Author Topic: USA calibration club  (Read 66302 times)

0 Members and 1 Guest are viewing this topic.

Offline SirAlucard

  • Regular Contributor
  • *
  • Posts: 57
  • Country: us
Re: USA calibration club
« Reply #550 on: February 24, 2018, 08:41:49 am »
Hey guys! So I've only currently have two DMMs a Cheap-O Innova 3300, and a AMPROBE AM-520, obviously I can't yet afford anything better. However I am a bit of a stickler when it comes to accuracy. Is it worth trying to calibrate these to be as accurate as possible? Or just wait until I get something better and just use these as is.
 

Offline cellularmitosis

  • Supporter
  • ****
  • Posts: 1029
  • Country: us
Re: USA calibration club
« Reply #551 on: February 24, 2018, 01:48:40 pm »
Hey there!

From a financial perspective, it definitely wouldn't make sense to pay a cal shop to calibrate these (that would be at least $200 per meter).

But the next time we start up a round of the cal club, you are welcome to join in!  Generally you'll be just paying for one leg of shipping a small transfer standard.  It looks like one is a 2,000-count meter and the other a 4,000-count, so anyone who has had a recent calibration would more than meet any uncertainty requirements to say if your meter was "bang on".
LTZs: KX FX MX CX PX Frank A9 QX
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 2590
  • Country: us
Re: USA calibration club
« Reply #552 on: February 24, 2018, 02:02:50 pm »
Hey guys! So I've only currently have two DMMs a Cheap-O Innova 3300, and a AMPROBE AM-520, obviously I can't yet afford anything better. However I am a bit of a stickler when it comes to accuracy. Is it worth trying to calibrate these to be as accurate as possible? Or just wait until I get something better and just use these as is.

You might want to think hard before you enter these waters.  Others have found this an expensive habit to support. 

Fortunately, the first steps are free.  You say you are a stickler for accuracy.  Have you fully defined what you mean by that?  If your meters meet their specification will you be happy?  Do you understand how the specifications are written and what they mean over the ranges of the meter?  Do you want each displayed digit of the meter to be correct?  Under what operating conditions?  Will performance verification at a small number of operating points be satisfactory to you?  What do you mean by correct?  Most of these questions don't have absolute answers, they are mostly determined by your personal needs and desires.  After you have answered these and other questions you will be better prepared to go further down this road.

If you do you will find much friendly support on this and other calibration threads.
 

Offline SirAlucard

  • Regular Contributor
  • *
  • Posts: 57
  • Country: us
Re: USA calibration club
« Reply #553 on: February 25, 2018, 06:38:18 pm »
Hey guys! So I've only currently have two DMMs a Cheap-O Innova 3300, and a AMPROBE AM-520, obviously I can't yet afford anything better. However I am a bit of a stickler when it comes to accuracy. Is it worth trying to calibrate these to be as accurate as possible? Or just wait until I get something better and just use these as is.

You might want to think hard before you enter these waters.  Others have found this an expensive habit to support. 

Fortunately, the first steps are free.  You say you are a stickler for accuracy.  Have you fully defined what you mean by that?  If your meters meet their specification will you be happy?  Do you understand how the specifications are written and what they mean over the ranges of the meter?  Do you want each displayed digit of the meter to be correct?  Under what operating conditions?  Will performance verification at a small number of operating points be satisfactory to you?  What do you mean by correct?  Most of these questions don't have absolute answers, they are mostly determined by your personal needs and desires.  After you have answered these and other questions you will be better prepared to go further down this road.

If you do you will find much friendly support on this and other calibration threads.

Lets just say at this point I'd be extremely happy in knowing that the readouts of my DMMs are showing me what I should be seeing. So if a 5.204 volt source where being measured, it'd come out as 5.204. The Innova 3300 has been rather old, so I calibrated it to the AM-520 reading because both did not want to agree, they were separated by a couple hundred millivolts. However I'd sleep better knowing that if read by accurate DMMs they'd all give the same output.
 

Offline vindoline

  • Regular Contributor
  • *
  • Posts: 209
  • Country: us
Re: USA calibration club
« Reply #554 on: February 26, 2018, 04:00:57 am »

Seems like I won an auction for 10 VRE3050 voltage references. I want two. The rest are available to club members for what I paid plus shipping.
https://www.ebay.com/itm/132499240091

Let me know.

Randy

Hi Randy, any update on the VRE3050's?  Thanks.

-John
 

Offline RandallMcRee

  • Regular Contributor
  • *
  • Posts: 222
  • Country: us
Re: USA calibration club
« Reply #555 on: February 26, 2018, 04:28:16 am »
Sure.  I have received several emails from an intermediate shipping agency saying “on the way”, “don’t panic” and “still on the way”

Randall


Sent from my iPhone using Tapatalk
 
The following users thanked this post: vindoline

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 2590
  • Country: us
Re: USA calibration club
« Reply #556 on: February 26, 2018, 06:08:06 am »
Hey guys! So I've only currently have two DMMs a Cheap-O Innova 3300, and a AMPROBE AM-520, obviously I can't yet afford anything better. However I am a bit of a stickler when it comes to accuracy. Is it worth trying to calibrate these to be as accurate as possible? Or just wait until I get something better and just use these as is.

You might want to think hard before you enter these waters.  Others have found this an expensive habit to support. 

Fortunately, the first steps are free.  You say you are a stickler for accuracy.  Have you fully defined what you mean by that?  If your meters meet their specification will you be happy?  Do you understand how the specifications are written and what they mean over the ranges of the meter?  Do you want each displayed digit of the meter to be correct?  Under what operating conditions?  Will performance verification at a small number of operating points be satisfactory to you?  What do you mean by correct?  Most of these questions don't have absolute answers, they are mostly determined by your personal needs and desires.  After you have answered these and other questions you will be better prepared to go further down this road.

If you do you will find much friendly support on this and other calibration threads.

Lets just say at this point I'd be extremely happy in knowing that the readouts of my DMMs are showing me what I should be seeing. So if a 5.204 volt source where being measured, it'd come out as 5.204. The Innova 3300 has been rather old, so I calibrated it to the AM-520 reading because both did not want to agree, they were separated by a couple hundred millivolts. However I'd sleep better knowing that if read by accurate DMMs they'd all give the same output.

Your AM-520 is specified to be accurate to +/- 1% and 3 LSD.  So when hooked to your 5.204 volt source and reading plus or minus 104 millivolts is "correct" (with a little more allowance for LSD).  In addition, they probably didn't work to get drift and temperature sensitivity to values that are factors smaller than this, so even if you adjusted your AM-520 to read exactly 5.204 with a source that is somehow know to be "exactly" 5.204, you couldn't realistically expect it to read that value next week or even tomorrow morning.   Of course you might have gotten lucky and your AM-520 may have zero temperature coefficient and zero drift.  But until you have a long series of measurements of a known stable source you can't know that.

I didn't look up the specs for your other meter, but 3 to 5 percent accuracy allowances are not uncommon, frequently with the added caveat that it is 3 or 5 percent of full scale.  It is entirely possible that your two meters can disagree by a couple hundred millivolts and both be within specification. 

This is the kind of thinking I was recommending you go through before you dive into this pool.
 

Offline RandallMcRee

  • Regular Contributor
  • *
  • Posts: 222
  • Country: us
Re: USA calibration club
« Reply #557 on: February 26, 2018, 06:18:28 am »

Quote
Sure.  I have received several emails from an intermediate shipping agency saying “on the way”, “don’t panic” and “still on the way”

Randall

Sorry--that was a little glib. Here's the tracking info. Hopefully, LAX means LA international which is not *too* far. Not compared to the UK anyway.
This is the first time I've looked at the tracking info since I don't believe in getting my expectations up. Often, the package sits at my local post office for a day or more.

 

Offline RandallMcRee

  • Regular Contributor
  • *
  • Posts: 222
  • Country: us
Re: USA calibration club
« Reply #558 on: February 28, 2018, 01:34:25 pm »

The VRE3050Bs have arrived in California. Ready to be sent out to you all...

I sent a PM to those that asked for them...if you have not received a PM from me--but asked for a unit--please send me a PM.

 
The following users thanked this post: cellularmitosis

Offline cellularmitosis

  • Supporter
  • ****
  • Posts: 1029
  • Country: us
Re: USA calibration club
« Reply #559 on: February 28, 2018, 03:53:45 pm »
Would anyone be interested in a simple VRE3050 PCB via OSHPark?
LTZs: KX FX MX CX PX Frank A9 QX
 

Online Andreas

  • Super Contributor
  • ***
  • Posts: 2048
  • Country: de
Re: USA calibration club
« Reply #560 on: February 28, 2018, 06:04:22 pm »
Hello,

> simple

I would at least look that I have a stress-less mounting. (perhaps on the polyimide part of a flex pcb)
I never thought that with a ceramic package I would get 200uV change with mechanical stress.
(tested with SMD-Version VRE3050AS)

with best regards

Andreas
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 3529
  • Country: tw
  • xDevs.com
    • xDevs.com
Re: USA calibration club
« Reply #561 on: February 28, 2018, 08:50:09 pm »
I never thought that with a ceramic package I would get 200uV change with mechanical stress.

Why not? Ceramic can be a great piezo-electric resonator ;)
xDevs.com YouTube | Have test gear documentation to share? Upload here! No size limits, firmware dumps and teardown photos welcome.
 

Offline Conrad Hoffman

  • Frequent Contributor
  • **
  • Posts: 882
  • Country: us
    • The Messy Basement
Re: USA calibration club
« Reply #562 on: February 28, 2018, 10:53:14 pm »
I'd hope they're using alumina, which AFAIK has no piezoelectric effect. If something else...?
 

Offline cellularmitosis

  • Supporter
  • ****
  • Posts: 1029
  • Country: us
Re: USA calibration club
« Reply #563 on: March 07, 2018, 06:57:31 pm »
Making good progress on a temperature regulation board which gets thermal-epoxied onto the bottom of a Hammond 1590B.  Soon the cal club will restart with temperature-regulated LTZs!   :-+

----

On another note, I'm also trying to cobble together some resistance transfer standards for the club.

I know the (used) meters which are popular among hobbyists have ranges based on either 10x, 20x, or 30, so I'll build e.g. the 10k standard with a 10k and 20k resistor in series, so that values of 10k, 20k, or 30k can be read from it.

Question for the group:  :-/O

Should I just make these the nominal value and keep it simple, or should I go with a slightly smaller / larger value and add a trimming circuit?  (e.g. should my 10K be 10K +/-0.1% or 9.99K + trimming circuit?)

(The example trimming circuit I had in mind to mimic was the ESI SRX-10k (see attached)).

Another way to phrase this question: are there any meters out there which have digital calibration, yet don't allow you to enter in an exact value?  I.e. is there a meter which forces you to calibrate against a 10.00000k resistor?, and won't let you type in 10.00123k?

I'm leaning towards just sticking with nominal values, because that simplifies the design, and keeps the cost down (including 4-wire connections for both the trimmed and un-trimmed values in each standard would get expensive when Pomona 3770's are $10 a pop).

Which brings up another question: for a 4-wire connection, is it important to have low-thermal-emf jacks for all four leads, or just two of them?  My intuition says that the sense leads are voltage-mode, but the force leads are current-mode, so a few uV of thermal EMF on the force leads doesn't matter at all (so use the Pomona 3770's on the sense leads, and use cheaper brass jacks on the force leads).  Thoughts?
« Last Edit: March 07, 2018, 07:08:25 pm by cellularmitosis »
LTZs: KX FX MX CX PX Frank A9 QX
 

Offline cellularmitosis

  • Supporter
  • ****
  • Posts: 1029
  • Country: us
Re: USA calibration club
« Reply #564 on: March 07, 2018, 07:00:51 pm »
derp, I forgot to attach the pictures of the heater progress:
LTZs: KX FX MX CX PX Frank A9 QX
 
The following users thanked this post: kj7e, hwj-d, zhtoor

Offline Conrad Hoffman

  • Frequent Contributor
  • **
  • Posts: 882
  • Country: us
    • The Messy Basement
Re: USA calibration club
« Reply #565 on: March 08, 2018, 12:37:20 am »
I'd go with no adjustments. Adjustments are another variable and can change during shipment or temperature cycling. They sow the seeds of doubt. AFAIK, there are no meters where you can't enter a specific value.
 
The following users thanked this post: TiN, cellularmitosis

Offline ap

  • Regular Contributor
  • *
  • Posts: 209
  • Country: de
    • ab-precision
Re: USA calibration club
« Reply #566 on: March 08, 2018, 01:22:23 am »

Question for the group:  :-/O

Should I just make these the nominal value and keep it simple, or should I go with a slightly smaller / larger value and add a trimming circuit?  (e.g. should my 10K be 10K +/-0.1% or 9.99K + trimming circuit?)

Which brings up another question: for a 4-wire connection, is it important to have low-thermal-emf jacks for all four leads, or just two of them?  My intuition says that the sense leads are voltage-mode, but the force leads are current-mode, so a few uV of thermal EMF on the force leads doesn't matter at all (so use the Pomona 3770's on the sense leads, and use cheaper brass jacks on the force leads).  Thoughts?

I would not add any trimmer, and make the adjustment resistor small. That way its stability is not important. You will anyway not have an exact (I mean <0.1ppm or so) accurate value. And what I would not do anyhow is readjust the standard, I would leave it as is and watch anual drift. Tells you much more. All relevant precision meters accept deviations from nominal to be entered.

And while you do not need EMF posts for the force connections, I would go with all low EMF, just for the sake of avoiding any connection errors. Actually, all precision meters have EMF compensation anyhow, so you could argue you can go with no low-EMF but standard binding posts.
Metrology and test gear and other stuff: www.ab-precision.com
 
The following users thanked this post: TiN, cellularmitosis, CatalinaWOW

Online bitseeker

  • Super Contributor
  • ***
  • Posts: 5730
  • Country: us
  • Lots of engineer-tweakable parts inside!
Re: USA calibration club
« Reply #567 on: March 08, 2018, 05:38:33 am »
I also vote for fixed resistor references (no trimmers) for simplicity, stability, and ease of use. Having 10k and 20k in series is a good idea for some flexibility.
You don't acquire TEA. It acquires you.
 
The following users thanked this post: TiN, cellularmitosis

Online Echo88

  • Frequent Contributor
  • **
  • Posts: 394
  • Country: de
Re: USA calibration club
« Reply #568 on: March 08, 2018, 06:51:39 am »
@cellularmitosis: Would you mind sharing the Heater-PCB? Looks nice for Hammond-cases. Is it a single supply design?
 

Offline cellularmitosis

  • Supporter
  • ****
  • Posts: 1029
  • Country: us
Re: USA calibration club
« Reply #569 on: March 08, 2018, 09:09:21 am »
@cellularmitosis: Would you mind sharing the Heater-PCB? Looks nice for Hammond-cases. Is it a single supply design?

Yup, designed to run off of the same 12V battery pack which will power the LTZ board:

https://github.com/pepaslabs/1590b-isothermal

https://oshpark.com/shared_projects/qRlewBCr

However, I'd hold off just a bit before ordering this board -- I haven't really verified its functionality fully yet.  I hooked it up the other night and logged the temperature, and it seemed to behave a bit strangely, in a way which I don't think I can simply shrug off as "bad PID constants".  So its possible there's a bug in the circuit.

Tonight I'm going to hook up the serial TTL output and see what the PID loop is doing, then I'll know for sure.

Also, there are a few changes I had planned already:  1) the LED is actually using the reset pin, which I think needs to actually be a reset pin if you want the bootloader to work correctly.  2) I think I'd like to move to using two FETs and just one current shunt, rather than one FET with power resistors above and below it.

But if you just want something to hack on, go for it!

(attached graphs are... overnight if I recall (at least 8 hours)).

Edit:  I managed to get the pinout backwards on the 78L05.  Which means I've been feeding my attiny85 about 8 volts.  Oops.
« Last Edit: March 08, 2018, 02:24:48 pm by cellularmitosis »
LTZs: KX FX MX CX PX Frank A9 QX
 
The following users thanked this post: Echo88

Offline cellularmitosis

  • Supporter
  • ****
  • Posts: 1029
  • Country: us
Re: USA calibration club
« Reply #570 on: March 09, 2018, 03:06:27 pm »
On another note, I'm also trying to cobble together some resistance transfer standards for the club.

I got lucky on ebay and scored a GenRad 1434 for super cheap which we can use as a donor for our resistance transfer standards!

The 1434 appears to be a later, cost-saving model compared to the 1432 and 1433:

  • The case is physically smaller.
  • They pulled some clever tricks with the rotary switch to get away with only 6 resistors per decade (see schematic).
  • Mechanically, the switches are vastly inferior to the older models.

My 1434 arrived with no knobs, one of the switches was stuck, and a few others are not so easy to turn.  The resistors look fine, but mechanically, this design does not appear to stand the test of time.  So its a perfect donor for our club!

Each decade has two ceramic cards with three resistors each, of a 2-value (20R, 200R, 2k, 20k, 200k).  Through clever switching, they can turn 6 2-values into 1, 2, 3, 4, 5, 6, 7, 8, 9 and 10-values.

This is ideal for our desired 1, 2, 3-value transfer standards, as we can simply parallel 2 of the 2-values to make a 1-value, then put that in series with the other 2-value to make 1, 2, and 3.

The 200k resistors follow the same scheme, but are on individual bobbins rather than cards.

These cards are just the right size to fit into the smallest water-tight aluminum case which Hammond makes, the 1550WQ.  I'll order some of those along with some silicone oil, and include a thermistor in each case.  I think we can get away without active temperature control for these, as there should be enough thermal mass that the typical +/-1F swing of a typical air-conditioned lab should be smoothed out.  I can also characterize the tempco of the resistors so that we can provide a correction curve using the paired thermistor.

I'm super excited!
« Last Edit: March 09, 2018, 03:13:13 pm by cellularmitosis »
LTZs: KX FX MX CX PX Frank A9 QX
 
The following users thanked this post: kj7e

Offline cellularmitosis

  • Supporter
  • ****
  • Posts: 1029
  • Country: us
Re: USA calibration club
« Reply #571 on: March 09, 2018, 03:11:54 pm »
Over the past year I've been a bit indulgent with my ebay habit, so I also have a 1432 and 1433 to show for comparison.  Mini-teardown time!

See also EEVBlog 461:



I also got incredibly luck with the 1432 ($39 shipped!), but paid dearly for the 1433 ($200 shipped).
LTZs: KX FX MX CX PX Frank A9 QX
 

Online Vgkid

  • Super Contributor
  • ***
  • Posts: 2374
  • Country: us
Re: USA calibration club
« Reply #572 on: March 09, 2018, 04:42:22 pm »
Does your 1432 have brass , or silver contacts?
If you own any North Hills Electronics gear, message me. L&N Fan
 

Online bitseeker

  • Super Contributor
  • ***
  • Posts: 5730
  • Country: us
  • Lots of engineer-tweakable parts inside!
Re: USA calibration club
« Reply #573 on: March 09, 2018, 04:56:14 pm »
I got lucky on ebay and scored a GenRad 1434 for super cheap which we can use as a donor for our resistance transfer standards!

The 1434 appears to be a later, cost-saving model compared to the 1432 and 1433:

Yeah, that's significantly different from the older ones I'm accustomed to seeing. Wow, good to know.

I also got incredibly luck with the 1432 ($39 shipped!), but paid dearly for the 1433 ($200 shipped).

I see. But the 1433 is such a nice design, both mechanically and aesthetically. You could look at them, instead, as having cost $120 each — a bit easier to swallow, perhaps.
You don't acquire TEA. It acquires you.
 

Offline cellularmitosis

  • Supporter
  • ****
  • Posts: 1029
  • Country: us
Re: USA calibration club
« Reply #574 on: March 09, 2018, 05:55:08 pm »
Does your 1432 have brass , or silver contacts?

bottom decades appear to be silver, top two decades (10k, 1k) are bare copper.

Yeah, that's significantly different from the older ones I'm accustomed to seeing. Wow, good to know.

It looks like things went even further downhill once IET took over.  http://www.eevblog.com/forum/metrology/teardown-standard-resistors/msg1295900/#msg1295900
LTZs: KX FX MX CX PX Frank A9 QX
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf