Prelude conversion project - some questions

[antiscab]

why not just use the display that comes with the ZEVA? it already displays charger data from the CAN bus

[jonescg]

And convert it into an LED flashing? Just thinking of a way of using the LEDs there, that's all.

[coulomb]

TC used to supply a pinout for the LED but this doesn't appear to be the case on the big charger.

I believe that facility disappeared with the 7-pin round connector. My understanding is that the new chargers are potted, and so are not amenable to mods like this would require, but I've not seen inside them.

[jonescg]

Hmm, sounds like one day a photodiode array detector and signal transformation might be applicable... in a round-about complicated way.

[jonescg]

Slowly populating the HVDC box. I noticed that there are two kinds of cable gland. Those that use a metric sizing, and some PG thing, which accommodates the same size cable as a 25 mm cable gland, but calls for a 28 mm hole... Serves me right for using what i have lying around instead of buying the right stuff. Water heater contactor is in, as well as the pre-charge relay (I'll be getting a mini-tactor shortly - this one has a 48 V coil, but it's good for size.

HVDC dist box populated.jpg (197.75 KiB) Viewed 1262 times

I'm slowly going through and making busbars, fixing wires to the fuses and generally trying to fit it all in. I'm always amazed at how much space cable glands call for. I've already exhausted the only available side of the box, and most of this is 50 mm2 power cable from the two battery boxes, and the output to the drive inverter. I still need to fit all the 12 V wiring as well as the CAN lead for the current sensor.

[jonescg]

Battery cooling loop was primed and circulated for ages trying to get every last air pocket out. I think it will just have to be what it is, and maybe the air pockets will work their way out as the car bumps down the road. SO far so good.

20190811_180834.jpg (240.46 KiB) Viewed 1243 times

I picked up a stack of 5/8" hose with 90' bends so I can put the inverter/charger/motor/radiator cooling loop together. Then we are a few late nights away from spinning it up!

[jonescg]

Battery cooling loop has been bled - I added green corrosion inhibitor as well. The heat exchanger is keyed into the plate and doesn't move.

20190812_211940.jpg (185.92 KiB) Viewed 1214 times

I will need to restrict the flow to the front battery modules as they will take the bulk of the flow otherwise. The other frustration is the way the coolant circulates, bubbles don't tend to find their way to the reservoir naturally.

[jonescg]

With a bit of help from @zeva I think we have a workable solution for the EVMS 3 and Greatland inverter integration.

The Greatland inverter is simply powered on by supplying a switched (Key ON) supply, but the inverter won't be ready to drive until the precharge relay (IGN) is engaged long enough to precharge the caps. At this point the inverter supplied 12 V to the drive contactor which closes and the vehicle is ready to drive. I'm routing the original inverter supply line to the Key pin of the ZEVA EMS3 and taking the Main Contactor output of the EVMS3 to the Greatland inverter, sort of inserting the ZEVA unit in the middle. This way if a cell is low, it will sound the warning, and effectively turn the key off. The high cell voltage situation should be managed entirely by CAN Bus as the TC charger is set up to work well with the BMS.

Charge control box3.png (97.66 KiB) Viewed 1154 times

And as indicated:

EVMS3 Greatland link.JPG (98.75 KiB) Viewed 1152 times

[jonescg]

Video progress report for today - power steering and brake vacuum pump are installed. Other stuff... not so much.

[jonescg]

20190817_173810.jpg (247.57 KiB) Viewed 1077 times

Here's the conundrum. If we try to fit everything in as planned, it will be extremely crammed and quite difficult to work on. Also, the heat exchanger and aircon might not fit at all. Moreover, the speedo cable is right in the middle of everything, so if we ditch that we need an ADR approved alternative.

Or we settle for just the 6 battery packs at the rear and live with a 270 volt nominal system. This is just over 18 kWh, which might give 100 km range at metropolitan speeds. If we do this, we can fit all of the gubbins under the bonnet with a bit of room to breathe, and we can use a decent cell cooling system. It also means it can be driven in under a week, rather than ... who knows how long.

Hope to discuss with Francess tomorrow afternoon to get some directive.

[jonescg]

Well the decision to make it a shorter range, lower voltage car was made. Sad, but I'll feel much better knowing that there's nothing rubbing of chaffing against something it shouldn't.

It will make for a less powerful car too, with peak current around 200 A and 270 V nominal, it's a mere 55 kW machine. But, hopefully we won't be exceeding the original 1100 kg mass so it will go OK. The gears will prove essential up the hill I think.

As I'd already cut and crimped the lugs on the main pack leads before this decision, I had to make a terminal box. Might be a useful spot for a voltmeter or something

20190818_165043.jpg (253.55 KiB) Viewed 1035 times

Probably worth putting the bonnet back on fairly soon so I can work out what fits where. It's not driving today, but no reason it can't turn a wheel by next weekend.

[brunohill]

Maybe you need a bonnet scoop. People may think it is supercharged.

[jonescg]

I can probably find a smaller box for the charge control box - it doesn't need to be this tall.

Bingo!
https://www.altronics.com.au/p/h0312a-r ... enclosure/

[jonescg]

I popped the top off the 6.6 kW TC charger, and yep, it's potted like crazy in there, so no repairs will be happening here I suspect. However I could bolt the DC/DC converter to the lid with ease:

20190820_212612.jpg (164.45 KiB) Viewed 938 times

20190820_195621.jpg (157.98 KiB) Viewed 938 times

[T1 Terry]

Well the decision to make it a shorter range, lower voltage car was made. Sad, but I'll feel much better knowing that there's nothing rubbing of chaffing against something it shouldn't.

It will make for a less powerful car too, with peak current around 200 A and 270 V nominal, it's a mere 55 kW machine. But, hopefully we won't be exceeding the original 1100 kg mass so it will go OK. The gears will prove essential up the hill I think.

That is the same voltage as the Prius PHEV pack in my silver Prius, I'm all eyes and ears now I wonder if I could transplant the electric drive part of a Prius along with the inverter to give me an 80kW drive motor with a 270v battery pack, complete with regen. It has plenty of pull up hill on electric only mode, even from a standing start .... maybe a project for the Jimny when the auto finally drops its bundle.

T1 Terry

[jonescg]

Spent a couple of hours with @reecho putting inverter / charger / motor cooling loop together and crimping a few lugs with the power crimper (so fun, so dangerous).
I need to pick up a few more bits and pieces before I can plumb it up. All the low voltage wiring will take some time to get right...

[Petie]

Hi friendws!
Before I joint the forum, I could view the drawings etc. attached to this blog now they are gone, I wonder why?
It says: You don't have the required permissions...
I have the (rough) plan to convert a car, small van or similar to an electric vehicle, so plans, drawings etc. are of a big help. May be you can tell me how to get out of that problem..
Kind Regards, your new forum member
Petie

[weber]

Hi friendws!
Before I joint the forum, I could view the drawings etc. attached to this blog now they are gone, I wonder why?
It says: You don't have the required permissions...

Welcome, Petie. Thanks for letting us know about this problem. It should be fixed now.

[jonescg]

Far out. If I'd known that using an aftermarket potentiometer style accelerator pedal would be this much work I would have kept the original mechanical cable one.
I have just gone to great lengths to remove the original pedal by angle-grinding the bracket in about a dozen pieces, and levering things off with a pair of pliers. Only to now find that the pedal I bought from Kelly Controls would need some serious fabrication to fit properly. All the while you're upside down, backwards, looking up into darkness with no room to move your hand without dislocating your shoulder...

[brendon_m]

Ahhh Honda's, such lovely cars to work on.

[antiscab]

to be fair this was also my experience with a Toyota aw11 mr2

[jonescg]

It was a productive weekend, but not as productive as I'd hoped. I genuinely thought we would be driving the Prelude by now, but a few things have held me back.

Started off well - got all the HVDC work finished so now the only task is to install the LV wiring. Since I lopped the cables off a bit short thinking we'd be putting two battery modules up front, I had to install a waterproof terminal box at this point. But it also serves as a convenient place to monitor the HV isolation from, so that's what I've done here. The wiring is only fused back at the battery pack with the main 200 A fuse.

New HVDC termination box.jpg (168.84 KiB) Viewed 769 times

I connected all the plumbing from the radiator to the inverter, motor and charger, and primed the pump. Again, it's not self-priming like I wanted it to be, but I got there in the end.

Plumbing and HVDC wiring.jpg (227.64 KiB) Viewed 769 times

The CAN bus is mostly done, with expensive Amphenol connectors on everything. I hope the EVMS3 doesn't have a problem with being in the middle of the run -

<120 ohm resistor - Battery module 6 - Battery module 5 - Battery module 4 - Battery module 3 - Battery module 2 - Battery module 1> - EVMS3 - <Current sensor - TC Charger>

HVDC box nearly done.jpg (271.71 KiB) Viewed 769 times

I haven't had a chance to fire it up and find out yet, but the run to the charger via the current sensor is short.

The charge control box is coming along... I've built a shield for the Arduino with a 100 k voltage divider for all 6 thermistors. The 5 V supply to all 6 should be up to the task - a total of 300 mA will mean some voltage drop, but as long as it's proportional it shouldn't mater.

Charge control box.jpg (264.13 KiB) Viewed 769 times

I also spent half the morning installing the heater switch on the centre console, and running the ZEVA screen cable up to the dash. The top of the dash seems the best spot for it, considering the centre console was already busy with compartments etc. It looks a bit like an NG tube

Heater switch installed.jpg (196.6 KiB) Viewed 769 times

ZEVA BMS screen wiring.jpg (181.03 KiB) Viewed 769 times

But the most miserable part of the day was spent upside down, in the dark, contorting myself to fit in the driver's side foot-well working on the placement of the new throttle. If I'd had known it would be this hard, I would have left the original throttle cable in place and made it work somehow. But here we are.

Evil throttle mount.jpg (195.01 KiB) Viewed 769 times

I had to remove all of the remaining bracketry from the original mechanical throttle. It would seem this was one of the first parts on the Honda assembly line, and everything else was built around it. Because there's barely 30 mm of space anywhere around it, and you would need to remove all interior trimmings to get half-decent access to it. In the end I put a metal sheet down and wailed on it with the angle grinder until I could lever chunks of metal off. Eventually I found enough free space to build a steel plate which will form the basis of the new throttle pedal mount. Being one of these from Kelly Controls:

Kelly-0-5V-Accelerator.jpg (135.18 KiB) Viewed 769 times

...it was made with a forklift or some other convenient utility vehicle in mind. Not a 1988 Honda Prelude with nary enough room to swing a cat. So I will devise a series of angle-steel brackets to make it work...

[jonescg]

I put a connector on the thermistor bundle wiring and plugged the Arduino in. A friend helped with some code (Quick Median library) which takes 5 samples and chooses the median value for the result. It means a far less spikey sampling (so I'm told). The readings from each thermistor varied by no more than 2 counts, but from one thermistor to the next it varied by up to 5 counts (out of 1023). So I will set the AC compressor to come on when it is a good 5 or 6 degrees warmer than ambient. It was giving counts of about 386-392 and the ambient temps were around 13'C.

I have been trying to post code, but it seems to cause a bit of a 403 Forbidden error...

Might be easier to add the file as an attachment?

6 thermistors.txt

(2.9 KiB) Downloaded 24 times

[francisco.shi]

Just a note on the averaging.
You can increase the resolution of the ADC by adding the values and keeping a total.
For example if you read the ADC 4 times you will be adding 2 bits to the ADC resolution. This is assuming you have some noise in your input which is higher than 1 bit.
Then if you divide the total by the number of readings then you get some thing like an average.
Instead of adding all of the readings and discarding each reading you can put them in an array and use a circular buffer arrangement so you only need to read one value each time and you still have the readings from previous values.
This will also work better if you run the readings at a fixed time interval (say 100 times/sec)
To avoid the output flickering you need to have some hysteresis. So say it turns on at 630 then it turns off at say 610 or something like that.

[weber]

You can increase the resolution of the ADC by adding the values and keeping a total.
For example if you read the ADC 4 times you will be adding 2 bits to the ADC resolution. This is assuming you have some noise in your input which is higher than 1 bit.

That's not quite right. The rule for dithering using noise, is that to get n more bits you have to sum 2²ⁿ samples, not 2ⁿ samples, because half the extra bits you get are just noise. So to get 2 more bits of resolution you have to sum 16 samples, then round off the low 2 bits. value = (sum+2)>>2

Also, with the moving sum idea, there's a variation where you don't need to keep old measurements in a circular buffer. You only need to keep the sum. For each new sample you calculate

• sum = sum - (sum+2)>>2 + sample
  value = (sum+2)>>2

That's a single-pole low-pass filter with a time-constant of 16 samples.