New Rear Cover

I built the original lower rear rack based on the dimensions of Concorde AGM 1280T batteries.  CALB 100AH cells are similar in height - but each cell will have a BMS module on top of it.  This means the original rack is just a hair too shallow.  To fix this, and to provide more support for the cells, I built a new lower rear rack cover - built out of 1/8" thick steel again:

The clamps are to hold the top angle iron in place during welding.  I used every clamp I own for this...  The angle iron adds significant stiffness to the rack, without adding a lot of weight.  You can also see the tabs which will be used to bolt the top in place.  This new top adds about 7/8 of an inch of extra clearance - just enough.

Top-Balancing the Pack

I did not do exactly "nothing" on the Jeep over the year.  In May/June I top-balanced the pack.  Balancing and monitoring of lithium battery packs is a hotly debated topic in DIY electric vehicle circles.  There are those who advocate what is called "bottom balancing" - discharge all the cells to a given voltage level (frequently, 2.7V) and then charge the whole pack in series until the first cell hits a "full" voltage level (often, 3.5V or 3.6V).  You record the overall pack voltage at this point, which becomes the "full" point.  Not all the cells are full, but they theoretically will discharge to the same low point, which is why this is called bottom balancing.  Bottom balancing advocates say that they can run safely without a battery monitoring system - they check the cell voltages occasionally - but really that just means that the *owner* has become the BMS.

Top balancing, as you might imagine, is the opposite of bottom balancing.  All of the cells are charged to the same "full" voltage level, and then monitored to make sure that no cell either goes too low during discharge, or too high during recharge.  With 96 cells, this requires automated assistance - the BMS.  In my case, I'm using the MiniBMS from Clean Power Auto.  It is simple, relatively inexpensive, and conceptually robust.  But it will work much better if the cells are all fully charged to the same level before hooking everything up.

To do this, I used a 25A nominal battery charger, with an adjustable voltage level.  Although the battery manual suggests that "full" is when the charge level hits 3.6V, I chose 3.5V to give a slight margin of error.  There are almost no amp-hours between 3.5V and 3.6V.  I hooked four cells up in parallel to charge - this fit well with my schedule, since it took about 18-20 hours or so for this 400AH megacell to fully charge, and I could just go out every morning and swap out the cells.  This went on for 96/4 = 24 days.  Here is one example, showing the target voltage set to 3.5V, and the charger putting out 13.7 amps:

The cells you see cropped to the right of the picture are from another project - the Toad Car - which I used in the 2013 St. Patrick's Day Parade in Fort Collins.  I used it to get some experience with lithium cells, and with the MiniBMS.  Those are 40AH cells, paralleled to get 80AH, and then put in series to get 25.6 volts.  Here is the Toad Car in the parade:

The Toad Car, and the City Council race, were another huge time investment which delayed resumption of work on the ElectroJeep.

Building the New Front Box

I'm catching up on a posting backlog dating back over a year.   It all started on the 4th of July weekend 2012 with removing the old batteries:

I also pulled all of the racks and cabling, since I'll be replacing the cabling (with orange high-flex 2/0 gauge welding cable), and refurbishing or replacing the racks:

Next, I built the front box.  It holds 30 cells, and fits inside the original front rack:

I welded a new front upper rack.  Because lithium cells are lighter, I used 1/8" thick steel rather than the 3/16" of the old lead-acid racks and boxes.  I originally was going to replace the front rack, so I built a new rack for the bottom.  I decided later to just use the existing front rack, and flip the whole rack so that the "bottom" becomes the top - this will be detailed in a later post.  Here is the completed rack:

I then created the battery box box from polyethylene, 1/4", natural color.  I started by cutting a single piece which covers the bottom, front, and back of the box.  The piece is then "V" grooved so it folds.  This starts with marking the path where the groove should go, and then clamping guides for a router to cut the path:

Here is the router, showing the "V" groove bit in place:

With the correct bit depth (not quite all the way through the plastic), it is a simple matter then to run the router between the guides, creating a nice straight groove:

Here is a test fit of the rack on the box, to make sure it fits:

And here is a test-fit of the cells.  I was originally going to use threaded rod to retain the cells in place, but it turned out to be impossible to get the rod through the sides when the rack and batteries are in place.  But the test-fit was with the threaded rod:

And that's what I got done on that weekend.  I also threw my back out as a result (what a drag it is getting old) which put a crimp (heh-heh) on further progress for over a year...

New battery layout

In the past few months, although I have not done a lot of building on the Jeep, I've done a lot of thinking.  I've figured out the exact battery layout I need - using the existing racks.  Here is the schematic layout:

There are 30 cells in the front rack, 3 in each of the under-seat boxes, 26 in the upper rear, and 34 in the lower rear.  This totals 96 cells - a little less than the 98 I was originally going for, but not enough to make a significant difference.

The front box sits in the original front rack, with some room to spare:

The lower rear rack is nearly filled with its 34 cells.  It will need some plastic inserts to retain the batteries in their positions:

Under each rear seat, there is not enough room to put the BMS on top, so it goes on a separate plastic insert/retainer on the side:

And the upper rear rack's 26 cells fit in the original rear rack with room to spare - which will allow for some extra components, such as a mid-pack fuse, the BMS controller, the amp-hour counter/sender.  It will also allow all cabling to be inside the boxes, which will eliminate high-voltage cabling from inside the passenger compartment:

Back at it

I've been busy with other things over the past year, but now I have a couple of weeks - and I think I can get the Jeep back on the road with its new LiFePO4 pack in that time.

I've created a Google Docs Spreadsheet of the to-do items - you can see it here and follow progress.  I'll also try to blog what I've done in the meantime, but the focus will be on completion not documentation.

Coming Soon...

...to an ElectroJeep near you (or, at least, near me):

A pallet full of something.  What could it be?

110 CALB LiFePo4 cells!  Each 3.2 volt cell stores 100 amp-hours.  The ElectroJeep will have 98 of them (the rest are for spares / redundancy).  Here is one possible layout (it fits in the existing battery racks):

98 cells yields a 313.6 volt nominal pack.  The charge voltage will be 352.8 volts, and the "empty" voltage will be 274.4 volts.  There will also be a BMS to keep any individual cell from becoming too full or too empty.  This compares well with the lead-acid 312 volt nominal, 371.8 volt charging, 275.6 volt empty stats.

Assuming 500 watt-hours per mile, and 75% depth-of-discharge, this should take the ElectroJeep over 45 miles range per charge.  My experience with the lead-acid was more like 10-15 miles per charge.  A significant upgrade, and almost 800 pounds lighter!

Rudman Mk3 Digital Battery Regulators

The ElectroJeep has been a little idle as a project as I've been engaged in other things - such as upgrading the Volt914 to 216V.  However, with that out of the way, I've returned to it.  If you remember, the BMS I installed originally was a PakTrakr (scroll to the bottom of this page) with Zener diode regulators.  This sort of worked - but I was very nervous about putting full amps into the batteries, and the PakTrakr has the unfortunate side effect of draining the first in its string of 6 or so batteries.  Not a good combo - several of the batteries were getting seriously out-of-balance, and the Jeep was sitting at 325V after a "full" charge.

So, I finally got around to replacing that system with a much better one, based on Rudman Mk3 regulators from Manzanita Micro:

These are designed to connect to each battery.  In addition, they are connected to each other and to the PFC-20 (soon, PFC-30) charger to tell it when the battery is getting full.  I installed the older analog model on the Volt914 - the digital model lets you adjust the cutoff voltage dynamically, so I can actually charge the batteries correctly no matter the temperature.

The first thing to do was to make plastic boxes to keep the regulators protected.  I started by cutting little corner chunks out of 1/2" polyethylene:

The idea is to weld together a box around these corners, which themselves become the mounting points for the transparent covers.  I had previously obtained a plastic welding kit, which included the welder as well as various plastic welding rods (including polyethylene):

The boxes are of various dimensions, to hold different numbers of regulators.  After calculating dimensions, I cut out the basic shapes from 1/4" polyethylene, and then cut bending tracks where the sides will fold up:

Four of the corner pieces go on each box.  Here is one box, from the inside (before folding).  Note that the corner chunks of the envelope have been removed here:

Here is the same box from the outside.  I used 3/4" self-tapping screws (although I also drilled pilot holes).  These screws are mostly to hold it together long enough to weld, although I never bother removing them:

The next step is to fold up all four sides and then screw them to the corner pieces:

The corners are then welded (and a bit of the long fold is welded as well, to add strength).  Holes are drilled for cable egress, and smaller holes are drilled for the nylon screws which hold the regulators securely in the boxes:

And here are all seven completed boxes.  One of them has some of the wiring attached:

The next step is to attach the connectors which will allow me to easily connect and disconnect the boxes for servicing.  I use WeatherPak connectors - you can see them here grouped in the order in which they will be used in each box.  In addition, you can see the 12-gauge wire I use to hook the batteries to the regulator.  100 feet of red and 100 feet of black were just plenty:

Here is all the wiring complete.  I call this the "14-armed monster".  If you look closely, you can see that the internal phone-wire style interconnects are also attached:

Next step is to construct connectors to attach to the batteries.  A note on safety: before connecting the ring terminals to the battery bolts, it is safest to insert the pins into the shells - this prevents them from rattling around and perhaps making a circuit with another battery post:

Another note on safety - all the tools used near the batteries have been wrapped in electrical tape within an inch of their life.  Note that there are still a few exposed metal places - particularly on the torque wrench - to prevent from interfering from the operation of the tool.  Nothing is 100% safe, but every millimeter that is covered in tape is a millimeter that will not close a high-amperage circuit - I've experienced that once, I never want to try it again:

So, with the process established, it is just step and repeat.  Here is the lower rear battery box all wired up.  You can see that each battery has both a red wire and a black wire.  In the connector, the red and black wires alternate - this makes it easier to verify that they are correctly paired, and less likely to make a catastrophic mistake:

Here is the upper rear battery box, complete.  Note that, before connecting the batteries to the boxes, I use an ohm-meter to verify that the connections go where I think they go - better safe than sorry:

Here are the under-seat batteries and their regulators:

And here are the front batteries and regulators:

The next step was to interconnect the regulator boxes with the 6-wire phone cord, and then verify their correct hookup by flipping DIP switch 6 on the charger and turning it on (with the amp knob turned to zero).  Here are the rear regulators - the yellow lights indicate that they are all talking on the regulator bus:

Here are the under-seat regulators glowing:

And here are the 8 front regulators in the big box glowing:

And finally the lonely little battery at the end of the string with its regulator glowing:

With the regbus wiring verified, and the boxes closed up again and everything connected to enable the full 312V circuit, it is time to charge!  For a test point, I disconnected the DC-DC converter and used its HV input lines as a convenient place to plug in a multimeter:

In the rear, I used a clamp-on DC ammeter to verify the current going in to the pack (the new PFC-30 has a built-in ammeter so this step will not be needed):

You'll note that the ammeter says "10.02" - that's not a full load.  As I mentioned before, the batteries were significantly out-of-balance.  So, a few of them got to full charge first.  The most problematic were two batteries in the front compartment.  Here you can see me blowing a fan on them to cool them down - this helps the regulator dissipate excess current more effectively.  The overheating regulators are the ones with the glowing purple lights:

There were a few regulators in the back that also got hot, so I took the covers off and applied significant airflow to them.  Note the hot air gun at the lower right - it was turned to blow cool air, not hot air:

A combination of lots of cooling air along with turning the charger down to output 3 or 4 amps finally allowed things to settle down and charge.  Here you can see the green lights of several regulators in the back glowing - this means the regulators are bypassing current but not overheating:

At this point, the pack is nearing full.  We've reach 371 volts or so, and the constant current phase is over.  Now the charger switches over to constant voltage, ramping down the amps as it counts down on a timer.  I set the timer close to the max - the batteries seem to require it.  You do not want to cut off this phase prematurely, or the batteries will not be charged.  Here is the voltmeter monitoring the constant voltage phase:

Finally, at the end, the timer completed and everything was done.  The green, yellow, and blue lights on the charger are normal, and indicate that everything finished OK:

And here is the pack the next morning.  333.5 volts is by far the highest the pack has achieved in a year, and a significant improvement over the 325 volts or so it was getting before.  This is an average of about 12.8 volts per cell, which is where it needs to be:

This process will need to be repeated a few times before the batteries stay more balanced.  To facilitate that, I will be installing fans on all the regulator boxes (so I don't have to grab all the household fans every time I want to charge).  In addition, I will soon be upgrading from the PFC-20 to the PFC-30 charger.  This will allow me to charge at up to 30A input at 240VDC - which translates to allowing me to charge at 16A at 370VDC - 16A is the recommended C/5 charge level of the AGM-1280T.

This process took several hours per night over the course of a week.  And I'm exhausted.  But it is nice to be making forward progress again.