Teardown: Standard Resistors

[manganin]

Information about the Otto Wolff NG series:

The standard resistor NG 100 consists of a wire coil which is hermetically sealed between two brass tubes. A nickel-plated brass tube of diameter of 102mm and height of 170mm serves as a protecting device which is locked at the top by an ebonite cover. The sealed coil is connected to the outside terminal through glass. The ebonite cover as well as the terminals on it are of exactly the same design as those of the N 100. The escape of the heat being poorer with the model NG 100, the admissible load is lower than the case of the model N100 and corresponds to the values of the N 85. The model NG 100-2a has an accurary of 0.005 percent; the other models have an accurary of 0.002 percent. Loading in air must be restricted to 0.1 Watt.

[Vgkid]

@ Manganin how did you find this information, I looked a few years ago, and found nothing. Were these oil filled, or just air filled? When I opened it up, it smelled of stale gasoline, and noticed that the internal walls of the brass conainer appeared to have corrosion on them.

[Vgkid]

For those curious, if I killed it. Nope
Ran it on the 3456 over a day, didn't note the temps.
Min: 10.00055k, max:10.0076k, avg: 10.00068, var 4.21988E^-3.
i need to look into a cheap datalogging thermometer.

[manganin]

Manganin how did you find this information, I looked a few years ago, and found nothing.

I have a full Otto Wolff binder somewhere, but all I could find now was a poor one page photocopy which had that short description.

Were these oil filled, or just air filled? When I opened it up, it smelled of stale gasoline, and noticed that the internal walls of the brass conainer appeared to have corrosion on them.

Air filled. But a few decades in an oil bath...

The sealed version is quite rare, probably because it was more expensive and didn't offer much advantage over the open version. The brass cylinder was less rigid than the ceramic and the closed container actually amplified rather than attenuated the pressure coefficient. Same thing with the temperature hysteresis.

I took a photo of one of my N 100 resistors (100 ohm) wound on a ceramic cylinder. No hard laquer, but elastic wax instead which doesn't stress the wire as much.

[quarks]

does anyone know what is inside this "Zero Ohm Standard"





http://www.ohm-labs.com/resistance-standards/100-0-standard.html

and explain the
    BALANCED TETRAJUNCTION DESIGN
    A TRUE ZERO-OHM RESISTOR
   

[manganin]

does anyone know what is inside this "Zero Ohm Standard"

and explain the BALANCED TETRAJUNCTION DESIGN A TRUE ZERO-OHM RESISTOR

http://www.google.com/patents/US3252091

[zlymex]

does anyone know what is inside this "Zero Ohm Standard"





http://www.ohm-labs.com/resistance-standards/100-0-standard.html

and explain the
    BALANCED TETRAJUNCTION DESIGN
    A TRUE ZERO-OHM RESISTOR
   

It's better than those in patent US3252091, all of which are not fully symmetrical. That is, when any two terminals passing thru a current, the other two might have none-zero potential.
A regular tetrahedron will do the trick but is difficult to implement. Another one I know is attached.
 

[manganin]

It's better than those in patent US3252091, all of which are not fully symmetrical. That is, when any two terminals passing thru a current, the other two might have none-zero potential. A regular tetrahedron will do the trick but is difficult to implement. Another one I know is attached.

In real life the simplified versions developed by ESI provide excellent performance. With practical current levels and reasonable copper thickness the connector thermal EMFs are a bigger problem than the error caused by the current distribution. And the shape can be fine tuned afterwards based on the measurements to compensate the limited metalwork accurary.

[quarks]

http://www.google.com/patents/US3252091

Looks very much like inside my ESI SR1010s

[manganin]

Looks very much like inside my ESI SR1010s

And if I remember correctly, there is even a brief junction error analysis in the SR1010 manual.

[babysitter]

Yes, there is. Giving approx. numbers for the junctions, iirc in the 50-150µOhm range depending on the used "ports".

[zlymex]

It's better than those in patent US3252091, all of which are not fully symmetrical. That is, when any two terminals passing thru a current, the other two might have none-zero potential. A regular tetrahedron will do the trick but is difficult to implement. Another one I know is attached.

In real life the simplified versions developed by ESI provide excellent performance. With practical current levels and reasonable copper thickness the connector thermal EMFs are a bigger problem than the error caused by the current distribution. And the shape can be fine tuned afterwards based on the measurements to compensate the limited metalwork accurary.

I've never in doubt the performance of those ESI simplified versions within their Hamons. It is the universal zero ohm that people may connect 4 terminals in anyway which requires the use of full symmetrical version.

A 2D fully symmetrical version(electrical sense) can be made like this.

[quarks]

@zlymex,

the 4 hole symmetry makes very much sense, but I wonder why they made all posts inline and not just use the same pattern outside of the box. At least that would avoid cabeling inside.

Also is there a reason to go for the triangle shape of the plate?
Im asking because a round disc with the same 4 hole symmetry pattern would be my first idea to build it

Thanks
quarks

[zlymex]

@zlymex,

the 4 hole symmetry makes very much sense, but I wonder why they made all posts inline and not just use the same pattern outside of the box. At least that would avoid cabeling inside.

Also is there a reason to go for the triangle shape of the plate?
Im asking because a round disc with the same 4 hole symmetry pattern would be my first idea to build it

Thanks
quarks

I didn't quite get the question, did you mean why they use a metal box?

It just because I draw the triangle easy, it's thinner than a round disc and will save material. The thinnest would be down to just 4 wires. Be careful though, a cross won't do.

[quarks]

I didn't quite get the question, did you mean why they use a metal box?

No my question was not about the metal box but about the arrangement of the connectors.

It just because I draw the triangle easy, it's thinner than a round disc and will save material.

if the triangle plate is not for a technical reason, I think with my limited lathe/mill skills, it should be easier to use round stock and achieve near perfect symmetry (also there are no plates in CuTe but only round stock afaik).

[Kleinstein]

The triangle shape should be a little less sensitive on how well the wire is connected / soldered. It could also make soldering more easy, as heat is not conducted away so fast.
Still I see the advantages in machining for the round form.

[quarks]

The triangle shape should be a little less sensitive on how well the wire is connected / soldered. It could also make soldering more easy, as heat is not conducted away so fast.

when I build it, I plan not to use any wire or solder and directly screw on the binding posts

[zlymex]

The triangle shape should be a little less sensitive on how well the wire is connected / soldered. It could also make soldering more easy, as heat is not conducted away so fast.

when I build it, I plan not to use any wire or solder and directly screw on the binding posts

I made this purely by hand: hand cut, hand drill, hand soldering, without any measurement of geometry when make, symmetry done by eyeball observation only, no adjustment/trim of any kind afterwards.
Thickness of the triangle copper is 1.5mm, length of each side is aprox. 21mm, diameter of the copper bar is 2.2mm, diameter of the wires is 0.8mm.
When I run thru a current of 6.2A into 2 wires of any of the 6 combinations(excluding reversal), I measure the other two wires with less than +-0.3uV voltage variation with my 34401A.
This means that I have achieved the kind of zero that OHM-LABS claimed in their specification just by this simple DIY.
 

[Vgkid]

Before reading through the last few post I assumed that either a sphere, or diamond would be needed to make a zeroe ohm resistor.

[quarks]

I made this purely by hand: hand cut, hand drill, hand soldering, without any measurement of geometry when make, symmetry done by eyeball observation only, no adjustment/trim of any kind afterwards.
Thickness of the triangle copper is 1.5mm, length of each side is aprox. 21mm, diameter of the copper bar is 2.2mm, diameter of the wires is 0.8mm.
When I run thru a current of 6.2A into 2 wires of any of the 6 combinations(excluding reversal), I measure the other two wires with less than +-0.3uV voltage variation with my 34401A.
This means that I have achieved the kind of zero that OHM-LABS claimed in their specification just by this simple DIY.

my guess was, that a perfect symmetry is the most importand part, but if it is that easy, I am even more interested to see what is inside the Ohmlabs product

[zlymex]

I made this purely by hand: hand cut, hand drill, hand soldering, without any measurement of geometry when make, symmetry done by eyeball observation only, no adjustment/trim of any kind afterwards.
Thickness of the triangle copper is 1.5mm, length of each side is aprox. 21mm, diameter of the copper bar is 2.2mm, diameter of the wires is 0.8mm.
When I run thru a current of 6.2A into 2 wires of any of the 6 combinations(excluding reversal), I measure the other two wires with less than +-0.3uV voltage variation with my 34401A.
This means that I have achieved the kind of zero that OHM-LABS claimed in their specification just by this simple DIY.

my guess was, that a perfect symmetry is the most importand part, but if it is that easy, I am even more interested to see what is inside the Ohmlabs product

Agreed that symmetry is the key. The bulky copper also plays an important role if not perfectly balanced. And you always have the option if trimming if all fails.

[lowimpedance]

Here is a PDF (pt1 and 2, split to fit the forums upload conditions!!) which you resistance 'tinkerers' might find an interesting read.
Its a scan from the ESI Engineering bulletins; this one being the Traceability of Resistance Measurements no.30.

[Vgkid]

Thanks for those, I wish IET Labs hosted the esi documents.

[quarks]

lowimpedance thanks very much for the "Traceability of Resistance Measurements no.30" document


I just looked through my ESI documents to see if I overlooked the details about 4-terminal junction like Figure 27 in your document


but only found this (which also shows patent from manganins post)


 

[zlymex]

IET SRL 1 Ohm. Same specification as Fluke 742A 1 Ohm but different internal construction.

The internal construction is actually the same. Multiple resistors connected in parallel with the copper equalizing plates. But in the IET version the whole package was resin potted which allows to use much cheaper non-hermetic resistors.

Here is a new photo showing the oil leakage from one of the SRL-1. It seems that the pot is oil filled.