[Success Story] Sporttrike 1.8l Toyota 1ZZ-FE [microRusEFI]

[Sporttrike]

Hello there!

We're a team of 3 (two studying mechanical engineering, one electric engineering) that is building a reverse-Trike with a 1.8l Toyota 1ZZ-FE engine taken out of a Toyota Avensis. The Trike is built for racing and accelerating fast, thus the name. We have suspension that is adjustable via hydraulics (using the power steering pump) that also adjusts our ride height. We're already at a state where we can drive a bit by making use of the starter motor and now we need to make our engine run.




We have tried using an Arduino Nano before and with a program we wrote ourself we did get the ignition to work but we're not 100% certain up to what speeds that would work reliably, if at all. We then kind of resorted to using the original ECU from a Celica T23, we did buy a cable hardness and the ECU for it but buying the necessary parts to make the actual ECU run would be pretty expensive as well as still restricing us in flexibility.
RusEFI looks perfect though! Our plan is to use the microRusEfi kit, 3D print our own enclosure and connect it to the wiring harness we have now.

Some more info on the engine: The crankshaft and camshaft are both using VR sensors, the throttle valve is mechanical. Our electrical connector to the VVT-i valves is gone. We have a MAF sensor. We do not have an o2 sensor at all. The only information we have on the knock sensor is that it has two pins, no idea on further specifics though.

We have a few questions:
Any idea what kinda of knock sensor that is and will we use it with the microRusEfi board?
Will we be able to use the MAF sensor or do we need something else?
How much cooling for the board is required?
Is it possible (and how much effort is it) to make VVT-i work? Would VVT-i even be worth replacing the electrical connector (30€)?
What kind of o2 sensor should we use?

Any other suggestions where to start, what problems we can expect to come across etc would be very welcome as well

[mck1117]

That's an awesome build!

Any idea what kinda of knock sensor that is and will we use it with the microRusEfi board?

microRusEfi doesn't currently support any knock sensing.

Will we be able to use the MAF sensor or do we need something else?

This is an area that we're are currently improving in the firmware, though it does "work" today. If you have the transfer function (voltage vs. flow) for your MAF, it should work ok. Without that, it's a bit harder to get bootstrapped, but likely still possible with some patience.

How much cooling for the board is required?

It probably needs more than is offered by a plastic enclosure. I wouldn't recommend using a 3d printed plastic enclosure. The aluminum enclosure we include does a pretty good job of heatsinking, as the PCB bolts directly to the case.

Is it possible (and how much effort is it) to make VVT-i work? Would VVT-i even be worth replacing the electrical connector (30€)?

It is certainly possible, yes. We currently only support one or two VVT trigger patterns, but more can be added as there is demand for them.

What kind of o2 sensor should we use?

A few of us have had good luck with the stuff sold by AEM. Their basic wideband sensor/gauge combos (like the AEM 30-4110) work totally fine, and their newer X-Series stuff is even better (but a bit more expensive). They let you interface to the ECU via CAN, instead of analog voltage, though rusEfi supports both methods.

[AndreyB]

Hello and welcome! Kudos for doing something really unusual! Are you going to spin that tire?!

The crankshaft and camshaft are both using VR sensors

Are you sure both are VR? Bad news microRusEFI has only one VR input since most vehicle have Hall cam sensor.

[Sporttrike]

If you have the transfer function (voltage vs. flow) for your MAF, it should work ok

We luckily do

It probably needs more than is offered by a plastic enclosure. I wouldn't recommend using a 3d printed plastic enclosure. The aluminum enclosure we include does a pretty good job of heatsinking, as the PCB bolts directly to the case.

Ok, then we'll simply use the aluminium case. It's not that costly anyways.

It is certainly possible, yes. We currently only support one or two VVT trigger patterns, but more can be added as there is demand for them.

A few of us have had good luck with the stuff sold by AEM. Their basic wideband sensor/gauge combos (like the AEM 30-4110) work totally fine, and their newer X-Series stuff is even better (but a bit more expensive). They let you interface to the ECU via CAN, instead of analog voltage, though rusEfi supports both methods.

We're just gonna do without VVT-i (for now) and have a look at what sensors AEM provides.

Are you sure both are VR? Bad news microRusEFI has only one VR input since most vehicle have Hall cam sensor.

Pretty sure, yes. The sensors have two pins and we measured the spikes in voltage (the following image is from our very crude oszilloscope app we built with an arduino, we used a diode to 5V for protection from the voltage spikes):

Could we make it work with a circuit that simulates a hall sensor when it detects the spike from the sensor?

Thanks for your help!

[AndreyB]

Could we make it work with a circuit that simulates a hall sensor when it detects the spike from the sensor?

50 years ago they put people on the Moon and returned them alive. You can totally translate VR into Hall with enough motivation.

[Sporttrike]

50 years ago they put people on the Moon and returned them alive. You can totally translate VR into Hall with enough motivation.

What I mean more specifically is if the software uses some metric like the length of the pulses. Of the only thing that matters is the timing of the signals then I'm pretty sure we'll be able to make it work.

What stops the port from working with a VR sensor exactly though? Is it simply a digital input or would it only need some electrical components to process the signal beforehand?

[kb1gtt]

About VR to hall, we have this board,
https://rusefi.com/forum/viewtopic.php?f=4&t=994

This might also be an option.
http://jbperf.com/dual_VR/v2_1.html

At cranking RPM's what is the length of time between the CAM pulses? Is that length of time above 85mS?

[Sporttrike]

Those boards look like a good solution, thanks!
At cranking speeds (iirc around 200rpm) the time between two pulses is about 160ms.

[kb1gtt]

Those boards use a MAX992# chip. The MAX chip has several "modes" in which it can be configured. Those boards are designed with a mode that includes a self tuning feature. Basically if you don't get a pulse in 85mS, it jacks an internal gain way up and causes noise issues. If it gets a pulse in less than 85mS it notices the high and low of the signal and self adjusts the gain properly. The MAX992# chips can be configured in a different "mode" but these PCB's are not done that way, and there is a lack of software support in other modes.

It sounds like this would be an issue for your CAM.

I understand your thinking MRE. Perhaps we can swap the CAM and CRANK signals, your CRANK has many pulses so even during cranking you'll have below the 85ms window of time. I think the MRE VR does not have this kind of issue. So if you wire your real world CAM to the MRE CRANK, then if you use one of these PCB's to make your CAM look like a hall, then send that signal to the MRE CAM, that might work.

I think that will work, but there might be software issues, or I might be mistaken about the MRE VR.

[AndreyB]

Yes ^^^ should work.

[kb1gtt]

Great.

So as a recap, the real world CAM VR is wired to MRE CRANK input. Then real world Crank VR is wired to MAX992# board. Then this board is wired to MRE CAM. Then you do happy dance, and unleash the beast contained inside your metal cage.

[Sporttrike]

Soo... the kit is on the way, and we'll try to make the VR->Hall thing work.

There is another thing we would like to try, and that is not using a normal exhaust - instead we'd like to pass the exhaust gases through short pipes out the side. For that we will need to use 4 separate (cheap, we don't really have the money to spend on even one $150 wide band sensor) O2 sensors. Is that something we can make rusEFI work with, without too much effort?
One advantage would be that we'd be able to control the air/fuel ratio for every cylinder on its own. It would also save us some weight and look cool, too.

Thanks for all the help so far!

[AndreyB]

junk yards are full of $9 used sensors but rusEFI does not have trivial cheap controller yet

[kb1gtt]

WO2 is not required for operation. That's mostly useful for auto-tuning.

[Sporttrike]

Ok, so we now have a new super sick looking exhaust,

a7f50556-cc36-4bb2-af2b-e0e76abefd6a.jpg

20ab7470-552f-4b55-98e3-aad7ff36d03b.jpg

and the cheapest O2 Sensor possible (8€ new). Our options now are: 1. Buying four more sensors and then probably read them out with a seperate little board and calculate an average to send into the RusEFI. or 2. We are already planning to make a removable manifold with a muffler, that can be "plugged" into our current "exhaust" to quiet it down for tuning and to combine all the exhaust gases to only use one O2 sensor. To decide about this we would have to know a few things: In case of option 1: Is it in any way possible to get around an extra board to deal with the four sensors, by making RusEFI deal with averaging them out? (This is however not that important, as a board like that shouldn´t be very hard to make). In case of option 2: Is it a good (or at least not engine destorying) idea, to put a single sensor onto the removable "quiet mode" exhaust and only use it for tuning and then run without the O2 sensor when on the track, where we would only use our "extreme straight pipe" setup. The biggest problem with that would probably be, that we would not only remove the Sensor (which should theoretically not be a problem, if the tune is right, right?), but would also completely change the exhaust setup (which we could not possibly account for in the tune, because we can only properly tune it, using the sensor)

Hope this is all somewhat understandable...

We´d be happy to hear some of your thoughts on this...

[AndreyB]

I have a feeling that you've got narrowband sensors and you are probably missing the fact that rusEFI does not use narrow band signal?

More information at https://github.com/rusefi/rusefi/wiki/do_i_need_wideband_oxygen_sensor

[Simon@FutureProof]

You will not get reliable Lambda sensor AFR readings from zoomie headers, there is often too much reversion along the tube that pulls 02 into the tube and causes an incorrect AFR reading.

Regarding the cam sensor issue, the toyota 1ZZ-FE uses a fairly common sensor configuration, I expect a hall effect sensor is available that will go directly into the existing hole and save the need for messing around with a MAX9926 chip.

[Sporttrike]

Thanks for the quick feedback.

So what I have now gathered from your responses is:

We absolutely need a wideband O2 sensor for tuning, and there is absolutely no way to use the narrowband sensor, we already have, right?

Also our headers are simply too short for O2 sensors at all, which makes total sense and I had already feared something like that. Installing sensors at the end of the Tubes would make them read the oxygen from the surrounding air, and installing them as close to the head as possible, would maybe fix that, but would probably burn up the sensors.

As I said we were already planning to lengthen the exhaust system, with a removable manifold and muffler, that can simply be slid into the already installed headers, where we could then install a single O2 Sensor. So we could essentially have a completely "normal" exhaust system, for tuning and to save our ears, but could also seperate it, to save some weight and have a cooler look. In this configuration, we would then of course have no O2 Sensor at all. This however should still work, if RusEFI can work in an open loop mode.

As far as I know, normal car ECUs only run in closed loop mode, during idle and low-medium load, and switch into open loop mode when you step on it, where it ignores the O2 sensor. Please tell me if I´m absolutely wrong on this. But I would assume that RusEFI also works this way, and that it should be possible to run completely without an O2 sensor, once tuning is done? Then the only question remaining would be, what has to be done, when switching between exhaust setups (with/without O2 sensor), and if it´s possible to do that without reprogramming every time?

Oh, and regarding the cam issue, our Electronics Engineer thinks he can deal with that relatively easily, so we will try using our current sensor first, but thanks for the tip , it´s always good to have options...

[AndreyB]

We absolutely need a wideband O2 sensor for tuning, and there is absolutely no way to use the narrowband sensor, we already have, right?

My native language is not English, I assume your native language is not English as well. Please read https://github.com/rusefi/rusefi/wiki/do_i_need_wideband_oxygen_sensor one my time, I have added more text. I hope that this would reduce the chance of confusion.

[Sporttrike]

Ok, just double checking before buying expensive sensors .

[Simon@FutureProof]

You can use the wideband to tune and then remove if you want to, watch out though as you may change the engine tune by changing the exhaust tuning.
You are unlikely to hurt an LSU4.9 sensor with it installed very close to the cylinder. I have had the pipe around the sensor glowing redhot before without long term damage to the sensor.

When swapping between exhaust setups the only thing you will need to do is to make sure it is absolutely sealed.
We have had good results with this on a turbo car and on experimental test rig setups.
https://granvilleoil.com/prodInfo?pID=45

No it is not rated to exhaust temperatures but it's still been present in the flange of the turbo car after a redhot dyno session. (literally red hot too).
Only thing to be careful of is to run the engine and cook the silicone before installing the WBO2 sensor to avoid any risk of contamination.

[Sporttrike]

We found a relatively cheap Bosch LSU 4.9 Sensor, which seems to be the best, most commen WBO2 Sensor. But the expensive part is the controller, as some of you already said...
Any suggestions for a cheap controller, that still works well?

[Simon@FutureProof]

Experience has show the best cost vs performance one you can just open the box and plug right in is the AEM X-type. About $150.

SLC also do some cheap stuff but I am not sure the extra time is worth the lower cost.

[Sporttrike]

Hey, so we've now mostly assembled the microrusefi board; could you tell us what way do the VNLD5090 chips need to go on? There doesn't appear to be any mark on it (except for the text)

[AndreyB]

https://rusefi.com/docs/ibom/micro_rusEFI_0.5.0.html

VNLD5090 sides are not the same - one edge is more sloped, that's how ST marks some parts. Datasheet has more info on that.

there is no software to run VNLD5090 on MRE 0.5.0 so at this point there is no point to mount those for you unless you plan to fix the rusEFI TLE8888 driver.

[Sporttrike]

Thanks!

[AndreyB]

Photo added to https://github.com/rusefi/hw_microRusEfi/tree/master/gerbers to show orientation anyway

[Sporttrike]

We got another question: We want to route the USB cable out the main connector so that we can leave the case closed.
According the pinout https://github.com/rusefi/rusefi/wiki/Hardware_microRusEfi_wiring we can connect the 5v lane of the USB cable to pin 19. Now that USB connection doesn't work and there's no 5V on that pin.
What can we change for this to work?

[AndreyB]

Code: Select all

Starting from version 0.5.0 this pin can be used as 5V input for USB connection through main connector (No need to open case to access micro-USB connector)

Just do it if you have MRE 0.5.0

[Sporttrike]

Thanks again. We found a problem with the cable. Now it works