Solving Amp Hour Meter Problems: Wiring Series vs Parallel

[mrjeffy321]

Hello,
the other day I got an hour meter to help me keep track of the amp hours I have put into my circuit, but I am not too sure how to hook it up.
it came with a page of instructions giving a few diagrams on how to connect it, the only 2 that made sense for my application of it was to either wire it in series or parallel with my circuit.
to me, series only makes sense, since in parallel, it wouldn't really be measuring the amp hours int my circuit, but rather just it's part of the circuit.
when I wire it in series, it works, I hear it clicking away, but it has so much internam resistance that it barely has any current at all left (about .01 amps), this is unacceptable. when i wire it in parallel, it works, (it clicks) and it counts the amp hours, but as I suspected, it is not correcly measuing what I need it to, since I can take a direct measurement of what the real current is and assume it was constant for 1 hour, an that is not what the meter shows.

is there some other option? am i right to think it should be wired in series with the circuit?

 

[chroot]

It'd be much more helpful if you provided us the make & model of your meter, or a link to its documentation.

- Warren

 

[mrjeffy321]

Ok, I didnt think that would mater too much.
It is a Model "TMI61D96 Engine Powered Hour Meter",


on the oppsite side of the page were 5 diagrams showing possible ways to connect it up, but I can't find a link to those.

 

[faust9]

That device is just a run time clock. It's not an amp-hour meter. It's designed to be connected in series with a source(across the kill switch assuming the kill switch is open while running). That particular device does not change its time base for different input conditions(operation range is 6 to 50V with a xstl oscillator). It simply clicks away the minutes that power is supplied to it(engine running).

 

[mrjeffy321]

Hmm.. well I must say, that is rather dissapointing news to hear.
but I guess it makes sense now that I think about it, when I tested it, I ran it for 1 hour and 10 minutes (I was 10 minutes late, that is why I had such an odd time) and it gave me a reading of about 1.1, so then I think you are right.


Do you happen to know where I can find that I am looking for then, meter to measure amp hours? or any other names to call it by?

 

[mrjeffy321]

So does anyone know where I might get an amp hour meter?
I haven't been able to find much online, just a couple ones that are too expensive with a couple more features then what I need.

Or if not an amp hour meter (which is prefered), I could substitute something else if it will get the job done, for example, if I assumed a constant voltage, then I could use a watt hour meter, since it will measure the total [kilo] watts hours,
watt hour = volts * amps * hours
so then divide by the constant # of volts, and tada!, but I can't do that since the voltage isn't constant enogh for that to work acuratly for me.

 

[chroot]

We used to use amp-hour meters from a company called "eMeter" on our hybrid electric vehicle. They are likely to be too expensive, but I don't honestly know of any other ready-made devices.

Your best bet might be to use an ADC to measure the voltage across a shunt resistor, and then use a small microcontroller running an infinite loop to perform the sum to get the running amp-hours. This will naturally require some programming, etc.

- Warren

 

[Averagesupernova]

I was contemplating on how you could easily make one after the first several posts in this thread. You could use a voltage controlled oscillator to run a counter. The voltage controlling the oscillator would be based on current obviously.

 

[mrjeffy321]

I am not opposed to the idea of making something to do this at all, I am up to the challenge.

Your best bet might be to use an ADC to measure the voltage across a shunt resistor, and then use a small microcontroller running an infinite loop to perform the sum to get the running amp-hours. This will naturally require some programming, etc.

This sounds complicated, but I suppose it would work. I would have no idea where to begin or how to make this although, so I would need a lot of guidence.

The voltage controlled oscillator sounds simplier, but again, I wouldn't know where to begin.


I am going to need a lot more help on how to make eithe og these things.
I wonder how much they will effect the circuit they will be in, how much will they drop the voltage, current, ...?

 

[chroot]

The VCO solution given by Averagesupernova is a good one -- the only drawback I can think of is that most VCOs can only generate a pretty narrow range of frequencies. It's definitely worth pursuing, though.

If you use a shunt resistor and measure the voltage across it, it will barely affect your circuit at all. The shunt resistor is basically a small-value resistor (only a few thousandth of an ohm, usually) placed in series with your device. The voltage across the shunt can be then be used to calculate the current through the device.

- Warren

 

[Averagesupernova]

I suspected the frequency range issue would come up. Here is my solution:

Use a triangle wave type generator using an op-amp. These can be made to have a fairly wide frequency range. They are based on using a current source and capacitor. Be sure to use a very low leakage type cap. Polyethylene I suspect is a good choice for this type of cap. Not sure, haven't worked with them in a while. It is important to keep leakage currents in general down with this type of thing. Zero current through the ammeter should yield NO pulses out of the oscillator. It is also imortant to know what the current range will be. This makes design easier when considering shutting the oscillator off when no current is passing.

 

[mrjeffy321]

not to make myself sound any more stupid, but could you please refrain from using many electronic abbbrevations, they only lead me to more confusion when I don't know what you are talking about, some I can decipher, other I cant.

As for the range of current in the circuit, I suspect it will be somewhere between (roughly) 1 amp and 5 amps, but mostly around 1 to 3 amps.

As for the actual use of this, "amp hour meter", how exactly will it be counting, if we go with the oscillator idea, then I assume we will cound the oscillations and from there determine the current. then won't there have to be some more stuff to keep the running total going?

Is there something I can be reading up on this, or parts to go get, I feel like I am contributing nothing right now at the design phase.

-Thanks for your help

 

[faust9]

you can use a small microprocessor. Don't let the term microprocessor scare you, they are actually pretty easy to make and use. AVR's (made by atmel) have a floating point so C can easily (and freely) be used or you could learn the 50 or so opcodes and program in asm. PIC's are another choice but lack a floating point thus C compilers are mostly proprietary. Swing bt AVRfreaks.com or subscribe to the piclist (Google either term for the exact sites).

I'd say, an 8 pin tinyAVR would do the trick nicely. You can use one pin as a frequency counter, another for simple resey(or simply drive the P/S to the chip low) and the remaining 4 pins for a simple LCD output. Less that 10 components to make what you need I'd say.

How accurate do you want your counter? If you don't mind a litte deviation go with a VCO(voltage controlled oscillator) chip. A 74HC4046, two resistors and two capacitor(I think that's all you need one cap sets the freq range and the other cap is used to clean the power supply voltage) and one very small resistor (0.1ohm) in series with your load. Feed the VCO output to your Micro/C add a cap to clean the P/S of the micro/C(put these caps as close to the chip's power input pin as possible). You can make a serial output using a max232 and 4 caps or just feed 4 line to a LCD(go to Earth LCD for cheap 2x24 or 2x40 displays--run about $7). You can run some LCDs in serial mode or add a serial in parallel out shift register to free up two of your micro/C i/o pins. Usa a 7805 and two caps (one on the input side, and the other on the output side) as your power supply.

Well, according to my write up your actually looking at 20 components but hey. Oh, you'd need an OpAmp between the 0.1 resistor and the VCO to boost the voltage to an acceptible range. Or you could go the cap to an OpAmp as mentioned above.

Instead of the VCO you could use an analog to digital(ADC) input on on AVR or PIC. Use the same 0.1 ohm resistor and OpAmp so your voltage as measured across the resistor varies from 0 to 5V (look up inverting opamp and you'll find a million web sites to help you design the correct amp). Feed this voltage to the ADC and there you go. I'd personally go this route because it would be real easy to do IMHO.

Good luck.

 

[mrjeffy321]

I don't think I have to hardware needed to program a microprocessor (ie. write it in C on the computer and send it over to the chip), although that option sounds attractive, I don't think i'll be able to do that. I assume that is what you meant.

As for acuracy, I would like it to be fairly acurate (as acurate as I can get it I suppose), precise to .1 or .01 ah if possible. as that what I can expect from any of these designs?

Simplicity in the design is good, so let's try to do it with as few parts as possible without comprimising its quality.
I will do whatever method you all think is best, but I do like the idea of an LCD display, and the reset option.

the voltage through the circuit will be around 3 to 5 volts normally, if this is of any concern.

if we go with The voltage controlling the oscillator idea [or any other idea], what kind of parts should I be looking for.

 

[NoTime]

If your circuit is a constant DC load then you could use the hour meter as is.
A little math will convert run time into amp hours.

For a variable DC load, many of the small uP (microprocessors) have built in ADC (Analog to Digital Conversion) so you could test current in your sense resistor directly (or prescaled with an opamp)

You could get a uP board that has built in comm ports for programming functions and talking to a PC.

 

[mrjeffy321]

the load in the circuit is not constant, it will vary, thus changing the current and voltage.

OK, then where can I get the type of microprocessor I need, and what type do I need to get?

At this point I don't care too much about what method I use, just as long as there will be someone to "hold my hand" all the way through building it, because I have no idea where to go from here.
If I had some instructions or something to work on that would be of great use so I can get to ordering the chip (or whatever is needed), unless
Radio Shack will have what I need.

I looked at that AVRfreaks.com site, but I don't see the but just a few 8 pin chips, no prices, and no place to sign up.

 

[faust9]

Here, I wrote a long response a few hours ago but Camino crashed when I wnet to Earth LCD.

Here's the micro/C I'd recommend:
http://www.atmel.com/dyn/products/product_card.asp?part_id=3175

Has 4 channel ADC built in. This little 8pin chip will do all that you need.

Here's the documents:
http://www.atmel.com/dyn/resources/prod_documents/doc2535.pdf

Here's the programmer(It's cheap and easy to use):
http://www.atmel.com/dyn/products/tools_card.asp?tool_id=2726

Which can be bought from:
http://www.digikey.com/scripts/dksearch/dksus.dll?KeywordSearch?Mpart=ATAVRISP&site=us

Sources for electronic parts(google these):
Jameco(good for hobbiest)
Mouser
Digi-key
Newark
Allied

Here's the IDE for programming your chip:
http://www.atmel.com/dyn/products/tools_card.asp?tool_id=2725

Here's the GCC-avr (for windows) C compiler:
http://www.avrfreaks.net/index.php?module=FreaksTools&func=viewItem&item_id=145

And some tutorials(registration required):
http://www.avrfreaks.net/index.php?module=FreaksAcademy

Source for LCDs:
http://store.earthlcd.com/s.nl/c.318770/sc.5/category.4/.f

Good site about serial/parallel/USB ports:
http://www.beyondlogic.org/

OpAmp stuff:
http://www.uoguelph.ca/~antoon/gadgets/741/741.html

you'll want a differential amplifier across series resistor

How ADC works:
http://hyperphysics.phy-astr.gsu.edu/hbase/electronic/adc.html

Standard LCD stuff(search for HD44780 for more information):
http://bray.velenje.cx/avr/lcd/lcd.html

AVR freaks registration:
http://www.avrfreaks.net/user.php?op=register&module=NS-NewUser

Don't let the name scare you. AVR's are easy to work with. Use the built-in ADC(analog to digital converter) and a little math to calculate the current through the resistor.

 

[mrjeffy321]

Wow, lots of information, just what I was looking for.

So here goes, I will need:

• MicroController/Processor
  -As you recomended, I'll go with Atmel ATtiny13, which I found here:
  ATtiny13
  but if you do a quick 'Keyword' search on that site, you'll see they have about 12 different varieties of that chip, which is correct, does it matter?
• Chip Programmer
  -Not really needed for the circiut, but I'll need it to program the chip before I can use it [funny how this is by far the most expensive piece]. As you said, I can get it here: Chip Programmer.
• LCD Display
  -I'll go with this 2x16 LCD screen (its the cheapest I saw), http://store.earthlcd.com/s.nl/sc.5/category.4/it.A/id.1297/.f
• OpAmp
  -I would have no idea what kind I should be looking for, how about this, http://www.alliedelec.com/cart/ProductDetail.asp?SKU=248-0027&SEARCH=OpAmp&ID=&DESC=LM324N
• .1 [ohm?] Resistor
  -Will just any .1 ohm resistor work, or does it have to be something special?
  For example, would this one be acceptable, or is there a better choice, http://www.alliedelec.com/cart/ProductDetail.asp?SKU=832-0349&SEARCH=%2E1+ohm+resistor&ID=&DESC=WW2+0%2E1+5%25
• Differential Amplifier
  -http://www.alliedelec.com/cart/ProductDetail.asp?SKU=568-4645&SEARCH=Differential+Amplifier&ID=&DESC=LM2902N

and the analog to digital converter is built into the chip.

Did I leave anything out, or include something I Dont need?

[This post took about 3 hours (including distractions) to write because of all the info that needed to be sorted through]

 

[faust9]

ATTINY13-20PI-ND

The very top MCU is a dip and runs up to 20MHz(though you'll want to run the internal 4MHz clock instead to free up I/O lines) and comes in a DIP(Dual Inline Package). There is another DIP8 (the number tells you how many pins the chip has) but its max clock speed is 10MHz for the same price as the 20.

You probably don't want to fiddle with surface mounting just yet so the dip is the best choice.

P=i^2R --use this eqn to determine what resistor to get.

P=(5A)^2(0.1ohm)=2.5W

You'll need a ALSR5F-.10-ND which is a 5W 0.10ohm wirewound silicone coated power resistor.

The OpAmp can be a simple 741 series DIP8. Use this as a reference:
http://hyperphysics.phy-astr.gsu.edu/hbase/electronic/opampvar7.html#c1

R1=100K 5% (or a 68K in series with a 50k trimmer--this is a multi-turn potentiometer)
R2=100K 5%(or a 68K in series with a 50k trimmer)
R3=1M 5%
R4=1M 5%

5% resistors will result in a max error of about .4v out of your amp(resistors are usually pretty accurate though). Make the amp, determine the error and program a correction factor into you AVR, or add a couple of pot's to your amp circuit enabeling you to adjust your amp for the correct output.

You could also add a 10K trimmer to your op amp between pins 5 and 1 which allows you to balance your opamp with zero volts applied to both inputs. There's a little leakage inside the amp resulting in up to 7mV(I believe is the standard) of output with a zero input.

Power supply--your going to need an external P/S (wall wart) and a 7805 voltage regulator http://www.google.com/url?sa=U&start=3&q=http://www.fairchildsemi.com/ds/KA/KA7805.pdf&e=747

Get som tantilum 0.1uF capacitors to clean the power supply at the ICs(as close to the input power pins as physically possible--remember, tantilums a polarized and will blow up if connected backwards).

Other than that, read the AVR docs, read about 741's(these have been around since 1968 there are 1000000's of pieces of information available about this great little chip) and you should be good to go.

Good luck.

 

[mrjeffy321]

OK, I found th MCU and .1 Ohm resistor that you were talking about.

I think I'll skip the trimmers for now (maybe I'l add those latter, but not now).

As for the "normal" resistors (R1, R2, R3, R4), I can just go down to RadioShack and pick those up (I know where to get it on my own ).
I also found the voltage regulator at radioShack aswell,
Voltage Regulator .

I found these Capacitors, are either of these OK?
Cap1
Cap2

Is this OpAmp OK?
-OpAmp

I don't really know what you mean by the power supply, do I need a separate one other than the main one I am using for my circuit?
Do I still need a Differential Amplifier?

Lets see did I miss anything?

 

[Averagesupernova]

Draw us a schematic please. I have never seen a time when someone is doing something similar to what you are and ends up with something wrong. Draw a schematic. Verbal is not good enough.

 

[mrjeffy321]

Here is a start, I am not comfortable enough to draw the connections between peices yet, but I think I have it in gererally the right order.


Also, excuse the terrible drawing, I know very few electronic symbols, so I had to improvise on a couple and draw them how they look, or how I think they look.

 

[faust9]

Here, I'll draw a p/s later. W1,2,3 are use for programming. W4,5 are the input and output pins from the current source to be measured. R5,6,7 are optional(you already said you were going to forgo these, but I put them in if you changed your mind). R1,2 are 68K as drawn or 100K if you leave out R5,6,7. NOTE: I mislabeled these as 1M within the parentheses. They should by 100k. I also did not show the decoupling capacitors because they clutter things up. I'll show thos in the P/S diagram.

Enjoy.

[edit]I'll also draw a debounce circuit for the reset in a separate file to unclutter things.

 

[mrjeffy321]

Wow, that makes mine look comically simple and totally wrong.
Hopefully, once I get all this stuff together it will all start making more sense to me then.

By the way, do you approve of the parts I mentioned [4 posts] above (Voltage Regulator, Capasitor, OpAmp)?

 

[faust9]

If your going to order parts online you might as well get them all from the same place(saves on shipping and radioshack charges out the yin-yang compared to most mail order distributors). If you do decide to get your parts from RS then you can find single and quad op amps there as well along with the resistors (except the .1ohm). Thos caps won't work too well for what they are intended for. You should always buffer the P/S of an IC with a tantalum cap. Tantalum caps are very stable and have excellent frequency response(they do a better job of maintaining DC on the power supply pin than aluminum caps). I don't know if the second cap in your post is a tant or not. The first is a ceramic and not really what your looking for. The monolithic can be used to decouple but you can get a bunch of tantalum decouplers from mouser or digikey for the same price as one from RS.

The support resistors for the OpAmp need to be 1/4W BTW.

A differential amp is simply a packaged op amp anyway. Differentials need external components so I've always just used suitible op amps instead of paying the extra money for a packaged unit--so you don't need a separate differential amp.

Your circuit voltage needs to stay as close to 5V(7805's do a real good job of maintaining voltage) as possible. Drop below 5V and your op amp will not work correctly(unless you find a low voltage op amp). The min voltage for the MCU to operate is 1.8V though if you want to use a low volt amp.

 

[mrjeffy321]

It seems that most of the peices I will order from Digikey.
I only suggested RadioShack, not because I inteneded to mail order those parts, but because I could just go down to the store and pick it up in person (no [direct] shipping charge that way).

When you say the "P/S of the IC", are you referring to the voltage regulator that is supplying the power to the ATtiny13 chip?

OK, here is the type of capasistor you recomend:
capasitor

and is that OpAmp shown above good?

 

[faust9]

Yes, when I say P/S of the IC I mean Vcc (The power entering the chip).

That is indeed the correct Cap.

And this is the op amp:
http://www.digikey.com/scripts/dksearch/dksus.dll?Detail?Ref=300134&Row=320202&Site=US

LM741.

 

[mrjeffy321]

OK, that was an ordeal just to get the parts picked out,
now that that is done, I'll order them tomarrow.

FINAL LIST
-MicroController-Atmel ATtiny13-20PI-ND (Digi-Key)
-Chip Programmer-ATAVRISP-ND (Digi-Key)
-OpAmp-LM741CNNS-ND (Digi-Key)
-.1 uFD Tantalum Capacitor-P2067-ND (Digi-Key)
-.1 Ohm Resistor 5W 1%-ALSRF-.10-ND (Digi-Key)
-LCD Screen-LCM51203LM3 (EarthLCD)
-2, 100K Ohm Resistor 1/4 W 5% - 271-1347 (RadioShack)
-2, 1M Ohm Resistor 1/4 W 5% - 271-1356 (RadioShack)
-Voltage Regulator - 276-1170 (RadioShack)

Right now I am reading a buch of those data sheets trying to grasp as much as I can and reading the tutorial on how to program the chip.

 

[faust9]

The Attiny12 doesn't have an ADC, only an analog comparitor. An ADC actually converts and analog signal to a binary number whereas an comparator looks ot on pin and then at the other. If pin A is higher than pin B the comparator writes a 1 to a register. If pin B is higher then a 0 is written. In your case, the resistor will always have a voltage drop if current is flowing so a comparator will not give useful information. You'll need a 13(I think these are in full production now) or a 15.

 

[mrjeffy321]

That was a typo, I mean a ATtiny13, not the 12.
Thanks for catching that though, it could have been a real mistake.

 

[mrjeffy321]

faust9, might I trouble you for the 2nd installment of the circuit diagram?
Also, upson further inspection of the first circuit fiagram I see a couple parts (resistors) that I didnt originally get, a 4.7k orm resistor (R8), and I am not too sure what the W's stand for or the 'PB's, or the part labeled 'SW' that has a symbol like an 'I'

 

[faust9]

faust9, might I trouble you for the 2nd installment of the circuit diagram?
Also, upson further inspection of the first circuit fiagram I see a couple parts (resistors) that I didnt originally get, a 4.7k orm resistor (R8), and I am not too sure what the W's stand for or the 'PB's, or the part labeled 'SW' that has a symbol like an 'I'

I'm sorry. I'll upload those tomorrow morning(I don't have the figs on my computer). the the 4.7k can be replaced by anything greater than 1k. I used 4.7K because it is a common value found in digital electronics. The SW is a PBNO(Push button normally open) switch. The W's are connectors(like bannana plugs--though those are a little big). The W's are for in-circuit programming. You'll have to get some 0.1 pin headers for those. Read the Atmel docs for the chip pinout. PB stands for I/O port.

 

[faust9]

Go to figure 5 of this document:
http://www.ee.washington.edu/stores/DataSheets/voltreg/7805.pdf

That is your power supply. Use a wall-wart. The outer part of the plug is + and goes to pin 1 and the center part of the plug is - and goes to pin 2 usually. Most wall-warts us a center negative scheme but not all. You can use any PS from 6 to 12 v and I'd estimate a minimum current rating of 200mA

Here's a VERY good demonstration fo switch bounce:
http://www.ece.utep.edu/courses/web3376/concepts/debounce.html

Here's a simple debounce circuit:
http://www.all-electric.com/schematic/debounce.htm

You don't 'need' the schmitt, avr's are designed to operate in some pretty extreme situations. A simple capacitor between the switch an ground will work fine(use a moderetly low pull -up resistor R8 though 4.7K minimizes current draw from the source and allows the cap to charge moderetly quickly).

If you were using a higher end processor then I'd say use the Hex-Schmitt to debounce--here it's not required.

 

[mrjeffy321]

OK, thanks.

Thats leaved me with some more questions.
whis wall wart you speak of, what is that? is it the MC78XX/LM78XX thing that is in the PS diagram or is that the voltage regulator? (is is that a stupic question because they are the same thing).
In the DC parameters circuit (figure 5), it show a .33 uF capacitor, should that one de tantalum too? how important is it for it to be .33 uF, how about 1 uF, or something close. this would ger inserted into the original diagram right between the button and pin 1 right (with another of the leads going to pin 2)?
Are you saying I need a separate power supple for this "amp hour meter" circuit, other than the power that I would be putting through my normal circuit?, I have to plug it in so to speak, for it to work?

I understand the purpose for the debounce circuit, but I don't see how I need one. you said earlier about that SW piece, the PBNO, is a switch, but I never understood who I need it, I don't see how I need a button for this to work.

 

[faust9]

1uF will work--it can be a cheap ceramic BTW. Also, it's always good to have a reset. And, P/S can mean the source from the wall to the reg or the portion of your ckt which eventually powers up all of your components.

The decoupling needs to placed a physically close to the processor as you can get it(I usually use DIP sockets and solder the cap under the chip on the socket between the power and ground pins. You can actually buy these pre-decoupled sockets--just make sure the cap is connected between the correct pins.