Kelvin Varley Divider [and Precision Voltage Source]

[amspire]

This is an interesting source of wirewound resistors:

http://www.surplussales.com/resistors/WireWound/WW5k-799k.html

[HLA-27b]

Just had a look at the ESI dekapot and the Fluke 720a and the card resistors are as wound as Conrad suggested. The others ...well cannot see inside the can!.
The single turn pots that ideed appear somewhat cheep on the pictures are wire wound and were made by CTS.
John

Many thanks John, you helped a lot. This taught me that there is no point in overengineering beyound the point that Fluke et. al. did. There is no need for example to be concerned with things like interwinding capacitance and high voltage behavior.

[fmaimon]

Only the best for me.

Brand is totally unknown, so I will have to see if they are good enough to equal the world famous "One Hung Low" resistors. 

http://www.uxcell.com/wire-leads-33k-ohm-metal-film-resistors-100-pcs-p-161865.html

I saw them on ebay (150707751358), but they don't ship to Brazil...

[amspire]

I saw them on ebay (150707751358), but they don't ship to Brazil...

Well let's see if they are good enough. If you need them, something can be worked out.

Richard.

[fmaimon]

I've sent them an email just to be sure. Maybe just nobody has asked them...

[lowimpedance]

I will do a seperate post on that Sperry today, because it is is very interesting.

I bet is would be really hard finding anyone who could make a switch today like that green 720a switch. Very consistant low resistance, high voltage as it has up to 1.1kv on it. Very low current leakage. If someone does make an equivalent, the cost would be incredible.
[/quote]

One last post regarding the special switch used in the 720a. There are some like the two in the Fluke, made by Leeds and Northrup, at work. I have scanned the info sheet for them and attached here for those interested in the details of such a switch.
John

[amspire]

What a brilliant switch! 1500V rating. 1 milliohm contact resistance.  10¹² ohms insulation resistance after 2 months in 100% humidity. Amazing.

Thanks John.

[Conrad Hoffman]

That's an impressive switch; never realized how good they were. L&N made some high resolution bridges for testing RTDs and such. They sometimes go cheap because nobody has a clue what to do with them. I disposed of one years ago. In hindsight, it probably had more value from the switches in it, than as a bridge. I don't think it was a KVD, just a big divider network, but I could be wrong. L&N made some bizarre stuff; their big DPDT knife switch for reversing bridge connections comes to mind. I also have a couple of their projection galvos. Incredibly ugly things, sort of like a miniature gray bathtub lying on its back.

[quarks]

I know this is quite old, but I wonder if there came out a final KVD and would be very interested to read/see it.
Thx quarks

[Spikee]

I am also interested if there is any progress .

[JBeale]

I'm also curious how this project came out, and also the quality of the 3W 3.3k 1% 50ppm/C resistors, which are being sold today at $15/100.

By the way, if anyone else is building their own KVD on the cheap, I also recommend Conrad Hoffman's 1996 article in "Electronics Now"  (plus voltage reference, and null meter). The link in the original post of this thread was on his own web page as separate JPG scans, but the whole series is also assembled at k04bb.com as one convenient PDF.
http://www.conradhoffman.com/mini_metro_lab.html
www.ko4bb.com Mini Metrology Lab.pdf

[amspire]

I'm also curious how this project came out, and also the quality of the 3W 3.3k 1% 50ppm/C resistors, which are being sold today at $15/100.

I think I did post about them previously.

The ones I got had a temp co of -45ppm/degC. The match from memory was probably less then 5ppm/C. The thing is this is still enough to cause a significant change (if you have better then a 4 digit meter) in resistance just by applying a significant voltage, such as a few volts.

In a kelvin varley divider, the two of the resistors in a divider chain connected to the switch for the next lower range dissipate less power then the rest so the temperature differential will cause a significant error. There is also the temp co of the resistance of the next lower range to add to this.

I had a plan to correct for the negative temp co with some copper wire would around the resistor.  What I found is that it is not just matching the temp co that is important, but if the copper wire is too thick a gauge, it heats at a much slower rate then the resistor. I would probably need to find something like 0.2mm diameter wire that is not that easy to get, and it would be an absolute pain to work with.

I think the best bet would be to put all the resistors of the upper ranges (except for the trimpots) in mineral oil (easily available from a chemist) and select for matching temp coefficient. This hopefully would help to keep the resistors at a uniform temperature, and also minimize local temp differentials on the resistor body when power is applied.

I would need to do some tests, but I haven't had the time lately. I do have a nice diecast box with a seriously good seal that I could use to mount the resistors in an oil bath.

Richard.

[PA4TIM]

The inside of my ESI 722 decavider ( 200 euro)

http://www.pa4tim.nl/wp-content/uploads/2011/05/720openkapot.jpg. Fluke 720A before restoration
http://www.pa4tim.nl/?p=2332. More pictures from my 720A ( got it for free together with a 750,760, 731, 332, 845 , guildline cabinet and some more goodies)

I have restored a GR1608, bought an expensive 1M 0.01% vishay for that. The original WW was dead. That is really thin wire, like a hair. The 1M turned out to be 100 Ohm low so I added a 100 Ohm that nicely zeroed the TC. But after installing and adjusting the rest the range was off. I removed the 100 Ohm and all was perfect. So maybe the old one was also 100 Ohm off .

[quarks]

...ESI 722 decavider ( 200 euro)

...my 720A ( got it for free together with a 750,760, 731, 332, 845 , guildline cabinet and some more goodies)

Hello PA4TIM,
you are indeed a really lucky man
I hope you get even more of these special goodies you can show us!

bye
quarks

[EyeDontKnow]

What would be nice, is if someone made a board for a KVD.
Given the advancements in surface-mount resistors (and lowering cost)....possibly a multi-deck switch that had solderable boards for each deck.....resistors soldered on each deck.

Maybe I'm being too optimistic or hopeful that there might be a demand for us retro old-school metrology geeks.  O0

I have several (new-old-stock) Shalco silver-plated 6 deck open frame rotary switches..........but I would love to see a more miniaturized opportunity.

The resistor accuracy could be up to the builder.
The ultimate being Vishay foil,  VSMP series, .01%, 0.2ppm types (or better). If bought in quantity.....the cost comes down.

What is really cool, is their "bulk metal foil" resistors can be specifically made at any value, to-order.....(I have not tried it yet)

[PA4TIM]

The Fluke and ESI can do 1000V. I needed such a resistor for a GR bridge and asked around. A company in the stated still makes them but 1 resistor of 1M would cost over 100 dollar. I then used a precision vishay through hole resistor with a very low tempco. That costed me 35 euro. You need a lot those for a KV.

SMD would be a lot cheaper but that limmits the use to 50V or 100V or so.
The KV's from Fluke and ESI have low EMF contacts and switches and everything is isolated very well to make sure leakage current stays low (10xE14 Ohm if I recall it right, and it was a hell of a job to get it within that specs again)

[EyeDontKnow]

Yes, my ESI kelvin dividers still do a great job.....over 20 years old ?
Clean and re-lube the contacts.

Often, old boat-anchor test stuff suddenly comes close to original spec just by cleaning the contacts (and replacing power supply caps).

I repaired a JRL voltage divider that had an open 15k resistor inside the oil-bath housing......no way to get to it. I used a Vishay RNC90y resistor as a replacement across the dead area...300v/0.6watt....that's the limit now.

[babysitter]

Some Kelvin Varley pron showed up during repairing the cheapest KVD available on eBay when i felt I need a little brother of my RV722 O0. There was a device on ebay USA employing a Electro-Measurement Inc. (later ESI) DV411, together with a power supply. Everything besides the Divider was removed, the remaining problem was the third decade switch was not latching to the digit positions . So i removed the three screws and opened the second can (my coworker from Taiwan was reminded of landmines  )

Overview picture

Above you see the switch without the latching mechanism on top,there is a inner and a outer ceramic wafer, the inner one is turning with the outer drive shaft and has the moving contacts. The outer part is fixed.

Under it the resistors that make up the rest of the KVD network. The second decade here is realized by a potentiometer with two fixed R in parallel, in the first can with 2 decade switches i would simply expect that it repeats.


The Potentiometer without its cap. Although sporting 270° range, it is really turning 360°. Asked the gurus at volt-nuts, they confirmed this strange behaviour is okay. 


The broken latching mechanism short after repair-the cogweel on the drive shaft is intended to be locked in positions by a ball-on spring. The cogwheel was loose, but to my surprise only soldered to the drive shaft.  Employing a agressive flux, that is what I reworked after repositioning the wheel on the shaft.

After putting everything back together in its housing it has shown to be a sucessful repair.
It divides on demand, Ratio is good for 4 decades and the 5th depending on operators skills
Cons: US screws! The handwheels had to kiss TORX 7 to align to the shafts, not the real US imperators imbus.

[Flump]

nice welding job there babysitter 

is it fully operational now and what do you plan do with it ?
I had some old decade boxes that had similar resistors inside.

can we get a pic of it all together ?

[babysitter]

lets call it soft welding, maybe ?

I am not going to open the can with the first two decades out of respect, and the wiring between the cans is not interesting.

It is working on first sight, 4 decades fine on 34401A, 5th user skill dependent.

As explained I removed the integrated power supply and the second potentiometer (Helipot 10 turn) from the big box. As I have 2 and a half labs (way too small home lab, my hobby basement about a km away and my workplace) this is my mobile ratio box. (The 7 decade part is a pita to move without a car and you just dont haul such things around.)

Teaching our trainee, simulating a NTC-resistor series circuit as temperature simulator are the things it is scheduled for in the next weeks.

[cellularmitosis]

Time to bring this thread back from the dead!   

Here's a question: do all of the switches need to be of equal quality?  I haven't found a definitive answer on this, but my intuition says that yes, they do.  Variation in contact resistance, or leakage should be just as detrimental at any decade, since they are all in series.

What sayest the forum?

[Kleinstein]

The less significant digit switches can get away with lower quality switches. There are parallel resistors and thus less current flows through the lower switches and less influence of the lower decades contacts.

This is even more so if extra shunts are used to keep the resistors in a smaller range: So not strictly the KV circuit, but from a certain point on the same value resistors can be used in the lower decades with an extra resistor in parallel to each decade. The first 2 decades are usually still higher impedance (classical KV) to keep the load reasonable.

[cellularmitosis]

Thanks Kleinstein, that makes sense.  Some current is diverted by the parallel resistance (blue), and some by the shunt in each decade (green), so not all of the current makes it to the final decade switch (red).

[Vgkid]

From the teardowns I have seen , a few milliohms are acceptable. 2-5 is the average for a good high quality switch.  As long as they are consistent , you can even parallel them up , though that price it may be cheaper to buy a used unit. Though I do want to test those Chinese ceramic rotary switches some day...

[lars]

What about thermal EMFs, could that be something to consider even for the less significant digit switches?
Lars