Author Topic: 10k precision resistor project  (Read 7984 times)

0 Members and 1 Guest are viewing this topic.

MasterTech

• Frequent Contributor
• Posts: 860
• Country:
10k precision resistor project
« on: February 22, 2018, 06:49:20 pm »
I'm in the mood of experimenting with the build of a 10k precision resistor. After reading the posts here about standard resistors I've decided to try my own design given the constraint that I'll rely on commercial items only.

This is the design, it will be inside a small metallic box filled with oil (except for the potentiomenter) with pomona 3770 binding posts.
The composite adjusting resistors vary between 9.1 and 10.7 ohm, so each degree of turn of the potentiometer (26 turns) accounts for around 0.2mOhms of change (given a linear curve which is not)

Let me know what you think or even if it is an idea worth trying or not
« Last Edit: February 22, 2018, 09:03:34 pm by MasterTech »

The following users thanked this post: doktor pyta

The Soulman

• Frequent Contributor
• Posts: 624
• Country:
• The sky is the limit!
Re: 10k precision resistor project
« Reply #1 on: February 22, 2018, 07:45:48 pm »
The trim range is to big, it will be a pain to adjust to read 10.000.xx? on your favorite meter, assuming that is the plan.
After that you're depending heavily on the stability of the trimmer and its accompanying resistors.

The stability of the main 10K resistor is the most important, others can be lower tolerance/higher tempco depending on how much the attribute to the end value, you can calculate what their influence would be.

I'd probably go for a low value r in series with the main 10K and use a high value r in series with a trim pot parallel to the main 10K.
Calculate for a minimum (leave some margin..) trim range.

Edit: It would be helpful to know it's intended purpose and how you are going to adjust it.
« Last Edit: February 22, 2018, 07:48:54 pm by The Soulman »

MasterTech

• Frequent Contributor
• Posts: 860
• Country:
Re: 10k precision resistor project
« Reply #2 on: February 22, 2018, 09:14:38 pm »
Reducing the trim range is something I still have to look. However note my comment about "given the constraint that I'll rely on commercial items only". Once you come up with a topology, like the one you suggested which is fine, and start crunching numbers you'll realize you need exotic values like 0.33 or 9.9 which are not readily available, and I do not plan to select among a purchase of 50 expensive resistors.

I plan to adjust it with a 3458A and later probably send it to someone to have it checked/retrimmed
« Last Edit: February 22, 2018, 09:26:43 pm by MasterTech »

RandallMcRee

• Frequent Contributor
• Posts: 322
• Country:
Re: 10k precision resistor project
« Reply #3 on: February 23, 2018, 04:33:37 pm »

I have been measuring a lot of 10K Z-foil resistors lately (VAR and S102K, not the Y series but same technology, similar datasheet). 6.5 digit measurements only, but tcr and hysteresis are readily visible.

So, yeah, 0.2ppm TCR typical is not that typical. Temperature hysteresis is real.

Personally, I would investigate PWWs like Ultrohm. For the same price you can buy several and pick a winner.

See Andreas' thread for measurements which, I believe, confirm my observations.

cellularmitosis

• Supporter
• Posts: 1046
• Country:
Re: 10k precision resistor project
« Reply #4 on: February 23, 2018, 04:51:49 pm »
I am just about to post my own results in Andreas thread.  The summary is that yes, 0.2ppm/C is not what I would call "typical"

LTZs: KX FX MX CX PX Frank A9 QX

MasterTech

• Frequent Contributor
• Posts: 860
• Country:
Re: 10k precision resistor project
« Reply #5 on: February 23, 2018, 04:57:57 pm »
So, yeah, 0.2ppm TCR typical is not that typical. Temperature hysteresis is real
hmm, because of this I was thinking of ovenizing the whole thing to 40C. It is something not common in standard resistors, but I prefer short term stability and would be interesting to see the 1-year drift for this kind of setup

cellularmitosis

• Supporter
• Posts: 1046
• Country:
Re: 10k precision resistor project
« Reply #6 on: February 23, 2018, 06:02:30 pm »
Drift may be higher if you ovenize at a higher temperature.  I'm planning to ovenize my resistance transfer standard as well, but at 25C.
LTZs: KX FX MX CX PX Frank A9 QX

MasterTech

• Frequent Contributor
• Posts: 860
• Country:
Re: 10k precision resistor project
« Reply #7 on: February 23, 2018, 06:13:31 pm »
Drift may be higher if you ovenize at a higher temperature.  I'm planning to ovenize my resistance transfer standard as well, but at 25C.
40C is just a bit higher than the max temperatures I expect where I live, thats why.

The following users thanked this post: cellularmitosis

MasterTech

• Frequent Contributor
• Posts: 860
• Country:
Re: 10k precision resistor project
« Reply #8 on: February 23, 2018, 09:26:09 pm »
I'll think I'll change the circuit for this one. This way the whole 26 turns of the 10ohm pot will create a change of 0.416 ohms, or 35mOhms per turn (considering its linear which is not)

Besides that I'll think i'll put the potentiomenter also inside the oil box with means to be adjusted from the outside

Edit: the schematic had a mistake
« Last Edit: February 23, 2018, 09:37:46 pm by MasterTech »

• Super Contributor
• Posts: 1058
• Country:
Re: 10k precision resistor project
« Reply #9 on: February 24, 2018, 12:31:10 am »
That just seems way too complicated when you consider all the internal terminations and whatever problems get added by the physical construction. I'd probably buy as many good 10k parts as I could reasonably afford, then select one that's just a tiny bit high, and apply a pad to that.

The following users thanked this post: TiN, 2N3055

MasterTech

• Frequent Contributor
• Posts: 860
• Country:
Re: 10k precision resistor project
« Reply #10 on: February 24, 2018, 12:56:08 am »
I'd probably buy as many good 10k parts as I could reasonably afford, then select one that's just a tiny bit high, and apply a pad to that.
That’s boring and same ol same ol, where’s all the fun of building something new then, you and Tin are no fun
« Last Edit: February 24, 2018, 06:17:50 am by MasterTech »

eurofox

• Supporter
• Posts: 835
• Country:
Re: 10k precision resistor project
« Reply #11 on: February 24, 2018, 01:28:46 am »
That just seems way too complicated when you consider all the internal terminations and whatever problems get added by the physical construction. I'd probably buy as many good 10k parts as I could reasonably afford, then select one that's just a tiny bit high, and apply a pad to that.

This is my option as well, I got Vishay 10K 0.005% 2PPM for 7 Euros per unit.
eurofox

quarks

• Frequent Contributor
• Posts: 821
• Country:
Re: 10k precision resistor project
« Reply #12 on: February 24, 2018, 10:40:00 pm »
It is always good to experiment with new designs, but why tell a good meant suggestion boring?

MasterTech

• Frequent Contributor
• Posts: 860
• Country:
Re: 10k precision resistor project
« Reply #13 on: February 24, 2018, 10:59:19 pm »
It is always good to experiment with new designs, but why tell a good meant suggestion boring?

Because that approach lacks my interest and greatly deviates from the objective of this project?

Vtile

• Frequent Contributor
• Posts: 972
• Country:
• Ingineer
Re: 10k precision resistor project
« Reply #14 on: February 25, 2018, 01:01:06 am »
May I ask why you use series pot?

At parallel configuration:
If your fixed value is A and the pot value is 100*A your drop area stays inside 2.5% of the fixed value in aprox. 65% of the control window (pot turns). You might want to calculate which one is better for your application series or parallel pot?
« Last Edit: February 25, 2018, 01:12:09 am by Vtile »

MasterTech

• Frequent Contributor
• Posts: 860
• Country:
Re: 10k precision resistor project
« Reply #15 on: February 25, 2018, 02:05:50 am »
I've simplified the circuit to eliminate a few resistors:

May I ask why you use series pot?

It mainly has to do with the availability of low TCR good quality precision trimmers, they are made for low Rs, like the Accutrim series from Vishay (attached PDF). Maybe there are others but with Mouser being my preferred distributor this is what they have.
Besides that the series configuration in this case seems ok, the series adjustment varies from 0.5 to 0.916:

The area of interest for adjustment of the composite series part is at 0.6667 ohm approximately (ie when the total resistance amounts exactly to 10k). That is not very convenient because by the way because the slope there is high. Still a 15ppm increase in resistance of the 26 turn trim at that point creates an overall change of 1.7uohms at the output terminals. So the sensitivity to its TCR is quite good.
Anyway I'll look If I can shift the area of interest to a place with a lower slope

EDIT: just noticed that lowering the 50M resistor to 20M, the adjusting point moves to 0.833 which is better.
« Last Edit: February 25, 2018, 02:23:25 am by MasterTech »

Vtile

• Frequent Contributor
• Posts: 972
• Country:
• Ingineer
Re: 10k precision resistor project
« Reply #16 on: February 25, 2018, 02:24:18 am »
Have you noticed this piece of information from accutrim datasheet: "through the wiper ± 50 ppm/°C"

MasterTech

• Frequent Contributor
• Posts: 860
• Country:
Re: 10k precision resistor project
« Reply #17 on: February 25, 2018, 02:33:54 am »
Have you noticed this piece of information from accutrim datasheet: "through the wiper ± 50 ppm/°C"
Yep, I'm still taking the 15ppm because those TCRs are stated for -55 to +125 and I expect them to be much better around 25º and also I've decided to place the trim inside the oil box (trimmable from the outside)
And still, this is the best trim I have right now

The following users thanked this post: Vtile

Svgeesus

• Contributor
• Posts: 36
• Country:
Re: 10k precision resistor project
« Reply #18 on: February 25, 2018, 03:07:45 am »
That just seems way too complicated when you consider all the internal terminations and whatever problems get added by the physical construction. I'd probably buy as many good 10k parts as I could reasonably afford, then select one that's just a tiny bit high, and apply a pad to that.

I understand the point I believe you are making about minimizing the effect of all those copper to solder to copper connections in the circuit.

cellularmitosis

• Supporter
• Posts: 1046
• Country:
Re: 10k precision resistor project
« Reply #19 on: February 25, 2018, 04:06:13 am »
LTZs: KX FX MX CX PX Frank A9 QX

The following users thanked this post: MasterTech

MasterTech

• Frequent Contributor
• Posts: 860
• Country:
Re: 10k precision resistor project
« Reply #20 on: February 25, 2018, 04:21:45 am »
Your simplified design is similar to the SRX-10k.

Its seems so, and if I had a custom made 9999.x resistor I could eliminate the parallel with 120Mohms, but commercial resistors are my constraint

TiN

• Super Contributor
• Posts: 3902
• Country:
• xDevs.com/live - 24/7 lab feed
Re: 10k precision resistor project
« Reply #21 on: February 25, 2018, 04:42:14 am »
TiN is boring?
Try to build 5700A from bits and pieces, then you call me boring

I'll just leave this here:

.
YouTube | Chat room | Live-cam | Have documentation to share? Upload here! No size limit, firmware dumps, photos.

cellularmitosis

• Supporter
• Posts: 1046
• Country:
Re: 10k precision resistor project
« Reply #22 on: February 25, 2018, 04:42:36 am »
MasterTech, I'm not sure I understand.  Your 10k Vishay resistors are the same as 9.999k Vishay resistors ordered from Texas Components (which can deliver in 2-3 weeks).  They are both "commercial".  They are both custom laser-trimmed to meet a spec.  The only difference is that with the 10k, they decided what value to laser-trim to, instead of asking which value you wanted.

In other words, they are all "custom".
LTZs: KX FX MX CX PX Frank A9 QX

MasterTech

• Frequent Contributor
• Posts: 860
• Country:
Re: 10k precision resistor project
« Reply #23 on: February 25, 2018, 04:57:25 am »
MasterTech, I'm not sure I understand.  Your 10k Vishay resistors are the same as 9.999k Vishay resistors ordered from Texas Components (which can deliver in 2-3 weeks).  They are both "commercial".  They are both custom laser-trimmed to meet a spec.  The only difference is that with the 10k, they decided what value to laser-trim to, instead of asking which value you wanted.

In other words, they are all "custom".
U are right, what I meant is that although all are custom manufactured some are stocked and sold thru distributors and others needs to be backordered. Basically I need a 9999.2, 0.2ppm resistor. For a production unit it may be worth it to order one, but for a one-time project I'm not sure....

MasterTech

• Frequent Contributor
• Posts: 860
• Country:
Re: 10k precision resistor project
« Reply #24 on: February 25, 2018, 04:58:34 am »
TiN is boring?
Try to build 5700A from bits and pieces, then you call me boring
I knew you'd bite

The following users thanked this post: TiN

Smf