Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

largest-farva wrote:

RCC says the cage is on the loading dock waiting for pickup, so maybe it gets here this week.

We also ordered our cage from RCC - don't be surprised if you don't get a tracking number for 2 weeks after the first time you're told the cage is ready to be shipped.

52 (edited by rozap_ 2023-08-28 12:07 PM)

Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

hey it moved, that's progress.

_rozap that is a very kind offer I will probably be taking you up on if we get the cage and are still committed to the October race.

sounds good, hit me up on the Lemons discord if you end up doing a pre-race wrench marathon https://discord.gg/EJWrv9KMZn

BSOD Racing, 1987 Fiat X1/9

Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

The clunking sound was either going to be a known issue, not enough clearance between a bolt head and the jackshaft u-joint, or the chain drive.

https://cdn.discordapp.com/attachments/ … /clunk.mp4
Welp, it's not the u-joint, so my sprocket alignment is off or something...

RCC wrapped the cage up and it's on the truck, but it won't be here for another two weeks. I think that between those two challenges our fate is sealed and we'll be showing up to Smells Like as spectators this year.

I would really like a re-do the chain drive too, with a hy-vo chain instead of roller chain.
https://www.tayloredge.com/reference/Me … tChain.pdf


Otherwise, progress is steady.
https://media.discordapp.net/attachments/1103701707105509467/1147262438224445490/image.png?width=888&height=665
https://media.discordapp.net/attachments/1103701707105509467/1147262437674975383/image.png?width=880&height=665
https://media.discordapp.net/attachments/1103701707105509467/1147262438736154755/image.png?width=898&height=665

Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

Getting pretty good at pulling the engine out.

https://i.imgur.com/ltoGGta.png
The u joint was making contact, but just grazing.

The theory is that something with the chain is binding up and snapping loose.

The sprocket that slips over the input shaft of the transaxle is supported by the input shaft's bearing on one end and an added support bearing on the other end. The support bearing isn't normally loaded, a hedge bearing.

For this first run I had that extra bearing in there with jbweld, because I didn't know another way to locate it accurately. As a bonus, if it got cracked up from driving up and down the driveway that would be some valuable evidence of unexpected forces.

https://i.imgur.com/1Dm5AdT.png
The good news is that there aren't any signs of trauma with the jbweld, so I will convert that to regular weld.

The other good news is that after spinning it at 3krpm for a minute (it doesn't bind up if it's just spinning with no load), everything spins freely.
It wasn't even that loud. Bet if I throw a strip of acetal in there it would cut most of the clatter.

https://i.imgur.com/TOhOVFR.png

Then more good news.
It looks like the master link has a slightly wider pin than the other links. (picture)
If that's what was causing the lurching and clunking then it's an easy fix.

But... the bad news is that the impact point and the master link line up with each other on every revolution of the chain.
The number of chain links and sprocket teeth are even multiples of each other. Doh! and/or oh well.

Kind of annoying knowing I could have just plowed through the clunking and settled it in with violence...
but I had to check on that jbweld bit anyway, so, onward.

Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

Hy-vo in oil or grease is a great way to go. GM FWD autos (among many others), from the TH425 to the TH125, use that style of chain to good effect.

Love this project as a builder of a bike engine car (Honda Z600 with VFR800 power - all shaft drive with an output shaft adapter smile )

2x Volvo PV544 (RIP '63) B20 power!
2007/2012/2013 Driver's Championship (what was I thinking!?) 146 races and counting.
5/24/24

Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

X-args wrote:

Hy-vo in oil or grease is a great way to go. GM FWD autos (among many others), from the TH425 to the TH125, use that style of chain to good effect.
<snip>

For future reference, fell into this website which is right in line with this awesome build thread:

http://www.epi-eng.com/propeller_reduct … issues.htm

Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

RCC wrapped the cage up and it's on the truck, but it won't be here for another two weeks. I think that between those two challenges our fate is sealed and we'll be showing up to Smells Like as spectators this year.

Bummer. Well, come by our paddock spot and say hi. This build has been entertaining enough that I owe you a beer.

BSOD Racing, 1987 Fiat X1/9

58 (edited by largest-farva 2023-09-09 10:53 AM)

Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

https://i.ibb.co/jJ5jykQ/ezgif-2-8853227d00.gif
It drives!

Video courtesy of my neighbor.
When I got back into the driveway he donated a fire extinguisher to the cause.

Ride Report:
  Clutch: Smooth as butter.
  Transmission: All gears accounted for.
  Engine: Rips.
  Brakes: Are shit.
  Alignment: Plows.
  Fuel system: Unsafe at any speed.
  Fires extinguished: 2
 
So there are some things to address.

Then auto-tuning ze lemon.
https://i.imgur.com/8oaBVwd.png

Happy Friday.

59 (edited by largest-farva 2023-09-18 11:01 AM)

Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

Log:

- An exhaust pop shot a bung plug out the unused exhaust hole and I had to go find it.

- The roll cage pipes got delivered. There was a cat sized hole in the box but nothing seems to be missing.

- There is this bent metal coolant pipe with a tee in it from off the fiesta's parts heap. It had/has a rubber cap on one outlet that looks like the foot from a metal folding chair. It blasted even further than the exhaust bung.

- The ecu is clumsily rev-limiting at about 4,700. (because the code I added is slow?)
https://i.imgur.com/MKI2jHH.png

Above 4,700 RPM the ecu can't run instructions fast enough to keep up with the engine and it looses track of the timing wheel ticks for one revolution. That translates to a negative RPM spike, it does its math wrong for a cycle and it misfires. Then the engine is no longer spinning too fast, nature heals itself.

https://i.imgur.com/yqzSNzw.png
Edit for clarity: In this chart, the Sync Loss Count is actually the crank timing wheel tooth # when sync is lost.

- The exhaust is piped all the way to the back of the car, but there is a ground clearance concern.

- The suspension bushings made out of hardhat material make this car feel like driving a big shopping cart. The bushings need to be greased.

Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

There's a max rpm setting in globals.h that can be lowered to make the ECU run more efficiently. I think it's 18k by default? It's used in the trigger wheel calculations. Not sure what you think the max reasonable rpm is but you could try turning that down a bit.

BSOD Racing, 1987 Fiat X1/9

Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

rozap_ wrote:

There's a max rpm setting in globals.h that can be lowered to make the ECU run more efficiently. I think it's 18k by default? It's used in the trigger wheel calculations. Not sure what you think the max reasonable rpm is but you could try turning that down a bit.

Thank you, thank you!. That is exactly the kind of thing I've went spelunking for for, but I missed that setting. I'll give it a try tonight.

I was hoping to find an -O0 or something dumb in the cflags, but no.
I'm not completely sure about my diagnosis either.

Here are some other things it might be:
- primary/secondary phasing
The signal conditioner converts an analog waveform into a digital signal.
https://i.imgur.com/vis0GDs.png
The fundamental frequency of the cam timing pulses is half that of the crank pulses. So while the TDC teeth of both timing wheels pass their respective pickups at the same time, one is generating a lower frequency pulse because it's moving slower. So depending on where it is in the rev range, the falling edge of the cam pulse can be ahead of or behind the falling edge for the crank pulse for the same tooth. Maybe this is happening at 4700? Need to get scope out and look.

- serial i/o Maybe the serial communication (which takes priority over ecu calcs) is too expensive?

- bad logic Speeduino decoders have to re-calculate the "end tooth" for each ignition channel every time the timing advance changes. This isn't common logic in the code, each decoder has to implement it. Maybe my implementation is bad and when the timing gets advanced past ?? degrees it calculates the end tooth wrong.

I'll probably have time to work on it mid-week. It's not worth dialing in the VE table if I can't get it past 5K. All the power is between 5K and 8K.

Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

largest-farva wrote:

- bad logic Speeduino decoders have to re-calculate the "end tooth" for each ignition channel every time the timing advance changes.

This has to be it...
https://i.imgur.com/PqT9Vbd.png
Look, the tooth angle is 15 degrees and as soon as the advance gets to 16 degrees, misfire.

Whew, don't have to optimize anything, just fix a bug.

63 (edited by rozap_ 2023-09-19 09:20 AM)

Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

The picture on your scope shows that you have a rising edge which is almost certainly wrong as those look like VR waveforms. Are you backprobing using the same polarity that it's wired to the ECU? If all of the above is true, afaik that's backwards. VR sensors need to be falling edge, as the voltage swing remains fast across the full RPM range. If it's wrong, you'll get timing drift at high RPMs, which could be your issue. If you have a max9266 conditioner you'll need to swap the VR leads and set rising edge in tunerstudio as it inverts the signal. With the mini dsc you'll need to swap the VR wires and set to falling edge in TS.

So ultimately, there are 4 ways to configure the primary trigger but only one is correct.

For reference, this is what mine looks like (on an ej20 which has a weird tooth profile). The crank and cams are flipped from yours. I just had the misfortune of going around in circles on this for a while, eventually finding that it was a integer overflow bug in the ej decoder code in speeduino.
https://i.imgur.com/RXKp3qn.png

BSOD Racing, 1987 Fiat X1/9

Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

So ultimately, there are 4 ways to configure the primary trigger but only one is correct.

https://i.imgur.com/lldNK1e.png

Yep, that's it.

Re: "old hillclimb car FREE!!!" - '78 Fiesta + Honda V-4 swap

https://i.postimg.cc/T3Xf1Pjc/monk.png
where updates

BSOD Racing, 1987 Fiat X1/9