Slow burn project- T1 Speeduino EFI

[frdd]

I use Nylon with Carbon Fibre, for printing car parts. It’s not easy to print on my ender3 (needed a full metal HotEnd).

Esun PA-CF is the stuff I’ve used so far.

I’m considering a printer upgrade to a Bambu X1C or a P1S (with extruder change to hardened steel) but it’s a fair cost. Worth it if you are selling the parts but hard to justify for a hobby

Thank you, I'll look, because it interests me.
I have an ender5+ and an SX2, and I've been into 3d printing for years, so if I can also make parts for my cars, that would be great!

[Bruce.m]

The feature you circled in the photo is not a fuel connection. It’s just a metal stub used by the clamp in the original fitment to support the filter. So you can ignore it

Ok perfect, thank you for all these valuable answers.
One last question and then I'll leave you alone (for the moment)...
Have you changed the diameter of the tubes between the tank and the pump? because the circuit after is 8mm, but before it is smaller?
Frdd

No problem with questions, go for it
I created this thread to share my unconventional approach.

For the tank outlet I’d suggest upgrading to the Mexican beetle 1600i part. It’s fits an 8mm hose. CSP sell a similar part. I used CB performance AN fitting instead, with a right-angle (elbow) hose tail fitting which clears the chassis frame head.

I also removed the tank filter (gauze) because the pump housing includes a mesh filter before the pump inlet.

[Bruce.m]

The elbow I used with the CB tank adapter is below.
You could also use the CB performance T fitting.

[Bruce.m]

Will be make a couple of changes soon so will add updates as I go. The planned T4 engine will not be fitted this year (still working on a T1 build for another car), so looking to make some minor improvements.

Part of that will be some improvements to the tune. I now realise that I misunderstood one of the speeduino settings. In the fuel table screen one of the options at the bottom is to use the ratio of 14.7 : the value in the lambda table as part of the fuel calc. Switching this on means the VE value doesn’t include the adjustment for whatever lambda value is in the lambda table. The default setting does. The downside of the default setting is any change to the target lambda value means a re-tune of the VE table. However with that setting “on”, if you have a good config of injector flow, dead time, etc. you can tweak the lambda table & the actual lambda measurement should follow it fairly closely.

That is the way ECUs such as Haltech work. I thought this wasn’t an option for speeduino & was considering requesting the devs add that as a feature.

[Bruce.m]

Dash wired up.
Mostly working but one bug. The ECU resets when the dash RPI powers up. This is designed in behaviour for the Arduino. Holding the reset pin high stops this so a small mod needed

[panel]

[Bruce.m]

There is a spot on the NO2C pcb for a jumper which didn’t come in the kit. That shorts the reset pin on the mega to 5V via a resistor. I added a small switch via flyleads to enable that. For firmware updates the switch needs flicked but otherwise it can stay on.

[Bruce.m]

Bit more of an update.

Dash Screen:
This has been troublesome. It’s hard wired currently via a USB cable. As mentioned previously, the engine cuts out when Tunerstudio handshakes because the ECU resets. Holding the reset pin high stops that but I’ve had occasions when the laptop or dash would not gets comms signal. More annoyingly I had severe misfire issues while the dash was connected. I had some luck today on a run with no misfire problems. I had disconnected an earth wire at the dash thinking it might be causing a ground loop & crank signal interference. Will see if it remains stable on a few drives.

Linkage:
My home made cable linkage wasn’t as accurate as I’d expected. The outer Bowden conduit compresses at the cable between the master throttle body & the slave side is tensioned. That means the slave TB opens less. I made a temporary fix last year to increase the pulley OD on the master side which somewhat compensated.
This year I removed the cable linkage & fitted a CSP IDF bell crank linkage instead. It wasn’t a simple fit because the TBs open the opposite way. I also needed to extend the throttle spindle to make it work. Took a couple of attempts to get it balanced but it’s now in fairly good shape.

[Bruce.m]

With the new linkage & a couple of logs I could see the usual issues with mapping ITB. I was using MAP as load for the fuel table but the vacuum signal is mostly above 80kpa up to 2k rpm. So the fueling was inaccurate.

Further up the rpm range, the MAP tuning works well so I wasn’t keen to start from scratch using TPS. So I used the second fuel table to create a TPS load map & specified a switch over at 75kpa. That looked promising & I used mega log viewer to generate the initial table. Then I realised the TS can’t autotune a second fuel table for speeduino. That was annoying! Luckily I figured out how to create a custom.ini file based on the Megasquirt TS config so an extra tab appeared in the autotune page. I added a custom filter on both Table 1 & 2 autotune so I could tune them simultaneously.

The tune is now much better although still has some rough spots below 2k on light throttle.

[Bruce.m]

For this year, given I won’t get started on the T4 build for another month or so, I made a small change. I fitted high lift rockers to get a bit more out of the engine (1641cc with stock crank / cam & ported stock heads). I got a set of the empi forged 1.4:1 rockers for a good price. They were fitted with the stock springs & shortened stock pushrods. This will be fine since it’s not going over 4.5k rpm

These are only a small tweak & typically good for 5bhp on a stock engine. With the ported heads I figured I’d get increased flow for part of the extra lift then the port would max out. Best guess, I’d thought maybe 20% more flow across the full valve cycle (20% more flow area “under the lift curve”). Based on the G-meter “dyno” which was 67bhp last year, I hoped to hit 80bhp. I did a run with the same settings as last time and got a bit more. Again given it’s a mostly stock engine with a tiny bit of extra CCs I’m happy with the outcome (until the T4 is ready!)

[Bruce.m]

[Bruce.m]

Been using the current engine through the summer. This was back in May on a cold, damp day.



The dash continues to be unreliable. Someone found a bug in the firmware that can cause the ECU to get stuck if a partial serial command is received. Hopefully the fix for that will make a difference.

I also had a bad oil leak that was traced to a cracked pushrod tube. I forked out for CSP sprung tubes for a fast, quality fix but they aren’t cheap.

[Piledriver]

Better off with a good shielded serial cable and stubby USB adapter vs long usb cable.
Bluetooth or better WiFi comms work surprisingly well, but typically get lower data rates than DB9.
(hardware data correction likely)

You can build and probably buy DB9s with ferrites on every pin in the connector (molded in) ... Not sure they work any better than one big one on each end.

Add a larger clip on ferrite at both ends for a bit of EMI suppression Use resistor plugs and wires, the spiral wrap EMI suppressive ones are good IME..

[Bruce.m]

It’s an area of development for speeduino currently as Dash screens are getting more popular. So expecting a more stable solution will follow.

USB is good as it has dynamic baud rate settings but the 5V rail & shield connected at both ends is a negative. The 5V wire isn’t fully isolated when the mega is powered via the gpio pins and the shield can cause an earth loop (ideally should only be connected at one end).

Serial is okay for short runs but can-bus is better. Wireless is also possible.

[Piledriver]

I pull the power pin out of the DB9 at the MS serial connection, as it is not in the spec and it damaged a BT adapter I had.

A good DB9 cable gives better results vs. an equivalent length USB cable.
(from MS manuals, also have tested this myself b4 going WIFI, used non-resistor plugs/wires as test, had big ferrite on MS3 mainboard cable MS end and MS3X end to supress resets. Never ending comm glitches with 6' USB cable. No issues with quality 10 9 pin RS232 cable and stubby USB adapter, talking to TS on my X300 thinkpad.}
ECU lives under back seat opposite battery. Laptop was on passenger seat.
Cables were what I had laying about.

Note RS232 timing requirements are an ~order of magnitude looser than USB2 even at 115kbaud.