DIYEFI.org Forum

Hello,

I am planning on making an implementation of FreeEms for my motorcycle. I want to do this to fix the faults of the OEM ECU. I am looking for some input on how to handle the various inputs and outputs while keeping in mide that this engine revs up to 14000rpm.

Engine: 786cc(800) 90deg V-4 EFI, each cylinder has its own throttle body

What I have for sensors:
Bosch TPS (dual resistive type)
Bosch 3 wire MAP sensor
Denso coil over plug igniters (3 wire type)
2 narrow band Bosch O2 sensors, both are 4 wire heated
4 denso fuel injectors (can not find any information on if they are high-Z or not, I will be measuring the impedence of these when I pull one next)
1 thermo-resistor used as an IAT
1 water temp sensor
1 pn type ambiant temp sensor

inputs of the third kind: bank angle sensor, side stand sensor, EKS all of these sensors are part of the emergency engine kill system

This bike is equiped with ABS and so has 2 ABS sensors and the standard VR speedo sensor.

Outputs:
the tach signal to the guages is a varible frequency square wave of a steady magnitude

V-TECH solnoid, currently either on or off, I want to make this varible via PWM, im looking at a min 3 stage engagement possibly more depending on how the valve reacts to this type of input.

varible intake gate, this is a flap on the intake box that is supposed to keep the intake speed more constant, since my eventual goal is to use forced induction this signal could be used for wastegate control.

ignition, 4 seperate signals, each with its own timing

Fan on/high/off, this is currently just wired for high and off, but the fan is capable of a lower speed.

Injectors: the current ECU uses a bastard version of sequential, I want to go full sequential

I have been looking at several ECU systems on the market, but non of them address the needs of my hardware and my wants as the user. I want to maintain the guage cluster so I will have to figure out the signals and comnication format, currently the guages talk to the ECU for all the information they display (LCD displays). I also want to convert from cabled thottle to fly-by wire and add a simple cruise control ability to the system. The ABS system has its own computer, but does get some information from the ECU. Once I know what it is getting then I will have to provide that to it to keep its functionality. I also want to keep this project as plug and play compatible with the factory harness as possible, this is for a few reasons 1) to allow users to keep their factory ECU as a backup incase the FreeEMS ECU dies for some reason( not that it would, but one can never be too careful) 2)Warranty. Honda will not warrant any bike that has significant modifications, the main failures of this bike are regulator/rectifier and connectors melting. Changing the ECU will cause them to cancel the warranty. 3) Make it so just about anyone can install one of these with a safe tune blown into the memory.

I encourage anyone to help and would someone please forward this on to Fred (as per his request to discuss hardware projects with him) as I have not yet gained messaging rights.

Hello and welcome!

The sensors look pretty normal. Narrow band isn't much use, you'll want to use a wideband, or two, to tune it if/when you get that far.

You'll want to leave the EKS stuff out of the equation at first and add it in later once you've got the basics working satisfactorily. Likewise for the ABS and dash cluster, you'll want to ignore them, for the time being, if you can. The tacho signal should be easy, though.

If the VTEC on the bike is normal small/big cam pin operated vtec, then I don't think you'll want to PWM it. I just searched, it looks like the standard type. What you can do with it, is leave it disengaged, tune the entire map, engage it, tune the entire map, look for the intersection points, and plot the switch around those points for a smooth transfer.

Can you explain what you mean by "bastard sequential" ?
Can you tell us what the fault of the OEM ECU are?
Can you post exact details on the cam/crank angle/position sensors? Here: viewforum.php?f=56

That last thing is key. Depending on the system, 14k may be easy or not available right now. That won't stop you from running it in the mean time while we optimise the firmware, though.

Finally, V4 with true COP + true sequential isn't possible right now without hacking in Sean's XGATE code. You can do that, but at that point, you get zero support from me. I should post a thread on this, it's becoming something that I have to state too often. Depending on the firing order and crank config etc, you may be able to run pseudo COP/wasted spark with sequential. You can definitely run semi sequential fueling, too. It's not as bad as some bigots think :-)

The last thing to say is that it's probably optimistic to expect a better experience from a development system than a finely tuned OEM system. FreeEMS is not mature yet. There is a LOT of change occurring every day in every aspect. So expect multiple revisions of hardware and gradual improvements to software.

Your hardware will also have to be more complex than average to deal with the bike's subsystems.

Re my request to discuss hw projects, I didn't necessarily mean in private. I prefer most stuff to be public - that way other's have the chance to answer and lower my work load in the process.

I hope that helps!

Fred.

A thread on why no support for Sean's XGATE stuff from me, for the time being: viewtopic.php?f=8&t=1700

I just wanted to say Hi and let you know there is a fellow VFR rider here.
I have a 98 (so not a vtec) that I bought new and have enjoyed many long miles over
the years.

I'll be very interested to follow your build and see how it progresses. Currently
I'm not looking to fiddle with the bike but if things go well for you I might give it
a shot.

Also your desire to use Drive By Wire is interesting since that is something
I'm wanting to dig into right now, but not for the motorcycle. I'm interested
in doing precise speed control for our ski boat.

Have you looked into what hardware is available for MC DBW throttle control?

Good luck

-Russell

Fred wrote: The sensors look pretty normal.

Just read about his BOSCH TPS with two outputs.
I know them very well, they're stock on fuel injected BMW motorcycles (R1100 ... R1150).
I like the throttle bodies of this bikes for different reasons for my projects but regularly replace the stock TPS.

The idea behind the two outputs is better resolution at small butterfly openings.
Perhaps this was an issue back in 1993 when the bikes came out, could even be that they had 8 bit ADCs back then.
With a 10 bit ADC I never had any TPS problems resolution wise.

Thus, I wouldn't recommend this sensor. Besides it is ugly large ...

Cheers,
GJ

lumpensack2003 wrote:The idea behind the two outputs is better resolution at small butterfly openings.

Are you absolutely sure about this?
I always assumed / thought that it was a "safety" / verification mechanism.
Thanks,
Huff

In case of the BMW bikes, I'm absolutely sure.
Attached, you find a datasheet of the questioned TPS (in German, unfortunately).
See "Kennlinie 1 und 2" on page 2.

Cheers,
GJ

Ha! Now I understand what you meant. I was thinking what Huff was, Huff and I worked with a VW unit that was for DBW and just had full range duals. Yes, you're right, this looks to be a minor pain in the arse, thanks for your insight. I'd not want to add special code to handle that, just like I don't for the GM dual bias temp setup, also quirky/weird. He could run a patched version, though. It's be straight forward to do. Or just retro fit another type, or use only one type and ignore the other, or something. Silly Germans! :-)