in February when I travel to the cal lab.
« Last Edit: January 24, 2017, 06:40:10 pm by VintageNut »
working instruments :Keithley 260,261,2750,7708, 2000 (calibrated), 2015, 236, 237, 238, 147, 220,  Rigol DG1032  PAR Model 128 Lock-In amplifier, Fluke 332A, Gen Res 4107 KVD, 4107D KVD, Fluke 731B X2 (calibrated), Fluke 5450A (calibrated)
 

Online RoGeorge

  • Super Contributor
  • ***
  • Posts: 1998
  • Country: ro
Re: DIY High Resistance bricks
« Reply #1 on: January 24, 2017, 07:04:39 pm »
Is it safe to use teflon?
I have no experience with such etalon devices, but I am asking because of this famous story from Bob Pease: http://electronicdesign.com/test-amp-measurement/whats-all-teflon-stuff-anyhow

https://www.youtube.com/watch?v=B4G3YPlO6Wg?t=13m4s

Offline VintageNut

  • Frequent Contributor
  • **
  • Posts: 521
  • Country: 00
Re: DIY High Resistance bricks
« Reply #2 on: January 24, 2017, 10:12:06 pm »
Thanks for the links. I just followed the Bob P link and read his article. In the Bob P stories, the teflon is placed where a charge can build up between HI and LO of the measurement. It makes sense that this can be a problem.

In my DIY guarded resistor assembly, the teflon will be between HI and GUARD which are at the same potential with a small offset caused by the input offset of the op amp that drives the GUARD signal.

There should be no problem with charge buildup since there will be no potential across the teflon. Also, there will be no movement of the teflon. The teflon will be trapped between the resistor and the guard tube.

We shall see. This will be a device that gets calibrated regularly  and will be used with calibrated instruments. Any issues should be identifiable.

I am contemplating where the guard tube should be and should not be. In particular I am wondering if the guard tube should only cover the resistor lead that will be attached to HI.



Is it safe to use teflon?
I have no experience with such etalon devices, but I am asking because of this famous story from Bob Pease: http://electronicdesign.com/test-amp-measurement/whats-all-teflon-stuff-anyhow

https://www.youtube.com/watch?v=B4G3YPlO6Wg?t=13m4s
working instruments :Keithley 260,261,2750,7708, 2000 (calibrated), 2015, 236, 237, 238, 147, 220,  Rigol DG1032  PAR Model 128 Lock-In amplifier, Fluke 332A, Gen Res 4107 KVD, 4107D KVD, Fluke 731B X2 (calibrated), Fluke 5450A (calibrated)
 

Online RoGeorge

  • Super Contributor
  • ***
  • Posts: 1998
  • Country: ro
Re: DIY High Resistance bricks
« Reply #3 on: January 24, 2017, 10:48:30 pm »
Thank you for taking the time to explain all this.

Offline Alex Nikitin

  • Super Contributor
  • ***
  • Posts: 1021
  • Country: gb
  • Femtoampnut and Tapehead.
    • A.N.T. Audio
Re: DIY High Resistance bricks
« Reply #4 on: January 24, 2017, 11:51:09 pm »
1) I would not recommend a Teflon tubing over the resistor itself. It would do more harm than good.

2) I am not sure an internal guard would be useful. You need to apply a voltage across the resistor to use it (and to measure it).

3) These resistors (like the 100G on the photo) have a very high  temperature coefficient (~800ppm/C according to Keithley) and a fairly high voltage coefficient as well. You can probably get about 0.1% accuracy at best if you have it measured at least in a couple of temperature points and apply a correction for the actual temperature, providing the voltage across the resistor does not exceed 10-20V.

Cheers

Alex
« Last Edit: January 24, 2017, 11:56:23 pm by Alex Nikitin »
 

Offline VintageNut

  • Frequent Contributor
  • **
  • Posts: 521
  • Country: 00
Re: DIY High Resistance bricks
« Reply #5 on: January 25, 2017, 12:19:46 am »
From a cursory inspection, it looks like this is the same resistor that is in my 263. On the lowest current range, the 263 uncertainty is 0.4% of full scale. That would be in keeping with what you are stating about the resistor.

Do you know of a better resistor solution for 100G?

In my experience, for high resistance, guarding is required. If you do not guard, there is a leakage from HI to LO in the air. What you want is all of the air around the resistor to be at HI potential. That discourages any leakage from HI to LO in the air. This is how guarded triax cable operates except the dielectric is not air.

1) I would not recommend a Teflon tubing over the resistor itself. It would do more harm than good.

2) I am not sure an internal guard would be useful. You need to apply a voltage across the resistor to use it (and to measure it).

3) These resistors (like the 100G on the photo) have a very high  temperature coefficient (~800ppm/C according to Keithley) and a fairly high voltage coefficient as well. You can probably get about 0.1% accuracy at best if you have it measured at least in a couple of temperature points and apply a correction for the actual temperature, providing the voltage across the resistor does not exceed 10-20V.

Cheers

Alex
« Last Edit: January 25, 2017, 03:02:48 am by VintageNut »
working instruments :Keithley 260,261,2750,7708, 2000 (calibrated), 2015, 236, 237, 238, 147, 220,  Rigol DG1032  PAR Model 128 Lock-In amplifier, Fluke 332A, Gen Res 4107 KVD, 4107D KVD, Fluke 731B X2 (calibrated), Fluke 5450A (calibrated)
 

Online TiN

  • Super Contributor
  • ***
  • Posts: 4113
  • Country: us
  • xDevs.com/live - 24/7 lab feed
    • xDevs.com
Re: DIY High Resistance bricks
« Reply #6 on: January 25, 2017, 04:14:16 am »
I think Alex meant that you can have Guard on chassis box itself and outer side of regular BNC. Since chassis is enclosed, it will be essentially same minus little LO terminal stub from resistor to LO BNC port.
YouTube | Metrology IRC Chat room | Live-cam | Share T&M documentation? Upload! No MB limit, firmwares, photos.
 

Offline Alex Nikitin

  • Super Contributor
  • ***
  • Posts: 1021
  • Country: gb
  • Femtoampnut and Tapehead.
    • A.N.T. Audio
Re: DIY High Resistance bricks
« Reply #7 on: January 25, 2017, 08:08:06 am »
From a cursory inspection, it looks like this is the same resistor that is in my 263. On the lowest current range, the 263 uncertainty is 0.4% of full scale. That would be in keeping with what you are stating about the resistor.

Do you know of a better resistor solution for 100G?

In my experience, for high resistance, guarding is required. If you do not guard, there is a leakage from HI to LO in the air. What you want is all of the air around the resistor to be at HI potential. That discourages any leakage from HI to LO in the air. This is how guarded triax cable operates except the dielectric is not air.

A somewhat better resistor is a glass encapsulated version - RX series from Ohmite (which I am using in my high resistance box). Expensive though. The best would be to assemble 20 of 5G 100ppm/C resistors in series using Teflon supports - even more expensive!

Cheers

Alex
 

Offline daqq

  • Super Contributor
  • ***
  • Posts: 1690
  • Country: sk
    • My site
Re: DIY High Resistance bricks
« Reply #8 on: January 25, 2017, 08:46:32 am »
Quote
These resistors (like the 100G on the photo) have a very high  temperature coefficient (~800ppm/C according to Keithley) and a fairly high voltage coefficient as well.
Just thinking - would it be possible to build a thermostated resistor if you can't get around the temp. co? Maybe have it heated to, say, 60 deg.C and calibrate it at that temperature? Aside from a little extra noise, would there be any negative effects?
Believe it or not, pointy haired people do exist!
+++Divide By Cucumber Error. Please Reinstall Universe And Reboot +++
 

Offline VintageNut

  • Frequent Contributor
  • **
  • Posts: 521
  • Country: 00
Re: DIY High Resistance bricks
« Reply #9 on: January 25, 2017, 01:15:47 pm »
I think Alex meant that you can have Guard on chassis box itself and outer side of regular BNC. Since chassis is enclosed, it will be essentially same minus little LO terminal stub from resistor to LO BNC port.

This is a good trick. It is only safe if the guard box is enclosed in shield/LO/GND box that prevents a user from touching the guard box. A busy calibration lab would not want to bother with an unsafe device.

My proposed guard tube is the micro-miniature evolution of the guard box idea.
working instruments :Keithley 260,261,2750,7708, 2000 (calibrated), 2015, 236, 237, 238, 147, 220,  Rigol DG1032  PAR Model 128 Lock-In amplifier, Fluke 332A, Gen Res 4107 KVD, 4107D KVD, Fluke 731B X2 (calibrated), Fluke 5450A (calibrated)
 

Offline Alex Nikitin

  • Super Contributor
  • ***
  • Posts: 1021
  • Country: gb
  • Femtoampnut and Tapehead.
    • A.N.T. Audio
Re: DIY High Resistance bricks
« Reply #10 on: January 25, 2017, 02:58:23 pm »
1) Sorry, I can not see any reason to use a guarded enclosure on a resistor. A grounded enclosure is enough to screen from the interference and stray fields and the parasitic capacitance reduction is negligible (unlike for a guarded cable where the difference is huge) and might be even counterproductive if the voltage across the resistor is reasonably high. I use a triaxial socket now on one end of a 100G resistor so a guarded cable can be used, but no internal guard.

2) A temperature stabilised at 50-60C enclosure would be a good idea especially because it would also keep moisture out, always a good thing in high impedance circuits. It would make the box a lot more expensive though.

Cheers

Alex
 

Offline Neilm

  • Super Contributor
  • ***
  • Posts: 1441
  • Country: gb
Re: DIY High Resistance bricks
« Reply #11 on: January 25, 2017, 07:00:50 pm »
I would not bother with the tubes around the resistor. Unless you clean the teflon VERY well you might very well find that it provides a significant leakage path. That resistor looks as if it is designed to be submerged in oil. These are usually designed to be cleaned very well then the oil provides the insulation. Welwyn do some thta are good to 40kV if submerged in good quality oil.

You don't say what voltage you are planning on applying, but if it is not too high I would just put a few turns of wire around the middle and connect to the guard terminal. If you are planning to put high voltage on it then I would get a resistor with better coating and use that.
Two things are infinite: the universe and human stupidity; and I'm not sure about the the universe. - Albert Einstein
Tesla referral code https://ts.la/neil53539
 

Offline VintageNut

  • Frequent Contributor
  • **
  • Posts: 521
  • Country: 00
Re: DIY High Resistance bricks
« Reply #12 on: January 25, 2017, 10:33:48 pm »
I would not bother with the tubes around the resistor. Unless you clean the teflon VERY well you might very well find that it provides a significant leakage path. That resistor looks as if it is designed to be submerged in oil. These are usually designed to be cleaned very well then the oil provides the insulation. Welwyn do some thta are good to 40kV if submerged in good quality oil.

You don't say what voltage you are planning on applying, but if it is not too high I would just put a few turns of wire around the middle and connect to the guard terminal. If you are planning to put high voltage on it then I would get a resistor with better coating and use that.

Typically, I work at 40V or less for safety reasons. The idea is for the teflon to be between the resistor and the guard tube. The potential between HI and GUARD is nearly zero; just the offset of the opamp that drives the guard. While any impurities on the teflon maybe a conduction path, no current can flow since no potential exists.

It is good idea to clean the teflon well. I will make sure to do that.
working instruments :Keithley 260,261,2750,7708, 2000 (calibrated), 2015, 236, 237, 238, 147, 220,  Rigol DG1032  PAR Model 128 Lock-In amplifier, Fluke 332A, Gen Res 4107 KVD, 4107D KVD, Fluke 731B X2 (calibrated), Fluke 5450A (calibrated)
 

Offline Alex Nikitin

  • Super Contributor
  • ***
  • Posts: 1021
  • Country: gb
  • Femtoampnut and Tapehead.
    • A.N.T. Audio
Re: DIY High Resistance bricks
« Reply #13 on: January 26, 2017, 10:39:16 am »
As a side note - there nothing magical about guarding. It works well if the potential difference between the guard and whatever is guarded is close to zero. It does not work well if it isn't. If you have 40V across a resistor,taking one end as zero volts, you'll have 40V at the other end and only 20V in the middle of the resistor, so the guarding tube at that point already doing more harm than good, increasing capacitive coupling from the "hot" end to the "cold" end. A grounded tube would make more sense as it will reduce that coupling. A Teflon tube over the resistor, as already noted, is essentially asking for trouble, unless there is a very specific reason to use it , for example in a very high voltage circuit where the air is no longer sufficient as an isolation.

Cheers

Alex

P.S. - by the way,  the Keithley 5155 series uses only plain BNC connectors, no triax, no separate guard 8) .

P.P.S. - for the best of both worlds, a distributed guard is possible if the main resistor is constructed from many smaller value resistors and a sectioned guard is driven from a low-ish resistance auxiliary divider, so the guard voltage in each section is close to the voltage over each resistor of the main divider. Not really applicable unless you have at least ten resistors in series.
« Last Edit: January 26, 2017, 01:20:36 pm by Alex Nikitin »
 

Offline VintageNut

  • Frequent Contributor
  • **
  • Posts: 521
  • Country: 00
Re: DIY High Resistance bricks
« Reply #14 on: January 26, 2017, 04:57:36 pm »
Hi Alex

Thanks for the discussion. The teflon tube is just a spacer for the guard tube. Leakage can only happen on the portion of the resistor that is not the HI lead. If the current is measured only at the input to the HI lead, then any leakage (from GUARD) into parts of the resistor that are not HI will not influence current measurement.

It could be argued that the guard is not needed around those portions of the resistor that are not at HI potential.

My opinion concerning the 5155 resistors' construction is that they would benefit from either having the enclosure at guard potential or some other guard arrangement. My experience with measuring high ohm resistors is that they behave differently when guarded then when not guarded. I think that the difference in behavior is leakage from HI to LO in the air. Just my opinion.




As a side note - there nothing magical about guarding. It works well if the potential difference between the guard and whatever is guarded is close to zero. It does not work well if it isn't. If you have 40V across a resistor,taking one end as zero volts, you'll have 40V at the other end and only 20V in the middle of the resistor, so the guarding tube at that point already doing more harm than good, increasing capacitive coupling from the "hot" end to the "cold" end. A grounded tube would make more sense as it will reduce that coupling. A Teflon tube over the resistor, as already noted, is essentially asking for trouble, unless there is a very specific reason to use it , for example in a very high voltage circuit where the air is no longer sufficient as an isolation.

Cheers

Alex

P.S. - by the way,  the Keithley 5155 series uses only plain BNC connectors, no triax, no separate guard 8) .

P.P.S. - for the best of both worlds, a distributed guard is possible if the main resistor is constructed from many smaller value resistors and a sectioned guard is driven from a low-ish resistance auxiliary divider, so the guard voltage in each section is close to the voltage over each resistor of the main divider. Not really applicable unless you have at least ten resistors in series.
working instruments :Keithley 260,261,2750,7708, 2000 (calibrated), 2015, 236, 237, 238, 147, 220,  Rigol DG1032  PAR Model 128 Lock-In amplifier, Fluke 332A, Gen Res 4107 KVD, 4107D KVD, Fluke 731B X2 (calibrated), Fluke 5450A (calibrated)
 

Offline VintageNut

  • Frequent Contributor
  • **
  • Posts: 521
  • Country: 00
Re: DIY High Resistance bricks
« Reply #15 on: February 19, 2017, 03:40:52 am »
Here are two of the three bricks that I assembled. They are 100M, 10G and 100G. These are the resistors that I had acquired from a stash of gear from a private collection.

The 10G is connected to my KE487 at the moment.

The 10G has a guard installed. The other two do not. I will see how they behave over time.

One end of each assembly is triax. The other end is BNC/coax. The intent is to have the voltage source connected to the coax end and the picoammeter to the guard side.

The assemblies will be calibrated in March. Looking forward to using these resistors to help calibrate my picoammeter and my 23X SMUs.
working instruments :Keithley 260,261,2750,7708, 2000 (calibrated), 2015, 236, 237, 238, 147, 220,  Rigol DG1032  PAR Model 128 Lock-In amplifier, Fluke 332A, Gen Res 4107 KVD, 4107D KVD, Fluke 731B X2 (calibrated), Fluke 5450A (calibrated)
 

Offline VintageNut

  • Frequent Contributor
  • **
  • Posts: 521
  • Country: 00
Re: DIY High Resistance bricks
« Reply #16 on: February 19, 2017, 04:04:08 am »
Here is the 10G resistor being measured on the 487. Voltage is adjusted for as close to 1nA as possible. The display reads 0.99999 nA. The resistance is the voltage setting of 9.970 V for a resistance of 9.970G.

Some display segments appear dim but that is a photography/persistence artifact. The display is bright and clear.

working instruments :Keithley 260,261,2750,7708, 2000 (calibrated), 2015, 236, 237, 238, 147, 220,  Rigol DG1032  PAR Model 128 Lock-In amplifier, Fluke 332A, Gen Res 4107 KVD, 4107D KVD, Fluke 731B X2 (calibrated), Fluke 5450A (calibrated)
 

Offline VintageNut

  • Frequent Contributor
  • **
  • Posts: 521
  • Country: 00
Re: DIY High Resistance bricks
« Reply #17 on: March 03, 2017, 12:49:32 am »
Here are the measurements of the high resistance bricks. These will be good enough to calibrate my picoammeter.

working instruments :Keithley 260,261,2750,7708, 2000 (calibrated), 2015, 236, 237, 238, 147, 220,  Rigol DG1032  PAR Model 128 Lock-In amplifier, Fluke 332A, Gen Res 4107 KVD, 4107D KVD, Fluke 731B X2 (calibrated), Fluke 5450A (calibrated)
 

Offline guenthert

  • Frequent Contributor
  • **
  • Posts: 334
  • Country: us
Re: DIY High Resistance bricks
« Reply #18 on: March 03, 2017, 01:08:26 am »
Am I to understand that you had self-made resistors professionally calibrated (characterized)?  Could you share which lab offers such?  (I'd like to know how much for such a service is charged, but the labs seem to not willing to share their rates and rather negotiate individually, which I find objectionable)
 

Offline VintageNut

  • Frequent Contributor
  • **
  • Posts: 521
  • Country: 00
Re: DIY High Resistance bricks
« Reply #19 on: March 03, 2017, 01:48:04 am »
Am I to understand that you had self-made resistors professionally calibrated (characterized)?  Could you share which lab offers such?  (I'd like to know how much for such a service is charged, but the labs seem to not willing to share their rates and rather negotiate individually, which I find objectionable)

This lab is not public. They are a captive primary standards lab for an organization that provides service and calibration. Like other members of this forum, I have access to such a lab as part of my day job but for me only once or twice a year. The operators of the lab kindly perform some calibrations for me. What they do for me would cost more than I could afford.
working instruments :Keithley 260,261,2750,7708, 2000 (calibrated), 2015, 236, 237, 238, 147, 220,  Rigol DG1032  PAR Model 128 Lock-In amplifier, Fluke 332A, Gen Res 4107 KVD, 4107D KVD, Fluke 731B X2 (calibrated), Fluke 5450A (calibrated)
 
The following users thanked this post: guenthert

Offline daqq

  • Super Contributor
  • ***
  • Posts: 1690
  • Country: sk
    • My site
Re: DIY High Resistance bricks
« Reply #20 on: March 04, 2017, 06:31:30 am »
Quote
Am I to understand that you had self-made resistors professionally calibrated (characterized)?  Could you share which lab offers such?
Any metrological lab would probably do it, you'd just have to pay for it. I'm looking at my local state metrology lab, it would probably cost around around 400 EUR.
Believe it or not, pointy haired people do exist!
+++Divide By Cucumber Error. Please Reinstall Universe And Reboot +++
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf