[rusEfi] Skoda 120L + ITB

[tAB]

Hello All,

First of all a little introduction of me. I'm an electrical engineer, but dealing with cars in my free time as a hobby. I repair everything including engine, suspension, electronic, bodywork and also doing paint jobs as well. In parallel I'm building my old Skoda's. I have a 120L which is in daily use, running ~150km/day. My other long term project is an 1000MB from 1967, but it will be an other story...

The subject of this topic would be the Skoda 120L from 1985. It has several modifications in the engine: bigger bore/stroke, camshaft, bigger valves, rockers with higher lift and finally got a cylinder head from the legendary RS. It's an 8 channel head, instead of the original 5 channel (2 intake, 3 exhaust). The engine is equipped with an ITB setup. The throttle bodies were coming from Yamaha R6. Intake and exhaust manifolds are self/custom made. Currently the engine is running with a stock ECU from Skoda Felicia.

The current setup looks like as follows:
- stock Felicia 136MPi ECU
- ITB's from Yamaha R6
- idle motor + TPS from Felicia
- injectors from Felicia, driven in pairs (custom fuel rail)
- HALL crankshaft sensor
- flywheel from Felicia (like here: https://rusefi.com/forum/viewtopic.php?f=3&t=1133)
- wasted spark ignition, no distributor
- MAP sensor hooked up to a merged line coming from each throttle bodies
- narrow band lambda sensor
- knock sensor

Due to the stock ECU it has it's limitations... Especially in the lower end, but at least it's running...
https://photos.app.goo.gl/2YPWrnz9KpsGNwc39

So, here comes why I'm here
I'd like to build and tune my own ECU. The idea is to build a model specific board with the stock connector, so I could reuse 95% of the original harness.
The base would be Frankenso board with some simplifications:
- less driver outputs
- no CAN bus
- 4 injectors only
- no need of MAX9926 because of HALL sensor(s)
- no thermocouples (or should I leave at least one?)
- only onboard STM32, no Discovery board

Additions:
- WBO control with CJ125
- ...we will see

And here are my first questions:
- Will the ITB's work correctly with MAP or I have to change to MAF?
- Is rusEFI able to handle my idle motor which is not stepper, but DC? (it has own pot also)
- Does it make sense to switch to sequential injection? (that case I would need to install cam sensor as well)
- Will the original flywheel trigger pattern work? Can it be implemented?

I'm sure, more questions will come later

Thanks,
tAB

[AndreyB]

Are there any MAPs on these ITBs? If there are MAPs it would be cool to get them working.

I will get you trigger decoder once your board is closer to reality that's not a problem.

Nice to meet you!

[tAB]

Yes, each TB has a dedicated vacuum port which are connected to a common MAP sensor. I get ~500-520mBar at idle.

[Thommm]

MAP has mixed results on ITB engines because of the very short runners with low volume. Some engines are tunable with MAP and ITBs, some keep struggling. Rusefi does have great options in software for MAP filtering (you can select min-max/avg/at a certain angle), so you have a fighting chance.


Usually though , Alpha-N is used with ITBs. That is based on TPS inputs

[tAB]

Usually though , Alpha-N is used with ITBs. That is based on TPS inputs

Yepp, Alpha-N is an option too, but I want to keep it as worst case scenario.
What about MAF? I'm planning an airbox for ITBs, so could fit MAF sensor later.

[kb1gtt]

I once created a PCB which would takes several MAP signals, and combine them electrically, reporting the lowest of all the signals. This was a potential alternative to the more typical check valves and tiny hoses all connected to one MAP.
https://github.com/jharvey/MultiMAP

If @ is up for it, I think that support for multiple maps is the way to go. AKA tune each cyl individually. Think of it like 4 ECU's in one, but with one crank, and one temperature sensor, etc.

Are you thinking of doing this in baby steps? I think you could get this working with a Frankenso, adapter PCB, with WO2 and other circuits installed in the Frankenso. If you get that working, then you could do a spin of the final circuits stream lining the PCB and using an OEM connector.

Do you know your connector and it's dimensions? Is is possible to make a connector PCB similar to the below?
https://rusefi.com/forum/viewtopic.php?f=4&t=507

[tAB]

I once created a PCB which would takes several MAP signals...

Yes, I found it already when digging around the possibilities.

If @ is up for it, I think that support for multiple maps is the way to go. AKA tune each cyl individually. Think of it like 4 ECU's in one, but with one crank, and one temperature sensor, etc.

Sounds good, I will keep the possibility from HW point of view.

Are you thinking of doing this in baby steps? I think you could get this working with a Frankenso, adapter PCB, with WO2 and other circuits installed in the Frankenso. If you get that working, then you could do a spin of the final circuits stream lining the PCB and using an OEM connector.

Yep, this would be the right way to do. I think I will follow this to save some time also. I'm totally new in Kicad (I'm an Eagle guy more than 15 years), so I'm struggling a bit...

Do you know your connector and it's dimensions? Is is possible to make a connector PCB similar to the below?

Did not checked in details, hope I can find datasheet for it. Otherwise I have to measure the dimensions from OEM ECU. The pinout I have.

[Thommm]

If @ is up for it, I think that support for multiple maps is the way to go. AKA tune each cyl individually. Think of it like 4 ECU's in one, but with one crank, and one temperature sensor, etc.

What problem are we trying to solve here? I'm a big fan for keeping stuff as simple as possible. Multiple tables, 1 per cylinder, results in more tuning complexity and possible engine vibrations due to different outputs per cylinder. A feature like this will also apply to a very limited amount of projects, so does it add to the vision of RusEFI.

Although the multiple tubes with vacuum hoses is not super fancy, it does do the job (if you want to avoid Alpha-N (which I understand :p))

[tAB]

...I think you could get this working with a Frankenso...

Meanwhile I ordered the Frankenso PCB (R0.5.3.1) and started to gather the main parts.

[tAB]

One question is still remains open from the initial post.

Can I use my current idle motor (DC motor)? Just use as a simple idle solenoid and drive it with PWM? Will I need the feedback pot?

[AndreyB]

Meanwhile I ordered the Frankenso PCB (R0.5.3.1) and started to gather the main parts.

Do not see your order, did you order a batch of boards at a fab show? How many boards and most importantly, what color?

[AndreyB]

Can I use my current idle motor (DC motor)? Just use as a simple idle solenoid and drive it with PWM? Will I need the feedback pot?

I simply do not know the answer.

[tAB]

Do not see your order, did you order a batch of boards at a fab show? How many boards and most importantly, what color?

Ordered 5pcs in a fab shop in simple green color.
I can share some of the boards if needed.

[tAB]

Can I use my current idle motor (DC motor)? Just use as a simple idle solenoid and drive it with PWM? Will I need the feedback pot?

I simply do not know the answer.

Noone used similar config till know? I thought this kind of idle motor is more common... Than it will be a nice challenge

[kb1gtt]

I think we have PWM idle solenoids, a DC motor is likely similar, from a software stand point. It might need a semi special PWM driver. Do you know the amps the motor desires?

[tAB]

The current is not known yet, but I guess it's not more than 4-5A. Will check it when I come closer to realization. Maybe an other output FET will be needed. I'm not afraid of HW side, but SW...

[kb1gtt]

You'll likely need a driver of some sort. I recall we configured PWM for a solenoid valve on a truck engine. So I know that can be done. I think the real problem is turning that PWM into a motor signal. Do you expect an analog feedback signal? I would expect that is also something that can be accommodated for. There have been some Electronic Throttle Bodies done, which are commonly a PID loop which reads a POT and generates a PWM.

[tAB]

As mentioned earlier, there's a position feedback pot for the idle motor.

Here we have two options:
- create a small circuit which receives PWM from rusEFI, drive the motor and handle the pot signal (no position feedback to rusEFI)
- drive the motor directly from rusEFI's PWM, feed back the pot signal to rusEFI and handle from SW

[tAB]

Just found, in the OEM ECU the idle motor is driven by TPIC0107
http://pdf.datasheetcatalog.com/datasheets/208/312259_DS.pdf

[kb1gtt]

Looks like this one can do it for $25
https://www.pololu.com/product/1451

The TPIC0107 could do 3A continuous. The above can do 12A, so it's a bit of overkill.

[mck1117]

I'm developing the electronic throttle body stuff on this guy: https://www.pololu.com/product/1212

A bit cheaper, and it's a chip that was designed to be an ETB controller. You're also guaranteed good support from rusEfi

[tAB]

I'm developing the electronic throttle body stuff on this guy: https://www.pololu.com/product/1212

A bit cheaper, and it's a chip that was designed to be an ETB controller. You're also guaranteed good support from rusEfi

It's too late, I've already ordered some VNH5019 to develop my own module
There will be a small PIC receiving the PWM signal from rusEFI, controlling the H-bridge and processing the feedback pot of idle motor. So everything will be done by this module based on the PWM coming from rusEFI.

[mck1117]

That works too! That way rusEfi can pretend it's driving a normal PWM idle solenoid, and not require custom code.

[tAB]

Just arrived

IMG_20190123_181340.jpg

IMG_20190123_181329.jpg

[AndreyB]

Next step is getting BOM parts

[tAB]

Next step is getting BOM parts

Most of them arrived already
...and I'm working on my adapter board to OEM connector. I will integrate the WBO and idle motor driver circuit also on this board.

[tAB]

Just finished with the proto of my adapter board.
It includes the WBO circuitry with CJ125 and the throttle motor driver.

rusEFI_Simos2P_adapter.png

(it was my first project in KiCad, just starting to figure out how it works...)

[kb1gtt]

Very cool. If you share it some where I can offer a review and make suggestions.

[tAB]

I do not want to put in too much effort. This board is just for proofing the concept. If works, I will integrate the whole stuff in one board.
If you are still interested, I can share the files.

[kb1gtt]

Yes still interested. Some comment I could make now,
-- If you aligned the via's with the Frankenso board, you could simplify the connecting wires.
-- Do you have 3D of your connector? If so adding the 3D to KICAD can prevent issues in the real world.
-- Can you print it 100% scale, then check the connector against the paper? This helps prevent issues. It is a bummer when the PCB arrives a month later and the via's do not align as expected.