Bosch K-Jetronic and ms2extra (FrankenCIS)

This forum is for the discussion of other projects on Megasquirt/Microsquirt hardware that don't fit into any of the other forums

Moderators: jsmcortina, muythaibxr

stevevp
Helpful MS/Extra'er
Posts: 40
Joined: Fri Nov 11, 2011 3:43 am
Location: Brisbane, Australia
Contact:

Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by stevevp »

Long story short I have a nice little Porsche 928S that is about to be pulled off the road for a refresh. Part of that process is a rebuild of the fuel system.
I've been running a microsquirt in the car for a couple of years now to drive the CANbus instruments and accessories transplanted during the original build.
During the planning of the refresh I thought "Hey, why not EFI?"

Well after some research, I then chickened out as the CIS system is working fine, but the silly side of me then thought "Hey, what about a digital WUR?"
A few weeks of internet research later and it appears the small number of previous attempts by others have either had flaws in control or execution or failed to kick off.

So with a mechanical design in hand and a Microsquirt in the vehicle, the logical conclusion of using the Megasquirt family for fine control to eliminate flaw #1 just seemed to make sense.

Coding has begun and is looking good so far.

PWM 0-100 on Inj 1 for K-Jet
PWM 50-100 on Inj 1 / 49-0 on Inj 2 for KE-Jet
**************************************************************************
SpareADC for WUR CP > 5 Bar
Barometric correction not required as system is naturally MAF in air plate
Injector at constant duty with variable PWM closed-loop referenced to kPA/CP linear table and proportional adjustment referencing req_pw1 from VE/pw calcs in MS

Mad? maybe :)
Useful? probably.

works? definitely!
Last edited by stevevp on Fri Oct 31, 2014 6:22 pm, edited 5 times in total.
Steve
elaw
Super MS/Extra'er
Posts: 2926
Joined: Fri Oct 16, 2009 6:20 am
Location: Wilmington, MA

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by elaw »

Sir... I bow to your greatness! :lol:

I know a lot of people hate it, but I've always had a lot of respect for the CIS system. Among other things, it's the only automotive fuel injection system I'm aware of that was able to be adapted for use in a certified aircraft - the Mooney PFM (which used a modified Porsche engine).

Question: are you using "old school" CIS, or the newer KE-type that has the DPR on the side of the fuel distributor? That would be the best setup to use, as it reacts to control current changes very quickly and has a lot of authority (enough to implement overrun fuel cut!). The downside is in some cases you need a special driver since the current in the DPR coil can flow both ways. I know that's the case with the Mercedes systems, not sure about others.

Edit: I just reread your post and realized you covered the K- vs. KE- difference. Impressive!
Eric Law
1990 Audi 80 quattro with AAN turbo engine: happily running on MS3+MS3X
2012 Audi A4 quattro, desperately in need of tweaking

Be alert! America needs more lerts.
stevevp
Helpful MS/Extra'er
Posts: 40
Joined: Fri Nov 11, 2011 3:43 am
Location: Brisbane, Australia
Contact:

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by stevevp »

Thanks, for what it is CIS is quite a brilliant system and it seemed a shame to throw it away.

Ok, base coding for K-Jetronic is done and running in closed loop with simulated inputs.
Control Pressure targeting is rock solid and fast to respond so I think its ready for system tests
INJ1() and INJ2() have been modified to switch between normal operation and PWM/Duty modes depending on the injType setting in the tune
page4_data.InjType, 0 = port injection / throttle body
1 = K-Jetronic/CIS
2 = KE-Jetronic/DPR

additional code has been added in parallel for PWM/Duty calculations based on the CP map in the first post above and some initial settings on the injector page

I'll be running an in house custom made metering block with ~120hz pulse width modulation of a single fuel injector for pressure control
This operates in closed loop with a 100psi digital pressure sensor and a target pressure varied by the MS final VE target.

As K-Jet handles the mass air changes the VE tables can start flat at 100 and then essentially be operated as an interceptor by adjusting +/- the 100 base wherever needed.
As the Bosch WUR (Warm Up Regulator, and an understatement on what it really does) references engine temp and manifold vac., The Warm up enrichment in MS could be replaced by or mixed with the control pressure base map, and the load axis of the MS VE table will replace the manifold vac function of the WUR. We also now gain the RPM axis and all the other MS tuning options.
Accel enrichment etc. from MS is definitely now valid but must be considered in conjunction with the K-Jet feature of acceleration enrichment at the fuel distributor due to the air metering plate over-swing.
Decel-Cut and Fuel-Cut on limiter has been coded for but is really a KE-Jet only facility
Last edited by stevevp on Thu Sep 18, 2014 1:33 am, edited 2 times in total.
Steve
piledriver
Super MS/Extra'er
Posts: 1679
Joined: Tue Oct 27, 2009 6:24 am
Location: Van Alstyne, Texas

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by piledriver »

I always wanted a land shark...

Interesting, I was eyeing my 86 Cabby (CIS+Lambda) with evil intent just the other day, holding up a CIS-E metering unit that I had previously run in my 914, simply using a remote pot and some manual DPR adjustment to set it up. (ran awesome)
(I need to make a longer cold start hose to swap at this point. That's it)

I have been threatening to use a MS to control CIS for ages, very interested if you have developed CIS-E/M DPR* code, willing to test, have a spare MS2E setup for full sequential. I figure it would have to go through a low pass filter and possibly driver.(120mA max)

(*The CIS-E and -M DPRs I tested appeared identical save for the -Ms are black and the mechanical zero screw is adjusted differently)
Always doing things the hard way, MS2 sequential w/ v1.01 mainboard, LS2 coils. 80 mile/day commuter status.
stevevp
Helpful MS/Extra'er
Posts: 40
Joined: Fri Nov 11, 2011 3:43 am
Location: Brisbane, Australia
Contact:

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by stevevp »

I don't have a CIS-E system or DPR here at present so I'll be doing the CIS-E code last - volunteer test subjects most welcome :)
From the specs I can find the DPR is 19ohms, and probably driven PWM with some form of series load in the bosch ecu, as without it duty at load would have to be around 5% or less to get down to the specified current.
DPR 19 ohms pwm (at 100hz?) with an unconfirmed range of 0-140ma with normal operation between 10-35
roughly rounded, 19ohms at 12v gives 600ma and at 13.8 gives 720ma so starting duty to get 40ma would be 40/720=0.055556 or around 5.5% duty
an additional 160ohms in series should bring it to 33.5ma at 50% duty so the factory DPR operation range would need to be established to choose a suitable series load.
The reverse polarity only seems to be for fuel cut so I'm guessing the planned INJ2 output and the H-Bridge setup could be ignored completely if deceleration/rev-limit fuel management was not a priority
Steve
stevevp
Helpful MS/Extra'er
Posts: 40
Joined: Fri Nov 11, 2011 3:43 am
Location: Brisbane, Australia
Contact:

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by stevevp »

Actually - thinking about it further - I think the current limiting might be possible in code without too much fluffing round via dual frequencies on the Injection driver similar to the Peak/Hold setup.
depending on the driver something like this should be possible.
Image
Steve
piledriver
Super MS/Extra'er
Posts: 1679
Joined: Tue Oct 27, 2009 6:24 am
Location: Van Alstyne, Texas

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by piledriver »

One could implement a dedicated fuel cut via reversing the polarity of the signal via a dpdt relay/diodes via a spare programmable output.
(On MS3 the INJ1/2 channels are programmable outputs, and do PWM well, not sure on MS2E)

Given the low actual current requirements, might be able to center tap it for +/- polarity with minimal circuitry.

E spec IIRC is +120mA/-80mA, CIS-M is +/- 10ma with the center point essentially at stoich, but AFAICT mechanically and electrically the -E and -M DPRs are identical except for the zero screw adjustment.
Always doing things the hard way, MS2 sequential w/ v1.01 mainboard, LS2 coils. 80 mile/day commuter status.
stevevp
Helpful MS/Extra'er
Posts: 40
Joined: Fri Nov 11, 2011 3:43 am
Location: Brisbane, Australia
Contact:

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by stevevp »

OK,
Coded for closed loop PWM control with a VE to kPA target pressure at 0.1 kPa resolution and it seems to be running well on the bench.
i'm still hunting through the MS2EXTRA mixture code and getting lost a bit with all the tmp1/tmp2 variables :) have I missed any obvious modifiers? or compounded any?
Ive got Gamma / Fuel (Flex) / EGO and I think I'm still looking for a way to nicely do % accel enrichment although K/KE-Jet has a built in compensation via the MAF plate overswing on rapid airflow changes so it may not be necessary.
Last edited by stevevp on Thu Sep 18, 2014 1:33 am, edited 2 times in total.
Steve
stevevp
Helpful MS/Extra'er
Posts: 40
Joined: Fri Nov 11, 2011 3:43 am
Location: Brisbane, Australia
Contact:

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by stevevp »

K-Jet won't need accel enrichment but KE-Jet will so accell will have to be integrated as a % rather than the current additional time in ms
Steve
piledriver
Super MS/Extra'er
Posts: 1679
Joined: Tue Oct 27, 2009 6:24 am
Location: Van Alstyne, Texas

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by piledriver »

I had no noticable issueswith ton my -E "minus" (mechanical only) setup, the metering arm counterweight /overshoot setup is AFAICT same as +Lambda and basic meters, that provides the "accel shot".

It also bolts on, only difference is the cold start bolt/fitting is the next size up with the 14mm head, and the hose needs to be ~6" longer for my app.

My long term goal is to graft in a 2L TFSI controlled by MS2E or MS3, but I could probably at least get the current motor swicthed to the CIS-E meter as a mechanical only (with baseline electrical zero control box if desired: it and harness exist) setup in an hour.

I have a G60 valve cover on it, and have the matching G60 fuel rail, but I'll give CIS a chance again if only to poke at its limits.
Always doing things the hard way, MS2 sequential w/ v1.01 mainboard, LS2 coils. 80 mile/day commuter status.
stevevp
Helpful MS/Extra'er
Posts: 40
Joined: Fri Nov 11, 2011 3:43 am
Location: Brisbane, Australia
Contact:

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by stevevp »

Thanks, I thought it should overshoot too, although the Bosch tech docs state the ecu adds fuel on accell.

My thoughts are, both systems overshoot but.
K-Jetronic also has a drop in control pressure as vacuum drops lowering the force against the piston and therefore the plate and allowing a larger movement
KE-Jetronic is constant pressure at the piston so the overswing would not change for load or temperature, the ecu then compensates.
so with that in mind even if we don't need it, It would be sensible to bring it in to the calculation, just set accel enrichment to 0 as a baseline and tweak if needed.

What I want to avoid is losing features from either CIS or MegaSquirt, unless one replaces the other or the concept is not valid e.g. MS Cranking settings for the K-Jet
Last edited by stevevp on Sat Jul 12, 2014 11:37 pm, edited 1 time in total.
Steve
piledriver
Super MS/Extra'er
Posts: 1679
Joined: Tue Oct 27, 2009 6:24 am
Location: Van Alstyne, Texas

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by piledriver »

Only a few models with K-Jet (Some Mercedes, maybe Delorean, possibly Porsche) had the WURs with any vacuum ports/bellows etc, so the overshoot was usually all you got on basic or +Lambda. (at least on VWs) Admittedly it DID vary from the varialble control pressure from the WUR, which makes the flap easier to move when low/cold, resulting inmore overshoot/richer accel when cold.

All K-Jets are true MAF based systems, so the temp and altitude compensation is ~inherent to a great extent.
Altitude compensation I could see, as the O2 % varies slightly, could help open loop mode.
The WUR ran at a fixed pressure once warmed up, so air density and flow were the ~only variables once the WUR was fully heated.
(This frerquently caused warmed start headaches)

...not arguing they didn't also program some accel shot in on -E/M and flavors, as they actually could with -E/M (and had flap position sensing pots), just saying it still seemed to work just peachy even sans computer in the app I used it on, with very minor tweaking.

Also, the air cones on -E/M I have seen were all linear, so it ran at a quite stable AFR without programmed load enrichment.
I considered adding that with a Hobbs switch (actually an SMC vacuum/pressure sensor and a relay, but I digress)
(also warm up etc, needed some help, but only for ~30-45 seconds on an aircooled motor, even well below freezing, tapping the throttle or even a manual colt start pulse or three was all it ever needed a 15F, although it probably could have used more air than the std thermal warm up air valve supplied)

Note the flap position pots are a common failure item on KE/KM, and they are very custom/oddball bits, so the only fix with the original computer was a new air meter.
Moe than one car landed in the junkyard as a result of the following sticker shock from what should have been a <$50 potentiometer...
(Mercedes sold it, theoretically replaceable, at least, but are NLA for many years. It was at least primarily used in leiu of a TPS for accel/aiflow rate change ~just as the TPS is used on SD setups... a replacement part would be handy. .. An MS2E/MS3 setup would be a cost effective replacement :lol: )
Always doing things the hard way, MS2 sequential w/ v1.01 mainboard, LS2 coils. 80 mile/day commuter status.
stevevp
Helpful MS/Extra'er
Posts: 40
Joined: Fri Nov 11, 2011 3:43 am
Location: Brisbane, Australia
Contact:

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by stevevp »

Coding done for the MS2 and Microsquirt sans accel enrichment at the moment.
DPR +- across INJ1&2 seems to be working fine as well. (no dual frequency current control yet i.e. pulsed highs within the main duty cycle)
I've done the DPR code as a pure final VE from MS to PWM steps in the same 0.1 resolution, so
- 100% VE equates to no drive on either channel,
- over a hundred increases drive on INJ1 leaving INJ2 at 0,
- less than 100 increases drive on INJ2 with INJ1 at 0
This will then need a H-Bridge driver or patching to the stepper idle chip on MS2 if it can handle it
K-Jetronic stays as closed loop control pressure targeting with VE to kPa translation

So in theory we have K & KE Jetronic capability with MegaSquirt.
It will be a couple of weeks before the prototype metering block can be run through CNC and we can run on to vehicle testing.

MS3 might be overkill in this application but the process should translate.
Steve
piledriver
Super MS/Extra'er
Posts: 1679
Joined: Tue Oct 27, 2009 6:24 am
Location: Van Alstyne, Texas

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by piledriver »

stevevp wrote:Coding done for the MS2 and Microsquirt sans accel enrichment at the moment.
DPR +- across INJ1&2 seems to be working fine as well. (no dual frequency current control yet i.e. pulsed highs within the main duty cycle)
I've done the DPR code as a pure final VE from MS to PWM steps in the same 0.1 resolution, so
- 100% VE equates to no drive on either channel,
- over a hundred increases drive on INJ1 leaving INJ2 at 0,
- less than 100 increases drive on INJ2 with INJ1 at 0
This will then need a H-Bridge driver or patching to the stepper idle chip on MS2 if it can handle it
K-Jetronic stays as closed loop control pressure targeting with VE to kPa translation

So in theory we have K & KE Jetronic capability with MegaSquirt.
It will be a couple of weeks before the prototype metering block can be run through CNC and we can run on to vehicle testing.

MS3 might be overkill in this application but the process should translate.
Groovy! :yeah!:
IIRC mainboard Inj1 & inj2 are each rated at 13A, so I suspect that will be sufficient.

It might actually get away with simple current limiting resistors, and/or 5v external supply.

I suppose it doesn't matter if you are feeding the DPR 12v in both sides via a pair of limit resistors.
What is optimal TBD, it may work better if the fixed injector is ON 100%.
(I can see it working either way, will build to suit)

If I wasn't up to my knees in another project I'd swap KE on today.
Always doing things the hard way, MS2 sequential w/ v1.01 mainboard, LS2 coils. 80 mile/day commuter status.
stevevp
Helpful MS/Extra'er
Posts: 40
Joined: Fri Nov 11, 2011 3:43 am
Location: Brisbane, Australia
Contact:

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by stevevp »

apparently functional but definitely "light fuse and stand well clear" version about to be real world tested
Last edited by stevevp on Wed Sep 17, 2014 7:37 pm, edited 1 time in total.
Steve
piledriver
Super MS/Extra'er
Posts: 1679
Joined: Tue Oct 27, 2009 6:24 am
Location: Van Alstyne, Texas

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by piledriver »

12v>pair of 100 ohm power resistors (or a pair of 200s) feeding the INJ1/inj2 legs with the DPR in between will work.

I used 100ohms as the current limiter with another 100 ohm pot for my manual control IIRC.
Always doing things the hard way, MS2 sequential w/ v1.01 mainboard, LS2 coils. 80 mile/day commuter status.
stevevp
Helpful MS/Extra'er
Posts: 40
Joined: Fri Nov 11, 2011 3:43 am
Location: Brisbane, Australia
Contact:

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by stevevp »

project updated

should now form a decent basis for
K-Jetronic with Digital WUR on INJ1
K-Jetronic - Lambda with dual maps #1 for Digital WUR on INJ1 #2 for Frequency Valve on INJ2
KE-Jetronic - control of DPR (INJ1 for forward/rich and INJ2 for reverse/lean via H-Bridge)
Last edited by stevevp on Wed Sep 17, 2014 7:37 pm, edited 1 time in total.
Steve
DKmike
MS/Extra Newbie
Posts: 1
Joined: Sun Aug 31, 2014 9:29 pm

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by DKmike »

:yeah!:

Hi guys Im new to forum stuff but I love what this one is all about.

Looking forward to getting this FrankenCIS going in a boosted Benz :D
piledriver
Super MS/Extra'er
Posts: 1679
Joined: Tue Oct 27, 2009 6:24 am
Location: Van Alstyne, Texas

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by piledriver »

I still want to try this but my spare MS2 daughterboard is still doing duty in in my MS3/MS3X box.
Always doing things the hard way, MS2 sequential w/ v1.01 mainboard, LS2 coils. 80 mile/day commuter status.
jsmcortina
Site Admin
Posts: 39585
Joined: Mon May 03, 2004 1:34 am
Location: Birmingham, UK
Contact:

Re: Bosch K-Jetronic and ms2extra (FrankenCIS)

Post by jsmcortina »

Steve, could you clarify what this has become?

Your website says "Megasquirt based" but then goes on to say that it runs on a Power PC. So which is it?

If it uses Megasquirt hardware, you can link to your site from here, but not otherwise.

James
I can repair or upgrade Megasquirts in UK. http://www.jamesmurrayengineering.co.uk

My Success story: http://www.msextra.com/forums/viewtopic ... 04&t=34277
MSEXTRA documentation at: http://www.msextra.com/doc/index.html
New users, please read the "Forum Help Page".
Post Reply