MIMA Pack Whack and rebalancing the battery Whack the pack to erase the cell memory and fix recals The MIMA pack whack could be a huge new advantage to owning MIMA for the many people struggling with limited capacity battery packs due to recalibration's. As many of you know, I have been using MIMA on my pack since 80K, and as I approach 140K, the pack still runs fully from top to bottom, on each cycle. Since I do not have recal's, it is something that I was not able to examine. On my trip to Hybridfest, I flew out to MI,to ride out with Ed Zandee,a very savvy auto technician who owns his own service center in Grand Rapids. Ed purchased a used 2000 Insight with 130K on the clock.Like many older Insights, he was getting a typical recal whenever his pack reached 5-6 bars from the bottom. Ed Installed a MIMA system the day before we drove the 350 miles to Madison. I drove some of the way, and got to experience recals first hand for the first time. Ed and I started to experiment and analyze what was happening. We noticed that the recal point seemed to move up the scale one bar as the pack temp increased.We also found that full regen could be applied with MIMA, even though the SOC guage was showing full and limiting the standard IMA regen. I have since confirmed that my pack will stop MIMA charging near the top bar, so it is synchronized with the SOC display. Most of us with MIMA will stop regen when the SOC is at the top, since we fear overcharging. This is where we do the PackWhack. Just let the ABC keep charging until it is limited, even though the SOC is at max. Drain it down till no assist is available, and you have given it a stop to stop full charge. On the way back, we again took turns driving,We decided to push past the SOC top end and using MIMA we continued to freely charge, for quite some time while the SOC was pegged at the top. A good pack will stop excepting charge from MIMA when full,even if the MIMA system request it, so this was a new experience for me. We decided to keep jambing charge into the pack to see if it would eventually start limiting, and eventually it did, but only after putting in the approximate equivalent of charging to about 6 more bars. We then switched on PIMA, and started running a lot of assist. We were drawing full assist for several minutes, before the SOC finally began to drop, and it dropped smoothly to the recal point, and then recalled as normal, but this time, even with no bars of SOC, we were able to draw the full 100 A of assist for several minutes as the photo shows, until the assist finally became limited. We ran the pack through the cycle another time, with the same results. We got back to Ed's place, feeling that the pack was outputting a lot more capacity than before. The real surprise came the next day, when Ed brought me back to Detroit for my flight home. The recals were gone. We cycled the pack several times, and the SOC ran from the top,right to the bottom. Ed will do some more test runs with his Honda factory scan tool and get us more information. Making some assumptions which still need to be tested, my read of this experiment is that the SOC limits that stop normal IMA action at both ends are different that the BCM top and bottom limits which are based on actual voltage measurements on both ends. When a pack starts having limited capacity issues due to the memory effect, the SOC in normal IMA mode will stop regen when the guage is at the top,even though the pack is really not charged fully. The MIMA forced charge is limited instead by the BCM when it determines that the battery voltage has reached the max. So this MIMA PackWhack as Ed and I began to call it, allows the pack to be fully charged past the SOC gauge limited point that the standard IMA is limited by, and is only stopped when fully charged, therefore doing the equivalent of a equalization charge on a lead acid based pack. I suggest that any of the MIMA owners that have recals give this a try. Just stay in an ABC forced charge, or manually do regen until the charge is limited even though the SOC bars are full. Then do assist with PIMA or MIMA until it starts to be limited on the bottom end. Do several of these cycles, then see what happens with your recals the next day. Of course you are doing this at your own risk, as it will take some time to see if this effect is temporary. This blog will be where we will put the results of further testing of this experimental procedure. All recals are not the same, so this procedure is for the standard regular recal at a specific SOC that seems to be the most common. fast pack rebalancing 101 Rebalancing off the grid Ron Hansen drove his space alien green insight up from NY, so we could try to rebalance his battery pack. He was having very frequent negative and then positive recals, and almost always saw a charge. The car was rebuilt from a salvage, so Honda would not do anything for him. The car sat for over 6 months which probably contributed to the problems. The IMA code of death kept coming back even if reset. I gathered up the required components. 220V variac 25A 400V diode bridge 300Uf 350V cap 3A dc ampmeter custom HV connection probe several DVM's We wired up the power supply and confirmed that I could get 0- 250vdc output range. We flipped the pack switch and found that the pack was at 154V. With the pack HV switch off, I attached my custom HV probe to the precharge resistor lead (+) and an alligator clip to the negative terminal. We raised the variac voltage until it was at the same 154V so we would not get a surge when the charger was applied to the pack by flipping the main switch. I made up a fan connector so I could run the battery fan if required from an adjustable DC lab supply. We also connected an ohmmeter to the PTC strip ends so we could watch the strip resistance change as the pack warmed up. I started with a 2A charge, so we could get to the 100% SOC point as quickly as possible. We watched the battery voltage raise, and tweaked the variac voltage to maintain the 2A charge.The PTC strip started our test at 22.9ohms, which represented about 85F. After 5 minutes or so, the PTC strip resistance started a gradual raise, and the pack voltage continued to raise.When the PTC strip got up to the 30 ohm range, I pulled the battery fan and used my non contact IR thermometer to see that the cells were at ~110F.I put the fan back in,turned it on and taped around the edges so it would draw air better. I used a digital thermometer to measure the outlet air temp, and found it at ~ 100F. The fan at 12V was really pulling a lot of air through the pack, so the PTC strip resistance rapidly began dropping towards the starting resistance, and the fan outlet air quickly dropped to ~88F. Conclusion, with the fan running at 12V, there will be almost no measurable temperature rise even when charging at 2 A. As the pack approached the 167V voltage we dropped the charge current back to 500MA, and left it there. The voltage of course dropped back a bit to 165V when the current was dropped, but after another 1/2 hour at 500MA, it pretty much stopped changing and settled at 168.9v. We determined that the best way to determine when you are at the 100% SOC point is when the pack voltage stops raising. A simple rate of charge vs rate of change of terminal voltage determination should be all that is required if an automatic charger/balancer were to be built. When the high rate charge pack voltage rate of change slows down, we switch to the trickle rate, and again wait for the rate of change to actually stop changing, then we soak for say 1/2 hour and then stop the charging. Ron drove back to NY, and found that the pack behaved much differently, and the charging would work much more normally. Unfortunately he will be not be able to drive the car for the next 3 weeks due to a medical issue. We will have to wait 3 weeks to get the full report. If anyone else in the area wants to give this quick rebalancing a try, contact me. (Posted 9/7/2008 by mikey) The HV panels can bend and will shock the S**T out of you Testing the HV panels The three HV panels have cured enough so I could start to play with them. I happened to lay one on the rear window of my insight in the full sun, and observed later that the PVC softened a bit when hot, and formed to the shape of the rear window. A few more test showed that even with the cells glued on, the silicone allowed the slight give in the bond required for the slight shape change.Cool. I will have to let it take the shape, then calk the cracks with silicone. This should help lock in the curved shape. I had to try connecting the 3 panels in parallel and series to see what they could do. I got the full 255MA SC, and 250 v OC that I expected. The series test got us up to 750 V, @ 85MA. I set up two small vices each with a graphite rod, and connected the + panel output to one, and the - to the other. The panels could sustain a continuous 1/8" arc, that could melt wires. I stopped the arc, and walked away to tend anouther experiment. Several hours later, I was going to clean up the panels, and without thinking, I grabbed one vice in each hand. Holy bat dropings, what a shock. I handled one at a time from there on out, with a very healthy respect for what 10%(cell efficiency) of 7.8 square feet of solar energy could do to me. I leave the leads twisted together on the panels now. They can't bite me that way. (Posted 7/16/2008 by mikey) Building the HV probe for the + attachment How to safely tap onto the battery pack + terminal If you looked at your car and said how can I connect to that bottom point where the battery pack positive input resides without shorting something out, or getting killed by the HV, here is how I did it. I looked on line to buy long clip on test probes, and did not like the selection or price, so I rummaged in my stuff pile, and came out with a piece of 1/4" OD nylon tubing, a piece of #14 romex, a 1/2" PVC rod, and a compression spring. I cut the nylon tube to the length of the required probe. I squeezed one end in a vice after inserting a 1/8 " rectangular spacer. This cold formed the tube so it was rectangular. I stripped the #14 wire, and formed a tight hook in the end.I cut a V in the long side of the nylon rectangle, to lock the probe onto the resistor lead. About 1" from the other end of the tubing I used an exacto blade to cut out a rectangle that was the same width as the wire was thick, so the wire slides in the slot. I fed the wire up through the tube and when it came out the slot, I pulled the hook into the full retracted position, and at the top of the slot, I made a 90 degree bend in the wire, stripped the end, and soldered on my red + wire. I put the PVC rod in a vice and drilled a .266" hole in the end to a depth that would just start to compress the spring when the assembly was put together. I used my milling machine to make a dog leg in the PVC rod, so the wire would lock into the dog leg. You should be able to carve the dogleg with an exacto knife, just don't cut your self. The spring is dropped in the PVC rod and the assembly is pushed together compressing the spring then engaging the wire in the dog leg. Once the operation is tested by pressing the PVC towards the nylon tube, a piece of music wire was bent and wound to a shape like a hypodermic needle finger hook. A wire tie was tightened on the nylon tube just below the fully compressed PVC pushers fully extended position. The finger hooks are covered with a small diameter heat shrink and the tightly taped in position with black tape.Wrap 4 or five wraps tightly, then one without stretching. Bend the wire over the PVC pusher, then tape the same way, 4-5 tight wraps, one no tension. Then using PVC pipe cement coat both tape jobs with a thin even layer. This will bond the black tape together so it will never get sticky or loosen up. One custom HV spring loaded attachment probe. (Posted 7/16/2008 by mikey) Building the HV charger/rebalancer Building a HV charger/rebalancer Finally got some more HV solar panels built. To make the process simpler and more reliable, I determined that a fixture to hold the cells in position while connecting wires and gluing with silicone was required. An 18.5" X 21" expanded PVC panel was drilled and fitted with 4-40 self tapping screws, with a pattern that positions the cells uniformly. The fixture will work up to 56 cells (HV panel). The cells are laid between the screws, and short stranded wires are used to connect the whole 56 cell array in series. The + and - end points have heavier wire. The rear of the cells each have four silicone bumps applied. and the backing PVC is carefully laid in position over the fixture and weight is applied. Overnight curing will fix the cells position and allow the whole fixture/panel to be turned over, and the fixture is removed. Silicone is applied to all cracks between the cells as well as around the edges, and the panel is left on a flat surface for several days of curing. The positive output is run through a fuse holder, and an isolation diode. This single panel charges the pack at a max of 85MA. A dead pack would take over 80 hours to fully charge with this low current.Three of these panels in parallel would take about 26 hours to fully charge a completely dead pack. Fortunately the packs usually are in the 20-80% SOC range. A pack that is at full on the Insight SOC gauge is about 80% charged, so it can still take ~ 1.3AH or about 15 hours of sunlight with a single panel, and a bit over 5 hours with a triple panel to reach the equalization stage. Longer is better, as we want to bring up any cells that are not full, and the low current will not damage the pack. (Posted 7/14/2008 by mikey) Insight Battery pack charger/rebalancer Danger this HV panel can electricute you Many Insight battery packs develop an imbalance. Sitting at a constant SOC for a long time without charging or exercise will allow the individual cells to self discharge due to internal leakage. This leakage is usually slightly different on each cell, so given enough inactivity time, the pack will develop a big difference in charge between the cells. The Insight system stops charging at 80 % SOC, so the imbalance never gets fixed by an equalization charge. If you fully charge a pack with a constant current of 85 -255 MA, this constant current will bring up the weaker cells by giving all cells a 100% charge.The cells are rated for 500 or more full cycles, so an occasional one or two will not effect the pack life. It is very likely that many packs that have become imbalanced, and were frequently recalibrating before throwing an IMA code can be revived by rebalancing. a solar pack charger that provides the required current at sufficient voltage to do this recharging and re-balancing is described here. The solar charger needs to be able to produce 180VDC to fully charge the battery.The panel output terminals with no load will be at 250VDC. this panel delivers about 15 watts of power. Using this procedure exposes you to the full pack voltage, so you can get electrocuted if you get your self across the HV, of either the pack, or the solar panels. I will be testing this system on some local people that have recalibrating battery packs. An AC powered charger/balancer is on the list of projects, but will not be as safe, or as green as this purely passive PV method. (Posted 6/23/2008 by mikey) Equilization test #1 Equilization charge at 300ma Ian and I took one of my 7 Seven subpack Prius based packs and after discharging it to lower the SOC by 10-20%, set up a constant 300MA charge system, and logged the voltages across each subpack. We spread the packs out so that cooling would be the same on each, and measured the temp of each cell when we took the voltage measurement. The initial condition showed a max of about .07V difference between the subpacks. The charge ran for 14 hours before I saw the sides of the cells start to bulge, but I never saw any measurable temperature rise. The Subpacks seemed to equalize starting at 7.85V/subpack, where all of the subpacks were at that exact voltage. I continued to charge up to 8.47V/subpack where all the subpacks remained at the exact same voltage. Next I will discharge the series string and log the voltages during discharge. (Posted 11/19/2007 by mikey) Battery pack rebalancing? HV taps WARNING this will fry you. WARNING, the batterypack can and will kill you if you contact the hot terminals, so do not try this if not fully comfortable with working on HV DC circuits. A look at the Saturday workshop efforts will show that Ian has rebalanced the silver Insights battery pack using the Triton smart charger. A lot of work, as the pack must be pulled and disassembled. James Frye from Reno, has duplicated the procedure for his consistantly recalibrating pack, and we await the results. Another approach to this rebalancing has been suggested by Armin Kusig who has been grid charging his battery pack each night with a gentle 300MA charge. Since the info on NIMH batteries would indicate that a low current charge of the batteries should not do any damage or substantially reduce the life of the batteries, it may be a better way to do a rebalancing. The battery pack current control/interconnect board allows access to both ends of the battery pack, and therefore is the best place to connect a series charger. The indicated points are only HOT when the main switch is on.We will be building up a constant current charge system to give this a try, and will report the results here. (Posted 11/17/2007 by mikey)