# 2.5V Refernce LM4128 problem

Discussion in 'General Electronics Chat' started by mcgyvr, Jul 9, 2012.

1. ### mcgyvr Thread Starter AAC Fanatic!

Oct 15, 2009
5,395
1,196
Having a problem here..
I'm using the attached circuit to measure -75 to 75V and output it to 0-5V for sampling by a micro (2.5V -/+2.5V).
Except instead of the 2.5V voltage as shown in the image I'm using a LM4128 2.5V series voltage reference. (personally never used a voltage reference chip before)

Reading negative voltages works just fine and the output of the LM4128 stays stable at 2.5V. (Vout goes from roughly 0V to 2.5V as Vmeasured goes from -75 to 0V as I want)
The problem is measuring positive voltages.. At about 10V into Vmeasure I'm getting 4.5V or so at the output of the LM4128 causing the circuit to not work properly. (basically as Vmeasure is going up positive Vref is also going up)
??help

Here are the voltages I'm measuring at Vout (don't have the micro attached yet..All measurements made with a fluke multimeter)
Vmeasured----Vout
Negative input voltages
0V=2.5V
-10V=2.17V
-20V=1.84V
-30V=1.51V
-40V=1.19V
-50V=0.86V
-60V=0.53V
(basically .33V per 10V) OK..Seems to work fine..

NOW..
Positive input voltages
0V=2.5V
5V=2.75V
6V=3.18V
7V=3.64V
8V=4.02V
9V=4.57V
10V=4.57V (stops here and won't go any higher)

If I replace the LM4128 with a 2.5V power supply all works just fine. Maybe the LM4128 is damaged? (edit..Just tried another LM4128 and no change)

• ###### Vref Problem.PNG
File size:
43.5 KB
Views:
28
Last edited: Jul 9, 2012
2. ### #12 Expert

Nov 30, 2010
18,093
9,683
Does this do you any good?

File size:
4.3 KB
Views:
26
3. ### mcgyvr Thread Starter AAC Fanatic!

Oct 15, 2009
5,395
1,196
No removing R4 doesn't help.
Or ..Are you also saying I need to reduce the resistor values too?

Last edited: Jul 9, 2012
4. ### #12 Expert

Nov 30, 2010
18,093
9,683
What I'm saying is that I completely ignored your circuit and designed one that will convert +/-75V to 0-5V where +75 gets 0 volts out and -75 gets 5 volts out.

5. ### mcgyvr Thread Starter AAC Fanatic!

Oct 15, 2009
5,395
1,196
No you didn't. It needs r4 to function properly I believe.
And that doesn't help me at all.

6. ### #12 Expert

Nov 30, 2010
18,093
9,683
Somebody else will give it a try.

7. ### mcgyvr Thread Starter AAC Fanatic!

Oct 15, 2009
5,395
1,196
o...k... ummm.. thanks I guess?

Still need help..
I just don't know why a 2.5V voltage reference IC is different than just a 2.5V regulator.
That seems to be the problem.. Why isn't the 2.5V reference behaving like a regular 2.5V regulator when positive voltages are trying to be measured.

8. ### mcgyvr Thread Starter AAC Fanatic!

Oct 15, 2009
5,395
1,196
anyone..anyone.. beuller..beuller??

Is it that this LM4128 voltage reference cannot sink current as its required with positive input voltages??
If so can someone recommend a 2.5V reference that can sink enough current for my application..
(maybe like MAX6100-MAX6107 or MAX6225)

Last edited: Jul 10, 2012
9. ### takao21203 AAC Fanatic!

Apr 28, 2012
3,785
502
I use LM7805 with 470 Ohms voltage divider for 2.5V reference.
Don't know if it can sink current.

10. ### MrChips Moderator

Oct 2, 2009
17,603
5,472
I wouldn't post my suggestion until I get a chance to test it on a breadboard.

11. ### mcgyvr Thread Starter AAC Fanatic!

Oct 15, 2009
5,395
1,196
Any suggestions are appreciated.. I think I'll order up some of the MAX chips or try to find another that can source/sink

12. ### mcgyvr Thread Starter AAC Fanatic!

Oct 15, 2009
5,395
1,196
How about just adding a 1k (1/16w?)resistor from the output of the 2.5V ref to ground to ensure a load is always there? Any possible problems with that?

When I add that resistor I get the proper output for both positive and negative voltages.

• ###### Vref Schem.PNG
File size:
51.4 KB
Views:
13
Last edited: Jul 10, 2012
13. ### MrChips Moderator

Oct 2, 2009
17,603
5,472
Here is my circuit:

U1 is a single supply rail-to-rail op-amp.

You can trim R1 for gain adjust and R2 for offset adjust.

14. ### Ron H AAC Fanatic!

Apr 14, 2005
7,049
675
Like most positive series references and regulators, LM4128 cannot sink current. Your load needs to sink more current than the measured voltage will source. 1K is OK.

15. ### mcgyvr Thread Starter AAC Fanatic!

Oct 15, 2009
5,395
1,196
yep just finished prototyping the complete circuit (way more than just the little bit I posted here.) and it all works perfect with simply adding the 1K resistor to ground..

Thanks again to all.

16. ### mcgyvr Thread Starter AAC Fanatic!

Oct 15, 2009
5,395
1,196
Thanks..looks very simple.. I'll give that a try. I really like the gain/offset adjustments.