Welcome to Manins' truck camper project

LiFe Battery

Overview

The camper was designed to accommodate up to three Fullriver HGL120-12 batteries. Since our objective is to power a 230 V induction cooktop, fridge and various accessories, with loads up to 3000 VA, I have replaced the two AGMs that came with TT30 with a Lithium battery as large as would fit in the space available (555 x 352 mm).

The Battery

20161123_120644.jpg

4 x 4 LiFe cell groups made up from 16 90 Ah Winston cells.

Because of its charging characteristics, a lithium battery is much less forgiving of over charging than a lead-acid battery. It requires a battery monitoring/manage- ment system. With help from T1 Terry Covill I have implemented one and it works very well. The major changes I designed and built are described below. The reasons for these changes and some implementation details are omitted here.

The battery, actually two 6 V packs in series, consists of 16 x 90 Ah LiFeYPO4 cells connected in groups of four in parallel. This makes up a 360 Ah 12 V battery. Winston cells have a good reputation and reliability. If they were available, the 100 Ah "(A)" cells from Winston would have given us 10% more capacity but no-one in Australia stocks them. The Iron Phosphate type of Lithium cell is very stable, is resistant to thermal runaway, and does not vent noxious gases when being charged. The Yttrium doping apparently helps with low-temperature performance. I use groups of four cells in parallel both to make good use of the space available and to enhance the voltage stability (the "balance") of the cells and overall pack. The cells, clamping end plates and interconnect links were purchased from EV Works since they were the only people who could supply 90 Ah Winston cells.

battery_links.jpg

Note that the cells can accept bolts with up to 12 mm of thread. All link surfaces were deburred, scrubbed with steel wool and smeared with electrically conductive silver-based thermal paste from Jaycar.

With end plates, strapping and the interconnect links in place, each 6 V pack weighs only 26 kg and is easily handled. A Blue Sea Battery Terminal Fuse Block with 300 A fuse is mounted on a terminal between the two 6 V packs, which are joined by 95 mm² cable. The fuse is a type of Marine Rated Battery Fuse (MRBF) with an Ampere Interrupt Capacity (AIC) of 10,000 A @ 14 VDC

20161123_130828.jpg

Cells strapped, connected to make 2 x 6 V = 12 V battery.

life_battery_fuse.jpg

Close-up of the 300 A fuse block and its connection.

5191.png
IMG_1292.JPG

The tag of the fuse block fatigue-cracked after ∼500 hours of use.

Update Feb17: After 25,000 km of travel, a lot off-road, the fuse block failed with a fatigue-crack at the junction of the terminal tag and the block itself. Made of copper, this tag is simply too thin and the wrong material. The 95 mm² cable must vibrate on corrugations so much that it flexes the tag, work-hardening the copper.

in-line_fuse.jpg

The 300 A fuse bolts directly to the battery terminal. A black plastic washer insulates the bolt from the cable lug.

My solution was to dispose of the fuse block completely, ask a friend to make me an acetal plastic top-hat insulating washer, and use a 46 mm M8 bolt to bolt the 300 A fuse and cable directly to the top of the battery terminal.


Battery Management

The battery is charged from the solar panels or from the vehicle alternator. Left to their own devices these sources will not significantly overcharge the battery so long as all cells are in balance. However, it is possible that individual cells may be sufficiently out of balance that they are damaged by the charging. This possibility is minimised by having four cells in parallel in each of four groups.

The lithium battery pack voltage and cell voltages are monitored by a Junsi Cell-Log 8M (or model 8S with data logging capability). This low-cost device can monitor the four battery cell groups at once, and can be set to sound an alarm and switch an open-collector transistor if a cell or the pack goes over or under set voltages. A length of 7-core trailer cable is used to connect each cell group, starting at 0 V connector, to a connector block of 5 A fuses behind the touch screen in the camper, and then to the Junsi 8M. All Junsi connections to the cell groups are made with M8 ring crimp connectors (crimped and soldered). These can be tightened very well — essential for reliable measurement by the Junsi device.

  • Black. 0 V. (manually switched to reset Junsi)
  • Blue. Group 1 (3.4 V).
  • Brown. Group 2 (6.8 V).
  • White. Group 3 (10.2 V).
  • Yellow. Group 4 (13.6 V).
  • Red. Temperature sensor red/black available for solar regulator.
  • Green. Temperature sensor black available for solar regulator.
IMG_5717.JPG

The battery with end plates and strapping is a close fit — vertical clearance is almost zero when the lid of the battery box is on.

The solar charging is set to cut out at about 13.85 V on the pack. That would be automatic if the battery temperature were to be above 36°C and is forced here by replacing the temperature sensor by a potentiometer set to about 2.3 kΩ — see the solar discussion.

Management at the cell level for the solar charging, and at cell and pack level for alternator charging, is done by the Junsi Cell-Log. The Cell-Log is set to alarm whenever the pack or a cell goes over-voltage [14.0 V for the pack, 3.45 V for any cell group]. It also alarms for under-voltage conditions [set to 11.5 V for the pack, 2.8 V for any cell group].

The open-collector output of the Junsi is fed to a simple delay circuit based on a 555 IC to provide sufficient hysteresis in the switching. Upon receipt of a signal from the Cell-Log, the circuit switches a DPDT relay and waits for about three minutes before checking to see if the alarm is still present. If so, it continues to activate the relay. One pole of the relay cuts the alternator charge. The other pole switches in an additional resistor to the temperature sense terminals of the solar regulator, causing the regulator to also cut charging (if it has not already stopped) if a cell voltage is out of range. The cycle repeats until the charge sources and/or loads change.

The circuit to provide this management function is here:

delay_circuit.jpg
delay_board.jpg

The delay and switch circuit board for charge control. The set point for the solar temperature potentiometer was initially 2.125 kΩ; its now 2.3 kΩ.

IMG_5940.JPG

Red (power on) and green (relay off) LEDs show normal operation of the circuit board behind the panel.

Here is the Junsi Cell-Log installed above the touch screen in the camper, with override toggle switch and monitoring LEDs for the delay board. Also a photograph of the circuit board, mounted in a plastic box. The circuit is built on strip-board and is approximately the same layout as the circuit shows. With the chosen component values (680 k resistor, 220 µF capacitor) the delay is 3 minutes, 10 seconds. Current consumption is 14 mA off, 45 mA on at 13.2 V.


auto_checker.jpg

I was manually cycling the 0 V line of the Junsi off and on most evenings. One screen of the Junsi shows the maximum (and minimum) voltage reached by the pack since the Junsi was last switched off. This was handy to see whether full charge had been achieved during the day. However, even using a microswitch with a good snap action was inadequate, causing inconsistent readings due to variations in contact resistance.

bms_components.jpg

As a check on the Junsi readings (something I have found to be essential!) a Turnigy DLUX LIPO Battery Cell Display and Balancer is in place, using the microswitch to cycle the 0 V line off and on. So long as the cell voltages are above about 3.0 V the Turnigy will display their voltages and if any differ by 0.01 V or more, the device will attempt to balance the pack by moving up to 50 mA for up to 3 hours between cells — this will not achieve much on a 360 Ah pack!


Update Nov16: During a 22,500 km trip through the outback of South Australia, Northern Territory and Cape York Peninsula on mostly muddy, corrugated or rough dirt roads, a few problems with connections to the battery came to light. Out-of-balance issues were being falsely reported by the Junsi Cell-Log, causing me to attempt to correct it. Upon strip-down the cause was plain: poor connector contact for the 0 V cable with Group 1 cells and a poor contact for 0 V line for the Junsi.

The battery was removed from the vehicle, each cell group was recharged to over 4.0 V and let sit to make sure all cells held above 3.5 V. Then all interconnect links were removed, cleaned and smeared lightly with conductive paste. The red dust that had covered everything was cleaned off the cells, the packs were reassembled and installed back into the battery holder.

20161122_140906.jpg

Red dust everywhere after a five month trip through outback Australia.

20161122_144422.jpg

Poor connection to 0 V cell group, chafing of interlink, and failed threaded screw hole for 0 V Junsi line.

20161123_181054.jpg

After cleaning and recharging, the battery is back in its holder with power and monitoring lines connected.


LINKS