DIY Electric Car Forums banner

1 - 11 of 11 Posts

·
Registered
Joined
·
91 Posts
Discussion Starter #1
Hi guys, I have my car running now (finally) and currently just have a multi meter for instrumentation. I'm running 96V which fully charged is around 104.5V. How do I convert a voltage reading to know what it means in terms of 3/4 charge, 1/2 charge etc?
 

·
Registered
Joined
·
571 Posts
Short answer: You don't. You can't.
A bit longer answer: With some battery chemistries such as lead acid, you can make a very crude approximation, but even then you need to stop the load and wait a little. Typical guidelines for a 12V lead acid battery when idle (no load): 12.8V = full, 12.0V = empty. For lithium, forget it.
 

·
Registered
Joined
·
91 Posts
Discussion Starter #3
Short answer: You don't. You can't.
A bit longer answer: With some battery chemistries such as lead acid, you can make a very crude approximation, but even then you need to stop the load and wait a little. Typical guidelines for a 12V lead acid battery when idle (no load): 12.8V = full, 12.0V = empty. For lithium, forget it.
Thanks for the reply. So I can roughly estimate that my system is 50% DOD at around 99.2V and fully discharged at 96V?
 

·
Registered
Joined
·
33 Posts
You can roughly estimate the State of Charge, but especially for the commonly used LiFePo4 type batteries its very very inaccurate.

If you use the LiCoO2 Batteries, your estimation will be more accurate, but still voltage based SOC calculation is clearly not the best choice.

Here the typical voltage curves of both battery types:


The common method of SOC calculation is based on current counting, so you measure and sum up the current that flows in and out of the battery.

Anyway, if you have to stick to the voltage method here some hints:
+ Take the voltage for SOC calculation only if the battery has settled, means not during or directly after usage.
+ Temperature and aging may have an influence on the battery voltage level.
+ Observe the individual cell voltages. Cut off the system if the lowest cell reaches a certain limit (e.g. 2.5V). For charging cut off if the highest cell reaches a certain limit (e.g. 3.5V)

Hope this helps,
Peter
 

·
Registered
Joined
·
4,007 Posts
I'd pay more attention to the sag voltage than the resting voltage.

Every lead batt is different, so you'll have to get to know yours, but for mine they sag to ~92% of nominal (110/120) under acceleration when full, 75% ~70% DOD, and ~60% ~80% DOD (you don't want to go that far).

Add a little math and you should be able to guesstimate where the 50% mark where you should be stopping is.
 

·
Registered
Joined
·
2,045 Posts
Hi guys, I have my car running now (finally) and currently just have a multi meter for instrumentation. I'm running 96V which fully charged is around 104.5V. How do I convert a voltage reading to know what it means in terms of 3/4 charge, 1/2 charge etc?
I decided when I got into my build that I didn't need any instruments that were not already in the car. I have a speedometer (so I dont spend a night in jail), a tachometer (because it was there), and the 12V original panel meter (so I can tell if the DC-DC has failed and I am running on the buffer battery) hooked up. For my SOC I just hit the trip odometer clear button when I turn the key after charging. The advantages of this are nobody who can drive a manual seems to be confused or alarmed by all the instrumentation. With LiFePO4 there doesn't seem to be a lot of reason for a voltmeter. It will tell you far to late to do much about it that you are out of energy. There doesn't seem to be a lot of reason to have an ammeter either. Amps is what you are commanding when you press the throttle. I am eventually going to turn the wattmeter technology I did in the early 90's for the electric RC aircraft fliers into a SOC meter to drive the fuel gauge but for now the odometer will tell me how far I have gone on this charge and that seems to have been good enough for the first 7 months.

As a scientist and engineer I can tell you I wanted all that stuff when I first started my build but I came to realize that none of it helps the mission of driving a car without distraction.

The JLD404 is the bargain instrument if you want to add something to tell you volts, amps, and AH for SOC all in one. It can even act as an emergency cut off for your charger on an over volt condition.

Best wishes!
 

·
Registered
Joined
·
33 Posts
Hi guys, I have my car running now (finally) and currently just have a multi meter for instrumentation. I'm running 96V which fully charged is around 104.5V. How do I convert a voltage reading to know what it means in terms of 3/4 charge, 1/2 charge etc?
All of the other posters outlined why you cannot accurately do what you are talking about. My post is interpreting your question or what's behind it. You wish to cheaply and accurately determine what you have left without spending a lot of money. Assuming you own a laptop this rather cheap voltmeter has usb conversion to feed a laptop voltage information along with other useful data:

http://www.ebay.com/itm/Digital-Multimeter-Voltmeter-Thermometer-Ohm-USB-C-Continuity-Frequency-Diode-/330935230606?pt=Motors_Automotive_Tools&hash=item4d0d46e08e&vxp=mtr

Caveat: I don't own one of these I just did a quick search to find a computer interfacing voltmeter.

You would need a very simple program to sample and store the information over time to calibrate the results into something useful.

Since measuring current is actually done by measuring the voltage drop across a "shunt" or known resistance (normally) you would need a shunt or I believe you should be able to measure voltage from one end of your longest primary wire (either plus or minus, makes no difference, just pick the longest cable so as to get the highest signal to noise ratio, i.e. voltage indication) to the other end of that wire and feed that to the laptop as the current indication. Sample it every second or so, store the amount and seconds information and there is your accurate fuel gauge.

In case this is confusing, here is the barnyard explanation of what the simple program would do: You connect your Harbor Freight $5.99 voltmeter across say the wire returning minus from you trunk up to the controller "minus" in the front. You extend the voltmeter black lead to the rear with a small wire and to the minus battery terminal. The red lead gets extended to the controller minus exactly where the cable you are "shunting" connects to the controller. Looks like you are doing something redundant doesn't it? Shorting across the cable or something? Just do it. The voltage is your "uncalibrated" current indication. THE ACTUAL VALUE IS NOT RELATED TO PACK VOLTAGE and it does not tell you your starting, ending or any other "voltage" in the system. This voltmeter is being used as an ammeter period. More current will make your "relative" voltage go higher. Now that is the "current " part of the equation. Let's build a program that keeps track of how much current you are using over time. Wait were in a barnyard and we want to use wood and bailing wire. OK so you get one of your friends to ride along with you and they have a watch (with a second hand and it is wound up and running), a pen, an old manual (lever you pull on the side to sum) adding machine since we don't want to steal any power/range from the system) and a clip board with graph paper.
First you have to calibrate the "system" so your "logger person" puts a zero in the box on the bottom left hand side of the paper. You start moving your EV using five percent of the throttle and while you hold that you both see .05 volts on the voltmeter. Your friend writes ".05" in the next box up from zero on the left of the paper. You go to ten percent, he writes ".1" in the 3 rd box above the others. This calibrating is done until at full throttle you see "1 volt" and that goes in the twentieth box above all the others. You have now calibrated your HF 6$ voltmeter to read in "relative" amps. The actual amount is not important for this purpose.

You make a commitment to drive at a steady state in one minute increments (Like as if you are the only person on the road and traffic signs and lights do not exists, or basically as if you are from Boston :))
Every minute the logger places a check in the box corresponding to that voltage he sees at that time on the meter and is above the "minute" line at the bottom of the graph moving to the right on the graph. Since he has a whole minute to work he adds up the boxes below his "entry" and puts that total into the adding machine. For the first two minutes you drive at 10% throttle so each minute equal two boxes under the checked box or 2+2=4 and 4 is the total on the adding machine for the first two minutes. For the next ten minutes you drive at half throttle and each minute is at ".5 volts" (or at the end of twelve boxes across, 2 boxes two "up" and ten boxes .5 "up") the adding machine shows 54 boxes below the line so to speak total. And so on until you run out of graph paper or battery. When your pack has reached discharge voltage(measured on another Harbor Freight 6$ voltmeter) (the one time the voltmeter is accurate as to what you have left which is zero) You can now look at the adding machine and see the total number of boxes which represents the total "charge" you have available to use in the future.
After you have the total number of boxes or charge you can just drive (after recharging fully) as you wish as long as you hold the throttle steady in one minute increments. Your friend will keep a tally of the boxes you have used and let you know when your total used is getting close to the original total you had available.
Now in the real world that kind of friend doesn't exist and one minute increments does not fit legal or defensive driving. We have to increase the sampling and totaling thing and leave the friend at home. So what to do?
A computer fed this "voltage across a shunt" can do all of these functions and that is what is done.

I should add that a cheap meter like the ebay one could also measure RPM and many other things mimicking a lot of the expensive EV monitoring equipment.
 

·
Registered
Joined
·
3,141 Posts
A major problem with using a battery cable as a shunt is the temperature coefficient of the resistance of copper, which is about 0.4%/C. The cable can easily span a temperature range of 0C to 50C so if it is calibrated at 25C the reading can be in error by 10%. This may be OK for a rough indication, but you can get a decent 200A or 500A shunt for about $20.

The digital multimeter with USB interface may meet your needs, but even better would be a two-channel USB digital storage oscilloscope such as this, for less than $30:
http://www.ebay.com/itm/2-Channel-PC-Computer-Digital-Storage-USB-Oscilloscope-/330731251939

But for $66 the following would be much better:
http://www.ebay.com/itm/Hantek-PC-Based-USB-Digital-Storage-Oscilloscope-6022BE-20Mhz-Bandwidth-New-/380654046939

The fuel gauge function can be done by logging just ampere-hours, but it may be better to monitor watt-hours, taking the voltage into account. However, for lead-acid batteries, the effective amp-hour capacity is reduced as a function of current, so a high current drain would need to factor in the Peukert effect, and watt-hour measurement is actually even less accurate than amp-hours, since much of the wattage is lost in the battery and not measured.

This would be a fairly simple and useful project for a PIC or Arduino board. Otherwise, there are dedicated ICs that function as battery fuel gauges and charge monitors, and they are quite inexpensive:
http://focus.ti.com/paramsearch/docs/parametricsearch.tsp?family=analog&familyId=412&uiTemplateId=NODE_STRY_PGE_T

I got a BQ34Z110PWR as well as an AMC1200SDUBR isolation amplifier from TI as free samples. But to use them you need to make a PC board and add various components, so the cost of the IC is negligible. The BQ34Z110 is about $4:
http://www.ti.com/product/bq34z110

This one is designed for a single 12V battery and drives a 10-LED bargraph:
http://www.ti.com/product/bq78412

Probably beyond the scope of this post, however. ;)
 

·
Registered
Joined
·
33 Posts
A major problem with using a battery cable as a shunt is the temperature coefficient of the resistance of copper, which is about 0.4%/C. The cable can easily span a temperature range of 0C to 50C so if it is calibrated at 25C the reading can be in error by 10%. This may be OK for a rough indication, but you can get a decent 200A or 500A shunt for about $20.

The digital multimeter with USB interface may meet your needs, but even better would be a two-channel USB digital storage oscilloscope such as this, for less than $30:
http://www.ebay.com/itm/2-Channel-PC-Computer-Digital-Storage-USB-Oscilloscope-/330731251939

But for $66 the following would be much better:
http://www.ebay.com/itm/Hantek-PC-Based-USB-Digital-Storage-Oscilloscope-6022BE-20Mhz-Bandwidth-New-/380654046939

The fuel gauge function can be done by logging just ampere-hours, but it may be better to monitor watt-hours, taking the voltage into account. However, for lead-acid batteries, the effective amp-hour capacity is reduced as a function of current, so a high current drain would need to factor in the Peukert effect, and watt-hour measurement is actually even less accurate than amp-hours, since much of the wattage is lost in the battery and not measured.

This would be a fairly simple and useful project for a PIC or Arduino board. Otherwise, there are dedicated ICs that function as battery fuel gauges and charge monitors, and they are quite inexpensive:
http://focus.ti.com/paramsearch/docs/parametricsearch.tsp?family=analog&familyId=412&uiTemplateId=NODE_STRY_PGE_T

I got a BQ34Z110PWR as well as an AMC1200SDUBR isolation amplifier from TI as free samples. But to use them you need to make a PC board and add various components, so the cost of the IC is negligible. The BQ34Z110 is about $4:
http://www.ti.com/product/bq34z110

This one is designed for a single 12V battery and drives a 10-LED bargraph:
http://www.ti.com/product/bq78412

Probably beyond the scope of this post, however. ;)

Dang it Jethro you left the barnyard
What was that puekard thang?
OK OK your ideas are better. I was just trying to demonstrate what the process was in a simple way. Probably wasn't to simple.

And actually I had not thought about the temperature change affecting the resistance. One thing too I would think transients could get into the system over that length of wire and cause somewhat erroneous readings.
 
1 - 11 of 11 Posts
Top