SMD TC??

Well, i know that the TC specs are "plus or minus" meaning the actual TC can be anything in that range. BUT. I wound up measuring the TC of four resistor types: 0.1% 25ppm/C Vishay, 1% 100ppm/C Vishay, 1% 100ppm/C Xicon, and 5% 200ppm/C Xicon. Despite this "wide" variation of types, they all measured NEGATIVE, and close to the spec limit.

My question: is this typical?

PPS: I expected a positive TC, closer to zero - especially for the quality Vishay types.

ual

,

As people get better control over their manufacturing processes, you can expect them to make a less variable product, and they are going to centre the variation they've got left at a point that maximises their yield.

No manufacturer I've ever run into guaranteed that the variation within the specification was a bell-curve centred on the nominal value, and there was time when tight tolerance resistors were just sorted out of the output of a lower tolerance production line - so no

+/-1% resistor was ever within +/-0.1% of it nominal value, and the 0.1% distribution looked like a segment out of a bell curve - and not always a segment all that close to the peak.

And no - I never seen this documented. It was one of many other instructive stories told to young engineers over a beer when I was much younger and a member of a group seen to be in need of that kind of instruction.

-- Bill Sloman, Nijmegen

It would have been easy to grab a few resistors and see for yourself.

--

John Larkin                  Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators

ual

,

I'd expect the TC to be pretty close to the limit and always the same sign for a specific type of resistor, since I would think it should mostly be determined by the physics of the material the resistor is made of

-Lasse

Nowadays everything is laser trimmed, and I suspect that parts are not binned post-trim. That's why 1% resistors cost about the same as 5% ones. We don't even specify 5% resistors any more.

--

John Larkin                  Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators

Yeah, i like the Susumu TC specs including the fact that at least SOMEONE carries them. There are Xicon 10ppm/C resistors, but no stock; ditto for all Vishay SMDs 15ppm/C or less (nevermind the cost gets very high).

Check; 40-50 years ago, typical quality 5% carbon comps ran 2% high on the average - and very "clustered" at that point.

In small quantities they may be the same price but not in the millions.

ctual

VE,

A couple of decades ago, we got a lecture from a Vishay expert on how they made stable low-temperature coefficient thin film resistors. One of the problems was getting enough resistance per square to start with, and to help that they built-in a lot of mechanical stress.

The substrates cost money and take up board space, so minimising the resistive area was necessary to keep prices competitive.

To get the temperature coefficient close to zero they then played off the thermal expansion of the substrate against the thermal expansion of the resistive film, which influenced the mechanical stress as a function of temperature. There was a lot of physics involved, but also a lot of deliberate choices.

-- Bill Sloman, Nijmegen

actual

C

IVE,

A few doesn't say much about the statistics. Normally what you grab out of stores all come from the same production batch, and they do tend to be more uniform than samples from different batches.

Spending an afternoon measuring enough resistors to make a informative histogram might have been interesting, but I never ran into a situation where it would have been a sensible - or justifiable - way for me to spend the time. As the boss, you can probably be more idiosyncratic.

-- Bill Sloman, Nijmegen

Vishay metal foil SMD resistors are generally available, e.g. Farnell/Newark/Element14 and others. (And yes the cost is very high, like $15 each). Down to 0.05ppm/C though.

--

John Devereux

actual

/C

TIVE,

A poor man's way of improving many circuits on the cheap is to use resistor arrays. Their tempcos are pedestrian but the elements track fairly well within the array. That's good enough, many times.

I also saw an article wherein the author measured the individual elements' tempco's, then combined them for minimum drift--that's kinda nutty (but he did get t/c insanely low, for that particular day :-).

-- Cheers, James Arthur

You do need to use a true common-centroid design for that to work, of course. The gradient and transient sensitivity can be pretty ugly.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510
845-480-2058

hobbs at electrooptical dot net
http://electrooptical.net

Doing interesting things is always worth one's time. Or at least, some peoples' time.

--

John Larkin         Highland Technology, Inc

jlarkin at highlandtechnology dot com
http://www.highlandtechnology.com

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom laser controllers
Photonics and fiberoptic TTL data links
VME thermocouple, LVDT, synchro   acquisition and simulation

he actual

pm/C

n.

GATIVE,

o

I've always had more interesting things to do than time to do them. Some of them have been both interesting and useful, and they tend to get priority.

Now I've got a parts list to finish....

-- Bill Sloman, Nijmegen

Yesss..and the cost is the same.

Hi James,

There are also clever ways of connecting the individual elements, and also multiple array packages, in such a way that all remaining errors are random and can be made very small by averaging. I can't find the exact details (anyone know?) but the technique is hinted at here (see fig 4).

Yes, that is a standard method by which precision resistor standards and dividers are constructed, AIUI.

--

John Devereux

Here are some plots I did recently of some resistors.

The Vishay Z201-1k and Z201-100R would be pretty good as lab standards. They do hermetically sealed ones to eliminate RH sensitivity, although these are the normal epoxy types.

--

John Devereux

the actual

ppm/C

on.

EGATIVE,

M t

Great app note--thanks.

Waayyyy back I saw an article, also Fluke I think, describing a REALLY cool technique for making precision dividers. It had to do with measuring resistors in series, then combining in parallel combinations. The math worked out such that you could make wonderfully accurate dividers.

I'm sure it'll percolate back later in the day. It was part of a precision voltage calibrator too, circa 1981.

-- Cheers, James Arthur