Pulled out the old drivetrain, removed a bunch of heat shields, skid plates, and spare tire cross member. Pressure washed the frame.
Got over my procrastination on that rear corner. It looks passable by my standards so now the passenger side of the tub is basically done. I need to fab up some rotted out body mounts but I’ve now moved on to the driver side which is both better and worse than the passenger side was.
Still procrastinating on sheet metal. Also procrastinating on taking apart the back half of the transmission to work on an adapter plate. Also procrastinating on removing the drive train from the chassis to see whether one of these battery packs can be made to fit.
I’ve been working on trying to talk to the BMS in one of the battery packs. Turns out each of the battery packs had drained down quite a long ways and since I don’t have a 350v power supply, I decided to charge the modules individually at very low current. The good news is the battery modules charge up to at least 42v (12s; I haven’t tried to go higher than 42v) and hold their charge for a week at least. Planning to do some load testing soon.
In the meantime, I’ve been trying to strike up a conversation with the BMS in each pack using this firmware.
My Saleae Logic analyzer is seeing the correct packets but the Lilygo CAN interface is not seeing any CAN packets. I think I’ve fried the transceivers maybe. Another board is on order. I’m pretty sure this is the case because it can’t see any CAN packets even from other CAN devices that are known working.
(id: 0x112 is a valid BMW SME status message).
I’m procrastinating on rebuilding the rear quarters. I’ve got 6 separate and not perfectly shaped pieces I’m trying to put together so I’m in avoidance mode. Found some battery packs at auction. 42kwh packs from a 2021 BMW i3.
Also got around to getting the transmission spinning using a 33v33A bench supply as well as a 12v3A bench supply for the controller.
I picked up some decent-ish front fenders on our recent road trip which meant I could sacrifice my existing front fenders. The front left fender wheel arch on the rear right quarter panel, fits not too bad. Could stand to drop a little further at the back end and I’ll have to abbreviate the front end of the arch and blend it into my ‘rockers’. See: “Not doing a museum level restoration”.
As far as the drivetrain part of the project goes, I’ve received the controller from Ireland, a case and connectors from Aliexpress, and an inverter wiring harness from California. Also got a throttle pedal from a Honda Fit. I need to start crimping pins and wiring up the controller and then I should be able to spin this thing up on the engine stand.
I’ve also started documenting the wiring connections between the inverter, controller, throttle pedal, and transmission.
I remember when welding patches and rebuilding quarter panels was fun. The end goal of this truck is not a restoration, but merely to have a structurally solid with non-rust-bucket vibe. As such, I can take a few liberties with how things are put back together, but it’s no less tedious.
Rebuilding the wheel-arches is another thing altogether. I’m trying my hand at some hammer forming with a wood/metal buck but I don’t think that’s going to turn out well. I used a 1-7/8″ trailer hitch ball to make a forming attachment for my air-hammer. Unfortunately, it wasn’t hardened so it just mushroomed at the end and also hardly made a dent in the sheet metal. Back to the drawing board.
The pieces of my drivetrain are coming together. What you see there is the transmission out of a Lexus LS600HL (known as an L110F), plus the transfer case from a 1992 Toyota Land Cruiser 80 series (known as an HF1A), as well as the Inverter from the Lexus LS600HL (bottom). The high voltage leads from the battery pack (I don’t have one yet) go into the inverter, and the two orange leads from the transmission plug into the inverter. There is a separate controller (I also don’t have yet) that will talk to the inverter and tell the transmission what to do.
There will be nothing connected to the front (left side of this photo). That’s where the V8 gas engine used to be.
Work continues on the FJ-55. The passenger side A-pillar is coming together. It’s on the rotisserie now and I seem to have acquired an L110F transmission. An HF1A seems to fit on the back.
A friend offered me a truck that he’s had on his TODO list for over 20 years. This truck has been sitting in his car-port, covered in various bits of detritus. After almost an hour of excavation, I was finally able to see the truck. Code-name: Terrible Idea.
What are my plans for it? I’ll save that for a later post because I’m not entirely clear yet. First order of business is to deal with rust and missing metal. Its biggest problem is the A-pillars that the front doors are mounted to are completely rotted.
The first decision I’ve made is not to do a museum level restoration. I want a truck that looks cool and still looks like an FJ-55 but a purist would immediately know something has been done to it.
First order of business was to tackle one of the A-pillars to see how doable it is. If those are too difficult, then the project is a no-go for me. Fortunately, they proved doable given the technology I possessed. I built the A-pillar out of some 2x3x.188 rectangular tubing. I mated it to some 1.5x3x.250 rectangular tubing to replace the almost completely missing rocker panels. I did a floor patch while I was in there. Lots of spot welds to drill out. Not sure I’ll go to that extend on the passenger side.
The drivers side front door mounts perfectly and lines up nicely and opens/closes cleanly. This project has now passed the ‘doable’ gate.
Next order of business was to dismantle the interior and build a rotisserie.
My truck is too long. The Ford website says 20’3″. The space between the front and rear garage doors is 21′ even. That doesn’t leave a lot of margin for error when driving in.
Enter Rat-ratgdo. This is a board that can link with Liftmaster/Chamberlain Security 2.0 garage door openers. It can read the state of the door, sensors, and control the door. With ESPHome firmware, it links easily to HomeAssistant.
I moved the obstruction sensor for the rear door up to the height of my bumper. When the bumper clears, the rat-ratgdo obstruction sensor switches from “Problem” to “OK”.
This allows me to create an automation:
When I drive in the back door, a light in the ceiling turns on as soon as I block the obstruction sensors and turns off again when the sensors are clear. If the light doesn’t turn on when I drive in, then I know that HomeAssistant isn’t working so then I just have to look in my rear view mirror and watch the blinking LED of the obstruction sensors which is slightly more error prone.
for I in 105 103 102 101; do cat $I/etc/vzdump/pct.conf | grep -v ^lxc | grep -v "^#" | sed -e 's/^/\-/' -e 's/: / /' | xargs pct create $I /mnt/pve/nfs-beernas-containers/template/cache/null_template.tar.gz; done
for I in 100 120 121 122; do mkdir -p /tmp/$I; lzop -dc vzdump-qemu-$I-*.lzo | vma config - > /tmp/$I/$I.conf; done for I in 100 120 121 122; do cat /tmp/$I/$I.conf | grep -v '#' | sed -e 's/: / /' -e 's/^/\-/' | xargs qm create $I; done
…and to avoid this pain in the future:
5 0 * * 0 touch /mnt/pve/nfs-beernas-backups/.timestamp 0 0 * * 0 find /mnt/pve/nfs-beernas-backups -name config.dump.* \! -newer /mnt/pve/nfs-beernas-backups/.timestamp -print | xargs rm 0 0 * * * sqlite3 /var/lib/pve-cluster/config.db .dump > /mnt/pve/nfs-beernas-backups/config.dump.$(date --utc +%Z%Y%m%d%H%M%S).sql