Here's a bizarre bit of military trivia: up until I parted with a company some 6-7 years ago we were making new equipment with old parts: our drawings said to use carbon composition resistors and use them we did... even though they were obsolete, out of production for years, and we were forced to pay a premium for old parts. Seems there are people out there who buy up obsolete military grade parts in the hope some will turn to (better than) gold.

It was easier that way:any and all component changes needed a ton of paperwork for each part on each unit. If a part was available from any (traceable) source then you were pretty much forced to use it.

So there were using ancient, out of tolerance parts on equipment to be installed on the latest generation of submarines, both the Sea Wolf and Virgina classes. Nothing too important, just the hot well level sensors that made sure there was enough water for the nuclear reactor.

Things may still be done that way, I'm just not in that loop anymore.

 

Sadly, that doesn't surprise me. The "Rules of Acquisition" are made by management/accountant/business types with absolutely no awareness of reality. They make rules that make sense in their world view under the assumption that everything outside of their world view works perfectly. To be sure, engineers and just about everyone else in the world have a tendancy to do the same thing.

They also don't realize that incentives matter and if you make it hard, or actually penalize, people for doing the right thing, then most of them will simply take the path of least resistance and do the wrong thing that is easy and that they don't get penalized for doing.

 

I just did a bit of checking and while I couldn't determine if carbon film resistors age the same way carbon comp do, I did find some stuff that makes me less confident that they should. Carbon comp resistors use various binders and fillers to achieve the different ranges. I suspect that carbon film resistors either don't use these or use less of them. So if the aging mechanism is primarily related to the binders and fillers, then carbon film resistors may be much more stable.

That's true. Carbon composite are terrible for ageing, especially in TV's and vacuum tube equipment where they get exposed to high voltages and high temperatures.

Carbon film is a much better (and more modern) process and don't age much at all, although exposure to high voltages will still cause some degradation of the carbon film resistors (causes increase in resistance over time).

 

RB, what is considered high voltage for these resistors? Would using them in a 100V/10V potential/voltage divider for a 339 comparator circuit be in the bad area for these?

 

Most well designed analog circuitry is not incredibly fussy about resistor values,so that 10% resistors did just as well as higher tolerance resistors.

Historically,1% resistors were just that--within 1% when they were manufactured.

It did not imply "High Stability",& although most "High stabs" were 1%,some were 2% & even 4%.
They maintained that tolerance in service,whereas the ordinary 1% resistors did not necessarily do so.

Modern manufacturing methods have increased the stability of ordinary resistors to the point where they match,or even exceed that of
the old "High stabs".

Early carbon film resistors made pretty good inductors,due to the practice of resistance adjustment by turning a
spiral cut through the film.
This may be the origin for the preference for carbon composition resistors in some specifications.

There are some pretty silly things done in the name of "quality assurance".
One place I worked at used 1.1 kΩ resistors in an integrating network with 47μF,20% tolerance Electrolytic caps.
We ran out of 1.1 kΩ resistors,so they stalled the whole production process-----even though we had plenty of 1.0kΩ & 1.2 kΩ in stock!

 

RB, what is considered high voltage for these resistors? Would using them in a 100V/10V potential/voltage divider for a 339 comparator circuit be in the bad area for these?

Most resistors are at least 100v rated, some 200v.

A 100:10 divider exposes the top resistor to 90v which should be ok.

If the question is spcifically related to these old "mud" type resistors I would simply not use them at all! New high performance metal film resistors are a couple of cents each and just a better choice all round.

When I said "high voltage" that was generally in regards to using the old mud resistors in vacuum tube apps where they might get >200v and failure over time was very common.

 

Thanks for the answer RB. They will be carbon film not composition. I have seen some threads here over time, that treat 100V as high voltage, so wanted to make sure.

 

As to why resistors over 1% are still sold: From my days in assembly I know that you'll see resistors with a tolerance wider than 1% quite commonly in mass-produced electronics. Even in extremely expensive consumer electronics gear most would consider 'Top Shelf' for the simple reason that when you're dealing with quantities in the Billions, with a big fat Carl Sagan B, every 100th of a penny you can save per-unit counts. If a manufacturer can use a cheaper part, it follows with the heartless inevitability of economics that they will use such a part. Though it seems laughable to those of us dealing in quantities of hundreds that 10% and even 20% tolerance parts even exist, to those who buy their parts by the freight train full they exist for a very good reason.

Once when I was assembling boards for a test engineering department at a major IC firm, in a quantity of about 50, I noticed that the engineer who did the board layout and generated the parts list (New Guy) had marked all the resistors to be used as 0.1%, there were some pretty exotic values we had to order out for, all the resistor pads were 0402, and all the capacitor pads were 0603, where there was easily room for 1206 parts in both cases. One of my guys who was more electronics savvy than myself mentioned that 1% or even 5% would have been fine for nearly every part, and that the odd values were also pointless. He said the engineer in question was likely 'having fun with the design' or even 'showing off' a bit. We assembled them as marked and when I dropped off the boards and the invoice, I popped into the department heads office on my way out, as we were on pretty friendly terms, and mentioned we could have saved them a few bucks on the run if the parts list hadn't been so odd, and that larger parts mean less billable time spent soldering them. I didn't think much would come of it as it was a small run and the extra expense was only a small fraction of the total cost. What followed was one of nastiest verbal reprimands I've ever witnessed in a business environment. I felt really bad I'd even mentioned it.

Corporations can be very serious about saving even tiny amounts of money, and as long as 5% parts are cheaper than 1%, and 10% cheaper than 5%, there will be a market for them.


Mike

 

I agree, except that I wouldn't characterize their decision to use the less expensive part as some "heartless" decision. I would do exactly the same thing if I were in charge and would certainly not consider it a heartless decision. Quite the opposite. If I could save 0.01 cent each off of a billion parts purchased each year, that's $100k of savings that can be used to give raises, increase benefits, make bonuses, pay out to shareholders, purchase new equipment, etc, etc. If I didn't make that decision (assuming no compelling reason favoring another decision) I would be failing in several of my fiducial responsibilities. Furthermore, by purchasing large volumes of low tolerance parts I both give the manufacturer of the part a market for something they might otherwise have to discard, at a cost, and reduce the demand for the higher tolerance part thus putting downward pressure on the price for everyone that uses them.

 

The new resistor manufacturing process of choice is carbon film, and this process does not produce 10% resistors so they are not offered for sale.

Cost is critical in high volume manufacturing, and designing the best cost effective product means using cost effective components; but, designing for low cost is harder than designing for above average performance.

 

I agree, except that I wouldn't characterize their decision to use the less expensive part as some "heartless" decision. I would do exactly the same thing if I were in charge and would certainly not consider it a heartless decision. Quite the opposite. If I could save 0.01 cent each off of a billion parts purchased each year, that's $100k of savings that can be used to give raises, increase benefits, make bonuses, pay out to shareholders, purchase new equipment, etc, etc. If I didn't make that decision (assuming no compelling reason favoring another decision) I would be failing in several of my fiducial responsibilities. Furthermore, by purchasing large volumes of low tolerance parts I both give the manufacturer of the part a market for something they might otherwise have to discard, at a cost, and reduce the demand for the higher tolerance part thus putting downward pressure on the price for everyone that uses them.

My use of the word was more aimed at the nature of the economics of mass production in general. The phrase 'it follows with the heartless inevitability of economics' was something an old friend used to say quite often when talking about large companies and their tendency to sacrifice quality more and more these days. It just sort of sprang to mind as I was typing my reply. In hindsight the word may be somewhat out of place in this context and I probably could have chosen more wisely. I tend to agree with most of what you've just said and retract the word, with apologies.

Mike

 

Within RCA during the 1950s an engineer would design a radio, and then the design would be sent to the product development dept. PDD would remove each component and check to see what effect it had on performance, no discernable effect and the component was removed from the design. So, the hatchet aproach to electronic design has been going on for many years.

 

I had a professor that worked for RCA in the 60s (maybe even the 50s) and he told me about a colleague who was one of the better paid engineers in the company and his sole responsibility was to reduce the cost of the flyback transformer by so many cents each year.

I would imagine that this guy was eventually one of the world's foremost experts on not only flyback transformers, but many of the associated things, at least to the degree they impacted flyback transformers, ranging from manufacturing processes, material properties, electromagnetics, thermal management, and so on.

 

Most resistors are at least 100v rated, some 200v.

Not true with modern thick film SMD resistors.

0201 rated 30V
0402 rated 50V
0603 rated 75V
​

I was surprised how low they are rated when I first saw it.

 

Not true with modern thick film SMD resistors.

0201 rated 30V
0402 rated 50V
0603 rated 75V
​

I was surprised how low they are rated when I first saw it.

The limiting factor in these sizes is likely size it's self.

0201 is 0.6mm x 0.3mm
0402 is 1.0mm x 0.5mm
0603 is 1.6mm x 0.8mm

I ran a business hand assembling runs of boards too small to bother re-tooling automatic place & solder machines. 0402 was a small as we'd even attempt to assemble by hand and even that was a time consuming affair. Only our 2 best girls could do it and they'd usually have to add the extra step of putting a meter across each part to make certain they hadn't bridged the pads. Just a tiny amount of excess solder and it'd flow under the part and there would be a short that the debuggers would spend all day looking for if we didn't fix it before the board went out.

As for smaller than 0402, I've only ever seen 0201 parts in person once, and it looked like a vial of coarse sand unless you were looking at it under a microscope or through a jewelers loop.

Mike

 

I never found 0402 that hard (I would today, given the state of my eyesight) and I've done 0201. I use a binocular microscope and a Metcal that I got at an auction. Best investments in lab equipment I've ever made.

 

Only our 2 best girls could do it and they'd usually have to add the extra step of putting a meter across each part to make certain they hadn't bridged the pads. Just a tiny amount of excess solder and it'd flow under the part...

Maybe you just needed more appropriate footprints for hand soldering?

 

Maybe you just needed more appropriate footprints for hand soldering?

You know, I'd never thought of that. It may be that the pads the engineers were using were geared more towards automatic soldering methods.

I never found 0402 that hard (I would today, given the state of my eyesight) and I've done 0201. I use a binocular microscope and a Metcal that I got at an auction. Best investments in lab equipment I've ever made.

It's not like 0402 was impossible, the other assemblers and I just didn't like having to do it. We had Unitron ZSB stereo microscopes and Metcal dual-output MX1's at every station. We did figure out a few months before the company closed down that using a fine point flux pen and solid solder rather than flux-core meant less bridging and made life easier for the assemblers.

The one thing I did like about 0402 was that we charged per man hour and boards with lots of 0402 parts were big money since even experienced assemblers like myself took easily twice as long to solder them than 0603 or 0805, which could be done under a simple magnifying lamp. (At least, before my eyesight went to hell. )

-Mike

 

If you used a 10k pull-up on an IC and it had a 10% tolerance you would never know it. For us analog guys we either plan circuits with very tight resistor/capacitor tolerances or build adjustments into the circuits. I am not sure what the commercial industry chooses but Mil-Aerospace adjustments are not permitted on most platforms.

 

If you used a 10k pull-up on an IC and it had a 10% tolerance you would never know it. For us analog guys we either plan circuits with very tight resistor/capacitor tolerances or build adjustments into the circuits. I am not sure what the commercial industry chooses but Mil-Aerospace adjustments are not permitted on most platforms.

A lot of analog circuits are not particularly fussy.
If your 10% tolerance caused a 10% change in gain in an amplifier it would still be less than 0.5dB.