Slow burn project- T1 Speeduino EFI

[Bruce.m]

I do wonder why nobody uses a voltage regulator circuit on the injector feed to keep it at a nominal value such as 12V. That would keep the flow stable in most situations.

[Piledriver]

Reasonably simple reason...
Linear voltage regulators have a voltage drop, usually .7v-1.2v.
If running on an old school generator or alternator, seeing less than 11v at times is pretty normal. (idle, at night)
...so you would be left with 9 or 10ish volts to work with on the hot side.

If you were only using it for injectors, it might provide a benefit, but even speeduino has voltage compensation for injectors.

Unless you rigged up a ~high current switching power supply to keep everything fed you would still need some sort of voltage compensation, both for injectors and coils (variable dwell for latter, is a thing on several systems, including some oem.

It might be worth it in some applications.
(That's essenially how electric cars charge their 12v utility battery.)

[Wally]

I do wonder why nobody uses a voltage regulator circuit on the injector feed to keep it at a nominal value such as 12V. That would keep the flow stable in most situations.

I noticed on this new to me ECU (Ecumaster, EMU Black) that they have a correction table for that as stock/default build-in. It also compensates injector dead-time iirc if Voltage drops, but still filling out tables atm, so I'am not fully up to speed with this ecu yet.
So, it seems newer ecu's are getting better. My DTA-ecu (in the 1303) is now probably 20 years old already and I did notice some improvements in the newer ecu. Basically they all still do the same though, but this EMU Black unit has for instance the lambda processing inside as opposed to an Innovate unit who does the lambda calculation. With this EMU unit you only use the 4.2 or 4.9 sensor and the ecu provides the signal itself. I think only then can you compensate for a voltage drop. In the DTA ecu (with an MTX or so as a lambda signal) I could always see from the AFR reading when I tuned on the headlights haha! Tricky when you run closed-loop at that moment..

[Bruce.m]

There is a correction curve in tuner studio, to set the dead time at various voltages. Without a bench power supply I found it difficult to get a stable voltage from the battery (engine off) to take measurements. I’ve seen an approach where you run the engine with a steady rpm and therefore steady lambda signal. Maybe 2k revs so I get full alternator voltage, then switch on all electrical loads & adjust the dead time setting to dial out any differences in lambda. Will see if that works.

[Piledriver]

The simple way to fix the issue for weird Lambda readings (reading reference gas perhaps) is use something like a Spartan3... with voltage or CAN output. (Or i2c IIRC, option)
Having said that, I never saw that issue, as all sensors are 5v referenced (regulated) and any good wideband should be too.

Of course I grafted a 100A GM alternator onto my T4 motor that puts out 14.1v at idle>redline 35/50W HID projectors and gas heater on or not.
Modern alternators also use different diodes in the bridges to keep any noise spikes down, keeps ecus happier. (secondary reason I went there)

[Piledriver]

Cheap "reference gas" can---empty metal can with hole in it just big enough to thread the sensor in,

put a couple holes in sides, start a car, hold up one hole to tailpipe for ~30 seconds... tape over inlet/outlet holes.

not perfect, will probably be on rich side, but should give you a stable reading for awhile.

For extra points weld an o2 bung on an old freon can for same purpose. becomes a tool.

[Paul H]

deadtime.png

This is a voltage compensation chart for a Bosch 350ccm
To get the exact deadtime you would hold the engine at load on a dyno and toggle between fire once per cycle and fire twice per cycle
then move all these numbers up and down until there is no change in AFR

[Piledriver]

...all while spinning the voltage knob with your third thumb.
...that high grade switching power supply is sounding better...

The ecu, high-z injectors and maybe the wideband with heater draw maybe 10A? 15 for a v8?
I know a lot of racers run 16v systems, that would be easy to regulate down, but needs a different battery setup or other work arounds.

A common lithium battery for a drill runs at ~18v nominal and you could use it as your damper with the right supply.
keep the "normal" 12v stuff as a power source for the supply, and to run starter etc.

[Paul H]

Voltage knob ?
I use a pot box with the EMU for raising and lowering a value or range and one click to switch between 1 and 2 pulse per cycle
High z injectors tend to draw 1 amp each
Some racers use 16 volt batteries because they don't run alternators

[Paul H]

Siemens Deka 850cc
All injectors have a similar curve so no need to pick out individual voltages
just raise or lower the whole thing.
If you are getting very low voltages sort out your alternator and battery

siemens 850.png

[Bruce.m]

Thanks Paul that’s really useful. The method of switching between 1 & 2 injector pulses I’d not heard of so that is a great tip.

[Bruce.m]

Weather has been poor lately so less trips out and the tune although far from perfect is good enough for spirited driving fun.

I will do some more work on the tune but it’s not really a high priority at the moment and other development work is less likely on this build.

The purpose of this build was to test / prove that a 3D printed manifold could work and enable a relatively cheap EFI build (with a lot more effort being the downside). The throttle bodies were also a success & using the full set up of injectors / fuel rails / harness was also a good shortcut. The linkage however is flawed and could do with a full redesign.

If I was sticking with this engine a cam upgrade would be the next step.

However the next project is a 1.6 type1 with CW crank & C25, (carbs) for another car. That will be followed by a type4 EFI upright cooling build to replace the 1641 in this topic. The 1641 will be kept as a backup motor. I’ll kick off a separate topic for the type4 when I get to it.

[wreck]

Measure the scat cam before you fit it , they are well known to be well out of spec . I have a C55 that has one lobe considerably different to the other three.
learning the tuning for me takes a little time , I've got the VE table close to where I'm happy enough but now am looking at getting the closed loop part working . I am amazed that such a cheap ECU works and uses the same board that I used when playing with fixed wing model planes and copters !

[Bruce.m]

To repeat what I posted elsewhere .

I had interference on the crank signal causing spark & fuel with the engine not running. Root cause was the wideband module. The earth is noisy & needed rewire away from ECU earth.

Currently working on the LCD dashboard screen.
Will update once it’s working & mounted

[Bruce.m]

Working on the dashboard design.

Screen will mount over the original radio slot with the blanking plate removed for wires & a mounting point. Will 3D print a rear shell with mounts.




Runs off a PI zero W2