# EEVblog Electronics Community Forum

## EEVblog => EEVblog Specific => Topic started by: mikeselectricstuff on February 12, 2014, 12:17:36 pm

Title: #579 - precision current source
Post by: mikeselectricstuff on February 12, 2014, 12:17:36 pm
Why use diodes to add a constant voltage drop on the output of a constant current source, when a resistor would work just as well?
Title: Re: #579 - precision current source
Post by: EEVblog on February 12, 2014, 12:31:00 pm
Why use diodes to add a constant voltage drop on the output of a constant current source, when a resistor would work just as well?

Yes, I had intended to mention that, but forgot, oh well. I had mentioned a diode in a previous video, and I wanted to show that you wouldn't get the expected  voltage drop at the lower current.
It's probably stupid to show that, but it is what it is.
Oh, and I was going to add some annotation too to mention that, but of course forgot to do that as well because I'm double stupid.
The final circuit I built already has a resistor instead.
Title: Re: #579 - precision current source
Post by: Jebnor on February 12, 2014, 02:36:51 pm
Heya Dave, I for one would like to learn about noise reduction in test setups. I've seen some people in a physics lab using grounded aluminum foil around their test leads to reduce noise.  I think it would be interesting to see how the pros do it.
Title: Re: #579 - precision current source
Post by: ecat on February 12, 2014, 04:50:30 pm
Why use diodes to add a constant voltage drop on the output of a constant current source, when a resistor would work just as well?

Yes, I had intended to mention that, but forgot, oh well. I had mentioned a diode in a previous video, and I wanted to show that you wouldn't get the expected  voltage drop at the lower current.
It's probably stupid to show that, but it is what it is.
Oh, and I was going to add some annotation too to mention that, but of course forgot to do that as well because I'm double stupid.
The final circuit I built already has a resistor instead.

It was interesting to see the diode in operation and find out why it worked and under what circumstances it appeared not to work. Time well spent imo.

Any benefits/problems fixing the rail to rail issue by running the op amp from a separate supply? Say, 15V (+/-?) for the ref and +/-18V for the opamp.

Title: Re: #579 - precision current source
Post by: EEVblog on February 12, 2014, 05:15:27 pm
Any benefits/problems fixing the rail to rail issue by running the op amp from a separate supply? Say, 15V (+/-?) for the ref and +/-18V for the opamp.

Should be none. You just need the negative supply for the opamp, which is not always convenient.
Title: Re: #579 - precision current source
Post by: ecat on February 12, 2014, 05:32:36 pm
Any benefits/problems fixing the rail to rail issue by running the op amp from a separate supply? Say, 15V (+/-?) for the ref and +/-18V for the opamp.

Should be none. You just need the negative supply for the opamp, which is not always convenient.

kk, thanks :)
Title: Re: #579 - precision current source
Post by: Dr. Frank on February 12, 2014, 09:27:23 pm
Dave, once again a very instructive and well done video on low value current sources.

The VHP100 donated by Vishay is even more stable, than you thought. They are real jewels..

First, this type is hermetically sealed and oil filled!
That removes any oxygen and humidty, and delivers an ultra stable shelf-life (no or small load) over time, as small as 2ppm/ 6years.

The non hermetically sealed ones, i.e. the molded versions, age around 20ppm/year, also shelf-life.

Second, the VHP100 resistors are even better regarding T.C. than the Z foil types.
The Z foil have 2ppm/K max only.

The VHP100 contain 2 resistor chips, one of the C-type, the 2nd of the K-type bulk metal foil resistors.
Those were the denotations of the older resistor technologies from Vishay.

K and C type have opposite sign of T.C., and are matched in this package, to give a much better overall T.C. compensation.

Therefore, the VHP100 have max. 10ppm change over 15..45°C, that is a max. T.C. of 0.3ppm/K.

Therefore, treat those resistors very carefully, they can serve you as real ultra stable resistor standards.

You only have to determine their exact values once, and then keep them away from loads > 10mW.
Best you assemble them in a case with 4 terminal Kelvin connection for best accuracy.

Frank.
Title: Re: #579 - precision current source
Post by: EEVblog on February 12, 2014, 11:05:24 pm
The VHP100 donated by Vishay is even more stable, than you thought. They are real jewels..

I have even better ones supposedly (or just as good). A couple of 4 terminal oil filled HZ ones of these at 10K that I showed a glimpse of:
http://www.vishaypg.com/docs/63120/hzseries.pdf (http://www.vishaypg.com/docs/63120/hzseries.pdf)

They are both part of their "secondary standard" range of resistors.
http://www.vishaypg.com/docs/49987/secstan.pdf (http://www.vishaypg.com/docs/49987/secstan.pdf)

And also some of these naked ones I showed in the microscope video. They cater to the audiophool market!
http://www.vishaypg.com/docs/63140/var.pdf (http://www.vishaypg.com/docs/63140/var.pdf)
Title: Re: #579 - precision current source
Post by: Dr. Frank on February 13, 2014, 12:01:43 am

I have even better ones supposedly (or just as good). A couple of 4 terminal oil filled HZ ones of these at 10K that I showed a glimpse of:
http://www.vishaypg.com/docs/63120/hzseries.pdf (http://www.vishaypg.com/docs/63120/hzseries.pdf)

Well, the VHP100 are supposed to be superior over the HZ series ones, as confirmed by a Vishay representative.

I bought 5EA of the  Z foil types, i.e VHP202Z, to use them as standards.
But initially they were a bit of disappointing for me, having relatively high T.C. between 0.3 and 1.0ppm/K.

They have 9.9998k, 0.001%, 80€ each.
Adjusted to exactly 10k by comparison against an SR104, in the end.
As each resistor assembly is equipped with a thermometer, I was able to determine each individual T.C. = dR/dT, which mitigates the unceratainty due to temperature.

Their ageing drift against each other is well below 1ppm over 4 years, until now.
Superb.

Frank

Title: Re: #579 - precision current source
Post by: Co6aka on February 14, 2014, 03:56:51 am
...but of course forgot to do that as well because I'm double stupid.

Not stupid, you've simply contracted CRS, which is among things that happen when you grow old (which is a damn-fool thing to do.) So, from now on, any day that you don't let the magic smoke out is a good day, and you did set a current limit on your PS. :-+

Anyway... What's the noise spec for the 10M version of the resistors? Prolly not an issue with 1uA and 0.05% though, but... :-//
Title: Re: #579 - precision current source
Post by: Wilbert on February 14, 2014, 08:40:51 am
Good stuff Dave! I enjoy and learn allot from your videos.  :-+
Title: Re: #579 - precision current source
Post by: digital on February 14, 2014, 07:07:13 pm
:=\Dave a really interesting video and subject matter,thank you
Title: Re: #579 - precision current source
Post by: alex.m on February 15, 2014, 02:47:46 am
Great video Dave.  :)

If you want to use this kind of circuit as a bias current for a sensor how can you use the same Vref for an ADC Vref pin?
Title: Re: #579 - precision current source
Post by: ohmineer on February 15, 2014, 07:36:20 am
Thanks for this series of videos!

That's the kind of video I would love to see all the time at EEVBlog.
I hope you spend more time this season recording/editing/publishing them.
Discussing about the secrets of a good design is more instructive than any theory masterclass.
Title: Re: #579 - precision current source
Post by: hgg on March 05, 2014, 09:17:55 pm
Hello Dave,

I really enjoyed your video series on the precision current source.

I saw in your video that, you lifted the ground of the reference
instead of buffering its output.  What is the advantage of that?

Thanks.
p.s. What happened to the mailbag Monday ??  :)
Title: Re: #579 - precision current source
Post by: bgoodr on December 13, 2014, 06:23:16 pm
Hi Dave,

Sorry to dredge up this old thread, but I'm trying to breadboard this constant current source, and need to confirm where the series of 1, 2 or even 3 diodes go.  Are they in series with the load resistor and ground?  And is the V- pin on the opamp also to go to the anode of the top-most diode?  The attached drawing shows how I'm thinking is the correct circuit.

Thanks,
bg
Title: Re: #579 - precision current source
Post by: sergey on December 14, 2014, 02:40:55 am
Hi Dave,

Sorry to dredge up this old thread, but I'm trying to breadboard this constant current source, and need to confirm where the series of 1, 2 or even 3 diodes go.  Are they in series with the load resistor and ground?  And is the V- pin on the opamp also to go to the anode of the top-most diode?  The attached drawing shows how I'm thinking is the correct circuit.

Thanks,
bg

It doesn't seem to be correct to lift the negative opamp rail with the diode voltage drop. The idea of diode is to lift the positive opamp input up a bit so it doesn't operate with voltages close to the power rails.

So from my understanding you need to use diodes in series with only the load, leaving the opamp rails as is.
Title: Re: #579 - precision current source
Post by: bgoodr on December 14, 2014, 05:23:43 am
Makes sense.  Find attached updated circuit diagram; is that what you mean?
Title: Re: #579 - precision current source
Post by: sergey on December 14, 2014, 05:42:50 am
Makes sense.  Find attached updated circuit diagram; is that what you mean?

That is indeed exactly what i meant :)
Title: Re: #579 - precision current source
Post by: jancumps on December 14, 2014, 07:06:05 am
It has been discussed in the beginning of this thread: you can use one resistor in stead of 3 diodes.
Because the current going through the load is constant, you can calculate what resistor you'd need to create the desired voltage drop using Ohm's law.
Title: Re: #579 - precision current source
Post by: diegosfb on January 23, 2015, 01:19:10 am
Hi guys;

One quick question. I tried to make one of these the other day, but since I didn't have the REF102 chip I used a REF195G instead (5v reference) and adjusted the resistor for the current I needed.

I know the precision won't be the same, but the thing is that it doesn't even work. When I measure the voltage reference between brief and "GND" of the REF195G I dont get the 5V I am supposed to get.
Is there any special thing about the REF102 that other voltage references like the REF195G dont have that makes this circuit only work with REF102?

PS: I also changed the opAmp for a OPA333 wich is quite a good opamp

Best Regards,
Title: Re: #579 - precision current source
Post by: RDbrvn on May 19, 2015, 04:12:26 am
Hi all!

I am using this precision current source to generate 10 uA current. Op amp is OP1177, REF is REF102. The current changes by 0.5% when the load resistance changing from 50 kOhm to 150 kOhm, but I use unscreened wires. The output current contains ac contribution that depends on the load resistance (approximately 0.1 V). How to get rig of this AC contribution? Is it a tycal accuracy for this current source? How can I additionaly improve the accuracy of current?
Title: Re: #579 - precision current source
Post by: sync on May 19, 2015, 05:04:51 am
I am using this precision current source to generate 10 uA current. Op amp is OP1177, REF is REF102. The current changes by 0.5% when the load resistance changing from 50 kOhm to 150 kOhm, but I use unscreened wires.
How did you measure the current?
Title: Re: #579 - precision current source
Post by: RDbrvn on May 20, 2015, 04:35:12 am
How did you measure the current?
I measured the current using the ammeter Keithley 6517B connected in serial with load. Also I measure the current throught a voltage drop on the precision 100k resistor connected in serial with load. The current drops with the increase of load resistance.
Title: Re: #579 - precision current source
Post by: sync on May 22, 2015, 07:19:32 am
IMHO 0.5% current change is a lot. I get about 0.06% from 1kOhm to 1MOhm load resistance with the attached circuit. Which I don't call a precision current source.

Have you checked your circuit for oscillations? I would also check if the REF102 output voltage is constant.