EEVblog Electronics Community Forum

Electronics => Metrology => Topic started by: dr.diesel on October 20, 2016, 06:24:48 pm

Title: New Precision Resistors
Post by: dr.diesel on October 20, 2016, 06:24:48 pm
A while back I ordered (4) precision resistors from a member here Mr. Pettis, 100R0, 1K, 10K and 100K, all at 10PPM.  These took a while to get, but he was upfront about that when I ordered and his communication was great!  I'm very pleased.

Anyhow, goal is basically a standard, age tracking, general fun.  These will be installed in a sealed metal enclosure, Pomona Microvolt Banana jacks, PT100 and some kind of small resistive heater.

I've been on vacation so no initial measurements till the DMM7510 warms up for a day or two.  Picture of how they arrived, not sure why only the 100R0 was labeled..

Big thanks to Mr. Pettis.   :-+
Title: Re: New Precision Resistors
Post by: lars on October 20, 2016, 08:45:36 pm
About a year ago I did the same as you with the same intent. Except for the 100-100k I also ordered a 1Mohm (+ LTZ resistors). All of them arrived quite soon with measured values. The problem was just that my measured values was not the same. Also the 1Mohm that was about 100ppm high in my measurement soon got open circuit after just a few tests with kelvin clips. The others was also high and seemed to get worse over time. I quite soon suspected humidity problems. But the answer I got was that it was impossible and it must be my measurements (skill). After a while a I got a mail that others also have seen this and the problem was the epoxy. A cure would be to bake the resistors. As I have ordered two 10kohms I baked one and it went down to almost the same as delivered but soon started to go up again slowly. As can be seen in the attached graph (scale is ppm) the time constant is months so a long test is needed to see this. My resistors is also without markings.

I have been promised to get a replacement but so far have seen nothing.

So for me these resistors are in the same league as the 2ppm volt refs on Ebay.

I really hope you have better luck and see forward for your results.

Title: Re: New Precision Resistors
Post by: plesa on October 20, 2016, 08:56:11 pm
About a year ago I ordered 26 resistors mainly for LTZ board. All of them has been measured for tempco. Tempco were exceeded on 5 pieces (10k and 100k ones). Replacement is going to be shipped soon.
Title: Re: New Precision Resistors
Post by: dr.diesel on October 20, 2016, 10:09:21 pm
I really hope you have better luck and see forward for your results.


Well, I won't start tracking until in the enclosure, but tomorrow we'll be able to eval the as shipped against my 7510 and 2010, both caled in the past 2 years.
Title: Re: New Precision Resistors
Post by: Edwin G. Pettis on October 20, 2016, 11:02:03 pm
Yes, as I have told these gentlemen, the epoxy I had acquired was supposed to work quite well, it didn't and its use was very short lived.  Replacements have been shipped at my expense for the resistors they said were not up to my standards.  Yes I also agree that it took longer than I would have liked to find a more workable solution, it is difficult to find and test epoxies which consumes a lot of time on my part.  Essentially, all plastic (epoxy type) resistors absorb water molecules at some rate, this rate is often specified under a specific condition which typically shows very small percentages of absorption, in reality the absorption rate is much higher than the data sheet reveals, hence trials must be done to determine how suitable a given compound is.  In the case of these resistors, the epoxy exhibited no significant water absorption over the trial period under 'normal' use conditions.  Suffice it to say that under conditions which are not 'normal' here, the epoxy showed a real good sponge effect which essentially applied stress to the windings, hence the apparent drift, this is NOT a change in TCR, stress can change the resistance measured and when that stress is removed, the resistance returns to original value.

Since this unfortunate episode, I have taken steps to eliminate such problems, previous resistors that I manufactured before this epoxy was used did not exhibit this particular problem which I was able to replicate.  I have made many resistors since I removed this goofy epoxy and none exhibit the humidity problem.

I might add that if humidity is a problem with components, including some plastic ICs, keeping the circuits energised 24/7 will minimize or eliminate this problem, also if humidity is generally high most of the time, I would suggest coating the circuit board and components with paralene which is fairly common and relatively cheap, this will stop any humidity from getting into the wrong places and causing problems.
Title: Re: New Precision Resistors
Post by: chickenHeadKnob on October 21, 2016, 08:55:08 am
Hello Mr Pettis

Would it be possible to buy your resistors "naked"? or if not naked then  with a hard plastic sleeve in which the resistor windings can float with freedom of movement?

I was holding off on contacting you to buy a set for various reasons on my end. I was harboring some inferred beliefs on the internal construction of your resistors  based on your posts on this forum, my assumptions:

I hope you can reply here rather than have a private communication as it would save you repeating yourself and be of interest to the many forum guests (lurkers) as well as members.

Thank you for your consideration and participation here.
Best regards

Title: Re: New Precision Resistors
Post by: zlymex on October 21, 2016, 12:29:38 pm
I'm worrying the humidity effect too, therefore for critical resistors, I always use hermetic(<100 Meg).
Title: Re: New Precision Resistors
Post by: Edwin G. Pettis on October 21, 2016, 04:34:04 pm
All of my resistors are made with various alloys of Evanohm to my specifications.

My resistors are welded with a proprietary method.

My resistors have always been made in essentially the same manner, the difficulty mentioned earlier was due to my supplier ceasing to supply smaller quantities of epoxy, nothing smaller than a gallon, hence I had to go 'looking' for another source.  That epoxy which caused the 'humidity' problem was used on very few resistors, I ceased using it as soon as the problem was revealed.  My resistors are not normally affected by humidity.

All wire wound resistors undergo some form of 'stress relief' no matter who makes them, this is due to the action of winding wire onto a mandrel which creates barrel hoop stress.  Commonly the resistors are 'baked' for some period of time (which varies with the manufacturer) and the results also vary with the resistor construction.  All resistors, whether wire wound, film, foil, ect. will still have some residual stress even after long period of baking.  This manifests as drift or hysteresis with temperature variations, it can be small or large depending on how well the stress relief worked.  Stress, in any form, applied to the windings (from any source) will cause a shift in resistance, sometimes misidentified as a change in TCR, this is incorrect, what you are seeing is the effect of stress being applied to the windings.  In the case of the 'faulty' epoxy I used, the epoxy acted like a sponge swelling with humidity putting pressure on the windings because the resistor is an enclosed unit, the shell working to hold in the swelling.  I also discovered upon opening up a few resistors that this epoxy, despite being rather thick, had crept past the bobbin wall by capillary action and had got into some of the top layers, this capillary action was not supposed to have happened because of the viscosity.  This had not happened with the previous epoxy used for sealing the ends of the resistor.  I do perform some conditioning of my resistors to relieve winding stress but it is different than other manufacturers do.

This faulty epoxy was used months ago and discontinued as soon as the problem was revealed, no resistors have been made with it since so this discussion about humidity problems is now a moot subject.

Due to characteristics of Evanohm, it is difficult to acquire large diameter wire with very low TCRs, so depending on factors, I may or may not be able to produce a low ohm resistor with a low TCR, this is usually somewhere in the 10 to 20 ohm range, I will tell you if I can't make a certain value.  On the high side, I usually do not make resistors much over 1 Meg, not because I can't do it but the cost of wire makes very large value resistors expensive.  In the past before wire costs went up so much, resistors up to 30 Megs could be made at a reasonable cost but very fine wire exceeds $18,000 / lb. or more, therefore I do not bid on very high values anymore.  All of the resistors I make are within 0±3 PPM/°C or less.

Humidity is not a problem with my resistors, using hermetic resistors is not only a waste of money but will not provide any improvement in the performance of an LTZ1000/A circuit.  Yes film/foil plastic cased resistors do have humidity issues partly because of the resistance element being so thin and tiny water molecules even in small numbers have a significant effect on those tiny traces, wire wounds do not have that problem.  As I mentioned previously, if you are having humidity problems with circuitry because of some components, it would be cheaper to use a parylene coating than buying hermetic resistors.
Title: Re: New Precision Resistors
Post by: 2N3055 on October 21, 2016, 05:25:59 pm
Is there a dip coating that would be alternative to Parylene ?
Title: Re: New Precision Resistors
Post by: lars on October 21, 2016, 06:18:21 pm
Hi Edwin,

Glad you have solved your problem with the epoxy. When did you do this? I had hoped to get a mail as I probably was the one that made you aware of this problem. You say that replacements have been shipped but I have not seen anything, hope it have not got lost, if so it will be the first time from US for me. I had also guessed you would send a mail telling that you shipped the resistors?

How did you test the resistors to be sure they are not humidity sensitive anymore?

I might add that if humidity is a problem with components, including some plastic ICs, keeping the circuits energised 24/7 will minimize or eliminate this problem, also if humidity is generally high most of the time, I would suggest coating the circuit board and components with paralene which is fairly common and relatively cheap, this will stop any humidity from getting into the wrong places and causing problems.

I am not convinced that running plastics ICs 24/7 or at elevated temperatures will solve the problems with humidity sensitivity. Some years ago I tested about thirty pieces of 10V refs in plastic packages to see if they were humidity sensitive. Half were run near room temperature and half were run about 15 degrees Celsius higher by heating in a box that was not tight. All were run 24/7 except some months to see the effect of power down versus 24/7. I run the test a couple of years to see long terms effects also. The humidity was about 20-60%RH in the room. All plastic packages were humidity sensitive about 0.2-0.5ppm/%RH (room) except one that was about 1ppm/%RH. If I remember correct it was AD587, REF102, LT1236, and some MAX and a mixture of DIP and SO-8. I had hoped that the warmed ones should have less humidity sensitivity but that was not the case. My thought afterwards was that even if the relative humidity in the heated box is lower the absolute water content is the same and the variations also the same and that could be the reason for the result.

Another note about plastic packaged ref ICs is that if you have a step in humidity some have a peak indicating more than one humidity sensitivity mechanisms and with opposite polarity cancelling each other in long term more or less but as they probably has different time constants they peak for a step. I once tested six AD587LN with different date codes (mounted on SVR-boards with sockets) and got very different results. One had almost full compensation for slow long-term shifts in humidity but others had mostly one time constant and up to 0.5ppm/%RH just after a week.

Title: Re: New Precision Resistors
Post by: Edwin G. Pettis on October 21, 2016, 06:55:08 pm
PTFE is another coating that is quite effective, check the catalogs of Digi-Key, Mouser, ect, there are quite a few sources for PCB coatings of various types and many of them can seal against humidity.  Some are spray on, others are brushed on.

Humidity and electronic components has been a problem for decades, old carbon comp resistors loves humidity, the equipment was left on 24/7 and was pretty successful at keeping humidity at bay.  More modern electronics do tend to run cooler so getting rid of humidity can require a different tact, if it is a problem, a coating such as I've mentioned does a great job against it.  My resistors construction has changed in some ways which severely limits the action of humidity and the epoxy has been isolated to a high degree from the bulk of the windings.  Short of hermetic, few components are humidity proof, they can be highly resistant, I have not found any significant effects from ambient humidity that would affect the operation of an LTZ circuit.  HP/Agilent/Keysight do not use hermetic resistors or hermetic ICs other than the LTZ in their modules.  You won't find any in a 3458A, it is not necessary.

If you are still worried about humidity, then the cheapest solution is to coat the PCB, hermetic resistors are far too expensive to justify unless you've got a pocket full of cash you want to relieve yourself of for no good reason.  Lars was one of two customers that received the problematic resistors, I've since shipped a lot of resistors since then and there has been no complaints.
Title: Re: New Precision Resistors
Post by: EmmanuelFaure on October 22, 2016, 06:10:02 pm
@ Edwin : For humidity sealing, what do you think about using wax? It has a quite low water vapour permeability, it's very cheap ($10/lbs), and components can be coated easily by dipping.

What about silicones?
Title: Re: New Precision Resistors
Post by: MisterDiodes on October 22, 2016, 10:07:47 pm
Most plastics are water absorbent by nature.  That means all plastic, epoxy, acrylic, urethane, silicone, etc. will absorb water, generally up to about 2% by mass.  Some more or less but never less than about 1.5%.  The type of plastic will change dimension very, very slightly depending on water content.  It is hard to predict what dimension will change on a plastic part without just taking careful measurements of a finished part.

Some material data sheets will show a plastics / epoxies with a low absorption rate - for instance .01% per 24 hours - but this is NOT the same thing as absorption proof.  It'll still absorb water, but take longer to do so.  Always.

Most spray-on / dipped conformal coatings will do a great job at really reducing most humidity effects but aren't 100% humidity proof.  They will really toughen up a PCB against dirt and abuse though, and some coatings are essential for high voltage applications.

The only PCB conformal coatings I know of that are really humidty-proof is a PTFE (Teflon) or Parylene coating applied by a vacuum vapor-deposition process.   This is not a hobbyist project - you send completed, assembled boards out to a coatings lab.  That's what I do for critical projects.  Pricing is not for faint of heart if you're doing one board.  But when you get the boards back they will have very high humidity tolerance.  They will also not be repairable unless you scrape off the coating, and sometimes that's not easy or impossible.  You really can't make any changes to a PCB board with this coating in place, but it's chemically tough.  That's why it's popular for industrial, military and space-flight applications.

The other method is a good air-tight enclosure with seals.  That works very well also and is an EMI / RF shield too.  Less expensive than some coating processes, and your board is repairable.

HOWEVER please note that 3458a's have been built for decades with plastic packages around the LTZ reference, and generally even the film resistors may or may not be in a hermetic package - and it doesn't make a lot of difference.  You still see very low drift rates year per year in almost every case regardless of how the components are packaged..

Why?  Because if the inside of your instrument package is always kept warmer (keep it turned on) it will by definition have a much lower, moderated humidity inside the case than outside - in general humidity effects are reduced or non-existent.  This also makes for much longer lifetimes in general of your IC's also without the thermal cycling of the chip package bonding leads (inside the package) , and keeps your Vrefs at a lower stress level.

Look at Fluke 732's - you keep these always powered on.  Part of the reason is for the Vrefs crystal substrate and resistor stability, part of the reason is to keep temperature constant and humidity in-consequential.  If you power cycle these, you're looking at a long de-stress time due to a variety of factors if you want to get it back to very low yearly drift rate.

My advice:  In general - Keep your precision equipment powered on always, if it designed for that (most are).  The warmer cabinet interior will do a lot to reduce humidity effects, and result in more stable devices.  The more you power cycle, the more you're adding stress to the components inside.  The same goes for any type of stable resistor, especially for precision DC - you really want that power dissipation inside the resistor to come to thermal equilibrium with the surrounding ambient conditions and stabilize - and that holds true for any PWW or Vishay Magical Hermetic or Diffused resistors or ???. 

It can take hours or days for any resistor to become truly stable under bias - and that IS NOT the same thing as doing a quick measure on a DMM. 

NOTE: To measure resistor at different humidity, you place them (under bias) in various saturated salt solution atmospheres.  It's easy to do but important to take your measurement under various actual current flow thru the resistor, with approximately the same thermal flow to ambient temp as your application.  Again, just measuring resistance value (with DMM) while it a certain humidity level won't tell you much, because you need to add the effect of resistor self heating at a realistic current flow - whatever that is per application.  That will give you the TOTAL effect of power dissipation and dimensional change due to humidity.  The problem with some "hermetic" resistors is yes - they have a can and maybe oil - but with even small current flows the power dissipation still causes stress problems on the substrate that you see as noise or resistance fluctuation.  This is why they sometimes don't give you 80X more performance than a $1 resistor - at least not in every case.

The humidity effect is usually the very least of your drift problems - although I have certainly seen this problem and you need to be aware of it.  Tempco, good tight ratio matching, low noise, repeatability and good aging characteristics I have found to be usually more important in the long run.  That's why PWW is a very good choice for DC and low freq - and go to some of the metal films and exotics for higher freq.

Title: Re: New Precision Resistors
Post by: MisterDiodes on October 22, 2016, 10:19:49 pm
...what do you think about using wax? It has a quite low water vapour permeability, it's very cheap ($10/lbs), and components can be coated easily by dipping.

What about silicones?

Wax used to be used sometimes, but it has problems - it never really becomes truly hard, and will still slowly flow year after year.  Especially watch out for hot weather and dust / insect accumulation!

Quality Beeswax can work in a pinch, but petroleum-based waxes tend to outgas, shrink and break down other sensitive parts long term.  Good beeswax can be more expensive than a better quality urethane though.

Other coatings are much better overall.

Title: Re: New Precision Resistors
Post by: acbern on October 22, 2016, 11:37:59 pm
HOWEVER please note that 3458a's have been built for decades with plastic packages around the LTZ reference, and generally even the film resistors may or may not be in a hermetic package - and it doesn't make a lot of difference.  You still see very low drift rates year per year in almost every case regardless of how the components are packaged..

The 3458A uses hermetic components where it is important, i.e. the voltage reference and the 40k reference resistor. The fact that other parts are non-hermetic is simply because their drift effect (e.g. arround the LTZ) does not contribute much, everything else is derived by digital means. There is no way that Parylene would be equivalent to a hermetic housing. Parylene is essentially used to avoid the effects of direct condensation and the fact that it is applied under vacuum avoids potential workmanship issues which you do not want in spacecraft electronics (e.g. someone erroneously covering a wire before spraying a board and thus preventing coating to cover it). Wirewound resistors, because they simply use more material, have a much lower drift (humidity reacting with the wires material) than e.g. non-hermetic foil resistors (epoxy issues is a different story, same with PCB materials becomming conductive). Non-hermetic foil resistors may drift in the 10ppm range pa due to humidity influences, still making them a solid solution for many precision applications. Beyond that, wirewould or hermetic parts (but not conformal coating) is needed.
Title: Re: New Precision Resistors
Post by: MisterDiodes on October 23, 2016, 05:30:31 pm
AS I noted before (I think - at least I meant to) - Parylene + sealed outer housing is considered equivalent to a hermetic package, and is accepted by military and DOD as such.  As far as Parylene by itself:  It will penetrate into very small gaps, less than 1mil wide.  In fact in some instances much smaller.  Since it is applied at room temperature, any effect on the circuit is minimal in terms of application process.

So in effect if you design your circuit for the coating process, it is very effective.  I.e. thru hole parts are excellent candidates.

I have never seen a Parylene-coated circuit respond by any horrendous amount to any real humidity test, even after a 24 month exposure test.  I have also seen the "hermetic can" on an IC add all sorts of thermal stress issues that cause nothing but problems, depending on thermal layout.  In the end, you use whatever process works - there is no one correct answer for every case.

I have also seen a regular dip-coating process work very well to reduce humidity effects down to a reasonable value.  From a profitable business standpoint, you use whatever process will get you the accuracy you need over time, expected longevity, etc.  Over-doing it is not going gain any real useful end result, and wastes time and money for everyone.  If customer needs +-25ppm per year drift, you shoot for +-15 Or 20ppm / year drift to stay in spec - and use the appropriate process.

Circuits like LTZ and such are very tolerant of hermetic or non-hermetic resistors - remember that stability built in to the circuit design itself.  You will see no or very little improvement spending a lot of money on LTZ resistors - and from a profit standpoint it makes absolutely no sense to over-spec the resistors.

NOW:  What you DO have to watch out for is something like a simple 7V to 10V op-amp boost circuit (if you're doing it that way).  You need a very, very stable resistor network in ratio mode, and here again the mechanicals and thermal layout becomes very important.  Any variation of resistor ratio is going to be gained up by the circuit.  Here again, hermetics may not be the best in every case, and well designed PWW's can work just as well and save a lot of money.  Depends on final application.  Sometimes a custom diffused or film resistor pair is called for to reduce size, digital methods, etc.  but at the cost of more noise.  Always a trade off.

RE 3458a - No, not all 3458a's are built the same.  Earlier models used different techniques for the ref resistor - typically a can used but considered what would be a commercial low-drift resistor sealed in a mechanical enclosure by HP - which is exactly what I was talking about above by adding your own enclosure.  That's the spirit of what I meant.

I know this because one of the projects I was working on as a young engineer was helping HP develop ultra-stable diffused-resistor building techniques for various parts of 3458a's and other calibration / test equipment at HP's test equipment division (when that existed).  Spent lots of time there.  Let's just say there are other technologies involved that are so expensive - that it becomes cost effective to just put a can around an off-the-shelf PWW or film resistor <Laughing> with a known atmosphere sealed inside.

In other words - if you do need a hermetic can, that is something that can be done by the end assembly process also.   

Point is:  Build your circuit, let it stabilize, and the best way to check for humidity effect is to run your circuit at expected operating temperature and current IN a thermal chamber WITH a saturated salt - controlled known humidity.  For instance we test at roughly 10%, 50%, 75% and 85%.  Sometimes it's a 24 hr exposure, sometimes 30 days, sometimes 90 days, whatever spec is.

Make corrections where you need to as you learn how your circuit works in real life.  You'll be surprised at what you don't have to do - and sometimes you'll certainly get an effect you weren't expecting that will need attention.  Sometimes what you thought was going to be a humidity effect really is solved by a better thermal layout, etc.

Running this test is far, far different from just testing a resistor on a DMM by itself.  You get a better picture of what's happening when real resistor currents, real self heating and real thermal flows are involved.

Title: Re: New Precision Resistors
Post by: EmmanuelFaure on October 23, 2016, 10:06:28 pm
This is very interresting., thanks you!

Could you tell me/us what's an approximate price for parylene coating? (Like $/sq inch of PCB).
Title: Re: New Precision Resistors
Post by: MisterDiodes on October 24, 2016, 03:19:47 pm
You'll have to contact your coatings lab for a quote - there are several online.  Be prepared to have a dummy sample board ready to send in to get a quote on, or 3D drawing.  Doing just one board might be cost prohibitive depending on application.  Usually the Parylene 'dimer' (the source of material that will be the source of the coating) will run anywhere from USD$500 to over $3500 per pound depending on qualities you need, but you pay for a small amount of that material transferred to your boards.  It also depends on if they run your boards with other customers, etc.  The lab will usually perform a wash in de-ionized water (they keep washing until no contaminates in rinse water - resistivity back in the 10s of G-Ohm range)) and then bake dry for several hours just before coating - so it will be in a very, very dry and clean condition.  You have to be careful to mask off what you don't want covered - because it can be very difficult to remove.  The larger the board and more surface area, the more it costs. The coating is done at room temperature, or slightly above.

You will select from several grades of the dimer - for instance if you need good crevice penetration, high dielectric strength, surface finish, etc.  There are lots of of options.

OR you can do a sealed enclosure for your board, or have your critical chip die mounted in a hermetic package.  Usually when you discuss with the end customer what this costs, and prove to them that Parylene can maybe work just as well, they will go with just the coating in the end.  Some customers (NASA, cal labs), university labs, etc) will just spec hermetic package throughout - and that's what you do, because that's what they pay for.  You look at individual packaging vs. a welded enclosure for the whole circuit.

OR - Just do a quality dip / spray coating and your circuit will still be MUCH more humidity-resistant than without it, even though not perfect.  For a lot of applications this is more than enough!

OR follow Pettis' advice and run your circuit 24/7 and keep the enclosure warm.  For hobby - PPM entertainment application that may be all you need.  Most of the time - unless your weather is changing 5% to 90% humidity a lot - that's fine also. 

It just depends on what you really need.
Title: Re: New Precision Resistors
Post by: Cerebus on October 26, 2016, 12:06:46 am
The coating is done at room temperature, or slightly above.

You've already mentioned that it's done under vacuum (or at least, under very low pressure) but I thought it was worth underlining as sometimes people don't immediately grasp the implications of this.

The first is that this will stress any hermetic packages you have on the board. That's probably not an issue but one ought to just double check just in case something is going to pop under vacuum.

The second is that many materials will off-gas under vacuum. Apparently solid materials  start turning into vapour. This can poison the coating process and can, of course, damage the off-gassing component. Again, you need to check  - component material specs in this case.
Title: Re: New Precision Resistors
Post by: MisterDiodes on October 26, 2016, 04:18:53 pm
I have never seen a vacuum - for a short period - have any effect on the usual components or hermetic package - these are usually sealed with a low pressure nitrogen, helium purge or vac inside anyway.  You DO have to be careful about say a mechanical assembly where if you pull the vacuum too quickly you might pull a lubricant out where you don't want.  Watch for this for instance on Bourns PWW pots around the shaft sea if you get the military type with a lightly greased shaft seal.  Treat them with care and they are fine.  Various processes don't always require a hard vacuum, it just depends.

These issues will be discussed at the coatings lab, and that's why the cost is variable.

You treat a coating process as a stress riser on your circuit - just as if you added a welded can - that will require a run-in stress-relief period.  So you get your coatings done early before you start your long thermal burn-in on critical circuits.

Generally, for very critical applications, you can spec a post-process SEM photo inspection to verify no pinholes, proper thickness, etc. We have even checked for water absorption by running the finished product underwater for months - and see a virtual zero mass gain when weighed out afterwards.

It works well, and as I said:  In many applications it is considered a virtual hermetic seal - sometimes this is done BEFORE the outer metal can is welded on for extra protection.
Title: Re: New Precision Resistors
Post by: MisterDiodes on October 26, 2016, 04:34:41 pm
I have one more thought on resistors / humidity protection - and this is a cheap, practical and classic way to do it - just something to add to your toolbox when working with possible humidity-sensitive stuff:

Toss in a desiccant packet into say your precision resistor enclosure and put the covers back on.

For instance, this was done on a lot of Precision resistance decade dividers.  We have a few 7-decade dividers built this way - circa late 60's.  All PWW resistors, none in hermetic packages, and these boxes are checked for NIST cal occasionally - and they are always within their expected +- 0.3ppm range or better - and have been perfectly fine for decades. 

BUT there is a purpose-built screen holder inside the cabinet to hold a 4oz desiccant packet.  We use the type that turns color when it's time to change, and we change it out once or twice a year.  You can bake the packet out in the oven and re-use it.  There really isn't a lot of airflow in and out of the cabinet, and the humidity inside is usually right around <30% or so.

While you change out the packet you might hit a switch contact with DeOxit or grease a shaft  if required but very rare and very sparingly.  That's it.

So that's another simple, practical approach that will work fine for some items that don't run warm, and we copy the idea for other boxes around the lab.
Title: Re: New Precision Resistors
Post by: branadic on October 26, 2016, 05:57:47 pm
I wonder if some Vishay VAR resistors with feedthrough capacitors (4 for 4W connection) in a copper case and the case filled with Galden could give some stable resistor without humidity problems. I think I need to check this some time...
Title: Re: New Precision Resistors
Post by: Andreas on October 26, 2016, 07:47:05 pm
Hello branadic,

according to my measurements they have a hysteresis that is not better (even worse) than epoxy-packaged resistors. (

It seems there is some paint/epoxy over the resistor which should be removed before packaging.

with best regards

Title: Re: New Precision Resistors
Post by: branadic on October 26, 2016, 08:01:21 pm
Thanks Andreas, have you ever checked what happens if you put them into some hydrophobic fluid?

If it is an epoxy only mechanical removing is possible, which should be avoided. Paint could probably be removed with some solvent, but could also dissolve the glue between ceramic substrate and foil?
Title: Re: New Precision Resistors
Post by: JS on October 27, 2016, 05:57:54 pm
  Just as a thought exercise, what's the case of resistors living under oil, like the fluke 720 first decade. I guess the humidity shouldn't get to the resistors and now instead of hermetic packaging only something able to contain the oil, which is much easier to stop than humidity in the air. Am I missing something? I'd still guess fluke did it just, or mainly, for thermal tracking there.

Title: Re: New Precision Resistors
Post by: babysitter on October 28, 2016, 08:10:30 am
Hi all,

i might have the chance to get stuff thru parylene sometimes in the future, maybe I can give a ride as a secondary payload so you can save setup costs.

@Edwin: Around what time did you start using the misbehaving epoxy resin and when did it phase out?

(No such complaints from my co-worker using those we bought, but the cellar the machine is running at has a quuite stable humidity and tempure.

best regards, Hendrik
Title: Re: New Precision Resistors
Post by: plesa on October 28, 2016, 09:08:36 am
And what about using e.g 3M/Scotch spray (
We were using it in humid environment for low leakage application.
Title: Re: New Precision Resistors
Post by: Edwin G. Pettis on October 28, 2016, 05:34:57 pm
Babysitter:  After acquiring and testing here where the humidity is usually in the range <10% to <35%, two orders ended up being built in August and September of 2015.  Orders in October were back to a previous resin which, while somewhat difficult to work with did not cause the humidity problem.  I might add that the resin performed mostly as presented in the data sheet except for the humidity absorption and a rather curious 'spongy' characteristic.  None of the previous resins had such a characteristic and perhaps this may have had something to do with absorption in high humidity.  After feedback from one of the two customers I simulated a higher humidity environment and found that the resin did indeed absorb humidity to the point of swelling if exposed long enough, while heating the resistors up and drying them out would return them to near original value, that did not stop absorption again under high humidity and apparently these two customers live in significantly higher humidity than I do.  In the mean time, I began looking for a new resin which is quite time consuming, trial and error does not go quickly, while I was filling a few 'special' orders with the previous resin, I did not want to do more orders until I had found a better resin.  Currently the resin I am using is working well but the handling of it has slowed down production, there has been no complaints.

Plesa:  In many cases you do not need to have a 'perfect' barrier against humidity as its effects are usually quite small, there are many dip and spray on coatings which are quite good at holding humidity at bay and are very reasonable in cost.  There should be something about humidity in the data sheet if the product is intended to help block humidity.  Parylene, PTFE and hermetic cans are used where the cost is justified, usually in harsh environments such as in industrial or aerospace.  To get that last tiny bit of humidity reduction can be comparatively expensive, hermetic sealing is labor intensive, I do not know of any automated method, I've done quite a bit of hermetic sealing myself over the years and even the big customers don't use it where it isn't absolutely needed.  As MisterDiodes mentioned, a bag of desiccant is cheap and effective, put in a small enclosed enclosure keeps humidity at bay.

JS:  Thermal tracking is a complex problem and the tighter the tracking requirements the more difficult it gets.  Simply enclosing resistors in an oil filled package does not always get the job done, it certainly helps and even oil can absorb some humidity depending on the type of oil and encasement.  Thermal coupling of resistors is not easy, even if they are on the same substrate (film, foil), in many instances the customer is looking for near perfect coupling which cannot be achieved, fractional PPM TCR tracking can be achieved but with difficulty and expense (nothing good comes for free unfortunately).  Oil is a decent thermal conductive median but there is a fairly long time constant involved to come to equilibrium and if the environment is changing that time constant may be a problem under certain circumstances.  If you need very good tracking but not fractional PPM, copper tape wrapped around the resistors can do a good job at fairly low expense.  There are few options available when going after <1 PPM/°C tracking, oil is one of them, other options are not production friendly.....they involve labor or other special procedures.  Various methods over the years have been tried with variable results, such as the resistor chips in the HP DVMs with limited success.

In short, trying to get to under a couple of PPM in any spec is going to get expensive and difficult, a lot depends on the resistors and the customer, just specifying a resistor ratio with tight tracking without considering the consequences (and some trial and error possibly) will very likely end in difficulties galore and friction between manufacturer and customer.  Every resistor house I'm familiar with has run into those problems with customers, even Vishay has failed with all their resources in some instances.  That magic resistor ratio for boosting output voltage to 10V with very tight tolerances is going to be difficult and fraught with problems, you can't just solder something in and expect near perfect performance, circuitry doesn't work like that.
Title: Re: New Precision Resistors
Post by: branadic on November 01, 2016, 12:08:59 pm
If it is an epoxy only mechanical removing is possible, which should be avoided. Paint could probably be removed with some solvent, but could also dissolve the glue between ceramic substrate and foil?

What ever it is, it seems to withstand IPA and Aceton.

So here are my thoughts:

Hermetic cases are mostly made of ceramic, metal or glas, plated with gold. Because the gold layer used to make it conducting is such thin CTE mismatch is not that dominating problem. However, these are no materials everyone can handle at home.
Brass is, as far as I was told, no good idea, as it is a wrought alloy and therefor not hermetically tight, while it is easier to process.
Because of its high heat capacity aluminium would be a good material. On the other hand, if you want to use feedthrough capacitors it's hard making aluminium solderable. Plasma treatment and PVD could be a possible solution, to deposit a Au layer. But be aware, there are also a lot of aluminium wrought alloys out there, having the same problem like brass.
A case made out of copper is a cheap and good solution, using soldering technics to close and seal it.

Parylene is a bad idea for what I know, because of CTE mismatch the coating will get micro cracks so again water vapour can enter.
There are other coatings that offer hydrophobic properties, but at the end of the day you only change time constant. The same with wax, while it's easy to handle it will not avoid water vapour in any case.
Silicone oil is also not the first choice. Silicone can absorb water and it's still the devil in most production lines.
Galden prevents humidity and because of its good heat conductivity keeps everything inside a case at the same temperature level. There are different types available, depending on your specifications, but you want to make sure it's liquid over the entire temperature range in your application.