Feasible? Convert Fluke 5440B A9 Into Stand Alone DC Standard

[denimdragon]

The title says it all. I have had this 5440B for several years and I have
attempted to repair it here and there, initially spending several days
replacing bad caps, burnt carbon resistors, burnt regulators, and even
meticously clean those damn relays 3 times (All of them  ). I have
revisited the calibrator at least 2 times every year, with some success,
only to have new faults appear. I was going to get rid of it 2 years ago,
but was advised against it as it is a very stable once working right, but
now, I don't have the luxury of time or space that I once had, so I need
to make use of it.

I'm in need of a very stable voltage reference, and since the 5440B
is very stable and has two aged SZA263s on board with a heater,
I am wondering how difficult it would be to use the A9 board as
a stand alone 13.2 volt reference, possibly trimming it later to 10V and
maybe 1.018 or 2.500 volts. But the 10V is of course the main target.
The other voltages would just be an added luxury.

I would like to do as little as possible to the original A9 board in order
to achieve this. I basically want to power the SZA263s, the Heater and
any absolute necessary supporting circuitry. I can machine a costume
enclosure for the complete board with power and terminal connections
on the outside and some type of temperature monitoring.

Does this sound moderately simple to do? Also, does it seem like
a good idea? It would save me quite a bit of money as I will still need
to purchase a second commercial 10V reference (732A/B).

I'm open to suggestions and comments. TIA

[bdunham7]

You'd probably be better off selling the complete but not working 5440B and getting something else like a 732A.  Dividing down your 13.2V to a stable 10V is not so easy either if you want a low output impedance.  And then you'd have to calibrate and characterize it anyway. 

You also might be overestimating your need for super-precision 10V references.  An environmental chamber, the ability to accurately measure temperature and humidity and a multi-channel logging method that has low thermal voltages are all things you probably should look at first.

[Rax]

You'd probably be better off selling the complete but not working 5440B and getting something else like a 732A.  Dividing down your 13.2V to a stable 10V is not so easy either if you want a low output impedance.  And then you'd have to calibrate and characterize it anyway. 

I agree with bdunham7. The 5440 series are well appreciated and sell very well (aside from their extreme size and weight and the associated hardship for the seller). But, with a bit of patience and persistence, you may be able to grab a 732 with hopefully minor issues with the money you'd get for a non-functional 5440.

While running the A9 PCA independently of the main unit may not be impossible, I don't think it will turn out to be trivial; how much effort are you willing to put into this graft is up to you, but I'd think it may not be the most optimal path forward to achieving your stated goals.

[denimdragon]

Ok. So I'm sold that it's not the most efficient way to move forward,
but now I'm thinking about keeping it after dropping of my 34401A
and Fluke 5450A to get calibrated today. I had a conversation with
the metrologist there and he told me that I should keep it and just
buy a 732A/B. He said that the 5440 and 5442 are really super
solid boxes. He also said he would calibrate it for a reasonable
price once I fixed it. It would be a bit because he had just sent
off his 732s. That gave me new hope so I guess I will start
Back on the repair path. Maybe I can get back on it at the
End of the month.

[bdunham7]

I'm thinking about keeping it after dropping of my 34401A
and Fluke 5450A to get calibrated today.

Two questions:  What age/firmware is your 34401A and are you getting data with the calibration?  If so and if it is a good lab, I'd like to see the data if you're willing to post it.

[denimdragon]

Two questions:  What age/firmware is your 34401A and are you getting data with the calibration?  If so and if it is a good lab, I'd like to see the data if you're willing to post it.

I'll have to check when I get it back next week, but I don't mind sharing
at all. I would guess it's rather old because it's branded HP. Actually
both of my 34401As are HP branded.

Here's the lab I use:
http://oscilloscopeservices.com/index.htm

They have a large prescence in the Houston area and are very reasonable. He told me to cal the 5440B wouldn't exceed $250ish. Once I get it back with my companies reference # I'll have access to the calibration information online in addition to my hard copies.

[denimdragon]

*Update*

Yesterday, we purchased a Fluke 732A. I hope it will be here next
week so I can inspect it to make sure it electrically sound before
dropping it off to be calibrated when I pick up the boxes I dropped off. I know that I need another reference (or two) to
compare it to, but that will have to wait until funds become
available again. Until then, I have set out to get the 5440B
up and running. Even though it isn't on the same level as the
732A, it should be a decent substitute until I can afford another 732A/B right?

Opinions?

[bdunham7]

Opinions?

Just as I predicted, you've abandoned your 50ppm goal and headed straight for 0.5ppm.

Yes, a working, calibrated 5440B would be about as good as you're going to get short of specific 10V standards like the 732x.

[denimdragon]

Opinions?

Just as I predicted, you've abandoned your 50ppm goal and headed straight for 0.5ppm.

Yes, a working, calibrated 5440B would be about as good as you're going to get short of specific 10V standards like the 732x.

 

Well, not exactly. I just want to make sure I have proper support equipment
for my little lab and projects. It may seem like overkill, but it helps me sleep
a bit easier.

[J-R]

Doug improved his 10V reference a while back: https://voltagestandard.com/001%25-10v-reference
Now you can easily power it externally for continuous operation and it has better temperature stability.
So I wonder if a few of these would be less hassle than the old Fluke gear and still be good enough?

[denimdragon]

Doug improved his 10V reference a while back: https://voltagestandard.com/001%25-10v-reference
Now you can easily power it externally for continuous operation and it has better temperature stability.
So I wonder if a few of these would be less hassle than the old Fluke gear and still be good enough?

I went to his site, but at the time, he was out of stock for the 10V
reference 

I ended up with this 732A

[J-R]

Hopefully Doug's working on a new batch, I need a tie breaker for my two existing 10V references and I've been saving up!

So I'll be that guy and ask: in the photo, which device is correct?

[TUMEMBER]

So I'll be that guy and ask: in the photo, which device is correct?

None, any.
There is no second photo showing the reverse measurement. What the HP 34401A measured measures what the HP3456A measured.
Determine the limits of the marginal error and you can estimate the Fluka condition unless you have an error certificate (current) for one of the meters. There is no need to mention the small fact that we do not know the environmental conditions.

[denimdragon]

Hopefully Doug's working on a new batch, I need a tie breaker for my two existing 10V references and I've been saving up!

So I'll be that guy and ask: in the photo, which device is correct?

Actually the 34401A is measuring the 10V output, and the 3456A is measuring the 1.018V output. I recieved this reference
3 days ago, so I'm just letting it run for now and longing the 10V output through Agilent DMM Utility. The lab that it's in has
a tempereture range of 21-28C°. 

I have a 34401A and Fluke5450A out for calibation right now and I should get them back next week. I will then connect
the freshly cal'd 34401A in place of the one which is now being used. I did this in order to get an estimate as to how well
the 732A is working and to make sure there's not a large amount of drift coming from the unit. So for, it is looking very
promising and I will post some data charts here as soon as I figure out how to properly display it in excel. The current
drift is -15uV over 72+ hours.

It is running on AC power only, as it shipped with no batteries. I've been researching the best battery replacements without
having to make alterations to the battery compartment. Once I have checked and refurbished the circuitry and installed fresh
batteries, I will let it run until it's stable again and then send it out for calibration.

[J-R]

Yes, I understood the separate 1.018V output, but it's derived from the 10V output so I was curious of the three devices which was considered the one to trust.  So you're saying none of those are calibrated, currently...

[denimdragon]

Yes, I understood the separate 1.018V output, but it's derived from the 10V output so I was curious of the three devices which was considered the one to trust.  So you're saying none of those are calibrated, currently...

No, not cal'd. Not that the 3456A needs it's. It's my most trusted unit. My second most trusted unit
is the 34401A that's out for calibration. The 34401A you see in the picture is "the new guy" (a couple
of weeks old). However, I have a few 34970A units that went out of cal last July, so I'll be hooking
them up to the 732A see what they have to say about the readings. Fun stuff 

Pardon the photo. I'm still a rookie at the photos on the blog

[Dr. Frank]

*Update*

Yesterday, we purchased a Fluke 732A. I hope it will be here next
week so I can inspect it to make sure it electrically sound before
dropping it off to be calibrated when I pick up the boxes I dropped off. I know that I need another reference (or two) to
compare it to, but that will have to wait until funds become
available again. Until then, I have set out to get the 5440B
up and running. Even though it isn't on the same level as the
732A, it should be a decent substitute until I can afford another 732A/B right?

Opinions?

Hello,
the 5440/5442 are superior DC voltage standards.
It's not feasible to somehow use the internal double SZA263 reference.
I don't know your specific problems with that box.
Is the INT CAL working?
For DCV calibration of your DMMs, you need a source for all ranges, therefore a 732A will not really help you, you would need a calibrator, instead.
In case the INT CAL works, you'll get a super stable 10V reference out of your 5440.
My 5442A drifts no more than 1ppm over 5 years, definitely (I switch it on only when needed).
The INT CAL also re-calibrates the gain constants of the higher ranges (22V, 250V, 1000V), so that they always are within their 24h spec, relative to the 10V range.
If you now have a calibrated 732A, you might calibrate your 5440 (if it's currently working) in its 10V range, and the INT CAL procedure does the rest, it's not necessary to re-acl the upper ranges.

Only disadvantage are the 2V and 200mV ranges, which have to be calibrated by a more precise ACAL DMM, like the 3458A, or a precision divider, like the 752A.

Frank

[denimdragon]

Hello,
the 5440/5442 are superior DC voltage standards.
It's not feasible to somehow use the internal double SZA263 reference.
I don't know your specific problems with that box.
Is the INT CAL working?
For DCV calibration of your DMMs, you need a source for all ranges, therefore a 732A will not really help you, you would need a calibrator, instead.
In case the INT CAL works, you'll get a super stable 10V reference out of your 5440.
My 5442A drifts no more than 1ppm over 5 years, definitely (I switch it on only when needed).
The INT CAL also re-calibrates the gain constants of the higher ranges (22V, 250V, 1000V), so that they always are within their 24h spec, relative to the 10V range.
If you now have a calibrated 732A, you might calibrate your 5440 (if it's currently working) in its 10V range, and the INT CAL procedure does the rest, it's not necessary to re-acl the upper ranges.

Only disadvantage are the 2V and 200mV ranges, which have to be calibrated by a more precise ACAL DMM, like the 3458A, or a precision divider, like the 752A.

Frank

Dr. Frank, I'm glad you stopped by to comment also. I was having a host of problems with my 5440B.

https://www.eevblog.com/forum/repair/fluke-5440b-psu-fault-check-gaurded-power/msg1509067/#msg1509067

I have started preparing to get it up and running again, especially since I have something I can
calibrate it with (732A). If you go and have a look at the link above, it will show some of the
problems I came up against. Any feedback would help since the 5442 is similiar to the 5440.

Also, does anyone know how long it takes the 732A to stabilize? I have heard months and I have
heard weeks. This unit is show -15uV on the 10V output over 93 hours of logging. If I am
understanding correctly, this -1.5ppm. Is that correct?

[bdunham7]

This unit is show -15uV on the 10V output over 93 hours of logging. If I am
understanding correctly, this -1.5ppm. Is that correct?

What are you logging it with or comparing it to?  And what has the ambient temperature been over that 93 hours?  If you see a 1.5ppm change it is difficult to know whether the source or measuring instrument has moved.  And yes, 15µV => 1.5ppm.

[Dr. Frank]

Dr. Frank, I'm glad you stopped by to comment also. I was having a host of problems with my 5440B.

https://www.eevblog.com/forum/repair/fluke-5440b-psu-fault-check-gaurded-power/msg1509067/#msg1509067

I have started preparing to get it up and running again, especially since I have something I can
calibrate it with (732A). If you go and have a look at the link above, it will show some of the
problems I came up against. Any feedback would help since the 5442 is similiar to the 5440.

Hi,
apart from some additional functionality, the 5442A is virtually identical in HW & SW.
The stability in reality seems to be the same as the better specified 5440, especially after years of use.
I don't know of any hardware related difference to achieve the better stability, maybe apart from a pre-selection of the reference.

I had a look into your error description, but there is too few details given, sorry. I assume you have one or several digital problems.
Have you checked systematically all the power supplies? Please provide a full list of all the test points, outguard and inguard which you might also find in the service manual.
Your error messages, do they arise from the internal analog and digital self test procedures, or are they appearing directly, when you power your unit on?
Does your unit output anything in OPERATE mode?


When I in school, 45 years ago, I have designed and built several measurement instruments, especially a 7digit precision universal counter, based on a grave of TTL ICs. Unfortunately, I have later torn it apart again for part salvaging, as it devlopped more and more loose connections.
Today, I really regret this deeply.

Therefore, I would never let go such a device like your 5440B. You can always fix it on a later point in time, when you have gathered more electronics experience and systematics in error analysis.

Also, does anyone know how long it takes the 732A to stabilize? I have heard months and I have
heard weeks. This unit is show -15uV on the 10V output over 93 hours of logging. If I am
understanding correctly, this -1.5ppm. Is that correct?

I second bdunham7's statements:
 
-15µV / 10V = -1.5ppm, that's right.
Against which other reference did you check this, and how was your setup?
How can you tell, that you are not only seeing the drift of your other reference?
(Or were these simply your 34401A or 3456A?)

The 732A is based on the SZA263 as well, therefore, I expect that your instrument would recover to its default output and its normal drift rate (~3ppm/Year) within a few days at most.
Whenever I power up my 5442A, it needs about 4h only to reach exactly its former 10V output, within +/- 0.2ppm, compared to my whole, very well characterized  reference group.

So I can tell, that my 5442A drifted less than 1ppm in 5 years, and that the SZA263 does not show any noteworthy on/off hysteresis or lengthy recovery time.

Have you made sure, that the 732As oven is running properly? Then, any room temperature changes during those 93h can only have an influence on that other reference you're using. The T.C. of the 732A is specified < 0.05ppm/°C.

If I use my 3458A for such a 24h comparison, I have to make sure that R.T. stays constant within +/- 0.2°C, which is the characteristics of my basement lab. A 3°C change of R.T. would already give 1.5ppm deviation for my 3458A, as it's characterized total T.C. is about +0.45ppm/K w/o ACAL. Your DMMs are on the same T.C. order of magnitude.


Frank

[denimdragon]

Hi,
apart from some additional functionality, the 5442A is virtually identical in HW & SW.
The stability in reality seems to be the same as the better specified 5440, especially after years of use.
I don't know of any hardware related difference to achieve the better stability, maybe apart from a pre-selection of the reference.

I had a look into your error description, but there is too few details given, sorry. I assume you have one or several digital problems.
Have you checked systematically all the power supplies? Please provide a full list of all the test points, outguard and inguard which you might also find in the service manual.
Your error messages, do they arise from the internal analog and digital self test procedures, or are they appearing directly, when you power your unit on?
Does your unit output anything in OPERATE mode?


When I in school, 45 years ago, I have designed and built several measurement instruments, especially a 7digit precision universal counter, based on a grave of TTL ICs. Unfortunately, I have later torn it apart again for part salvaging, as it devlopped more and more loose connections.
Today, I really regret this deeply.

Therefore, I would never let go such a device like your 5440B. You can always fix it on a later point in time, when you have gathered more electronics experience and systematics in error analysis.

Also, does anyone know how long it takes the 732A to stabilize? I have heard months and I have
heard weeks. This unit is show -15uV on the 10V output over 93 hours of logging. If I am
understanding correctly, this -1.5ppm. Is that correct?

I second bdunham7's statements:
 
-15µV / 10V = -1.5ppm, that's right.
Against which other reference did you check this, and how was your setup?
How can you tell, that you are not only seeing the drift of your other reference?
(Or were these simply your 34401A or 3456A?)

The 732A is based on the SZA263 as well, therefore, I expect that your instrument would recover to its default output and its normal drift rate (~3ppm/Year) within a few days at most.
Whenever I power up my 5442A, it needs about 4h only to reach exactly its former 10V output, within +/- 0.2ppm, compared to my whole, very well characterized  reference group.

So I can tell, that my 5442A drifted less than 1ppm in 5 years, and that the SZA263 does not show any noteworthy on/off hysteresis or lengthy recovery time.

Have you made sure, that the 732As oven is running properly? Then, any room temperature changes during those 93h can only have an influence on that other reference you're using. The T.C. of the 732A is specified < 0.05ppm/°C.

If I use my 3458A for such a 24h comparison, I have to make sure that R.T. stays constant within +/- 0.2°C, which is the characteristics of my basement lab. A 3°C change of R.T. would already give 1.5ppm deviation for my 3458A, as it's characterized total T.C. is about +0.45ppm/K w/o ACAL. Your DMMs are on the same T.C. order of magnitude.


Frank

I am using the 34401A to measure the 10V output on the 732A. I am a aware that
the 34401A is not up to the task of measuring the 732A to it's specifications, but I
am attempting to check that there is nothing really out of the ordinary happening
with the reference circuit. I plan on refreshing the power supply circuitry and installing
new batteries in hopes of restoring the reference to it's former glory, but I wanted to
have an idea of how it was performing before repairs.

I've started working on the 5440B again. I'll just take my time and work on it a little
bit everyday until I really understand how the circuit works and can figure out what
the problems are. I will be sure to add more details in my posts about the repair
in that thread.

It also looks like I will be needing a better meter for logging.

[alm]

It also looks like I will be needing a better meter for logging.

I'd say a second DCV standard (like a working Fluke 5440B) would be a better strategy. Even a 8.5 digit meter like HP/Agilent/Keysight 3458A is not good enough to test a standard like the Fluke 732A to its limits. You'll notice Frank only wrote about using the 3458A for a 24h comparison, not for any long term drift monitoring. Measuring the difference to another 10V source like the 5440B, another F732A or a well built LTZ1000-based circuit will give you lower uncertainties. Ideally you have at least three.

There's nothing wrong with using a DMM as a sanity check, but be aware that you'll only be testing the standard within the limits of the DMM. So the drift you measured might be -1.5 ppm +/- 25 ppm, so almost as likely to have drifted up as down. There is value in knowing the standard didn't drift more than 26 ppm, but it's not exactly a performance verification.

[guenthert]

I'd say a second DCV standard

And so it begins

[denimdragon]

I'd say a second DCV standard (like a working Fluke 5440B) would be a better strategy. Even a 8.5 digit meter like HP/Agilent/Keysight 3458A is not good enough to test a standard like the Fluke 732A to its limits. You'll notice Frank only wrote about using the 3458A for a 24h comparison, not for any long term drift monitoring. Measuring the difference to another 10V source like the 5440B, another F732A or a well built LTZ1000-based circuit will give you lower uncertainties. Ideally you have at least three.

There's nothing wrong with using a DMM as a sanity check, but be aware that you'll only be testing the standard within the limits of the DMM. So the drift you measured might be -1.5 ppm +/- 25 ppm, so almost as likely to have drifted up as down. There is value in knowing the standard didn't drift more than 26 ppm, but it's not exactly a performance verification.

I'm definitely going to get the 5440B running. Volt-nuts continue to say great things about the 5440/5442 in terms of stability.
I also have nearly everything to build one of the LTZ1000 references from here. I just need the board and the Vishay resistors.
They seem to be difficult to acquire from what I have read.

When I think about it, All I would need is and AC Reference standard and I could cal just about any of my meters here at home.

Yes, it's pretty much a sanity check. I wanted to make sure I didn't get a horrible unit. The 732A, the 5440B, and a good LTZ1000A
reference circuit would be a check and balance system until I can get 2 more good condition 732As. That will take a bit more time,
again. I still wish I could have aquired that Guildline Reference. I'm not sure why my partner didn't grab it. I think he wanted
the unit to come cal'd with calibration paper work (proper 17025...?) for a year from the company but they wanted extra money
for it. Oh well. I have enough to keep me busy for a while in my "spare time". I'm going to go fire up my old 5440B repair post.
I hope that's not a bad thing. I don't want to start a new post with the same exact topic.

[denimdragon]

I'd say a second DCV standard

And so it begins

...yeap. Bitten by the volt-nut bug...