Propeller as a precision volt meter / amp-meter
Reach
Posts: 107
I would like some suggestions here. 
I have decided to place a few shunts on a couple of battery banks to monitor current consumption.
Battery voltage range needs to be 0 - 16 V perhaps higher 0 - 48 V.
Im thinking about using delta sigma but am not sure of its ability to measure a wide range of voltage and remain precise -down to mVolts for the shunt to read current.
Keep in mind this will be installed on my boat so marine environment. How would you go about this? Please no HALL sensors I want to stick to shunts.
Thanks in advance.
I have decided to place a few shunts on a couple of battery banks to monitor current consumption.
Battery voltage range needs to be 0 - 16 V perhaps higher 0 - 48 V.
Im thinking about using delta sigma but am not sure of its ability to measure a wide range of voltage and remain precise -down to mVolts for the shunt to read current.
Keep in mind this will be installed on my boat so marine environment. How would you go about this? Please no HALL sensors I want to stick to shunts.
Thanks in advance.
Comments
http://www.microchip.com/wwwproducts/Devices.aspx?product=PAC1720
These are pretty ideal as volt meter / amp-meter, as they have
* Dividers for Battery voltage included
* Skewed gains on the I and V channels, designed for mV shunts on I
* easy to talk to.
Only the small package is a slight drawback.
That is outstanding thanks. Ill read over this data sheet it looks promising.
The data is a little vague, but the main issue seems to be pins more than electrical
SENSE+ is the Vsource pin, but common mode range is 0-40V
So normal use is High side connection (just move the shunt ?), where both I and V on one channel are read.
That said, the 1720 does have two channels, so you could use one for V+ and one at GND for Current.
Some series resistors on battery read-off pins would be a good idea, as you want some oops/slip protection to avoid literal melt-downs...
Does this involve a Blue Seas battery fuse system?
That's the only valid reason I can think of.
But even then, a jumper from the fuse to a distribution bar will
allow you to pick up everything but the starter motor.
And I doubt you would want to include that.
Looks like im going to have to design a circuit to do this - I swore there was an IC out there
Read #5, you can use a PAC1720, one channel for each task.
There are other current/voltage monitor devices out there, just check their common mode ranges.
I've a datasheet on an in-circuit device that can handle up to 50A.
If this is of interest, I'll wade into my pile of paper after work.
Current up to 100 Amps for now. But Voltage is more important to me. If I can read a voltage range of 0 - 16 V in uVolt increments id be happy. This might not be possible for one IC.
Sure, pick any of these with i2c showing
http://www.microchip.com/ParamChartSearch/chart.aspx?branchID=11022&mid=10&lang=en&pageId=79
I'm confused.
If you're attempting to measure the current, why are you concerned about a voltage range of 0 -16V?
For a current of 100A, 16V across the shunt would indicate a 160 milli-Ohm value. Is that the value of shunt you intend to use?
You need to add a simple resistive divider.
Another choice is
http://www.ti.com/product/lmp92064
SPI, Low side, but voltage channel is 2.048V, so again you use a divider of a couple of precision resistors.
? the ICL7106 has only a 2V signal range, so still needs external voltage divider.
but it can only do 40v with external ranging resistors, which is the same as all the parts linked above.
- so the ICL7106 is nothing special - others are easier to talk to and better precision.
I can't understand why you don't go with your original suggestion of a sigma delta A/D circuit right onto the propeller. All it takes is 2 properly calculated resistors and a 0.1 cap. You could put two such circuits on the propeller, one ranged for volts, and another ranged for current. Or add more yet to get several ranges for current.
Not sure why you might want to measure microvolts though. That requirement might put you too deep into the noise for a simple approach.
Really not that complicated.
Cheers,
Peter (pjv)
The parts are also available in thru hole, so simple to prototype, and a 12 bit adc provides a full scale accuracy slightly better than 0.025%. Don't need better precision than that for this application.
I really wanted to implement this circuit to measure many voltages on my ship --> my engine RPMs, engine temperature, and engine oil pressure, and later if I take the engine out I could use the same circuits to measure solar and wind, etc. Basically as this develops it seems that I want to make a multi-port Volt meter that revolves around the propeller.
- Op amps can be used to amplify low level signals such as those from current shunts and pressure or temperature sensors.
- Voltage dividers can reduce high voltages to levels the adc can measure.
- RPM is already a digital signal so all that is needed is to count the pulses for a fixed period of time and do a simple calculation.
The prop can easily do everything you have listed and more. All it needs are some op amps for the low voltages and resistor voltage dividers for the high voltages.
Kwinn
I disagree with you completely. I spent some time today looking for ADCs that will swallow 16 Volts of analog DC and happily send that on an i2c bus running on 3.3V. Is there a IC that does this? Sure but I could not find one, yet. Most ADCs are happy reading something like ~9V to 5V and under.
This is true, as all the gauges on my boat shoot out a voltage to reflect there respected values. RPMs (linear voltage), Oil pressure (linear voltage). These voltages range 0 - 16 V.
Im leaning towards sigma/delta now but an off the shelf IC would be nice.
? You seem to be meaning, with no added components, but the PAC1720 give in #2 can do exactly that, over 2 channels.
It is less common for ADC's to include the resistive dividers, but some in the battery monitor area do, and those also have tens of mV scale choices, for Current Shunt reading.
Mostly, the Generic ADCs give designers a high input impedance, and expect then to choose the two divider resistors to scale.
If you want to work with current shunts, then expect to use low offset/drift opamps, if you want to keep the origin errors to low levels.
http://cds.linear.com/docs/en/datasheet/4151fe.pdf
http://www.ti.com/lit/ds/symlink/ina300.pdf
Interesting device, but it looks to be only a Current-Sense Comparator
- no Analog output pin ?
There needs to be some input signal conditioning whatever chip is used. Time for some old-school op-amp circuits?
CA3140 and dual/quad friends can do rail to rail. For volts, use two 1% resistors in a voltage divider as mentioned previously. Maybe a divide by 6 if it is 16V into a 3.3V device?
For current, use an op amp as a differential amplifier - four resistors and then you can set the gain, as well as not having to worry if it is high side or low side as it is looking at the difference, not the absolute voltages. 50mV input, so set gain at maybe 60.
So for volts, it is two resistors. For current, it is one 8-pin op amp and 4 resistors. Board is flexible because different resistor values can be used to set the gain and the divide value.
I'm drawing up a board at the moment - should have something in the next 48h.
Several of the IC's suggested by others will swallow 16VDC and more, and send that on an I2C bus along with the mV measurement across the current sensing resistor, but those are intended for DC power monitoring and not very useful for what you describe in post 22.
Most of the adc on the market are intended to be powered from a stable supply in the 3V to 5V range, and handle inputs in the 0V to 3.3 or 5V range. They require additional components such as resistor voltage dividers for higher voltages and op amps for lower voltages.
A propeller could do this using a sigma/delta or multi channel I2C adc with resistor dividers to reduce the higher voltages, and op amps to amplify very low voltages.
Lets say you use an external 12 bit adc and the 3.3V supply as the reference voltage.
A 0 16V input signal would go to a voltage divider consisting of a 15K and 3.3K resistor.
That would scale the input voltage to the adc to 0-2.88V, which the adc can handle.
Multiply the adc output by the calibration factor, place the decimal point, and output the reading. While the adc inputs would be limited to a single voltage range the resistor values and calibration factors could be varied to suit the individual input ranges.
good choice of values as that also limits the current into the propeller pin should you put more than 16V on the input. (Consider a hypothetical where you chose 150R and 33R - same ratio but would feed too much current and zap the chip, plus draw too much wasteful current from the battery). And at the other extreme, 1.5M and 330K would not match the imput impedence of a sigma delta, unless maybe you used an op amp voltage follower. 15K and 3K3 is a good choice.