«home

Batteries for the Sparrow

When my Sparrow could no longer make it to work and back at speeds greater than 20mph, I had to figure out how to fix it.  I had read warnings against having an "unbalanced" battery pack. The manual from Corbin recommended replacing all of the batteries as soon as you were unable to drive the distance you need to go. It looked like I was going to have to spend about $2000 on a new set of batteries.

I am basically cheap, and was resistant to buying a whole new pack. In order to keep my Sparrow as cost effective as possible, I wanted to wring every mile I could out of each and every battery. Lacking an understood danger, I set about trying to determine how many batteries I needed to replace.

My "battery experiment" is to replace only the weakest battery or two when the pack voltage drops low enough that I don't make it home, or that I fear that I may not make it home the next day.

Identifying the weak batteries

Finding the weak battery is not easy. At first I tried to measure the voltage of each battery in the pack while it was sitting in my garage. This proved they were all within a few tenths of a volt of each other.

Next, I went to the local BatteriesPlus store where they could test the "Cold Cranking Amps" of each battery. I assume their meter measures the voltage with no load, and then measures it again with a several Amp load. The meter was not very big, so it could not be drawing many amps for very long and be able to dissipate the heat. They did not know exactly what the meter did, but it said each of my batteries could deliver over 1000 CCA which was above the spec. for a Yellow Top.

I decided I needed to drive my car until I had exhausted at least one battery and my voltage was sagging. Once in this state I could check the voltages of each battery. After driving 10 miles, my voltage was down to about 120V. As soon as I stopped, the voltage would rise back to 160V, and when I put my voltmeter on each battery they were all within a few tenths of a volt of each other.

I knew I could only test my batteries under load to figure out which were the weak ones. I messed around with using the emergency brake and chock blocks to try and hold the car still while I pressed the accelerator to draw enough current to lower the pack voltage, but nothing worked. I was going to have to drive the Sparrow to create my load.

I jammed a probe from my meter under the battery connectors of a battery under the seat, replaced the seat, and with my meter in my lap, floored the Sparrow down the street and when the pack voltage dropped below 130V, I read the meter.

I went around the block 7 times that day, once for each battery under the seat. After each trip I would move the meter leads to a different battery. Sometimes the leads would fall off and I would stop part way around the block and put them back on.

Using this technique I found one of my batteries was dropping to 2 Volts, and another was dropping to 8. The rest were above 11. I replaced these two batteries and was able to drive to work and back again. I was lucky the weak batteries were under the seat where my test leads would reach and I could identify them.

A more practical test

A few months later I couldn't make it work and back without the voltage dipping below 150 V. Then a few days later I could not maintain the speed limit. It was time to test the batteries again.

I was ready this time. I had bought a new voltmeter that would lock in on the high or low voltage it saw while it was hooked up. I also bought alligator clips that could be snapped onto the test leads. I could clip the test leads to the bolts on the battery connectors, set the meter to watch for the low voltage, drive the Sparrow for a block or so. While driving I would keep one eye on the road, and one eye on the ammeter on the dash. I decided 100 Amps would be the load I would use for all tests.

After holding the current at 100 Amps for a second or two, I could stop the car and safely check the low voltage captured by my new meter. If I was checking the voltage of a battery under the seat the meter would be on the floor of the cab. If I was checking a battery under the hood, the meter would be tucked in under the closed hood while I drove.

This plan was fine except for two problems. First, it takes about 2 hours to check all of the batteries this way. Second, there is one battery (#9 in the diagram below) where you cannot reach both terminals. Math can be used solve these problems.

The real battery test

There is no rule that a meter can only measure the voltage of one battery at a time. If I could put my meter across 6 of the batteries, I would know what their total voltage was. I could use this to figure the average of those 6. Comparing this to the pack average, I would know if the weak one was in that set of six or not. There is an assumption here that there is only one weak battery.

In one measurement I would eliminate half of the batteries from my search. If the weak battery were in the six, I would repeat the experiment on 3 of those 6, and eliminate 3 more from the search. Then I would test 1 of the remaining three and eliminate one, or two batteries from the search.

Due to the layout of the batteries you can't easily test 6 at once. I would test the 5 I could under the hood (#4 through #8). If the weak one was there, I would continue the binary search under the hood. If they were strong, I would test the string of batteries #10 through #13 under the seat. If those were strong, the problem was in batteries #1, #2 or #3. This is sort of a "binary search".

In 4 test trips, I could identify the weak battery. Often, this was the only testing I would do before replacing a battery. Sometimes the battery I identified was not weak enough to be responsible for the voltage drop. Based on the readings, I would sometimes look for a second weak battery. Only once did I make 12 measurements.

Hardware Solution

In October of 2005 I built a circuit that eliminated the testing.

Sparrow battery numbering

The negative terminal of #1 is at ground
#4 through #9 are under the hood
#4 is actually stacked on #8 and #9

Battery Replacement Log

	Installed		Retired		Life		miles per battery pack		Price	Notes
Battery	date	mileage	date	mileage	months	miles	optimist	pessimist
1	10/1/2001	300	5/3/2003	3700	19.0	3350	22,100	3,350	-	Factory replacement of pack. Marked "Blemish" and "Recond"
2	10/1/2001	300	5/3/2003	3700	19.0	3450	22,100	3,450	-	Factory replacement of pack. Marked "Blemish" and "Recond"
3	10/1/2001	300	9/27/2003	5000	23.9	4650	15,275	4,458	-	Factory replacement of pack. Marked "Blemish" and "Recond"
4	10/1/2001	300	9/27/2003	5000	23.9	4750	15,275	4,542	-	Factory replacement of pack. Marked "Blemish" and "Recond"
5	10/1/2001	300	10/12/2003	5200	24.3	4900	12,740	4,638	-	Factory replacement of pack. Marked "Blemish" and "Recond"
6	10/1/2001	300	9/4/2004	7200	35.1	6900	14,950	5,869	-	Factory replacement of pack. Marked "Blemish" and "Recond"
7	9/27/2003	5000	11/7/2004	7770	13.4	2770	13,873	5,933	135	BatteriesAreUs.com.  In garage for 14 months
8	10/1/2001	300	7/16/2005	9482	45.5	9182	14,921	6,789	-	Factory replacement of pack. Marked "Blemish" and "Recond" Died suddenly. I got towed home.
9	10/1/2001	300	8/15/2005	9738	46.5	9388	12,269	6,877	-	Factory replacement of pack. Marked "Blemish" and "Recond"
10	10/1/2001	300	8/15/2005	9738	46.5	9488	12,269	6,899	-	Factory replacement of pack. Marked "Blemish" and "Recond"
11	11/7/2004	7770	10/9/2005	10130	11.0	2360	11,617	6,701	135	BatteriesAreUs.com.
12	10/1/2001	300	11/20/2005	10541	49.6	10241	11,094	6,810	-	Factory replacement of pack. Marked "Blemish" and "Recond"
13	10/1/2001	300	12/28/2005	10800	50.9	10500	10,500	6,737	-	Factory replacement of pack. Marked "Blemish" and "Recond"
14	10/1/2001	300	1/12/2006	10840	51.4	10540	9,787	6,899	-	Factory replacement of pack. Marked "Blemish" and "Recond"  Died suddenly. After few hours of cool down I could drive home.
15	10/1/2001	300	3/18/2006	11231	53.5	10931	8,881	5,408	-	Factory replacement of pack. Marked "Blemish" and "Recond"
19	10/12/2003	5200	3/18/2006	11231	29.2	6031	8,881	5,408	135	BatteriesAreUs.com.
25	11/20/2005	10541	8/12/2006	12341	8.7	1800	9,208	6,793	139	BatteriesAreUs.com.  Had been returned
26	12/28/2005	10800	8/26/2006	12362	7.9	1562	8,253	6,010	137	Interstate Batteries
24	10/9/2005	10130	8/26/2006	12362	10.5	2232	8,253	6,010	155	BatteriesPlus
27	1/12/2006	10840	9/9/2006	12555	7.9	1715	7,586	6,034	139	BatteriesAreUs.com.
22	8/15/2005	9738	9/9/2006	12555	12.8	2817	7,586	6,034	155	BatteriesPlus
28	3/18/2006	11231	10/24/2006	12840	7.2	1609	7,410	5,891	147	Interstate Batteries (or maybe AutoZone)
18	9/27/2003	5000	6/22/2007	13986	44.8	8986	7,736	6,058	135	BatteriesAreUs.com.
30	8/12/2006	12341	8/25/2007	14516	12.4	2175	7,108	5,792	142	Interstate Batteries
31	8/29/2006	12362	8/25/2007	14516	11.9	2154	7,108	5,792	148	Interstate Batteries
16	9/27/2003	5000	8/25/2007	14516	46.9	9516	7,108	5,792	130	BatteriesAreUs.com.
33	9/9/2006	12555	9/9/2007	14639	12.0	2084	6,904	5,673	148	Interstate Batteries
32	8/26/2006	12362	9/15/2007	14712	12.6	2350	6,691	5,567	148	Interstate Batteries
34	9/9/2006	12555	9/19/2007	14747	12.3	2192	6,612	5,482	148	Interstate Batteries
17	5/3/2003	3700	11/24/2007	15255	54.7	11605	6,272	5,502	130	BatteriesAreUs.com.
20	9/4/2004	7200	11/24/2007	15255	38.6	8055	6,272	5,502	135	BatteriesAreUs.com.  In garage for 14 months
21	7/16/2005	9482			28.3	5773			155	BatteriesPlus
23	8/15/2005	9738			27.3	5517			155	BatteriesPlus
29	3/18/2006	11231			20.2	4024			170	AutoZone
35	10/24/2006	12840			13.0	2415			159	Interstate Batteries
36	6/23/2007	13986			5.1	1269			159	Interstate Batteries
37	8/25/2007	14516			3.0	739			159	Interstate Batteries
38	8/25/2007	14516			3.0	739			159	Interstate Batteries
39	8/25/2007	14516			3.0	739			159	Interstate Batteries
40	9/9/2007	14639			2.5	616			159	Interstate Batteries
41	9/15/2007	14712			2.3	543			159	Interstate Batteries
42	9/19/2007	14747			2.2	508			159	Interstate Batteries
43	11/24/2007	15255			0.0	0			159	Interstate Batteries
44	11/24/2007	15255			0.0	0			159	Interstate Batteries

The disturbing trend is that nearly all of the batteries I bought from August of 2005 to March of 2006 have failed in about a year, while older batteries are still in use.  It could be that Optima batteries do not last as long as they used to.

Battery life

At this point I have gone 15255 miles and have replaced 31 batteries. An optimist might say the 13 batteries in the car are good as new and so I have gotten 6,272 miles per pack. A pessimist might say the 13 batteries in the car are trashed and so my mileage is 5,502 miles per pack.

This graph shows how the optimistic and pessimistic estimates of my battery life have changed over time.

Conclusion

Corbin was probably right in recommending that all of the batteries be replaced at once. All this testing is not practical if you are running a business. You can spend an hour or two testing each time a battery fails. Replacing one battery takes about an hour if it is under the hood, or a half an hour if it is under the seat. If I were to take my sparrow to a garage, the labor would be much higher if I replace the batteries one at a time than if I replace all of them at once. It might be about the same price to replace the whole pack every 3500 miles as replacing the batteries one at a time but get more miles per battery.

Many impractical activities are called hobbies. Testing and replacing batteries has become mine.

I believe that replacing batteries one at a time is the cheapest way to run my Sparrow. Having many old batteries in the pack means my range is always just enough to get me to work and back, and not much more.

I feel that batteries are all about equal until just before they fail. Once one battery has more internal resistance than the others it will heat up more and fail faster. Once a battery has less range than the others, its percent of discharge will be greater than the others and it will fail faster.