Kelvin varley resistor porn

[ketil b]

Hi

Just got this of the eBay and thought you guys might like a look at some old-time resistors.

There's a bit of an offset error, but nether my meter or my source are calibrated

Thanks

ketil

Edit : fix the spelling error in the subject

[ketil b]

And the resistor

ketil

[TiN]

Sweet, thank you for teardown 

[elex_enthusiast]

Vintage porn beauty!!!  ..

[krivx]

Do you mind if I ask how much? The new prices on Dekapots are pretty shocking

[ketil b]

Do you mind if I ask how much? The new prices on Dekapots are pretty shocking

Not at all it was $96 au + $65 shiping.

Ya gota get in quick.

Ketil

[Theboel]

just thinking how to calibrated this beauty ?

[retrolefty]

Wow those are great pics.

 What was the retail price and year?

[ketil b]

just thinking how to calibrated this beauty ?

The quick answer is you don't, or there is nothing inside that can be altered.

The best you can do (and I suspect that this is how it was done back in the day), is you get a very good voltage reference and a very good DMM and make a list of the error for each dial setting. Then apply that error every time you use it

Ketil

[AF6LJ]

Very nice, good workmanship.

[ketil b]

What was the retail price and year?

I don't know, there is no dating information on the unit but ESI still makes them although you will have to get a quote.

Ketil

[PedroDaGr8]

I have FOUR of these badboys. Need to throw a couple of them up for sale sometime soon.

[Theboel]

just thinking how to calibrated this beauty ?

The quick answer is you don't, or there is nothing inside that can be altered.

The best you can do (and I suspect that this is how it was done back in the day), is you get a very good voltage reference and a very good DMM and make a list of the error for each dial setting. Then apply that error every time you use it

Ketil

I think its the best way

[linux-works]

I bought one of those at my local surplus store (halted, HSC in the bay area):







and an old genrad divider that I got on ebay a few years ago:







the genrad was $100 or less; but the dekabox was $25 iirc.  it was sitting behind a glass display cabinet and after a few visits there, I finally decided to buy this just for the heck of it.

[K5HJ]

What happens when you set one of these to divide by zero?
Does it blow up?

[PedroDaGr8]

[Alex Nikitin]

What happens when you set one of these to divide by zero?
Does it blow up?

It probably would (blow up) if you could set it to divide by zero.

 

Cheers

Alex

[Vgkid]

What happens when you set one of these to divide by zero?
Does it blow up?

if you set it to zero, you should get a few microvolts out. Due to reading the voltage developed across the switches. That is in series with the resistors. But it is a guess.

[linux-works]

What happens when you set one of these to divide by zero?
Does it blow up?

you get infinite gain.  but you get it all at once and it happens way too quickly.

[linux-works]

What happens when you set one of these to divide by zero?
Does it blow up?

It probably would (blow up) if you could set it to divide by zero.

 

Cheers

Alex

ha!  genrad gear, from that era, laughs at you.  you and I both age, but this stuff never ages.  lol

[AF6LJ]

What happens when you set one of these to divide by zero?
Does it blow up?

You Don't Want To Know. 
They did it to one at Atlantis. 

[Gyro]

Interesting to see how they arranged the staggered wipers on the General Radio Unit.

Does anyone know what a non-Kelvin Varley decade divider is called? I have a Muirhead D-801-D four decade "Voltage Dividing Resistance", it is externally the same as a Kelvin Varley divider, ie. It's a potentiometer but uses a different operating principle:

Instead of each decade bridging two resistors on the previous one, on the Muirhead each decade replaces one resistor on the previous one... It actually has roughly twice the number of resistors per decade (apart from the last one). Each switch has 2 wipers, one selects from a chain of up to 9 resistors below the tap and the other selects from a chain of up to 9 resistors above. Eg, when set to 5 it selects 5 resistors below and 4 resistors above - the next decade replacing the 'missing' resistor.

This arrangement means that there are nearly double the number of resistors (18 matched resistors per decade) and subsequent decades must accurately match the resistance of one resistor on the previous decade (unlike the Kelvin Varley).

Does anyone know what this arrangement is called? I would guess that it was built in the '50s so well after the Kelvin Varley divider was invented. Accuracy is very good, still within about 0.02% on ratio as far as I can measure.

Here's a link I pulled off the web for general description and external photo. I'll need to open it up if anyone is interested in internal photos.

http://rochesteravionicarchives.co.uk/collection/general-purpose/4-decade-voltage-dividing-resistance-box

[Vgkid]

Varley voltage divider. Will update in a few minutes(at lunch).
Google is returning it as a plain resistive divider(no fun there).

[Gyro]

Thanks Vgkid,

I suspected (feared) that it would have been invented by either Kelvin or Varley - of course Google completely swamps out with Kelvin Varley references. I've looked in my collection of vintage engineering books but no joy.

[TimFox]

just thinking how to calibrated this beauty ?

The quick answer is you don't, or there is nothing inside that can be altered.

The best you can do (and I suspect that this is how it was done back in the day), is you get a very good voltage reference and a very good DMM and make a list of the error for each dial setting. Then apply that error every time you use it

Ketil

I think its the best way

The ESI and General Radio Kelvin-Varley dividers use fixed resistors and cannot be calibrated further.
The Fluke 720A has trimpots in series with the individual resistors, and a Wheatstone bridge that can be switched to measure each of them separately.  It's not necessary that the resistors be precisely 10,000 ohms (for the Fluke), but that they all be equal.
If you measure the voltage ratio with a DVM, make sure you are on a voltage scale with very high input impedance.  Most DMMs have a finite input impedance for full-scale values above 10 VDC, due to an input voltage divider.  The output resistance of the divider is large (66k maximum for the Fluke 720A, which has 100k input resistance), and a 10M load resistance will have a noticeable effect.