Any resistor series, that are stable wrt time/temp, but not precision values

[julian1]

I want to make resistor dividers, that have low drift over time and have high temperature stability. If calibration and trim can be handled with monotonic/precision DACs - then part tolerance on measured R values is less important.

I discovered Vishay VPG http://au.mouser.com/search/refine.aspx?Ntk=P_MarCom&Ntt=113825677 but the prices are a bit alarming to a newbie.

Great temperature stability of 0.2 PPM / C, but I really don't need 0.01% value tolerance.

Are there any less expensive resistor series on the market that are designed to meet this kind of need?

[acbern]

Your question is very unspecific. What is low drift over time and high temperature stability? What do you need?
These epoxy encapsulated VPG resistors may be above 10ppm/year in drift. Better drift-wise are usually wirewounds, but standard types typically in the 3-5ppm/K range.

[T3sl4co1l]

There are some resistor packs with precision ratios and only modest overall tolerance, but it really isn't a problem, for macroscopic networks, for all elements to be accurate in the first place.

If you need high stability, you might as well get the high accuracy too.

Tim

[Cerebus]

Vishay have the NOMCA series, 8 resistors in a surface mount package, under £4 GBP each in small quantities. These are the best bang-for-buck that I've found so far in matched resistor networks.

25ppm tempco overall, 5 ppm tempco tracking
0.1% basic tolerance, 0.05% ratio tolerance.
Stability 0.05%/1000hrs, ratio stability 0.015%/1000hrs.

All the resistors in the package are the same value but you can make a set of dividers by using parallel and series combination that are going to take advantage of the matched ratio tempco and stability ratings while being on the same substrate. Only available in 1k, 2k, 5k, 10k, 20k and 50k off the shelf, but Vishay will make them to order in any value but of course that would come at a price

Depending on your design, it might take a bit of cunning to make the particular layout of these suit your purposes. Again, depending on your design, you might get better overall performance out of these than out of say, several Vishay dedicated MPM dividers which have 2ppm ratio tracking each, but obviously wouldn't have tracking between separate packages.

If you need better stability or tracking performance than these then you are going to have to step up in price.

[julian1]

you can make a set of dividers by using parallel and series combination

It's obvious when explained, but I wouldn't have immediately thought of this.
 

take advantage of the matched ratio tempco and stability ratings while being on the same substrate.

The idea being that if they're on the same substrate - then they're at the same temperature, so temperature effects should cancel in a divider configuration. Much like thermally-bonding a npn and pnp pair. Thanks.
 

[Cerebus]

you can make a set of dividers by using parallel and series combination

It's obvious when explained, but I wouldn't have immediately thought of this.
 

take advantage of the matched ratio tempco and stability ratings while being on the same substrate.

The idea being that if they're on the same substrate - then they're at the same temperature, so temperature effects should cancel in a divider configuration. Much like thermally-bonding a npn and pnp pair. Thanks.

If you go down this route, think carefully about where the resistors physically are in relation to each other. That is, if you've got one that you know is going to dissipate some heat, make sure that any resistors that have to work with it in a divider are physically close. So, if you make a divider like this

       |
      R1
       |
    =====
    |   |
    R2 R3
    |   |
    =====


(R2 and R3 are meant to look paralleled there) make sure that R2 and R3 are on either side of R1 on the actual substrate so that any dissipation is shared around as equally as possible.

[zlymex]

Vishay have the NOMCA series, 8 resistors in a surface mount package, under £4 GBP each in small quantities. These are the best bang-for-buck that I've found so far in matched resistor networks.

25ppm tempco overall, 5 ppm tempco tracking
0.1% basic tolerance, 0.05% ratio tolerance.
Stability 0.05%/1000hrs, ratio stability 0.015%/1000hrs.

All the resistors in the package are the same value but you can make a set of dividers by using parallel and series combination that are going to take advantage of the matched ratio tempco and stability ratings while being on the same substrate. Only available in 1k, 2k, 5k, 10k, 20k and 50k off the shelf, but Vishay will make them to order in any value but of course that would come at a price

Depending on your design, it might take a bit of cunning to make the particular layout of these suit your purposes. Again, depending on your design, you might get better overall performance out of these than out of say, several Vishay dedicated MPM dividers which have 2ppm ratio tracking each, but obviously wouldn't have tracking between separate packages.

If you need better stability or tracking performance than these then you are going to have to step up in price.

That's the idea. And NOMCA use Tantalum nitride (Ta2N) for the resistive element.

Datron claimed they use statistical resistors of Ta2N elements in their 7000 voltage reference for the step up of 7V to 10V, but actually was not. What they actually used is TDP16031002 which made from Passivated Nichrome. They specified very similar though.

I got one TDP16031002 cut open by a file.

[ebclr]

What about this

http://www.linear.com/product/LT5400

[Cerebus]

What about this

http://www.linear.com/product/LT5400

Excellent resistors but relatively expensive. UK 1 off prices for LT5400 vary from £10 GBP to £17 GBP depending on value and grade compared to £4 for the NOMCA. Spec wise the LT5400 tempco ratio is ~5x better, but the other specs are only sightly better or roughly equivalent to the NOMCA specs.