PrevPrev Go to previous topic
NextNext Go to next topic
Last Post 08 Mar 2015 12:14 PM by  Reanimotion
The Pelican Thread Summary
 16 Replies
Sort:
You are not authorized to post a reply.
Author Messages
Reanimotion
System Designer
Basic Member
Basic Member
Posts:124


--
08 Mar 2015 11:25 AM
    The full thread on the development of the FrankenCIS system can be found at http://forums.pelicanparts.com/911-...encis.html
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 11:46 AM
    Bosch specs for the original frequency valve 0 280 150 300
    242cc @ 4.7 bar (supposedly 190-200 @ 3 bar)
    the Andial injector was around 50% larger flow from memory
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 11:47 AM
    07-18-2014, 12:46 PM
    The plan is to begin with the K-Jetronic basic (no lambda)

    I've updated the MicroSquirt in the test vehicle with the special firmware and an o2 bung is being installed shortly. I'm waiting on the injector metering blocks to be made and have most of the required parts for the following initial steps.

    1. logging and comparing the control pressures against the wideband AFR with the system unmodified.
    2. Introduce the bypass injector to the existing WUR - test and verify range of control and resolution / response

    Once this configuration passes in a normally aspirated vehicle and a brave volunteer with a turbo can verify it works in a boosted environment we can move on to the alternatives.

    next -> Standard WUR and K/lambda frequency valve

    then the D-WURs and complicated combinations.

    I'll be staying with the MicroSquirt throughout this project as I believe it's the best balance of cost and capabilities. It's tiny, well made, waterproof, has automotive grade wiring and connector, and probably above all else, is cheap.
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 11:49 AM
    07-30-2014, 04:57 PM
    Parts are starting to arrive for physical testing

    Fuel Distributors that will be used

    1. 0 438 100 027 cast iron Porsche 928 non lambda

    2. 0 438 100 088 alloy Mercedes V8 with lambda FV capable ports and restrictors
    07-30-2014, 04:57 PM
    3. 0 438 101 018 alloy Mercedes V8 with CIS-E DPR valve control

    The first two are WUR based system units the third is a standalone CIS-E. and I still suspect that #3 the CIS-E will be the preferred option for full system control, and the FV only option #2 may be the closest to the AIC and therefore simplest solution.
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 11:50 AM
    08-07-2014, 11:37 AM
    I'm using an Innovate LC-2 so yes it should be happy running closed loop as it's a supported and recommended unit for the MegaSquirt. Map too.

    The issue I think, creating the lack of success with other projects so far, is every attempt has used bespoke electronics rather than an existing EFI controller.

    In the same way AIC have modified their additional injector controller for a fixed frequency and variable duty for CIS we have just altered the MegaSquirt per revolution variable pulse width to a fixed frequency variable duty, meaning we retain all of the VE/AFR mapping and target features. Translating the CP targets to VE or Duty% means from a tuning standpoint the MegaSquirt almost behaves like normal EFI with some minor differences in concepts depending on the application.

    So in theory it should be fine, the big question is will it work in a real world installation
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 11:51 AM
    08-07-2014, 01:06 PM
    No modelling has been done as the largest unknown variable is the CIS hardware itself and how it responds to pressure changes.
    It is well known the MegaSquirt can handle everything from daily commuters to turbo drag cars, Bonneville racers and superbikes spinning at ridiculous revs so the electronics should be fine.
    We are programmers and engineers who love cars but are not "Tuners" so the MegaSquirt was chosen as better people than us have already done the fuel strategy development. Our test vehicle is a big lumpy, normally aspirated V8 so gross fuel volume can be logged against AFR and control pressure although boost response is obviously out of the question. I'm certain we cannot do any worse than the existing WUR and at least match the AIC hybrids and similar installations but it will certainly require a few brave and patient souls who know what they are doing to experiment on the dyno to prove this works for the forced induction targets.

    What we do gain is the additional control afforded by the MegaSquirt system. Whether or not the WUR based CIS systems allow us to implement that advantage is still unknown.

    from information gathered:-

    The WUR itself is limitation point 1
    - a drop in control pressure and therefore richer mixture is supposedly fast enough
    - increases in pressure are dependant on the flow through the restrictor in the fuel distributor so have a lag according to Thierry25
    - there appears to be a practical limit at the upper ranges of control pressure where CP gets close to System Pressure and therefore ceases to have an effect according to TurboKraft during their tests of the UTCIS

    Lambda/FV is limitation point 2
    - again a restrictor in the FD is in play here separating System pressure from the lower chamber pressure and the range of control is not yet known
    - general information from previous modifiers seems to indicate FV has precise but limited control and WUR had imprecise but large impact on mixture

    The mechanical side of the CIS system is the base map and this add on becomes the fuel trim. Scope and range has not been properly quantified by anyone yet that I can see.

    CIS-E/DPR
    This third generation CIS fuel distributor has been redesigned with single point electronic control only and is the only option that can physically change states quick enough for Bosch to do overrun fuel cuts without air bypasses or other mechanical patches - meaning fast response up and down and accurate mapping is available with this integrated unit.

    Overall the CIS system has fuelling control limitations, any add on will inherit those limitations and have a positive or negative overall impact depending on tune and implantation. Building this system and bolting it on then going for a squirt will more than likely make something go bang without development and tuning. I would certainty implement a safety spark cut trigger on AFR going over or under safe limits as part of the install.
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 11:53 AM
    08-07-2014, 04:04 PM
    Further to the speed of pressure change discussion, and the relationship to available flow and volume in the control pressure circuits, a few factors may be important.

    Pressure lines - on the 928 the WUR is around two feet from the FD so on the V8 all lines are small diameter steel tube, on the 911 the special Cohline nylon flex lines appear to be the norm, as they are a larger volume but shorter in length, line flex should not be an issue.

    The UTCIS originally used a stepper motor which required time to spin between set points, the change in state of an injector is virtually instantaneous.

    The UTCIS also utilised an analogue mechanical pressure sender. I've not had one of the exact units apart but based on similar units with he same cannister sizing there would be a small amount of pressure dampening as the diaphragm filled or emptied in response to changes in pressure. I'm using a digital sender with a dramatically smaller sensing chamber and therefore faster response with an advantage in accuracy.

    How important is any of this? unknown and probably not worth verifying as long as the end result performs well enough.

    As for the control valve or injector. Bosch used a 242cc injector for the Frequency Valve and the Andial kit that bypassed the WUR was 300+ but when Bosch released CIS-E or KE-Jetronic it was controlled by a DPR which I suspect is both high flow and fast response. We have one on the way for testing too
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 11:53 AM
    The frequency valve morphed into CIS-E with the DPR and the control mechanism is basically the same and is therefore worth investigating. If the range is sufficient it could be the key to full control without having to swap fuel systems it also might be possible with some creative plumbing to convert to CIS-E levels of control via the FV. Because there are basically three generations of CIS and GEN1 we know cannot do Gen2/3 control methods there are still a number of options for all intended levels of silliness.

    I think the issue to date has been one of tradition - the WUR was the thing we tuned so the add on solutions played within the same area. Until we get some fuel running through these things we won't know which of the combinations is the best option
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 11:54 AM
    08-19-2014, 09:08 PM

    I returned from interstate today and the prototype blocks were waiting on my desk
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 11:56 AM
    08-23-2014, 07:24 PM
    First validation run during install and sensor calibration
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 11:58 AM
    09-07-2014, 01:21 PM
    ok first run today with the prototype electronic WUR.

    Way too much control pressure swing at current control levels
    2-5% duty cycle change at the injector can swing up to 100kpa or 1 bar

    need to bring it down to 1/10 or 1/100 control steps and retest as the flow through the standard WUR was a guess

    Good bad results if that makes sense as I now have an idea of the happy range of the duty cycle i.e. approximately 60-85% for a 240cc injector
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 12:01 PM
    09-08-2014, 11:26 AM
    A constant flow (non-pulsing) solution is the aim, but along the way we want to verify and allow existing solutions and DIY designs to be used if the results are acceptable.
    So if a small damping chamber smooths out the CP and still retains 1/10 second control or better the injector version should be retained as an entry level DIY option

    The trace below is with 1cc of air introduced to the metering block and it self bleeding over 60 seconds. no other changes
    As you can see the CP is quite stable at the left and varying by 4kpa or 0.04 bar, with a good quick response to two throttle blips.
    At the right as the air bleeds off we lose the damping and begin the hard swings
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 12:04 PM
    09-14-2014, 06:06 PM

    added this little fellow


    so we now have

    FD -> Sensor -> damper -> wur -> return to tank

    and received this nice and happy result



    Ok so for the moment the 6 prototype blocks require some massive reworking to integrate a damper, but if you are happy to re-plumb and use something like the bosch unit Alan has in the thread earlier and you can get the pressure sensor upstream then the block is fine or just a bosch frequency valve would do nicely.
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 12:06 PM
    09-22-2014, 09:36 PM
    ok Gentlemen we have good control and +/- 10kpa flutter at the sensor that's .1bar or 1.4psi




    This is with the factory damper in place and our sensor in line but away from the wur

    so now the plumbing is

    Fuel Dist -> short hard line -> Damper -> short hard line -> Sensor -> long hard line -> FrankenCIS thingy

    the exact same setup initially with the sensor as the inlet to frankencis was +/- 100kpa

    I suspect (and will test later on) that we may not need a damper as long as the sensor is at the FD end.



    Sooooo....

    Looks like we have a winner
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 12:07 PM
    The bottom of Post #1 now contains the following list for those interested
    I will keep it updated as we go, So Always Refer to POST#1 for the current status

    MicroSquirt (tm) http://www.diyautotune.com/catalog/...p-131.html

    FrankenCIS (tm) Metering Block
    Injector - Delphi FJ10409 Fuel Injector (Single) For Dodge Chrysler Eagle 1993-1997
    CP sensor - from Ahead Instruments (eBay store http://stores.ebay.com.au/Ahead-Ins...lectronics

    MAP sensor - Denso 1bar
    IAT sensor - BOSCH 0280130085, PORSCHE 99360611400 SEAT, SKODA, VW 058905379, FAE 33226
    Engine Temp sensor - Bosch 996 unit
    WBo2 - Innovate LC2
    (to suit a particular application the four sensors above can be substituted with anything compatible with MegaSquirt)
    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 12:13 PM
    11-25-2014, 10:30 AM


    Reanimotion
    System Designer
    Basic Member
    Basic Member
    Posts:124


    --
    08 Mar 2015 12:14 PM
    11-25-2014, 06:26 PM


    You are not authorized to post a reply.