A shuntless AH meter?

[Patriot]

While toying with my N-scale electric train yesturday, it dawned on me that perhaps an AH meter may be out there which doesn't need an inline shunt.

Now, I'm no expert, but this is just an idea.

Case in point: I was trying to figure out what was wrong with my train, so I had my little mult-meter hooked up to the tracks to see if my transformer box was working or not. As I adjusted it, the current reading varied appropriately as my train sped up and slowed down. (Turns out the tracks just needed to be cleaned.)

Then, I thought to myself, why don't we have any AH meters like this without a SHUNT?

Is there something out there that will measure voltage, detect current draw, and take the readings, divided by time start/stop to give you an indication of AH used, etc?

It may not be as perfectly accurate as the shunted Cycle-Analyst, but with an auto-start detector, it could be pretty close, I would think.

Something like this could allow the use of larger gauge wiring to and from the controller without being constricted down to 12ga and going through a shunt. Not that it's a huge loss, but it does allow for easier installation, by simply tapping into existing wires without cutting them.

 

[Russell]

A conventional multimeter uses an internal shunt to measure current. The only other way I know of to measure current without a shunt is using an inductive pick-up.

-R

 

[Lapwing]

Try an e-bay search for " LEM "

Hall effect current transducers. They come in all sizes and specifications. This is a company with excellent technical data on products, past and present.

 

[rkosiorek]

IC size hall current sensors like the Allegro ACS754 and ACS755 are available. need to supply them with a 5V supply. output can be interpreted on a Voltmeter. also one lead would have to be cut to insert the sensor. but this is not a shunt.

you could also roll your own using a clamp on ferrite core like one of these:



in the gap you can place a linear hall sensor like the A1302.



you will need a 5V supply and possibly a simple op-amp circuit to buffer the signal but it would be an effective clamp on current transducer.

both of these choices would be a lot cheaper that a commercial clamp on current transducer like the LEM. those cost like $100USD each.

so a shunt is often a cheaper and easier way to go.

here is my home-made version of a clamp on sensor using a ferrite core i liberated from a computer monitor cable.

View attachment 1


reads up to 100A reasonably well. as the NASA engineers used to say "Good enough for farm use only"

rick

 

[Mr. Mik]

I bought one of these kits (in Australia):

http://www.siliconchip.com.au/cms/A_30669/article.html

I have not put it together yet, though.

The 150A max current limit is no problem if you have regenerative breaking, it will demagnetize the ferrite core and should be good for much higher currents then.

 

[marshy]

reads up to 100A reasonably well. as the NASA engineers used to say "Good enough for farm use only"

So Rick - how "well" is "reasonably well"? Any idea how accurate the readings are - including in lower current ranges (e.g. 0-10amp)? I'm thinking of trying a DIY hall effect sensor myself (ferrite core and readily-available sensor), ideally looking for some encouraging results first.

or Mr Mik....what measurement performance does your kit claim?

marshy

 

[rkosiorek]

accuracy depends on many variables. drift and stability in the op-amp chosen, linearity of the hall sensor, stability of the resistors, offset drift in the amps, etc.

it would be most accurate around your calibration point. but using typical common 1% metal film resistors and comon op-amps like the LM358 linearity will be better than 3% using something like the A1302 hall sensor. the A1301 will provide better linearity but a narrower range.

accuracy will also depend on the accuracy of your voltmeter. but all things taken into account if you know the limits of your interface the results are very good.

my circuit was designed for a max of 25A, 100mV per amp. it was calibrated for 1A. the variation between using it and the internal shunt on either my Fluke or B&K DMM's is less than 25mA with a 5A current. I think that is pretty good for some junk drawer parts. if i had used a precision 5V regulator instead of the LM7805 (that i have lots of) and better low drift precision op-amps i'm sure the results could be better yet. a 9V battey will last for about a week of continous use.

as it stands it works well as a clamp on meter that i can hook up to a datalogger with little or no trouble. if i want better accuracy i use an adapter built around an ACS754 100A current sensor. but that needs to be inserted into one of the supply leads like a shunt.

rick

 

[marshy]

Thanks a lot for the info rick. I was worried that "garden-variety" hall effect sensor solutions just weren't up to the job, but it seems like they do have potential. Will give it a try!

marshy

 

[rbelisle1]

please please please,,,, someone post a schematic and a parts list!!!!!!!!!!!!!!!!! draw it on a napkin if you have to, but please post it....

bob in phx

 

[Patriot]

What I'd really like to know, is if I could pull the shunt on a "Watts-Up" meter, and attach leads to the hall effect sensors to get the same signal to the meter for fairly accurate information.

 

[rkosiorek]

What I'd really like to know, is if I could pull the shunt on a "Watts-Up" meter, and attach leads to the hall effect sensors to get the same signal to the meter for fairly accurate information.

sure you could.

a shunt is a very low value precision resistor. most often in the renge of 0.001 to 0.020 ohms. Ohm's Law tells us that Votage (E) Current (I) and Resistance are related in the following way E=IR. if the value of the shunt is known and you can measure one of the others the third value can be calculated. for example if we have a 0.001R shunt and when a current is passed through it we measure 1.0V across the shunt, then the current can be calculated. I=E/R = 1.0V / 0.001R = 1000A.

this is the way that the WATTS UP reads amps it reads how many volts are dropped acros the shunt resistor. the maximum voltage for most shunts is usually quite low 50mV or 75mV for full scale readings on the meter.

a hall sensor produces a voltage as well but it is usually higher than that for a shunt. full scale will be normally be 2.5V. since this voltage is higher than what a shunt will produce you could use a voltage divider to scale it so that the outputs will match. just need to know how many mV per A the shunt produces and how many mV per A the hall sensor makes. the 2 resistors in the divider will be in the same ratio.

all that circuitry means more expense. that is why most meters use a shunt. it is simple and easy.

rick

ps id did a schematic on a napkin. but i can't figure out how to upload it. (no scanner)

 

[billvon]

please please please,,,, someone post a schematic and a parts list!

You might find something like this easier:

 

[rkosiorek]

at $20.00USD each the price for the commercial LEM sensor is not bad, lower than i expected actually. wiring would have to be passed through it before it is completed and it needs regulated 5V supply to work. also the hole for the wire is only 3.2mm (1/8") which sort of limits the size of the conductor. like you could not pass a 10gauge stranded cable through it.

the circuit to do you own is really simple:


to set it up is pretty easy. connect the circuit to a DC meter. set it to a mV range. 200mV is the lowest on most meters. R4 is a user control used to zero the meter. set it to it's midpoint. set the R6 "calibrate" trimmer to it's midpoint. then adjust the R3 trimmer to 0V on the meter.

now to calibrate the amps you can use several different methods. one method is to calibrate it against a known walue. for example if you have a CA hooked up . you cold clamp this meter over the same wire and adjust the R6 Calibrate trimmer to match.

if you want to be more scientific you could set up a calibration circuit using a reference battery, a known resistance and some magnet wire. i used a length of 17AWG wire. i looped 50 turns of the wire through the center of the ferrite core. i connected it in series with a 25R 25W resistor i had handy. i attached it to a freshly charged 12V SLA battery. when i measured the battery voltage i found it to be 13.15V when connected to the resistor and the coil. the current going through the coil was 13.15V / 25R (1% resistor) = 0.526A. the wire loop through the sensor acts like a 50X amplifier since it was looped 50 times. so 50 X 0.526 = 26.3A. the meter reads 1mV per amp. so i adjusted R6 until the meter read 26.3 mV.if it reads - instead of + amps on the meter reverse the polarity of the 12V battery. once calibrated you can remove the 17AWG wire.

since the meter will read + and - amps. i suggest that you mark an arrow on the sensor to show which direction will read positive on the meter.

with the values given the meter will read several hundred amps.

rick

 

[rbelisle1]

Now thats what I am talking about!!!!!!! perfect schematic, perfect technical writing.... You should consider a job with Heathkit... oh wait, they are out of business... well then submit this circuit to popular electronics... oh wait,, I think they are out of print... Just showing my age a bit!!!!!!!!

again, thanks for this very cool design!!!!!!!!

bob in phx

 

[rkosiorek]

did you know that the first Heatkit that started the whole thing going was a parasol wing airplane?

built a few of their ahm radio kits back int the 60's and 70's. and i do miss popular electorics, radio&electronics and the rest of those hobby magazines.

they still have ELEKTOR in europe and Silicon Chip in Australia. in fact the idea for the clamp on meter was from an old article from Silicon Chip that i read years ago.

rick

 

[rkosiorek]

looking at the schematic i realize i made a mistake in the 5V precision regulator. a revised schematic is attached.



mV/A sensitivity is set by the gain of op-amp U3b. the gain is set by the ratio of the trimmer R6 + R11 to R7. with the values given to change it to read 10mV/A instead of 1mV/A the setting of R6 would have to be changed. with the numbers given in the previous post for a 1mV/A sensitivity i set the reading to be 26.3mV on the meter. to set for 10mV/A i would adjust R6 for a reading of 263mV.

the values given should be able to provide sufficient adjustment range to cover 10mV/A. if they do not experiment and reduce the value of R7. change it from 18K to 10K.

the limit for the current readings will be limited by the output swing of the op-amp. since it is already 2.5V at oA that only leaves about 2.5V worth of swing. at 10mV/A that allows for a reading of 2.5V/10mV 250A max. in practice it would be less as the LM358 does not quite swing rail to rail.

also it will be a bit sensitive to picking up noise in the leads connecting the A1302 to the rest of the circuit. I would keep the lead short and use a shielded cable like that used for stereo microphones. also C3 - 100n capacitor should be mounted to the A1302 itself.

rick