I need to sense current of a power supply. Sense resistor is before the pass transistor, where I have 45 V appr.

I've selected one differential amplifier apt for up to 100 V

Please check here:

http://www.national.com/ds/LM/LM725.pdf

After trying for days to make it work, by mistake I found that to work properly, the supply feeding the opamps must be independent of that from the power supply.

Since what I need is its Vout to help in limiting the current I am sensing, here are my questions:

a) Why in heaven is this happening?

b) How to relate the Vout of this amplifier to the rest of the power supply circuit where I need to limit the current I am sensing?

It seems that the separate supply for the opamp and the use of its Vout in the circuit under sensing is contradictory?

CAn anyone help me?

Gracias

 

Well, is your supply pumping out a regulated or unregulated +/- 14 volts to the diff amp? If it's pumping out raw DC it might be 5Volts but with a huge tolerance. This tolerance could be higher than the 0.5 volt swing 13.5 - 14.5 volt swing. Thats the only thing that came to my head when you said you powered the diff amp seperatly and it worked. You might want to check it with a scope, while it's connected to the opamp and pulling current.

As for your second question, I don't understand what you are trying to do "relate the Vout of this amplifier to the rest of the power supply circuit " what specifically are you trying to compair Vout to the power, Vout to the current being pulled?

I may not know the answer to your questions but I do work with hardware engineers every day, and I sit right next to guy who designs our power supplys.

- Matt

 

Matt, thanks for replying.

a) Shame to me but having done my testing more than one year ago, right now I am not sure if I actually used a split supply. I think so, but better I check it and then come back.

b) Once having a Vout proportional to load current, I intend to use it to limit maximum current. Vout and some variable control voltage, managed with a PIC + DAC would (through a transistor, perhaps?) do the limiting. I foresee an opamp also involved.

Is any chance that the guy next to you, suggests the basics of this? I like experimenting but not being an engineer, he could save some frustrating false starts for me.

Just in case: it is a dual power supply so I will duplicate the setup for each rail for independent limiting with proper sign reversing.

Variable control and consequent regulation is already solved and under control of the PIC.

Vout is going (abt.) from 1 to 45 V. Current probably no more than 2 or 3 A, depending transformer to use.

Thanks for your time.

 

Well, he's out until monday unfortunatly. But I'll still ask around, if you have any documentation on a suggested layout / application layout just send me a link or email it to me at matthew.davis@mu.edu. I have a pretty good idea of what you are trying to do, and I bet there is AtoD IC that can easily do this AC voltage comparison. I'm actually looking on analog devices site right now (www.analog.com) for a design I'm doing. Maxim (www.maxim-ic.com), dallas semi conductor might also be another great spot to look for a solution. Those are the 2 big ones that I know of off the top of my head to start looking at.


Also you have aol IM msg PsychoticDavis, I'm on all the time.

- Matt

 

Matt,

Tks again. I am frantically trying to recover all the papers to upload something for you somewhere. Please give me time.

It would be during Sat / sunday.

BTW, I try to go "discrete" because here in Argentina, all those good chips are very expensive and not easy to get.

Gracias.

Agustín

 

Matt, hi again.

After not finding the circuits and my notes I went to the hardware, thanks God, intact.

Here is what I did to test again this diff. amp. together with my power supply:

a) Dual voltage used +/- 15V for the opamps.

b) Circuit in http://www.national.com/ds/LM/LM725.pdf (apt for up to 100v) implemented with two OP77.

c) Sensing resistor (0.47 ohms) applied to (+) and (-) inputs of the circuit. Different high wattage resitors applied as load for the supply to vary the load current.

For five different currents I took the five following measurements (all in mV):

a) Vsense
/
b) Vout with separate grounds
c) Vout with common grounds
/
d) Vout with separate grounds inputs reversed
e) Vout with common grounds inputs reversed


1170 / 1186 1155 / 1203 1239

246 / 244 218 / 250 279

127 / 126 93 / 129 158

87 / 87 53 / 91 122

45 / 44 7 / 47 81

My preliminary onclusions:

a) Sloppy testing (as I did the first time) is as bad as no testing at all. Shame to me.

b) My assumption that I needed separate supplies was wrong since extending the range of currents, Vout started to come closer to Vsense.

I noted that between each pair of Vouts there is an almost fixed difference of appr. 30 mV and of course with lower Vsense that difference distorts in such a big proportion the Vout value, what brought me to a wrong conclusion. (My fault: not testing this for the whole range).

If you look at the "reversed" measurements, the sign of that difference, is opposite.

c) It is worth to note that even in the high end of the range, Vout with separate grounds (red value) is really close to Vsense.

I presume that all boils down to poor CMRR or ...?

Besides possibly unmatched resistors in the feedback networks, what else whould you look at for? Ground loops or poor grounding? The diff. amp is soldered in a perfboard.

OK, that's all for now. Please let me know your opinion.