Resistance Standards (Leeds and Northrup)

[Rax]

I have the opportunity (this is from a friend) to buy some decade resistance standards, made by L&N, model numbers 4020, 4025 etc.

I don't know much about these units. How useful are these for performing calibration checks on instrumentation? What is their level as normative standards? If they're good enough, is there some specific methodology to be followed given their physical/electrical construction etc.?

Thank you for your input in advance.

[edpalmer42]

FYI

[Rax]

FYI

Interesting, and good to have, thank you.

[Wallace Gasiewicz]

Standards like this are great to have around when you need to calibrate something.
I wish I had them.

Even if the Cal date is old, they tend to be very accurate and quite good for non professional use.

[Rax]

I guess, to dial this in into my mind, how shall I think of these relative to, say, something like this: https://www.mouser.com/ProductDetail/Vishay-Precision-Group-Foil-Resistors/Y1453100R000V9L?qs=adduyb8LSTG0THZHLXzYtw%3D%3D?

Given the above Mouser part can be bought for about $30, offers .005% tolerance and .2ppm TC... So easily attainable as an electronic part. Not cheap, but hardly an investment.

How are they comparable?

[Andreas]

You cannot compare that.

A standard is usually designed to have maximum stability over life time.

A resistor in plastic package (epoxy) is
a) sensitive to humidity (Epoxy swells if air humidity increases -> value change)
b) stability over life time is lower

For a "standard" I would use at least hermetically sealed resistors.

With best regards

Andreas

[edpalmer42]

I would call the L&N 40XX resistors 'standards', L&N does too.  But they're not hermetically sealed.  You can open them, inspect, change the oil, etc.  Have the definitions changed to the point where they don't qualify as standards?

[alm]

I guess, to dial this in into my mind, how shall I think of these relative to, say, something like this: https://www.mouser.com/ProductDetail/Vishay-Precision-Group-Foil-Resistors/Y1453100R000V9L?qs=adduyb8LSTG0THZHLXzYtw%3D%3D?

Given the above Mouser part can be bought for about $30, offers .005% tolerance and .2ppm TC... So easily attainable as an electronic part. Not cheap, but hardly an investment.

How are they comparable?

One is a standard resistor, the other could be a component to make a standard resistor. You'd need an enclosure, (low thermal-EMF) binding posts, a way to mount it with a minimum extra stress, possibly testing since VPG is not exactly known for tight maximum specs, but rather overly-optimistic typical spec (there's a huge thread about measuring temperature coefficients on this forum).

The L&N will likely have a better long term drift, while the VHP part will have a lower temperature coefficient. The Rosa-type resistors were kept in temperature-controlled oil baths to stabilize the temperature. The VPG resistors are not hermetically sealed, so will be sensitive to moisture.

I personally prefer standard resistors with a lower temperature coefficient (like Fluke 742A), but they are much more expensive and much less available. The L&N resistors generally seem to be in the $100 ballpark.

I would call the L&N 40XX resistors 'standards', L&N does too.  But they're not hermetically sealed.  You can open them, inspect, change the oil, etc.  Have the definitions changed to the point where they don't qualify as standards?

Funny, Rosa in the initial publication on Rosa resistors called them sealed:

When the coil is finally adjusted, the case is nearly filled with pure oil that has been freed from moisture, and the top screwed firmly into place. To make the joint perfectly tight, shellac is usually put into the threads before screwing up.
[...]
Sealing the standards in this way possesses several distinct advantages. First and most important, of course, is the protection against the changes in resistance due to the absorption of moisture by the shellac (or varnish or other protective covering). [...] Hence the resistances are not only protected from the serious fluctuations from day to day, and the still greater seasonal changes due to moisture, but also from slower changes due to oxidation of the manganin and the possible sudden changes due to accident.

In figure 15 on page 429 Rosa shows data that the new resistors perform much better over seasons (humidity changes) than the previous open Reichsansalt resistors:


Now the oil is probably not as low moisture anymore: it can actually turn acidic due to oxidation and is recommended to replace it. And the shellac seal is probably no longer intact. But it was certainly designed as a hermetically sealed replacement of the previous Reichsansalt resistors. Here are two later papers that summarize the history of the Rosa resistors and its predecessors / successors:
The NBS Ohm - Past - Present - Future
Stability of Double-Walled Manganin Resistors

[Conrad Hoffman]

IMO, the L&N standards are as good as any normal voltnut needs, but you need a good thermometer for the well. Definitely change the oil- various posts on the L&N standard here if you search.

One caution- the value of the ohm changed by 495 ppm in 1948. If you have old resistors and have them measured/calibrated, the lab may try to convince you they're defective, when they're actually still close to the manufactured value.

[edpalmer42]

Interesting.  That Rosa paper crashes my Foxit PDF reader!  I tried a couple of different copies that I found online with similar results.  This copy: https://digital.library.unt.edu/ark:/67531/metadc66524/ works okay and it's easier to read because it has a white background.

You roused my curiosity so I dug out my L&N 4040-B resistor and opened it up.  It uses a rubber O-ring for a seal so I don't think it ever had Shellac in the threads to create a hermetic seal.  Other than that, the internal construction is very similar to the construction described in the Rosa paper.  I also have a Rubicon resistor that looks almost identical externally and internally to the resistors shown in the Rosa paper.  Not only was it not sealed when I got it, it actually leaked and the seller drained the oil out to prevent a shipping mess.  It also appears to use an O-ring for sealing - or it tried to! 

Maybe in the time between the Rosa paper (1908), and the Rubicon and L&N resistors (1940s?), better oils and sealants were developed that didn't absorb water?   

In the May 9, 1941 issue of Science, L&N has an ad for both N.B.S. type resistors (40XX models) and Reichsanstalt type resistors (42XX models).

Ed

[alm]

For a long time L&N sold Rosa-type resistors for larger values > 1 Ohm, and Reichsansalt-type resistors for < 1 Ohm. In an oil bath, the Reichsansalt-type has better heat dissipation because the oil can freely flow in and out of the resistor, rather than relying purely on heat conduction.

[TiN]

I've tested my L&N 4040-B while ago. It has rather poor temperature coefficient α = +8.15 ppm/°C and β = -0.57 ppm/°C². So if one to use it as is in non-temperature controlled environment and typical lab - it's good only up to 5.5 digits or so. These resistors really want oil bath to unveil their maximum performance.

Since Rax was looking at VPG, here's a quick and ugly build of standard out of VHP202ZT 9998 ohm resistor.



Just take hermetic VHP resistor, add random metal box, sprinkle some low thermal copper posts, print and crookedly attach label (gotta make it official!)



And don't forget to measure it



VHP202ZT box has typical linear temperature coefficient α = +0.468 ppm/°C and β = 0.002 ppm/°C². Most of VHPs have TCR alpha from 0.5-1.5 ppm/°C, very few are better. One would need to buy and test 20+ resistors to find one that is "typical 0.2ppm/°C".

[Rax]

Since Rax was looking at VPG, here's a quick and ugly build of standard out of VHP202ZT 9998 ohm resistor.

Resurrecting this thread. Illya - thank you for this info. Is there an article on your site detailing this build?

I'm still looking for a path forward and have no idea what my feasible options are. I am looking for a reliable way to verify DMMs to 6.5 digits with either an affordable commercial solution, or a DIY resistance standard of sorts.

[Villain]

Since Rax was looking at VPG, here's a quick and ugly build of standard out of VHP202ZT 9998 ohm resistor.

Resurrecting this thread. Illya - thank you for this info. Is there an article on your site detailing this build?

I'm still looking for a path forward and have no idea what my feasible options are. I am looking for a reliable way to verify DMMs to 6.5 digits with either an affordable commercial solution, or a DIY resistance standard of sorts.

For verifying 6.5 digit meter you dont need to put much effort at all. Buy one hermetically sealed VHP resistor per range you want to check and put it in a box with low emf binding posts. Often very high and very low ranges are not used a lot so maybe you can keep it to the most important.
Get it characterized buy a friend or someone on the forum once in a while, that will do.

[MrYakimovYA]

Buy one hermetically sealed VHP resistor per range you want to check and put it in a box with low emf binding posts.

Excuse me, what if one take non-low emf binding post and cover them with something to obtain temperature equilibrium? It takes some time but result must be the same as with low emp binding posts. Am I right?

[DavidKo]

Why no to do it properly, but I will choose the standard 3/4" distance between posts (in Fluke / IET front post configuration or in ESI SR1 configuration).

[MegaVolt]

Excuse me, what if one take non-low emf binding post and cover them with something to obtain temperature equilibrium? It takes some time but result must be the same as with low emp binding posts. Am I right?

If the pairs of materials are exactly the same.

[alligatorblues]

Since Rax was looking at VPG, here's a quick and ugly build of standard out of VHP202ZT 9998 ohm resistor.

Resurrecting this thread. Illya - thank you for this info. Is there an article on your site detailing this build?

I'm still looking for a path forward and have no idea what my feasible options are. I am looking for a reliable way to verify DMMs to 6.5 digits with either an affordable commercial solution, or a DIY resistance standard of sorts.

For verifying 6.5 digit meter you dont need to put much effort at all. Buy one hermetically sealed VHP resistor per range you want to check and put it in a box with low emf binding posts. Often very high and very low ranges are not used a lot so maybe you can keep it to the most important.
Get it characterized buy a friend or someone on the forum once in a while, that will do.

No matter how you slice it, you need to find the exact value of the new resistance standard.

[RandallMcRee]

Excuse me, what if one take non-low emf binding post and cover them with something to obtain temperature equilibrium? It takes some time but result must be the same as with low emp binding posts. Am I right?

If the pairs of materials are exactly the same.

Seebeck coefficient is defined as S= delta-V/delta-T. Thus as Mr. YakimovYA hints at, if the difference in temperature is zero the emf will be zero. Seems right to me. Now, I would not want to be the one claiming that any temperature difference is exactly zero since temperature gradients in modern test equipment are legion. But, there may be a time and place where one could say, yes we kept the difference to zero.
https://en.wikipedia.org/wiki/Seebeck_coefficient

[czgut]

May be the best way would be to use both of them: L&N as long term Standards and VHZ based resistors as working standards.

Long term stability of L&N 4000 series 1 Ohm ...10kOhm Resistor Standards is from +-few ppm/50years to +-100ppm/50 years.
Most of my L&N 40xx resistors (13 pcs) have  long term drift below 1ppm/year.

By characterizing temperature dependence of R  R(T) in 20'C..30'C you can get repeatability of R(T) about 1 ppm (0.6ppm std dev).
Due to their construction and oil filling, they are not good as AC Resistor Standards (above 0.1Hz).

I am using selected L&N 40xx as long term R standards (some have drift below 0.1 ppm ..0.3ppm/year),
and Vishay VHZ based  resistors as working standards (with low TCR and better AC characteristisc, compared to L&N standards).
As Primary Standard I am using L&N Thomas type 1 Ohm ( 0.05ppm/year) + Hamon transfers.