# Ideas needed on how to design circuit

Discussion in 'The Projects Forum' started by ghines, Dec 9, 2009.

1. ### ghines Thread Starter New Member

Dec 9, 2009
9
0
Hi Everyone,

I'm hoping someone can come up with a couple of ideas how I can accomplish this task.

I have a DC postive voltage range of 1.2 - 2.6 that needs to drive a 1mA moving coil meter. +2.6 volts needs to read zero on the meter and +1.2 needs to read full scale on the meter. ie: +1.9V will read 50% on the 1mA meter.

I have a spare LM358N on the PCB that I would like to use if possible to accomplish this, but it's not necessary.

I have the voltages +/- 15V available on the PCB.

Any ideas how to simply do this would be appreciated.

TIA

Greg

Last edited: Dec 9, 2009
2. ### steveb Senior Member

Jul 3, 2008
2,431
469
The first thing you need to do is understand the grounding. Is your 1.2 to 2.6 V range referenced to the ground in your circuit, or is it isolated? Your question implies that the voltages are referenced to the ground, but it's good to think about these things and be sure. Assuming this is true. Here is one approach.

You can configure your meter to go to full scale at 1.4 V applied voltage. See this section to do that.

Then you can setup one of your opamps to output 2.6 V as a reference voltage for the positive lead of your voltmeter. Your measured voltage then goes to the negative lead of the voltmeter. Note that the 2.6 V reference voltage needs to be relative to your circuit ground.

Depending on the required measurement accuracy and the stability and accuracy of your power supply, I recommend using a voltage reference to establish a precise 2.6 V reference point.

It's too bad the voltage reference isn't 2.5 V because then you could use a 2.5 V reference chip. You can still use a 2.5 V reference, but then you have to amplify it slightly, since 100 mV error may be too much for the manual zero adjust. You can use an AD1508 shunt reference that will give 1.225 V, and you can amplify with a noninverting opamp amplifier with resistors of 10K and 11.3K. This will give a 10 mV error. Or you can use a 1.25 V reference and a 10K with 10.7K resistor. This gives a 12 mV error.

Errors at the 10 mV level should be in the range of the manual screw set adjustment on the meter movement.

3. ### ghines Thread Starter New Member

Dec 9, 2009
9
0
Yes, referenced to ground.

The problem with your suggestion is the 1.2 - 2.6 V range is nominal, so in reality it's probably 1.2 - 2.6 +/- 0.1V from unit to unit. But, I suppose I could still do it this way if I had a variable voltage source for the positive terminal of the meter.

The 1mA meter already has a function, see attached and needs to be somehow switched between being the present function of respresenting the 0-1V source to also being used to represent the 1.2 - 2.6V as previously discussed. Suggestions appreciated. Relay, FET??

Thanks again for your input it's appreciated.

Greg

File size:
7.1 KB
Views:
29
4. ### ghines Thread Starter New Member

Dec 9, 2009
9
0
Steve, If my calculations are correct your suggestion of placing 2.6V on the positive terminal of the meter and feeding 1.2-2.6V into the negative meter terminal will not give what we require.

Sure, an input voltage of 2.6 will make the meter read zero, but 1.2V will not make the meter read FS but rather 140% FS.

5. ### steveb Senior Member

Jul 3, 2008
2,431
469
Can you post your calculations? Did you read the link I gave above? You are basically saying that 1V (voltage difference across the terminals) will bring your meter to FS, which implies a 1K ohm resistance for the meter. So, you simply put a resistor in series with the meter to increase the meter range so that it goes to FS at 1.4V. You can see by inspection this should be a 400 Ohm resistor.

${{1.4\;{\rm V}}\over{1400 \; {\rm \Omega}}}=1 \;{\rm mA}$

By the way, if you make this a 330 ohm resistor in series with a 100 ohm potentiometer, you'll have the ability to scale the FS voltage to compensate for variations. You can also use the same idea in generating the 2.6V reference. You then have three knobs to calibrate with: offset adjust, range adjust and then the zeroing screw on the meter.

Last edited: Dec 10, 2009
6. ### Pich Senior Member

Mar 11, 2008
119
5
Greg
Set up half of the LM358 as an inverting amp and place the meter in the feedback loop then you would adjust the resistor at the inverting input to read full scale and add a potentemeter at the non inverting input set to 2.6 volts. When the voltages at both inputs equals 2.6v the current in the feedback loop will be 0.

7. ### ghines Thread Starter New Member

Dec 9, 2009
9
0
Pich,

Is the attached what you mean?

Greg

File size:
12.5 KB
Views:
29
8. ### Pich Senior Member

Mar 11, 2008
119
5
Yes. The 1k is not necessary, also to protect your meter add ~ 10k in series. You can experiment with that.

9. ### ghines Thread Starter New Member

Dec 9, 2009
9
0
Pich, Thanks that all work fine.

Another question if I may. The 1mA meter is presently being used as per attached and I need it to operate like we previously discussed at the press of a button. ie: representing the 0-1V input and secondly 1.2 - 2.6 volts in the inverting mode.

What would you suggest is the best way to switch between modes? FET's, relays or something else.

Thanks again.

Greg

File size:
11.9 KB
Views:
13
10. ### Pich Senior Member

Mar 11, 2008
119
5
This circuit is probably what you are looking for, something went wrong so I apologize it it was posted twice.

File size:
38.4 KB
Views:
33
11. ### ghines Thread Starter New Member

Dec 9, 2009
9
0
Thank you so much Pich, that's exactly what I wanted.

BTW - What does "Key = Space" mean?

Once again. Thanks.

12. ### Pich Senior Member

Mar 11, 2008
119
5
For simulating the switch, toggling open closed with space bar
The battery voltages 1v and 2.6v you see at the inputs are only for simulating and not part of the circuit I forgot to take them off.
Good luck

13. ### steveb Senior Member

Jul 3, 2008
2,431
469
Good work, but one comment as follows.

I disagree with using a voltage divider directly off the +15 V supply. Granted, you haven't given any specs on required accuracy or tolerance/noise/loading on your power supply voltage, but that approach is not ideal. It's better to run a precision 5 V reference off the 15V supply, and then voltage divide off of the reference voltage.

14. ### ghines Thread Starter New Member

Dec 9, 2009
9
0
Two more questions if I may.

1. Is R2 really necessary, because if it wasn't you could get away with only a DPDT relay instead of a 3PDT?

2. Why the strange value for R2. Why not a standard value of 1K5?

TIA

Greg

15. ### steveb Senior Member

Jul 3, 2008
2,431
469
Of course you need R2. This sets the current to make your meter go full scale with 1.4V. The value is not strange at all. Note that 1.4 V divided by 1400 Ohms equals 1 mA, which is your meter full scale current. That OPAMP is functioning as a voltage to current converter, also called a transconductance amplifier.

A better question is whether you need to include a variable resistor as part of that 1400 Ohms. You mentioned that the voltages might vary from unit to unit.

Overall, you've been given a good starting point, but you should think about ways to improve the circuit for your application. You have provided very few real detailed specifications, or description of the application. So, only you can optimize the circuit.

Things to consider are as follows:

1. Using a voltage reference to stabilize the 2.6 V voltage divider output.
2. Using capacitors to filter noise and give a better DC reading.
3. Using FETs or analog CMOS switches, rather than a relay.
4. Using potentiometers in series with some resistors to allow calibration.

And, you may discover more once you build the circuit.

16. ### ghines Thread Starter New Member

Dec 9, 2009
9
0

As for your four points for consideration, I would like to use an analog CMOS switch rather then a relay. Could you recommend a couple for me that would be a good replacement for the relay?

Thanks again,

Greg

17. ### steveb Senior Member

Jul 3, 2008
2,431
469
The old standby part used to be the CD4066 years ago, but the switch resistance is too high for your case. The newer parts, such as Maxim examples (see attached), are a better choice. You can review these to see which package and type is best for your needs. This document also has a nice explaination with it.

File size:
114.2 KB
Views:
22
18. ### ghines Thread Starter New Member

Dec 9, 2009
9
0
Thanks Steveb.