ISO 120/121 Connections

Chuck,

I do not recommend using the ISO120 or ISO121. These are expensive devices that are not recommended for new designs. For any present users who may read this, there is no intent to make these devices obsolete, however they are not recommended for new designs.

I think that the INA148 would be a better choice for your application. This is a difference amplifier with high common-mode voltage range. It would measure each battery voltage differentially, level shifting that voltage to ground reference.

If, for some reason, you need to use an isolation amplifier, consider the ISO124. Understand, however, that an isolation amplifier requires an input-side power supply voltage. Where would you get this power supply voltage?

Regards, Bruce.

Bruce,

  I've been looking at differential amplifiers; they definitely seem like what I need. The INA148 says it can take a 75V difference at the +5V supply, and up to 200V at the +/- 15V supply. 

  Is it 75V from (-)Vs to ground, or 75V from (+)Vs to ground? 

  At max charging voltage, the highest I'd expect to see the negative terminal of the highest potential battery is +72.5V over pack ground, and the positive terminal would be about +87V over pack ground. I'd like to use this chip with one +5V supply instead of resorting to the +/-15V supply to get the 200V protection.

  If I *do* need to go the +/-15V supply range, I'll need to move from the single 12V battery I was hoping to use to a 24V power supply. Our vehicle happens to have a 12 and 24V bus for other applications, but it's not the preferred route I'd like to take. If it is required, I'm looking at the DCP022415DP,U to supply the INA148 with the +/-15V, as I believe the data sheet specifies you must use a dual supply for the 15V source. 

  My question about the DCP022415DP, then, is if I am allowed to tie the output common to the input ground. All the reference circuits in the datasheet show the output ground separate from the input ground, but again, the end result I'm trying to go for is to get everything to a common ground so I have a common point of reference.

  I'm sorry for asking for so much assistance. I am a mechanical engineer, and the minutia of circuit design is still new to me. 

Thanks very much for your help

  Sincerely,

  Chuck Saunders

  Veteran, USN 

Chuck,

You cannot use a 5V supply because the INA148 is a G=1 device. Measuring a 12V battery, you need a positive power supply that is a couple of volts higher than the measured voltage to allow an accurate output.

I suggest that you use the 24V supply with no negative supply (V- pin of the INA148 is grounded). I don't see a need for the DCP power converter. This will give you plenty of common-mode range. It will allow measurement of a pretty wide range of battery voltage from less than a volt to over 15V. The Vref terminal is connected to ground so your output voltage is ground referenced.

The first battery in the stack needs no amplifier, right? Its voltage is already ground referred. Each additional 12V battery up the stack is measured differentially with the input pins of an INA148 and its output is now ground referred.

Is that clear?

Regards, Bruce.

Bruce,

  First, thanks so much for your help. Your replies are always prompt and insightful. 

  If I could use the existing 24V bus with no voltage regulator that would be great. I read where the data sheet says max differential voltage between (+)Vs and (-)Vs is 36V, but the data sheet also says to use a dual 15V (+/-) for the 200V common mode protection. Is it safe to use 24V for (+)Vs and ground (-)Vs?

  Everything you've written has been very clear, I just want to make absolutely positive it's okay to hookup 24V directly to the IC without risk of damaging it. 

Thanks again for your assistance,

  Sincerely,

  Chuck Saunders

  Veteran, USN

:EDIT:

  I would just like to add that I am absolutely going to add the INA148 to protect each battery as I am designing this board for use on a vehicle to be tested by Marines. I am positive I will not be there every time someone connects the batteries, and I don't want to risk frying the board because the batteries were connected out of order. 

  Chuck

Chuck;

Having an INA148 monitor each battery, including the first one, is a good idea as it will provide buffering to the battery outputs and prevent catastrophic failures if the outputs are accidentally shorted.

Bruce provides good guidance for your application.

Thanks, Neil. Good point.

Chuck,

Yes, it is safe to put 24V from V- to V+. It's the total voltage that matters in this case. It's not necessary to regulate this voltage if it is reasonably stable and does not have large voltage spikes present.

Another point worth mentioning. The inputs of the INA148 do draw some current from the battery under measurement. The current in each input is approximately the measured voltage on that pin divided by 1M-ohm. I believe you have seven 12V batteries in series so the top battery would have 7*12/1M = 84uA load due to the INA148's measurement current. The bottom battery in the stack has the most current. It tallies up to 48-times the current of the top battery (actually 49, adding the last amp on the bottom). This is a load of approximately 4mA. It's not a lot but could influence shelf life if it sat around for a long time (not sure of the capacity of your batteries).

Regards, Bruce.

 Bruce,

  I'm not entirely sure what you mean by the bottom battery has the most current. If the current draw is the measured voltage, wouldn't the chip monitoring the highest potential battery draw the most current? The pins would (nominally) be at +60 and +72V; the pins on the lowest potential would be at +0V and +12V. 

  We're using six 120Ahr batteries, so even at 4mA we'd never notice. 

Thanks again,

Chuck Saunders

Veteran, USN

Chuck,

I wish I had taken a picture of the drawing on my white board. I had to draw it out to see where the current was flowing. It makes an interesting arithmetic progression.

It's true that the inputs on the top draw the most current but the bottom battery must supply current for all the inputs connected to it and all those above it in the stack. It adds up.

Regards, Bruce.

Chuck,

Here is a diagram of all the currents. The current is all scaled to I = 12V / 1M = 12uA. So the current in the bottom battery in the stack is 48 * 12uA = 576uA.

I made a mistake in a previous posting calling the current in the bottom battery "approximately 4mA."  (I mistakenly multiplied the top battery's current, 7I, by 48.)

And on this Veteran's Day 11/11/11, thank you (and all veterans!) for your service.

Regards, Bruce.

Bruce,

  I wrote up the whole circuit and ran it by the project lead. He's adamant about everything on the high voltage side being completely isolated from the low voltage monitor; the high-common mode differential amplifiers are not sufficient.

  So, I revisited the datasheet for the ISO120 isolation amplifier. Figure 10 shows a suggested circuit for a high voltage battery monitor that uses the battery pack as the power supply for the high-voltage input for the isolation amps. I don't want to do this as I'm worried the constant power draw will skew my voltage measurements when the battery pack is resting.

  I'm planning to use wide-input 15V isolated DC power supplies to run the isolation amplifiers on the high voltage side, but I'm not sure if I need to provide a +/- Vs, or if a +15V input is sufficient and I can tie the -Vs pin to the high side ground. I'm not anticipating any negative voltages. I'm looking at about 12 dollars per isolated power supply (and I'll need one or two per battery), so I really don't want to provide a -Vs if I don't need to. 

  Also, you mentioned that the ISO120/121 series are not recommended for new designs. Is there a similar product line that is?

Thanks so much,

  Chuck Saunders

  Veteran, USN

Chuck,

I recommend that you consider ISO124 or AMC1200.

ISO124 is a more like the ISO120. The AMC1200 is a newer device intended for current shunt measurements in motor control, though I think it could be used for your application. The input voltage range is quite narrow and operating voltage much lower. You would need to divide the battery voltage down and buffer it to drive the AMC1200. Still this may be the best approach because of its 5V power supply.

Regards, Bruce.