Author Topic: T.C. measurements on precision resistors  (Read 396314 times)

0 Members and 2 Guests are viewing this topic.

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2378
  • Country: de
  • Sounds like noise
Re: T.C. measurements on precision resistors
« Reply #75 on: July 06, 2014, 08:46:30 pm »
What are the thoughts on resistors made by Burster?

Precision and High Precision Resistors Series 1140, 1150, 1160, Model 1178

We haven't talked about them yet. They say that they can achieve <2ppm/K with selecting materials and Zeranin, but also do some "meticulous artificial aging procedure".

I've been told that in a quantity of 10 a single resistor (1142) is about 23,13€.
« Last Edit: July 06, 2014, 09:01:43 pm by branadic »
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline AndreasTopic starter

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #76 on: July 06, 2014, 09:48:28 pm »
What are the thoughts on resistors made by Burster?

We haven't talked about them yet. They say that they can achieve <2ppm/K with selecting materials and Zeranin, but also do some "meticulous artificial aging procedure".

I've been told that in a quantity of 10 a single resistor (1142) is about 23,13€.

Hello branadic,

I also thougth of them. Another option could be Powertron.

My personal opinion is: If you can get a matched set of resistors for a voltage divider
with even narrower specs or a special resistor value this could be fine.
For a single resistor with low TC I see no advantage against a Z201 with similar price for a single unit.

On the other side I stumbled over the Riedon T.C. curves:

http://www.riedon.com/media/pdf-tech/TemperatureandTC.pdf

what the hell does the statement "The TC adjustment is per charge at the reference temperature of 25°C (T0)." really mean?
I fear my english is too bad to fully understand what they want to express here.
And what does this mean together with the "picture 2" with the mean value on -1.8ppm/deg?

Ok I know that T.C. curves of resistor manufacturers are never true since they test only at exact 3 temperatures.
25 deg, -55 deg and +125 deg. All other values inbetween are interpolated.

But for me the interesting part is the behaviour around 25 degrees.
I do not want to do precision measurements around -55 deg on north pole or in the sauna at 125 deg C.

By the way I´m still working on the theme but have no spectacular results at the moment.
In the mean time I found a systematical error in my calculations.

But since the (non-linear) error of 0.008 ppm cannot be seen in the curves it´s not worth to mention.

On the other side today I tested a buffer amplifier between multiplexer and ADC.
There is also no significant change against unbuffered mode.
But the buffer will allow me to use much larger resistor values (than the 1K) together with the ADC
without running into trouble with non-linearities.

With best regards

Andreas



 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2378
  • Country: de
  • Sounds like noise
Re: T.C. measurements on precision resistors
« Reply #77 on: July 06, 2014, 09:57:25 pm »
Quote
Another option could be Powertron.
Powertron - A VPG Brand

Do you mean a special series?
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #78 on: July 07, 2014, 01:49:47 am »
To Branadic,

The Burster resistors you linked to are made out of three versions of Manganin, the two newer versions sports some improved characteristics, mainly TCR,  over the much older Manganin but also retains a lot of the same characteristics.  Given their small size and power rating, I'd wager there is a ceramic bobbin inside as these power ratings with a 'plastic' bobbin would make the resistor run on the hot side which these alloys don't like.

Properly treated and handled, Manganin can make a very good and stable resistor, the main limitations are only lower values ~<100K are available because of the limitations on the size these alloys can be drawn down to.  As I've mentioned elsewhere here, these alloys do not like being manhandled very much so care should be taken to observe the ratings and conditions of use.

Manganin is an old and proven alloy, it is still used in low ohm resistors, particularly in shunts and to some extent low ohm standards, but in most respects Evanohm and its multiple versions have superseded Manganin in standards, both working and primary.  Only the quantum based resistance standard is superior if you have a few hundred thousand dollars laying around for one and don't forget the liquid nitrogen to keep it nice and cool.

I will leave to the buyer whether or not these resistors are worth the cost.
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #79 on: July 07, 2014, 02:45:24 am »
Hello Andreas,

Can't say I've heard of Powertron either, I will look them up.

As to Riedon, I just came back from a visit with them in May, I have seen the statement you are asking about and frankly, I'm not sure what it means either.

On your point about TCR curves and manufacturers always 'lies' about them....I must pose an objection here, at least in my case.  You are generally correct, the standard procedure is to measure the TCR at the three cardinal points and this can (and in some cases does) result in covering up aberations in the TCR curve.  I won't go as far as to say it is intentional but I'll leave that conclusion up to the reader.

I do know of some resistor manufacturers who do not 'intentionally' hide TCR warts, it is more a case of ignorance of the warts, which one is worse depends on the circumstances.  I would not use a universal stroke of the pen to say "all manufacturers never tell the truth" as it is invariably not true.

When I say my TCR is flat, that is exactly what it means,  it is difficult to obtain such a 'curve' without all of the correct pieces of the puzzle.  Many PWW resistors do have fairly flat TCRs around the 25°C cardinal point, that is relatively easy to do, it is the rest of the temperature range that causes the problems.  The 'room' temperature is what everybody tends to use as the reference and after the so-called stabilization bakes have been done, that particular cardinal point is where things tends to settle down and look good.

While the three cardinal points are the usual ones requested, other points can and are specified, I know of at least one military specification that called for seven points on the TCR range.

Depending on the resistor type, TCR may only be sampled in the beginning to establish maximum and minimum values for TCR, commercial resistors generally are not TCR'd unless requested at additional cost.  The same thing is generally true for industrial grade resistors, military and aerospace are usually the only ones who put TCR measurements into the specification and those must be done all the time.
 
 

Offline Dr. Frank

  • Super Contributor
  • ***
  • Posts: 2377
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #80 on: July 07, 2014, 05:18:48 am »
@Edwin G. Pettis:

The Burster resistors are either made of the standard Manganin, i.e. CuMn12Ni2 and T.C. 10ppm/K, or for their precision types, they use a different alloy, called Zeranin (brand name from Isabellenhuette for this alloy and their wires), i.e. CuMn7Sn2.3, which is selected by Isabellenhuette already to have T.C. <3ppm/K, and a "flat", parabola shaped T.C. curve between 20..60°C.

Obviously, Burster then makes some further selection or treatment, I do not remember that any more correctly, I fear.
I bought many custom specific Zeranin resistors from Burster 25 years ago, for my precision current sources.


Which alloy (with < 1ppm/K ?) do you use for your resistors?
I assume that you also have to buy resistor wires, with a special alloy, but also from wire manufacturers like Isabellenhuette, or whoever delivers such specialties in U.S.

Frank
« Last Edit: July 07, 2014, 05:51:26 am by Dr. Frank »
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #81 on: July 07, 2014, 01:52:13 pm »
I examined the data sheet from Burster last night and all the claims appear to be reasonable with the Manganin alloys.  I could not find any mention of the conditions under which the claimed long term stability were tested.  There should be specifications as to the temperature and power conditions under which the long term stability is tested.  Usually, the only long term stability specification I've seen for any Manganin resistor is at the calibration temperature and low power levels.   It would be of some help if Burster specified its long term stability test conditions.
 

Offline robrenz

  • Super Contributor
  • ***
  • Posts: 3035
  • Country: us
  • Real Machinist, Wannabe EE
Re: T.C. measurements on precision resistors
« Reply #82 on: July 07, 2014, 02:03:41 pm »
Hello Mr. Pettis,  Thanks for all your excellent input on this thread and others. It seems to me you may have some valid input on the topic of Tellurium Copper.
 
I am of the opinion that Tellurium copper is used purely because of its machinability with a side benefit that it has slightly better corrosion resistance. Since the conductivity of Tellurium copper is only 90% of ETP copper it could be argued the thermal EMFs will be greater in a Tellurium Copper to ETP Copper connection than a ETP Copper to ETP Copper connection.  I think this is indicated by ETP copper being used on low EMF spade connections and Tellurium copper used only when the part has to be machined as opposed to stamped.
 
Some think that there is something special in the electrical properties of Tellurium Copper that make it exceptionally low thermal EMF compared to pure Copper. This is helped by all the high end test gear touting “Tellurium Copper” input connections as a sales buzzword.  Can you debunk this myth or tell me why I am wrong?

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #83 on: July 07, 2014, 02:33:26 pm »
Dr Frank,

With improvements in alloys come big improvements in price, I can currently pay as much as $33,850 per pound for the smallest wire size, 0.0004" diameter that meets my specifications.  Larger wire sizes, of course, cost less but still are not cheap, a 0.0063" diameter wire costs $239.71 a pound.  Material prices have skyrocketed in the last 5 years.

While Evanohm has been with us since the early 1940s, it has also undergone changes over the years, most have come in the last 25 years with about 5 improved versions, each having its own advantages, including higher costs.  I use three of the alloys depending on which particular specifications are required.  The newest alloy of the group has some terrific characteristics but it also more difficult to handle.  This is the one I use if the customer requires particularly tight TCRs, tight tracking TCRs and long term stability, it also costs the most.

I cannot use cheaper materials or I would not get the same results.  If the best performing resistors are required in a design, it is very foolish to buy cheaper, inferior resistors, that is why my PWW and Vishay's VHP series resistors cost more, performance comes at a cost.  Cheap only gets you trouble.
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #84 on: July 07, 2014, 02:51:28 pm »
To Robrenz,

Thank you for the compliment.  While I am not 'technically' a machinist, the use of tellurium copper alloy is generally used because copper, being a somewhat soft metal, does not machine nicely and tellurium gives it just enough 'stiffness' to machine smoothly.

Interestingly, Tellurium copper is rarely ever seen without gold plating on all of the binding posts and other terminals for obvious reasons.  Frankly, I agree with you that Tellurium is not used because it has 'special' electrical properties but because of its machining properties, a gold to gold connection would produce little EMF, the Tellurium copper becomes irrelevant.  Manufacturers have taken advantage of this 'myth' to charge rather high prices for items like Tellurium binding posts, sometimes into three figures (>$100)  each.....a pure ripoff.

Tellurium does have its uses but not for the usual reason.
 

Offline AndreasTopic starter

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #85 on: July 07, 2014, 06:22:06 pm »
Powertron - A VPG Brand

Do you mean a special series?

Hello,

I have to admit that it is some years ago since I had a look on their components.
Against formerly you now cannot distinguish the locally produced parts (mostly w.w.) and the vishay global parts.

Perhaps I will find something about the parts in my notes.

With best regards

Andreas
« Last Edit: July 07, 2014, 06:25:46 pm by Andreas »
 

Offline AndreasTopic starter

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #86 on: July 07, 2014, 06:39:59 pm »

On your point about TCR curves and manufacturers always 'lies' about them....I must pose an objection here, at least in my case. 


Hello,

Ok you are right it is a standardized procedure. And since all manufacturers do the same stuff there is a certain comparability.
On the other side I cannot understand how they generate the 0..60 degree coefficients out of the 3 cardinal measurement points.
(The 0..60 degree value would be that what comes closer to precision applications).
I have found no document about this up to now.

So they do not "lie" in a direct manner. But I feel that they dont tell you the whole thing.
It´s a bit like the "miles per gallon" on your car from manufacturer (or the exhaust gas values)
which is only true for the standardized test but not valid on the "Autobahn".

With best regards

Andreas

 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #87 on: July 07, 2014, 06:43:53 pm »
I was just looking at the Powertron website, Vishay again and it has the hallmarks of Vishay here and there too, they just can't resist putting their own 'spin' on things even though it probably didn't need it.  It is a relatively 'young' shop, founded in 1990, it may have been a local resistor house, bought out by Vishay or Vishay just founded a new one for the region.

The resistor offerings look similar to most other resistor house lines and the data sheets appear to have the hand of Vishay in them as important details tend to be put in some out of the way place.  For instance, the 100/SM/PC series looks an awful lot like Ultronix parts.  At the head of the data sheet, it plainly claims TCRs of +/-2 PPM/°K, looks impressive.  Further down the sheet on page 2, they finally spell out the 'standard' part specifications and the TCR is +/- 10 PPM/°K, the 2 PPM/°K is a special order.....not quite so impressive.  The entire data sheet would pass for an Ultronix data sheet if the header was changed, although Ultronix used to manufacture a much wider range of wire wound parts before Vishay got their sticky paws on them.  But, since Ultronix was owned by Vishay, they could do whatever they wanted with it.

The rest of the film/foil resistors look just like Vishay commodity parts, I didn't see any thing like a VHP in the lineup.  All in all, I don't see any compelling reason to buy from Powertron over the other shops.

I also found this line in their temperature dependence and TC paper:  "The TC adjustment is per charge at the reference temperature of +25°C (T0)."  I'm still not clear about the usage of the word 'charge' here and elsewhere in the document.  It sounds like they are referring to possibly a change in temperature but the usage of the word does not seem to be consistent, any more so than the Riedon usage.
 

Offline ltz2000

  • Regular Contributor
  • *
  • Posts: 103
Re: T.C. measurements on precision resistors
« Reply #88 on: July 07, 2014, 06:55:01 pm »
Hello Mr. Pettis

A few years ago I got some 0.04" enamelled manganin wire. It is probably 50 years old, 40 at least. I think building a current shunt would be a good way to use it.

But the problem is how to properly solder manganin wire to copper wire/plate. I thought it would be simple, but after reading this thread I am not convinced at all. All advise very welcome because the limited amount of wire doesn't allow much experimenting.

First, how to clean the wire ends to be soldered. Sand paper? Scraping with a knife?  When I got the wire, I was told that the manganin oxidises very fast and needs to be soldered immediately.

Is it possible to use a standard 60/40 solder? I am a little worried if the flux meant for copper (for example Multicore Crystal 400) works with manganin. The person who gave me the wire used some special flux but doesn't remember the details anymore. That was very long time ago and long before the soldering technology we have today.

Would it be a good idea to heat sink the rest of the wire with metal clamps to limit the annealing to maybe a half inch? Which is only a small part of the total resistance and TC.

Thanks in advance.
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #89 on: July 07, 2014, 07:39:47 pm »
To Andreas,

You are correct in that the much more limited temperature range of Manganin, et al, does require a different TCR procedure than the other wider range resistors.  The cardinal point still remains 23°C (or 25°C as the case may be) as the reference point.  Because Manganin, et al, has a hyperbolic curve for TCR (the newer alloys, particularly Zeranin, does flatten out the curve some around the cardinal point but it is still primarily a hyperbolic curve), this requires that the TCR be measured at points within its specified operating range.  Generally, the initial TCR curves are measured at 5°C intervals above and below the cardinal point to establish the basic curve characteristics, once that has been established in a repeatable manner, the batches are then only checked at three points, usually these are the points where the TCR curve intersects a specified reference line.  This line may or may not be at a zero TCR point depending on the 'average' TCR curve established for that particular alloy type.

Yes, this does require careful direct measurements (no statistical lying here) since the temperature range is restricted but the TCR measurements should be done carefully in all circumstances to insure accuracy no matter what the measurement range is.  I also agree that in the case of resistors which have a non-linear TCR characteristic, the three cardinal points are very inadequate to convey the necessary information about the curve, in this manner, warts may be hidden from the customer.

Well made resistors will stay close to the established TCR curve(s) within a normal variance, any that deviate outside an established norm should be considered faulty.  Unfortunately, the customer usually does not have such information and must rely on the manufacturer to do it right.
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #90 on: July 07, 2014, 10:49:38 pm »
To Itz2000,

If your Manganin is 40 or 50 years old, you may have an enamel which comes off pretty easily during soldering, those were fairly low temperature coatings unlike today's high temperature coatings.  The old coating evaporated during normal soldering and did not require a separate operation to remove it before soldering.  Soldering must be accomplished using standard rosin core 60/40 solder and the soldering must be done as quickly as possible as the temperature is well above normal operating temperature range for Manganin.  It certainly wouldn't hurt to use a heatsink such as forceps. 

Manganin does oxidize but you should be able to wet the wire and get it soldered before oxidation takes place, if you do encounter oxidation problems, use a small amount of liquid rosin on the wire before soldering.  Using a fine abrasive or knife edge to remove the coating/oxide does work, just try not to nick the wire with the knife but this should only be necessary on modern coatings, Manganin is fairly soft.

If you are soldering the wire to something bigger, such as a copper plate, large terminal, ect., first prepare the larger material and get its surface soldered, then you can quickly solder coat the Manganin wire, reheat the larger solder joint, insert the wire, remove the soldering tip and cool off the solder joint quickly, this should result in a nice clean joint and minimal heating of the Manganin.
 

Offline ltz2000

  • Regular Contributor
  • *
  • Posts: 103
Re: T.C. measurements on precision resistors
« Reply #91 on: July 08, 2014, 11:27:57 am »
Thank you very much for your answer and sharing your expertise.

I think I could try removing the oxidation and the enamel with fine sand paper dipped in the liquid rosin ("wet sand paper"). If I understood correctly the rosin protects the wire from oxidising (at least for a short period of time)? Do you think that would work and do you have any specific liquid rosin that you would recommend for manganin wire?


And finally to the original subject of this thread:

Almost 20 years ago I measured the temperature coefficients of 10 or maybe 20 bulk metal foil resistors. If I remember correctly the resistors were the old yellow/orange Sfernice. I still haven't been able to find the documents, but I keep on searching...
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2378
  • Country: de
  • Sounds like noise
Re: T.C. measurements on precision resistors
« Reply #92 on: July 08, 2014, 07:23:39 pm »
Quote
I think I could try removing the oxidation

We use something that is called "micro acid", a delute acid, to remove only the oxide from copper traces before end finish is processed. Maybe something similar (natrium persulfate) could be used?
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline AndreasTopic starter

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #93 on: July 08, 2014, 08:29:41 pm »
Hello,

I have updated page 1 with UPW25#1 data.

With best regards

Andreas
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 2378
  • Country: de
  • Sounds like noise
Re: T.C. measurements on precision resistors
« Reply #94 on: July 08, 2014, 08:59:14 pm »
Interesting, obviously this part is out of spec, even if it is one third of UPW50 in price.
Computers exist to solve problems that we wouldn't have without them. AI exists to answer questions, we wouldn't ask without it.
 

Offline AndreasTopic starter

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #95 on: July 08, 2014, 09:29:03 pm »
Interesting, obviously this part is out of spec, even if it is one third of UPW50 in price.

I don´t think that you can say that.
the T.C. in datasheet is +/-3 ppm/K "typical" and +/-5ppm/K max.
And the measurement conditions are different. 8-45 degrees (continuous)
against -55, +25 and +125 degrees cardinal values.

With best regards

Andreas
 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #96 on: July 09, 2014, 12:58:57 am »
To Itz2000,

The liquid flux I've been using is from Radio Shack, #64-021, a 1 oz plastic tube, officially called, "rosin soldering paste flux".  It isn't a paste, it is liquid, I don't know who the dummy was on that one.  I've had this tube for some years, I'm sure that other electronic outlets such as Mouser, Allied and Newark still carries this stuff, should be listed right with the solder.  I usually just dip the tip of the wire into the flux about 0.2" (50.8mm) and then directly solder it to the connection.

I'm sure, if this Manganin is that old, the enamel will just melt off during soldering, if it doesn't, then use some fine sand paper (crocus cloth) on it, it doesn't have to be wet, just a little quick rubbing and it should come right off.

By the way, it was Sfernice that broke Felix Zandman's patents on the film resistors, unfortunately for them.
 

Offline AndreasTopic starter

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #97 on: July 10, 2014, 09:46:52 pm »
Hello,

updated UPW25 #2 on page 1.
T.C. again around -4.3 ppm/K like UPW25 #1.

There is one remarkable thing:
UPW25 #1 has a rather high resistance value (around +620 ppm against Z201 #1)
UPW25 #2 is at +10 ppm
UPW50 #1 is around -100 ppm from Z201#1

So if the T.C. values of the 2 UPW25 were not so tight together I would have thought of some kind of production error on UPW25#1.

With best regards

Andreas



 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: us
  • The plural of anecdote is not data.
Re: T.C. measurements on precision resistors
« Reply #98 on: July 10, 2014, 11:04:11 pm »
Hi Andreas,

You likely received two resistors from the same production batch, although chances that you would get the same TCR from two resistors is a bit unusual if by chance.  Remember that the claim of 0 ±3 PPM/°C is for 0°C - 85°C temperature range (according to their data sheet) and as I recall you were told that the TCR 'curls' beyond those lower temperature limits to 0 ±5 PPM/°C.  These TCR measurements  of -4.3 PPM/°C within your 8°C to 37°C clearly places the range of your measured TCR outside the stated 0 ±3 PPM/°C limits.  While that TCR is within the 0 ±5 PPM/°C, it isn't within the lower limits, I'd say that makes these resistors defective in either case, but since they hedged their bets and said the 0 ± 3PPM/°C is 'typical', that lets them off the hook by a technicality.

« Last Edit: July 11, 2014, 12:06:16 am by Edwin G. Pettis »
 

Offline AndreasTopic starter

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: T.C. measurements on precision resistors
« Reply #99 on: July 11, 2014, 04:14:49 am »
You likely received two resistors from the same production batch,

Hello Edwin,

both UPW25 carry the same date code 1244.
So its very likely that they are from the same production batch.

With best regards

Andreas
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf