Electrical 2 – Low voltage & accessories

This is a page of fixes that could apply to future projects or OPC’s (Other People’s Conversions). It’s how I went about making car-controls “better” and connect them into a digital controller / PLC so I can add new features through programming. There’s not a lot of planning in the sequence of features on this page, it’s just The List.

Overall layout

The basic layout of the low-voltage system is that there is a central control-unit/PLC in the back, and then 5 fuse/relay boxes throughout the car to control the actual bits and pieces. There are 2 boxes in the front for lights / horns / pumps / DMOC control. Then there’s 2 boxes in the back for tail-lights, trunk open / close, charging control etc. The last box is inside the cabin and deals with the dash / gauges, door-locks and wiper-motor. I probably could have put all of it in 1 giant box somewhere, but the cabling became just unmanageable.

In the process of EV converting the car, I have started over on its low-voltage wiring. The original harness is/was 65+ years old by now, the connectors were rusty and the insulation was dry & crackly, and it just did not seem to be something I wanted to rely on for a daily car. Adding the new cruise control, the new interval-switching on the wiper-motor, the new indicator and tail lights, the new trunk open/close feature and everything needed to control the electric drive-train & charging would have more than doubled the existing wiring anyway, and rather than splicing into an archaic cable trunk I’ve started over.

.

PWM-booster / accessory control

This is a component that I’m using in several places to operate lights & gauges, so I’ll put the description up here.

The way to make a motor spin slower/faster, to make a bulb brighter/dimmer or to make a gauge go up & down is to put more or less power through it. One way of doing that is to pulse the power really fast (20,000 pulses per second) and make each pulse longer or shorter. If there are 20,000 pulses per second then each pulse is 50 milliseconds. If the pulse pushes power for 10% of that time (5 msec) you get 10% power so 10% speed / brightness / fullness. If the pulse pushes power for 50% of that (so 25 msec) you get 50% power. For sure, not all instruments react quite so perfectly, some gauges might need 30% to do anything and max out at 72%, but same idea. This pulsing of power to “dim” things is referred to as Pulse Width Modulation or PWM.

To make this happen, I have made a small module / circuit-board that I can pulse with the PLC (or any small signal-generator board) and that can then handle larger power. Call it a booster for PWM. I’ve made 2 versions, one that pulses +12V (for lights) and one that pulses GND (for dash-gauges). The module itself is the same size & pin-configuration as a standard Bosch-style car relay, so it will fit into regular standard relay-boxes etc. I had to pick a size, and this seemed like a good idea?

Once in a 3D printed sample of a housing (the white one is the PWM-booster, the black one is a Bosch relay) it looks like this. The housing gets filled with epoxy-resin for solidity, dust- and waterproofiness etc through a hole in the top, and that also insulates the exposed PCB traces on the bottom.

The schematics for the boosters is this, if anyone sees a mistake or improvement then please let me know.

Fuel & temp gauges

This is the first application of the PWM booster. The schematic for the original fuel-sender/-gauge circuitry is as follows:

I went ahead and measured out the resistor values inside the gauge (for completeness and to have an idea of how much power I would need to apply). The original setup draws the gauge “input-pin” to GND via the fuel-sensor, so I’m using a GND-pulsing PWM-booster to simulate that fuel-sensor through the PLC. With that, the circuit turns into

So, as far as the dashboard is concerned, there is no change to anything and the original gauge displays the battery status. The PLC has a little lookup table to define what % pulse-width corresponds to what level on the gauge (it was not a linear curve). The temperature gauge has the same basic set-up, just a different lookup table of values.

Speedometer & odometer

Windscreen wiper control

Cruise control

The original cruise control design is very clever, and very “old” / limited. They knew what they wanted to achieve (hold speed), and they went the shortest route to get there. The head-unit has a pull-function to “set” a relay that makes cruise control work, and the brake pedal re-sets that relay. The driver turns the knob to set the speed through a mechanical cable & spring setup (so it does not look at the current speed/pedal, it just knows what the knob is set for). It works well enough.

For the new version, wanted something that would look at the current speed/pedal and go from there. It seems obvious now, all modern cruise-control set-ups do it this way, but the “knob-position=speed” way of working was how the first 10-or-so years of cruise control units did their job.

To make the modern version work, I need a set of separate digital signals coming into the interface that then connects to the CAN-bus interface. I’m going for on/resume, +1mph, -1mph, off, and cancel via the brake pedal. Not as nice as a modern unit, but still kind-of close and it fits within the four directions/kinds of motion I can get out of the OEM head-unit.

To do that, I’ve re-designed the inside of the head unit to take out the cable-adjuster and install a series of spring-loaded toggle-switches. Turn left now clicks -1mph, turn right clicks 1mph, pull on the knob is on/resume, and push is off/reset. The interface electronics handle de-bouncing the contacts and feed the button-pushes into the two CAN-controllers.

This insert leaves the original housing intact, the bulb can still get in for the back-lighting, and the original knob sits where it’s supposed to sit. All in all, not un-elegant. I still need to verify the stiffness of the front- back springs so the mechanism centers nicely and has a good “finger-feel”, but the 3D printed prototype works just fine to write code. Once it’s all been tested I’ll go ahead and make a brass or aluminum version of the whole thing, but for now it’s another feature done.

Tail lights

On the OEM version of the car, the tail-fins are massively over-stated and the tail-lights are (in my opinion at least) very much under-stated. The front has double headlights plus turn & fog-lights, it’s all lit up in the grandest way, and in the back there’s really not much of anything. While the bullet tail-lights do look cool and have a lot of intricate detail in the chrome etc, they do feel pretty skimpy as the only lights in the back.

Rather than just bolting on some extra tail-lights, I’ve turned the large reflectors in the rear bumper into a set of tail-lights. The PLC can then control the intensity of those, so they could even be break/tail lights. The only question there is whether the reflector-part will still allow enough light through to be bright enough to function as a break-light.

The new lights are made from a sheet of PCB material and pieces of regular hi-density 12V LED light-strip. I looked at doing a “regular” PCB and a bunch of LED’s, but it got kind-of expensive and I still did not get a lot of LED’s into the housing while still being able to hand-solder them. The LED strip on the other hand has a ton of LED’s per inch and can be mounted vertically so I was able to get to maybe 5x the number of LED’s and still have something that is so-so easy to make (pouring the epoxy resin is a messy operation). The +12V PWM-booster can (easily) handle both tail-lights, but I’m using 2 so I could theoretically make the LED lights blink as directional lights or shine brighter as add’l brake-lights. Too many options, and the PLC just gained a few lines of code.

The OEM light is held in place with a few long screws / threaded rods, and I re-made one of those out of a piece of tube so I can feed the new wire out of the housing without drilling etc. The LED light fits into the original tail-light housing “inside” the plastic reflector, and it only adds 1.4mm to the thickness of the reflector so there is no noticeable height difference when the tail-lights are assembles. Also, the original reflector-piece is unchanged and there’s nothing new in front of it, so nothing here is violating the rules around rear reflectors either. I think the result is pretty cool:

Indicator lights

The OEM car comes with the typical 50’s indicators, as in in the front they’re white and in the back they’re part of the red lights. Those red lights are low-lit as tail-lights, bright-lit as brake-lights, and they flash as indicators. It’s how it was, it worked, but it’s not much.

With the new LED tech out there I’ve gone to 2-color bulbs in the front indicator housing and the backup-light housings (white/orange) via a 2-lead light socket. This way I can have self-contained orange blinkers, and in the front I can have Daytime Running Lights in the indicator lights (maybe thru PWM control, maybe just straight) while in the back the white half still works as the reverse-lights. Of course, seeing as it’s now all under PLC control, the front can also use original white blinkers & no DRL, and the rear ones can be just-backup-lights (and the PLC routes the blinking back to the red bullet-lights). All original, or all enhanced, all depending on mood-of-the-day and what the local street-legal inspector tells me is required.

Trunk open/close

Wiring diagrams

I have put in kind-of a color-coding on the wiring to try and keep things more or less organized Blue wires connect a device/relay and a logic controller (usually the PLC). Yellow wires run from a relay to the device. Green wires run between the PLC and the CAN-controller. White wires run between switches and devices that are not on PLC control (like the up/down on the radio antenna). (Separate) red wires are 12V from the low-voltage battery, (separate) black wires and exposed braided straps are “ground” on the 12V battery and chassis.

Most wiring is single-wire / independent wiring, however, a few sets are combined into a 4 or 8-strand cable to try and get a proper shielding set-up going. The DMOC connects to the CAN-controller using a 4-strand white shielded cable, and to the motor sensors using an 8-strand shielded cable. The wiring from the gas-pedal to the CAN controller is another 8-strand shielded cable. The wiring for the radiator sensors and fans is an 8-strand unshielded black cable because it was convenient.

For the drivers side front relay box wiring

Passenger side front relay box wiring

And the lay-out of the actual boxes

Drivers side rear relay box wiring