High side Current sensing

[nura100]

Hello Forum,

A Happy New Year to you all !

its been quiet some time since i visited the forum, a lot of new things are happening.

coming to the point, i need help in measuring the High Side Current sensing between charger, battery and load.

i can sense low side current between load and the battery but i need to measure the current between charger and battery and feel high side is the only way to go .... or i am wrong.

i have seen high side current sense amplifiers like INA169/170 from TI, MAX4080/81 from MAXIM also some from Linear etc.

but they are 1) very expensive by the time i get them at my doorstep
2) non of them are not second sourced

i know that now-a-days Rail-to-Rail input output op-amps can do the job with +-1% error but these too suffer from the same problems mentioned above. Hall effect based sensors too join the list.


i need help to do the job with normal op-amps, or if both low-side sensing is possible between charger-battery-load.

the sensed voltage is read by a micro with ADC ( no differential input allowed, only referenced to gnd)

Micro is powered by battery. current to be measured is around 10-15amps


Thanks in advance

Regards

Arun

 

[Roff]

What part numbers are on your list of "normal op-amps"?

 

[MikeMl]

What is the voltage between the high and low sides? How much current are you willing to waste in order to make the measurement? Can you go through a junk box of resistors and match two sets of two to better than 1%?

 

[nura100]

Thanks for replies


@ roff

What part numbers are on your list of "normal op-amps"?

to begin with LM324, LM358, TLC2272

@ MikeMl

What is the voltage between the high and low sides? How much current are you willing to waste in order to make the measurement? Can you go through a junk box of resistors and match two sets of two to better than 1%?

Lead acid Battery charger voltage can be varied from 14V to 84V(i have dimmerstat/variac on the primary of charger trafo) to charge 1 to 6, 12V batteries in series. Rsense is 0.01R, 10 watts and i can get 1% resistors and can live with 100ma - 200ma error in 10A current reading.


also i want to know if low side sensing is possible for both charger-battery and load-battery


Arun

 

[MikeMl]

I'm assuming that want your data aquisition system referenced (grounded) to the low side. High-side monitoring requires a differential amplifier which has a common-mode range of ~90V, which is hard, but not impossible to do.

There is no reason why you couldn't do low side monitoring. Look at the attached diagram. The only minor complication is that the voltage developed across the shunt goes both positive and negative with respect to ground. This meters the net current into/outof the battery. Many OpAmps allow their inputs to go up to -500mV with respect to their Vee pin (which will be at ground).

 

[smanches]

Are AD8227 instrumentation amps available and worth the price? I've used these for high side sensing before.

 

[MikeMl]

Are AD8227 instrumentation amps available and worth the price? I've used these for high side sensing before.

The common mode input range of the AD8227 is (-Vs - 0.1)V to (+Vs - 0.8)V. The max differential between -Vs and +Vs is 36V, so to do high-side sensing at 84V, you would have to put resistive voltage dividers from both sides of the shunt to ground. These resistors have to be matched to ~0.1%, so if you are going to go to all this work, you are effectively build a custom instrumentation amp, so why buy one?

 

[nura100]

@ MikeMI, Thanks for providing insight and diagram

Actually, I did a micro based Digital voice recorder using an AVR and SD card and have used an op-amp to amplify the Microphone's input voltage, the supply to op-amp was at 5V and i used a divider to hold reference of Vin+ of op-amp at 2.5V, so that the positive audio signal can be between 2.5V to 5V and negative signal between 2.5V-0V. and the ADC's Vref was set at 5V. so when no audio signal is present (silence) the ADC would read 2.5V

do you suggest i do a similar thing for low side sensing of charger-Battery-load ?

can an LM358 do the job ?


Thanks

Arun

 

[MikeMl]

Right, you want 0mv out of the shunt to map to 2.5V at the AD input; -XXmV to map to +5V, and +XXmv to map to 0V. Gain and Offset...

The LM358 has the common mode input range, its output will pull near 0V, but it will only pull up ~4V on the high end. A CMOS rail-to-rail in-out OpAmp like a TLV2371, TLV2461, TLV2770, OPAy340, or OPAy363/364 would be better.

 

[nura100]

yes, i have TLV2252 rail to rail opamp and want the op-amp to output 2.5V when no current flows and towards positive (5v+)rail when current flows from load to the battery and towards Gnd rail (0v) when current flows from battery to charger.

i need a gain of 10 ( 100mv amplified to 1v when 10 amps is flowing thru Rshunt), can you show me how to connect the op-amp in above diagram you have provided. the micro, adc, op-amp are powered by single supply from the battery.

thanks

Arun

 

[MikeMl]

Gain required is 25. This provides -25.

 

[nura100]

Thanks a lot for your time and help,

in the above chart, when -8A flows thru Rshunt, Vout of op-amp shows 4.5V and when 8A flows, Vout is 0.5V, can it be done the other way round ? will swapping the INV and N.INV pins of Op-amp do the job ?

Also i came across the following diagram which level shifts and scales the input negative voltage to 0V-5V, can you comment on this, please...




Thanks

Arun

 

[MikeMl]

Only by adding a second inverting amplifier. It is very easy to invert the signal in software inside the PIC.

 

[nura100]

Hello Mike,

Sorry for being buggy, but can you explain how you are setting the reference to 2.5V for the op-amp in your diagram, i find 470K and 20K pot for offset adjustment, where or how are you dividing the 5V+ to 2.5V ?

the gain is set to 25 by 10K and 250K which i understand. i'll be using my AVR's internal VREF which is at 2.56V, so 1.25V will be middle voltage when no current is flowing. what changes should i make in the circuit ? should the gain be set to 12.5

Thanks

Arun

 

[MikeMl]

When the voltage across the shunt is zero I wanted to get the opamp output to 2.5V. With respect to the non-inverting input (the pot wiper), the gain is 1+250K/10K = 26. So the input would have to be 96.2mV.

The total resistance from 5V to gnd is 470K+20K = 490K, so the current is 5/490K = 10.204 uA. The resistance from the bottom of the pot is 0.471*20K = 9420. The voltage at the wiper is 96.1mV, which is close to what I calculated above. I first created a two resistor voltage divider, and later converted it to the divider with the pot in the bottom. When doing that, you want the wiper to be near the middle at the desired set point.

I'll let you rescale everything for your A/D.

 

[nura100]

I'll let you rescale everything for your A/D

Thank you Mike, eagerly waiting for you reply


Arun

 

[Roff]

Thank you Mike, eagerly waiting for you reply


Arun

I think that was his reply.
He showed you how to center the output at 2.5V. I think he expects you to show a little initiative in centering the output at 1.28V.

 

[nura100]

Thank you Roff for putting some sense in me, i totally misunderstood Mike and was waiting for his reply, simultaneously i have been trying simulating in proteus and think i got it to work but i am not clear if Gain is 25 or 26

mike first wrote:

Gain required is 25. This provides -25.

later in his explanation, he mentioned the gain as 26 (the resistor values tell it is 26), also the TL2252 Rail-2-Rail op-amp doesn't output full +5V when -10amps is sensed, it freezes at 4.5V approx.

can i use LM358 and use 12V supply so that it can happily show full 5V ?

Arun

 

[Roff]

In Mike's circuit, the inverting (signal) gain is -25. The noninverting (offset pot) gain is +26.

You said your ADC reference is 2.56V. Why do you need the op amp output to swing to +5V?

 

[tcmtech]

Am I missing something here? This seems like a lot of run around to do basic current sensing for a battery charger. My physics books say current is equal on both the relative positive and negative of a battery or power source.

If you have a battery charger that puts out more amps on the positive line than what is returning on the negative line I would like to know how!

For my personal application I would just use a Small series resistance on the negative common line and use a simple Op Amp to boost the voltage reference off of it to the levels the ADC needs if it is not capable of reading mv levels directly.

I have used off the shelf digital voltmeters before that are based on ICL7107 IC to make digital amp meters just by having the + input (protected of course) of the ADC get its reading off of the positive side of a low ohm resistance in series with the common return.
Being the actual IC power and - input referance for the ADC on the IC's where on the lower side of current sensing resistor that made the + input see a positive voltage that was proportional to the returning current.