Precision Voltage Reference / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Precision Voltage Reference

1 2 3 next

Move Thread

LAN_403

IntraWinding

Fri Nov 27 2009, 02:19AM

Registered Member #2261 Joined: Mon Aug 03 2009, 01:19AM
Location: London, UK
Posts: 581

Since getting a 6.5 digit meter - Solartron 7150 plus - I've been researching what's the best voltage reference I could build without spending a fortune.

The LTZ1000 seems to have taken the crown, but it's pricey. Are there any challengers or clever older designs that do as well or better?

Bored Chemist

Fri Nov 27 2009, 06:52AM

Registered Member #193 Joined: Fri Feb 17 2006, 07:04AM
Location: sheffield
Posts: 1022

A Weston standard cell?

Dr. Slack

Fri Nov 27 2009, 08:59AM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

Ooh, that one looks quite good, I was just starting to do the same research myself.

Once you are ovening the active components, there's a lot to be said for minaturisation. It mitigates thermal gradients, you would be very hard pushed to match that performance by building something from discretes on a board and putting it in an oven. Without an oven, you don't stand a cat in Hades chance.

Pay the money and enjoy the performance.

It's not really relevant for a DVM, but if you want a low noise reference as well as a low drift one, then the lowest noise you can get in the region where larger and larger electrolytics become silly is a 50% discharged lead-acid accumulator. It drifts, so it has to be referenced to a buried zener like the LTZ1000, but it's quiet.

IntraWinding

Fri Nov 27 2009, 02:24PM

Registered Member #2261 Joined: Mon Aug 03 2009, 01:19AM
Location: London, UK
Posts: 581

The sort of thing I'm thinking of at the moment is using a large collection of cheap Zener based references and averaging their outputs. I figure I don't need to worry about thermal drift as I can put them in an oven, but long term drift is a real problem. Another trick would be to regulate the circuits supply from the output of the reference itself so power supply coefficient of drift is removed.
Another thought is that if operating references at high temperature results in more drift due to rapid ageing, what about using a tiny Peltier device to keep the reference at a constant low temperature, say 0Â°C

I'm also interested in high performance ovening techniques.

Steve Conner

Fri Nov 27 2009, 03:01PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I've got one of those 7150 plus meters as well. Seems to be a lot of military surplus units kicking around the UK.
What do you want the reference for, checking the meter's calibration?
(Mine appears to be 18 years out of cal :( )

WaveRider

Fri Nov 27 2009, 03:52PM

Registered Member #29 Joined: Fri Feb 03 2006, 09:00AM
Location: Hasselt, Belgium
Posts: 500

Google "temperature compensated bandgap reference" for lots of background. Analog devices has a series of precision references that are used for analog-to-digital conversion.. The usually cost no more than a few bucks and are pretty good. (16 or 24 bit ADCs need good low-noise precision references!)

Proud Mary

Fri Nov 27 2009, 05:30PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

I was wondering whether you meant something superior to the ordinary two-transistor band-gap reference, or the buried zener technique.

I have found the inexpensive REF-02 +5V reference/temperature transducer good for my very modest needs.

If you really feel you must, you can make the cutest little oven using 2N3055 as the heating element controlled by a glass bead thermistor, or other thermal transducer.

A selection of 0.1% tolerance resistors is another good way of checking out the 'goodness' of a multimeter.

Don't forget that stable though your reference source may be, your meter readings will drift with temperature, unless the meter too is placed in a stable thermal environment.

Stella

EDIT AFTERTHOUGHT: May I suggest - as I have often suggested before to other members asking similar questions - that you look FIRST to your application and then design the supporting circuitry, supplies &c according to its needs.

For example, had you specified "A +5V reference is required with less than 0.01% drift over -20 to + 70 C, with such and such supply rejection, and all that good stuff, then the answer is actually contained within the question

Electronic design is inherently teleological - i.e. design considerations spring from the needs of the end user, the application, the conclusion without knowledge of which no meaningful, coherent design is possible. The end justifies and dictates the means,
though there may be more than one way of reaching it, each of which paths can be arranged in an economic hierarchy to determine which should be chosen

IntraWinding

Sat Nov 28 2009, 03:11PM

Registered Member #2261 Joined: Mon Aug 03 2009, 01:19AM
Location: London, UK
Posts: 581

Thanks for all the replies.

Yes, those meter keep turning up. They're very old and in my experience burn out within a few hours of being powered up, but having repaired one, and having two others to fix, I'd like to get them calibrated.

The good news is you can calibrate them from the front panel without removing the lid. The bad news is that getting a meter calibrated is expensive and they drift depressingly quickly.

Wouldn't it be nice if I had my own voltage reference that, hopefully, some kind person could calibrate for me and which was so well made it kept its calibration sufficiently for several years use with my meter.

From what I've read I think this would be pushing the limits of non Josephson based voltage references, but I have the advantage that I can afford to construct a system that wouldn't be considered economically viable industrially since I get my own labour for free!

Hopefully my meter is reasonably well compensated for thermal effects.
(Compensation is the way to go, but go on, what's that 2N3055 oven circuit then?)

Looking at figure 1 on the LT Application Note 82 - 'Understanding and applying voltage references', I see that 6.5 digit equates to about 0.5ppm, so I would ideally like to hold a voltage to significantly better than that for several years.

Steve Conner

Sat Nov 28 2009, 03:19PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

When I was a kid, I thought that if a car's speedometer went up to 120mph, that meant it could do 120mph.

Now I'm older and wiser, and I also know that a 6.5 digit DMM isn't necessarily expected to have 6.5 digits of accuracy to go with its 6.5 digits of resolution. The actual accuracy is specified in the manual.

My 7150 plus has failed to burn out or drift noticeably in the 4 or so years I've owned it.

Proud Mary

Sat Nov 28 2009, 03:45PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

IntraWinding wrote ...

Wouldn't it be nice if I had my own voltage reference that, hopefully, some kind person could calibrate for me and which was so well made it kept its calibration sufficiently for several years use with my meter.

From what I've read I think this would be pushing the limits of non Josephson based voltage references, but I have the advantage that I can afford to construct a system that wouldn't be considered economically viable industrially since I get my own labour for free!

Aren't you being a big silly-billy here, sweetheart?

Since joining 4HV several years ago, I've always been at the forefront of calls for improved metrology, since - in my view - measurement is at the heart of all that we do as scientists - perhaps all that there is to distinguish us from other types of
mystical thinker.

And So! Looking to my admittedly controversial view that the most efficient electronic designs are designed teleologically (a view I have imported from rational philosophy) I would ask you for details of your device that depends upon a Josephson junction reference for reliable, repeatable operation.

You could have put it another way, of course, and asked "I want to explore the accuracy of my instruments and work them up as best as I can" and you'd have put yourself in range of some practical help.

Why not set about your calibation in a more rational way where other members could easily help you with small gifts of 0.05% resistors and so on?

Stella Maru

1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob