Simple 10V reference project - LT1021

[Jay_Diddy_B]

Hi group,

I am sharing the design and build of a 10V battery powered reference based on the LT1021-10. This is a DMM-check level of reference. I wanted something that I could take to a meter, instead of taking the meter to the calibrator.

Inspiration






This application circuit from the datasheet shows the LT1021 being used in shunt mode. The circuit can be modified to the positive version.



This application circuit shows how to implement a restricted range trim circuit.


Schematic



This is the schematic. The current source is implemented using a green LED and and BJT. The forward voltage of the green LED is around 2.2V. Minus the Vbe of Q1, 0.7V, results in 1.5V across the 510 resistor. This results in a 3mA current source.




PCB

The PCB was designed using KICAD version 5.










Build



A 3D-printed base was attached to the PCB.

Testing



The trimmer was adjusted for 10.000V


This is not the world's best reference, but it can be used for checking DMM's on the 10V. It is relatively easy to build.

Regards,

Jay_Diddy_B

[Jay_Diddy_B]

Hi group,

The LTspice model shows that the diode illuminates when the output is valid:



Regards,

Jay_Diddy_B

[zlymex]

A more simple way is the plug-in type that eliminates manual switch and leads.
Red box is a micro switch.

[CalibrationGuy]

Doesn't the DMM check version have diode(s) in their compensation network to minimize the tempco? Your circuit is simple, but the quality of the LT1021 makes the project suitable for many DMM checks. You have to start with a calibrated meter to check it initially, though.

TomG.

[edavid]

What is the reason for choosing shunt mode?

[zlymex]

What is the reason for choosing shunt mode?

It requires a more economical current source compare to a voltage regulator in series mode.

More over(not applies here though), it requires less voltage supply, say 10.5V in shunt mode, while it requires 11.5V(recommended, 11V at least) in series mode.

[Jay_Diddy_B]

Doesn't the DMM check version have diode(s) in their compensation network to minimize the tempco? Your circuit is simple, but the quality of the LT1021 makes the project suitable for many DMM checks. You have to start with a calibrated meter to check it initially, though.

TomG.

I implemented the trimming as described on the datasheet. The datasheet suggests a diode when using the 5V version of the LT1021. The DMMCheck uses the 5V version.


At https://voltagestandard.com/ you can buy a voltage reference based on the LT1021. Considerable value is added by burning in the references for 2000 hours and calibrating them with a 3458A DMM.

Yes, you need to get it calibrated one way or the other. I have a Fluke 732A, and a Datron 1281. I can get it close enough for most applications.

Jay_Diddy_B

[Jay_Diddy_B]

What is the reason for choosing shunt mode?

It requires a more economical current source compare to a voltage regulator in series mode.

More over(not applies here though), it requires less voltage supply, say 10.5V in shunt mode, while it requires 11.5V(recommended, 11V at least) in series mode.

zlymex is correct. In addition, the LED provides two functions, it is part of the current source but also indicates if the battery voltage is too low.

Jay_Diddy_B

[iMo]

.. also the green led is the lowest noise among the LEDs, afaik, much less than when a zener there in the current source, people say..
On the other hand I would be skeptical to use an LT1021 in plastic package while targeting 5+ digits.
It is a reference created many decades back for 12bit ADC/DAC mainstream usage (as any others except perhaps LM399)..

[mawyatt]

A more simple way is the plug-in type that eliminates manual switch and leads.
Red box is a micro switch.

Here's something we did along these DMM checking ideas. Also served as a AC rms waveform as well as the DC reference.

https://www.eevblog.com/forum/testgear/ac-rms-dmm-tests/

Best

[Jay_Diddy_B]

.. also the green led is the lowest noise among the LEDs, afaik, much less than when a zener there in the current source, people say..
On the other hand I would be skeptical to use an LT1021 in plastic package while targeting 5+ digits.
It is a reference created many decades back for 12bit ADC/DAC mainstream usage (as any others except perhaps LM399)..

The green LED has a -2mV / C temperature coefficient, this provides a first-order correction to the current source.

I had some plastic LT1021 in my collection of parts, so I used one. There are better options.

Jay_Diddy_B

[zlymex]

Here is mine.

[iMo]

With my first 7106 meter I wanted to save battery money, so I made DIY "Finger ON Switch with Automatic OFF Timer".
It was wired somehow like see below, no mechanical switch needed and it had automatic OFF timer.
Worked fine, the ON time depends on which finger you will use.. 
C1 is foil..

[zlymex]

With my first 7106 meter I wanted to save battery money, so I made DIY "Finger ON Switch with Automatic OFF Timer".
....l..

Great!
I once made a similar LED switch for kids as well which is finger on/finger off/auto off.

[iMo]

For people starting with electronics I should have added to the "Your Finger" label above following:

"Tap here shortly with Your Finger"

 

[Jay_Diddy_B]

A more simple way is the plug-in type that eliminates manual switch and leads.
Red box is a micro switch.

zlymex,

I see that you have implemented this. It works great! The battery should last a long time.

This would not work for some of my meters. My Datron DMMs all have binding post connections.

Regards,

Jay_Diddy_B

[zlymex]

A more simple way is the plug-in type that eliminates manual switch and leads.
Red box is a micro switch.

zlymex,

I see that you have implemented this. It works great! The battery should last a long time.

This would not work for some of my meters. My Datron DMMs all have binding post connections.

Regards,

Jay_Diddy_B

You are absolutely right. I plan to use two paralleled switches in the next version that can be seen from my last photo, one is momentary the other is alternate. Or to use finger switch.

[GigaJoe]

if you looked at schematic of 1021 , it already has current source inside , so design as shunt basically doesn't make sense for me personally , rather exercise ...
probably better redesign for a  compensation .

i wouldn't be so pessimistic about plastic case ..  Korean designers using ADR421 in soic for its 200000 dmm  ,  still reasonable