Newbie Precision Reference Calibrated with the 3478A Dinosaur

[hgg]

Hello,

I have just finished the Precision Voltage reference I was working on.   ...Finally!
A used HP 3478A that I scored from ebay for $99 arrived just in time for calibrating my reference.

The ebay seller was kind enough to calibrate the voltage range with a voltage standard he had.
Resistance measurements look pretty accurate.  I cannot say the same for its current range though
and I don't like the lack of uA range.

The multimeter arrived in a relatively good condition apart from some dust that has settled on the
board permanently...  I really like this multimeter!  A joy to use.  The display has a nice contrast,
but it lacks a back light.   A definite future mod.  The on board battery reads 3.042V which is very
low.  I would have to replace it very very soon.

Some photos:












Its has a Rev. C board.

The reference is based on the MAX6325 chip with an output of 2.5V
Very simple circuit with a chopper amp and two precision resistors.

You can calibrate multimeter voltage range with the 2.5V, resistance with 1K & 100K
and current with 2.5mA & 25uA.

Here it is:


(I will have to agree with Dave on that.  Just look where they did print their serial number...!







I chose these test points because I think that there are much better than the U shaped ones.
The top is a cone and the multimeter leads have a larger area of contact and at the same time
they don't slip.

The circuit also has a low battery indicator which I think is a must for a portable voltage
reference in order to be 100% sure for its accuracy.  The green LED starts to fade when the
battery reaches 8.1V and its completely off at 7.9V

The trimmer pot is a Bourns.  It is pretty good but if anybody know a better quality one please
let me know.




I have been running it for 3 days now and it has finally stabilized at 2.50005V
The last digit is a 5 because I prefer the voltage to be on the positive side. 
I don't like .99999s...   

I think its accurate enough to calibrate my UT61E or similar multimeters.
Now I would have to see its long term stability.

The UT61E was 3mV out and the UT136C was 5mV out.  Although the resistance measurement
of the UT1236C is painfully slow, it was spot on with the 1K and a couple of counts out with the 100K.
Its current measurement was spot on!  2.50 mA & 25.0 uA!

The leads you see in the picture above are a set of ProbeMaster leads that I saw in a Mailbag
episode.   They are simply perfect.  Thank you Dave !!  (and the kind person who sent them)

Thank you for helping me finish another little project.   
Regards,
George.

[Richard Head]

I've got the same meter. I agree, the lack of a backlight is a definite drawback. I also have a benchtop Fluke 4 1/2 digit that I use more often because the vacuum florescent display is more pleasant to read.
Beware, I have read that you have to supply a temporary backup voltage if you change the onboard battery or you will lose the cal settings.

[hgg]

Indeed, I've heard.  I will replace the battery with the power on.  My soldering iron is not ground
referenced.  I will also include a backup battery just in case...  One user left the following comment
in a youtube video which I don't understand:



"+GBPPR2 I don't think you need to worry about that SRAM battery getting disconnected,
according to the manual (Agilent still supports this). I'd suggest putting a big cap in parallel
with that or'ing diode since removing the battery could cause a rail bounce".

What exactly is a rail bounce??

[Andreas]

The trimmer pot is a Bourns.  It is pretty good but if anybody know a better quality one please
let me know.

Hello,

Perhaps you like to choose one of the AccuTrim series of vishay?

http://www.digikey.de/product-search/de?v=804&pv1=112&FV=fff40004%2Cfff80338&mnonly=0&newproducts=0&ColumnSort=0&page=1&stock=1&quantity=0&ptm=0&fid=0&pageSize=25

With best regards

Andreas

[Richard Head]

Rail bounce is just a term for a transient created on the rail due to a high momentary current that flows due to capitances etc. The track inductance and parasitic capacitances can cause an oscillation.

[hgg]

@Richard Head: So, removing the backup battery can cause a negative or a positive transient?

@Andreas:  The Bourns trimmer I used has a 100ppm/C while the one you just suggested has 10ppm/C
Not bad.  An order of magnitude better!   Andreas thanks again for your suggestion and your help with my
questions.

[Andreas]

Not bad.  An order of magnitude better!

Hello,

The price is also at least an order of magnitude better.
The question is if you will really need this.
If you use the trimming pin of the reference the influence of the trimmer TC
is already scaled down by at least a factor of 100 or even higher.

"The external trimming does not affect tem-
perature stability."

On the other side. My MAX6350 both have a relative constant ageing drift (at controlled temperature) of about 15 ppm / year.
So it would be interesting how your MAX6325A behaves.

With best regards

Andreas

[rdl]

Wow, those AccuTrim resistors from Vishay are very nice.

"smooth and unidirectional output"

15ppm/C for $10-$12

Hello,

Perhaps you like to choose one of the AccuTrim series of vishay?

...

With best regards

Andreas

[hgg]

"The external trimming does not affect tem-
perature stability."

Do you mean does not affect at all, or that the amount is very small?

Thank you.

[Andreas]

Hello,

the citate is from the datasheet.

I think that you / one will not be able to measure it.

with best regards

Andreas

[hgg]

Great!   Thanks again.

[hgg]

Just an update.

The 2.5V reference is rock steady.  17 days have passed and its still at 2.50005 Volts. 
It might be more accurate than 5 digits but I cannot measure that.
I have replaced the old battery of the 3478A and luckily I did not loose the calibration data...  :-)
The battery was the following: http://www.antoniadis.com.gr/product_info.php?products_id=1097
which fitted perfectly.

When the reference was ready, I calibrated both the UT61E and the UT136C.
The 61E has a very fine adjustment pot but the 136C has a single turn adjustment pot.
Despite that, after 17 days the 136C reads 2.500V but the 61E now reads 2.499V ...

So the 61E indeed has a drift problem !

The UT136C although is slow as a donkey, proved to be very accurate and stable.
(Quite accurate and repeatable current measurements especially with Frankie's leads
and some very low resistance fuses.)

[cellularmitosis]

I've got the same meter. I agree, the lack of a backlight is a definite drawback. I also have a benchtop Fluke 4 1/2 digit that I use more often because the vacuum florescent display is more pleasant to read.
Beware, I have read that you have to supply a temporary backup voltage if you change the onboard battery or you will lose the cal settings.

I posted my backlight mod for the 3478a at the end of this thread: https://www.eevblog.com/forum/projects/hp-3457a-lcd-backlight-mod-(also-for-3478a)/

I used EL tape, which turned out to be only just bright enough, and unfortunately the controller makes an annoying audible whine.

Soon I think I'll add a backlight kill switch so I don't wear out the EL tape while using the meter to log data over extended time periods.

[hgg]

Hi,

I've seen your mod.  Not bad actually.  Brightness level for my taste is perfect.
Is that a transparent el-tape?

[Dunk_c]

Hi

Is this beautifully crafted reference available for purchase?

[hgg]

Hi,

Its difficult for me to find the parts again, but I can send you one of the
printed circuit boards and the terminals if you like.

Thanks.

[Dunk_c]

Hi

I am a beginner and have just purchased a max6325cpa and some 100ohm and 5kohm resistors (0.05%, TCR 5-10ppm) to make a simple and cheap check for my meters.  Until I saw your beautifully designed board I was going to simply join the chip point-to-point, and put the whole lot in a box with some 2 mm banana plug sockets for my probes.   I haven't made it yet.  What are the hard to find parts?

[hgg]

Hi,

Sorry for the delayed reply.  I don't have an internet connection here...

Anyway, what I meant was that it was not easy for me to find some of these parts here in Greece,
like the precision resistors for example, so if you have all the components I can send you a pcb if
you like.

Regards,
George.

[Dunk_c]

Hi George

Please email me with purchase details and a component list.  Much apprecaited.
duncan

[hgg]

Just an update,

I've checked again the reference with the 3478A and after 5 months its still showing 2.50005V
Either its very stable or both the reference and the 3478A are drifting together.


26-10-2014

I had though a problem with the 3478A.  When I received it I had to change the transformer tap to 240V.
It also had a 220V option but here we have 230VAC so I chose the 240V. 

In the service manual it says that for option 345 which is 240V/50Hz I have to use an 125mAT fuse.
I used a Schurter F125mA/250V and it was working fine until today.  Now it blowed all the 125mA fuses
I've tried.  Do you think its because I was using the fast type and not the time delay type?

Now I am using a 250mA.

[tautech]

Do you think its because I was using the fast type and not the time delay type?

Most likely.
Unless a fast blow fuse is specified I would use a STD fuse for mains.
Has it a Linear or SMPS power supply?
Smoothing caps on PSU could cause the inrush.

[hgg]

It has a linear power supply and the specified fuse for 230V is a 125mAT,
so I assume the T stands for Time Delay?

[SeanB]

Fast fuse will blow, that is why they use a slow blow one, otherwise the fast fuse has to be very overrated to handle the surge within it's time envelope. Buy a pack of slow blow ones and use them, if it does blow them then you might have either high mains, a faulty mains switch or power outlet, or a faulty mains cable or finally a faulty meter.

[hgg]

I will try some slow blow ones and see what happens.
Thanks!

[hgg]

I am not sure that Utilboard can export to other programs but here are the schematics of the circuit:


vRef2.5V

I can send you the Gerber files if you are interested.