Back to MembersPage/NanassyPeter
MembersPage/NanassyPeter/CoolantSensor
MT button for WBO2 proposal
mde02 works for (freeair?) calibration, but inconvenient with an untuned engine that stalls often => The WBO2 keeps running even if the engine stalls, if you don't cut ECU power.
Currently the Manmde02 can be issued from the MegaTune internal terminal, but indeed a pain (and don't forget to say "bye" in the end). So:
MT needs an "mde02" button to force the WBO2 ALWAYS ON (even if RPM=0). Good idea. Helps cold start tuning.
- config13=06 # CONTROL_STRATEGY bit2=1 will disable MAP multiplier; bit1=1 is WBO2 enable
- mmttg40
- mmttc06
URGENT : make this bit adjustable from MegaTune
Tables were crashed in the KARKUS car,so i had to reupload the 1045 firmware,which worked flawlessly.Unfortunately in a really hefty run we lost out trigger wheel.I was a real flash when sparks came fom under the car.We were not able to find the wheel,so another unsuccesfull raceday.
After 3 races with 5 coils and with the proper trigger signal the 5 cyl High Output engine can be revved to 7500-7600 instead of the previous 6000 with dizzy.Nice numbers,but we have an URGENT need for even more dwell under boosted situations,and at higher revs.
My proposal is a total rework of dwell calculation,supporting overlapping dwell,with the added feature of BOOST dependent added dwell.
After searching some info on CDI, i found some interesting opinions about our systems capabilities.Because i do not really know a thing about HW,please tell me,is this true???
quote from efi101 forum
Pantera EFI
Joined: 12 Feb 2005
Posts: 421
Location: So. California
New postPosted: Tue Sep 12, 2006 7:30 am Post subject: " EVENT " Time based system
- Sorry, you have made a " Time Based System " this is EVENT driven.
- sound scientific, right? Think about "time based" and "event based" systems:
- every system has time (that's how the processor and timers operate: with clockticks) and
- every system has events (of course, every trigger input or the actuation of spark is an event itself)
- A system to neglect time (that you could perhaps call event-driven if such existed) would need very many trigger pulses like 360/crankrot as InputTrigger/NissanTrigger and use CDI ignition to allow operation (1 degree resolution at best). Even here there is time, because RPM would be calculated by the const/time_between_teeth, but at least timers would not be used for ignition actuation (it would be possible though very unnatural to use degrees for injector-pulsewidth instead of milliseconds).
- all other systems (essentially every system that works with missing-tooth or simple trigger wheels) are time-based as well
- sound scientific, right? Think about "time based" and "event based" systems:
- In an event-based system there is no way you can predict the " start of coil charging " BEFORE the event, the previous event occurs.
- you naturally ask: before what event ? The system is full of events (when engine is rotating). There is always a trigger event that is the very last before something is due (eg. the desired ignition advance, say 19 degrees BTDC). Since you don't want to choose between 22 or 16 degrees BTDC (if the wheel teeth are coming that way as 6 degrees apart in a 60-2; even worse with a 36-1 or 12-1 or 4 evenly spaced), the only choice is to measure time and extrapolate to achieve reasonable output. This "time-based" behaviour makes it possible to do better than 6 degree timing precision (eg. 1/4 degree).
- Sure, some here may state this is untrue, though they just do VERY SLOPPY dwell control.
- Dwell control is essential for inductive type ignition.
- how can it be untrue, when it actually doesn't mean anything ? Using undefined (and actually confusing) terms, drawing conclusions without any logical reasoning. Referring to "the event" without anything indicating which event he's talking about. That's what happens when trying to state something and look scientific ;-) (in hope that the audience will not think over it just believe the conclusion)
- The Atmel chip that VEMS uses has limited power, not enough timer/counters to do this type of work well.
- This is absolutely bullshit
- 16 million instructions/sec allows calculating the fuel equation >2000 times every second (required appr 50..100 times)
- 160000 instructions are available for every cam-rotation at 12000 RPM (required: appr: 5000)
- Timing clock is actually just 4 million cycles/sec (4 usec) that is 1/4 degree at 12000 RPM. Appr 100x more powerful than complete and well working fuel injection systems from earlier times.
- Here, with this type of EMS, the best answer, the only answer is to use a CD type of coil charging. The charge time should be under or about 1ms.
- CDI has it's advantages, this is true
- Look at this problem from the point of a distributor triggered eight pole ( 8 ) event system. Then divide the time to do the work ( dwell ), add firing time ( coil off time ), now you see the window.
- absolutely right, with 8cyl distributor, your best choice is CDI (very low charge-time ignition transformer is also an option, and only if RPM is max 9000: available time for dwell is appr. 1.5msec)
Lance
I contacted AIM for their DASH communication protocol.
Looks like they are very helpfull,and they want to know what communication we would like to use,here the email from AIM:
"Peter,
Yes, we can provide you with a protocol template. Would the
communication be RS232 or CAN?
Michael Jaynes
AiM Sports, LLC"
This is a VERY nice dash,supporting this would be a really good feature.Please help me to answer his question,thanks,Peter.
- Communication would be RS232 at the moment, at least if connected to same port as computer. What is the protocol with LCD-connector, that I don't know.
HERE IS the RS232 protocol:
http://www.vems.hu/files/MembersPage/NanassyPeter/AIM_support/AIM-ECU%20protocol.pdf
and the email from AIM:
Peter,
The attached PDF contains the CAN and RS232 protocol, though the CAN is really the RS232 wrapped in CAN... Using the provided template will give you instant compatibility, whereas a customer can select the 'AiM' template under the ECU connection. Should your needs be beyond what is provided, you can send your template and we can incorporate it into our software, in due time of course. This may have added marketing benefit to you, whereas ANY customer setting up an ECU connection, or considering the option, would see 'VEMS' in the drop down list.
Michael Jaynes
AiM Sports, LLC
So please check it programmer guys,and tell me if you have any request from AIM.
//pnanassy
Nice work on obtaining the protocol specs Peter, i'll look into how easy it is to add dash support to firmware, however for the future it would be very nice indeed if they provided support for us. - DB
- Tell them that if they would use same protocol as Megatune it would be available to thousands of MegaSquirt users world wide... //TKalske
I recently found a very nice Water/Methanol Injection supplier
www.coolingmist.com
Im thinking about buying a basic kit,and using it to cool my charge air,and also add some methanol to have some more ponies.
They have a control box
( http://www.coolingmist.com/detailmain.aspx?pid=controller)
which forces the pump to have more or less flow,as required.Note that this system does not have INJECTORS.I would like to use vems instead of their control box.
Is it possible to control the pump with VEMS so i can have less flow at lower boost levels,more at medium,and full flow at max boost?
- Under 100kPa no flow is required!
- it would be possible to control the pump with PWM signal. The pump - terminal would be connected to an injector output, and a high current flyback diode applied (across the pump)
- like BY399 + small 0.2 ohm resistor or 0.5m thin wire, 7..8 of this in parallel
- but the pump should always be connected to ground for fire/safety and this arrangement voids that. Maybe a 18..20V transient diode (along with a 20A fuse that is there anyway!) would help with this.
- little point in this (or even dangerous) without good fuel-pressure feedback
- it would be possible to control the pump with PWM signal. The pump - terminal would be connected to an injector output, and a high current flyback diode applied (across the pump)
Here is an email from coolingmist:
"coolingmist@comcast.net saját magam részére
Részletes fejléc 14:49 (1 órája)
The controller controls the spray by forcing the pump to spray more or less flow based on the controller settings and the amount of boost or 0-5V input.
Thanks!"
//pnanassy
Thinking by typing here, so please correct me if I get the wrong end of the stick...
Do you need an analog type output for this? I'm not sure if VEMS has any - I think they're all digital. However there are electronic tricks that can be played...
Do you have some spare VEMS injector outputs? I wonder if you could configure the secondary injection to work as you need - ie give a PWM signal proportional to load (and therefore boost) that can then be smoothed with a suitable capacitor to give a pump drive signal - like a switch mode power supply.
I'm not sure if the current options with the secondary injection will let this work... I've not used this feature (yet).
Otherwise, a nasty way to do it would be to make up a 2 bit digital signal from the two misc outputs - so you can derive a 3 speed (and off) pump signal (assuming you use Gray encoding)
I'm sure someone more qualified than me can poke holes in these and think up a nicer idea...
dnb
Ok,here is a datalog of how my ECU behavies under IGNITION ON only.I realyzed this problem long ago:PW and sparks drops to zero suddenly and goes back to normal cranking mode.Note that i do nothing ,no cranking or anything else.
any solutions would be appretiated.
- I have no solution for the problem, but I have the same thing but only seen PW goes zero. I must check my logs for spark drops. And if I remember correctly the same PW zero effect can be seen on LCD.
So far,so good,i uploaded 044FW and did a try,looks like works great.Further testing needed tough.Thanks for the fast response!
//pnanassy
Pending
- MAT ignition retard.
- custom oil temp/pressure gauge with V3.3 and LCD
should check the relevant page, IIRC these are not yet implemented
GenBoard/UnderDevelopment/StagedInjectors
- 5 primary injectors, driven sequentially
- and another 5 staged. (note: 2 free injector drivers on this 5cyl engine, because fuelpump switched from an injector channel as well). Document the flowrate of injectors - the ratio of primary and secondary is very important because it must be configured
We have 10 Audi S3/TT high impedance (14ohm) 420ccm "shorty" injectors.So basically we run the same in the staged group as in the sequential "main" injector group.
- Possibly 5 connected (while the FET and EC36xx pin would be happy to drive 5 injectors, highz or properly configured lowz, but this could abuse the flyback.
- or 2+3 split. This is recommended. It's OK to switch them simultaneously though, like h[0]=xx xx xx xx xx xx xx 60 where 60 (decimal 96) is sum of 20 and 40 (decimal 32 and 64) to switch injF and injG (see GenBoard/Manual/DigitalOut/Table)
Sorry ,this section is totally chinees for me....
Please explain it a bit more.
Is it possible to configure fuel pump out to another "thing",to have all 8 injector FETs? How this works on a V8 engine ,where all 8 injectors have to be driven independently?
- I have 8 separately driven injectors with the fuel pump fed separately. See MembersPage/DavidBlades/FanTest for my config.
Anyway,is there any difference between:
1 inj FET driving 2 injectors
VS.
1 inj FET driving 3 injectors??
What injector behaviour can i expect if i drive 2 injectors from 1 inj FET,this way i do not need staging,all injectors will work all the time.There is NO NEED for lean idle,this is a race project.Im basically looking for options here....
A: We have driven 3,4 and 5 high-z injectors on the same outputs on a few cars. With the powerflyback board you will most likely have no problem with driving those high-z injectors in pairs. They aren't too big. //Emil
How can i insert the new ALS functions (MembersPage/GaborRacz/NewAlsLaunchAndOthers) in the 1.0.30firmware and Megatune??
Q
I still experience some unwanted EGO behaviour sometimes,the EGO correction just switches off (no correction at all),even tough it would be needed!
Please tell me which MegaTune variable to set and which direction??And what is this all about?
1.0.35
Firmware 1.0.35 (compile for testing: read: never released) got phased out (almost immediately after uploading) in favor of 1.0.36, since an accidental testing line prevented 1.0.35 to switch off fuel pump "config.engine_off_delay" after engine stall (eg. an accident).
Note: I think you control fuelpump with p259ch5, NOT with P259/6.
Serial port / cable problem suspected, as this are the same tools that have been in use for a while. If your tables were corrupted during upgrade, you can try to generate the config and upload the .mtt files through Hyper Terminal instead. When you do next upgrade, run the included bat files in the order described in README . //Emil
Hi Emil,
i uploaded my mtt files (made by Marcell) through Bray Terminal.
And yes,i use a USD=>Serial adapter cable, bought 2 different types (not pl2303 type from WebShop!). Unfortunately trashgrade windows drivers are very common, and often the cables are trash too (like the cypress-m8 type). There are plans to make MegaTune communications more robust with trash cables. Until than, use good cable.
The firmware upgrade and verification uses checksums and warns even today if there are any problem.
NOW THIS IS nerve-wrecking:
IAC - make a subpage for your IAC, it's scattered through several pages
- in 1.0.35 (experimental) the Idle valve cliks like crazy!!!!
Just make a proper report with:
- your PWM solenoid connection
- firmware version
- the mcd/mct dumps as well (mcd at least with 2 different IAC speed settings, as it's vital in your case).
This allows checking your config and reproducing (maybe in the experimental testing snapshot you used, MegaTune doesn't set the right iac variables ? iac config changed). Taking note of the above takes 15 minutes. Without publishing the above, you're wasting your time, noone will be able to help you efficiently.
- See EasyTherm on how to hexpatch the airden into any firmware in 2 mins.
- I have used stock Audi idle valve (AAN) with same (same as 1.0.35? this is a contradiction) firmware (connected to INJFET), and it controls idle very nicely
- IAC actuator speed set to 5.
- Valve needs quite high DC%, some 50%ish needed to open it even a little bit. IAC REF table is set to 50..90% across the temperature range
Miska,DC depends on how your throttle body is set (i dont want to run the IAC vale when engine is warmed up).
- AFAIK there is NO reasonable report from anyone (this is no report either, as it lacks basic information too). Where is the Fero report you refer to?
IAC frequency changed from very high to configurable
IAC wasn't noisy with 1.0.30, but i had resolution problems. In 1.0.35 (with unpublished config!!!) it's noisy and does NOT work
- Before, IAC ran at a very high frequency (this or DC is what the Audi IAC wants).
- after request (mostly from BMW guys: starting an Audi-BMW war here, hehe :-) the iac speed was made configurable
- the Audi solenoid is also very noisy, if it's driven at too low frequency it will tick very loudly. Maybe the problem is misconfiguration (maybe MegaTune vemsv3.ini - remember this is experimental - or usage: when you write the minimum required info, it's gonna be easy to tell)
HW change should NOT be needed - so go the proper route first and publish the required stuff (see mcd/mct above). Besides, (but remember, config first!), this is you can try ...
If you have powerflyback, Jörgen recommends trying an external (BY399 or Schottky) diode with enough current rating for the solenoid (around 2A for most Audi IAC's, 4-5A for the A8 engines IAC). The powerflyback makes the solenoid close quicker which is the exact opposite of what we want for the IAC. Even if you don't have powerflyback I recommend using the Schottky flyback diode on the solenoid as it has a fraction of the voltage drop of our normal low voltage flyback. -Jörgen
Don't use non-released firmware unless you know what you are doing!. There are no known problems in 1.0.36. The only things we know for sure from the above:
- you're using trash cable
- forget to apply your temperature calibration
- and seem to claim things without publishing even the minimum required info (eg. config!).
Strange that you get these problems with your serial adaptor. It's well known in the industry that USB cables cause problems, but not to this extent.
- 1, Please check that your config was uploaded correctly after the upgrade, with mct/mcd, or check in megatune that all the IAC settings are sane. Serial problems during download and upload of config can cause a lot of headache. I'll pick up a USB->serial converter and try the tools at home using this. I recommend Hyper Terminal for Manmcd/mct and mtt transfers when you get problems with the bat-files. If you dont get this straight, i recommend that you disable IAC and set idle on your throttle body instead until you can sort the IAC settings out.
- 2, The temp sensors are off because of the new firmware testing release doesnt contain your custom patched temp scaling. (Marcell, could you please patch 1.0.36 for him again, with the same calibration and upload?)
- 3, We do our best to test firmware before sending them out for release, these last 2, 1.0.35 and 36 are testing firmware's and not fully released. I will personally try this firmware on a BMW Turbo tonight. Only slight changes from 1.0.34 that this bmw is running already. I am though not able to test IAC output tonight, as we never use that in our racecars here.
Ok,thanks.
I have 2 of this shitty cables,and after an hour of suck!n,i installed cable #2.with this,the upload of firmw.36 went smooth,and then i uploaded my mtt's.i also looked into MT and everything was fine.IAC can be disabled of course,but to be honest,i need some validation that it works fine.
Good to know that it worked better with the other cable, are they branded, so you can warn other people for the non working cable?
Since 1.0.34 (that i tried in car without problems), the only change is that it now is less likely to leave coil active if loosing trigger at high RPM.
Also minor changes to megatune .ini. As i wrote before, will report as soon as i load this into a car. This BMW is 80 km away, so i likely cant report until approx 24.00 tonight.
//Emil
We installed the direct inputs of Launch and ALS.It worked fine for weeks,but now sometimes (1 from 30 trys) the pressed ALS switch just SHUTS DOWN the ECU!!!
Whats happening?
Happened only 2 times,and i had to concentrate on driving away,and since then i couldnt reproduce the sympthom
It's likely a wiring problem, because ALS activation hasn't been observed to shut down ECU.
Actually, "shut down" is only possible if it's a wiring problem.
If it's a firmware problem, it could reboot but not shut down.
Refining what "shut down" means in this case:
- LCD backlight goes off ?YES
- this indicates there is a HW (likely wiring) problem. The LCD backlight is NOT controlled by the processor so if it goes off, it indicates that power was cut or shorted
- can you measure DC voltage on EC36-pin28 (5V) ?
- Does that drop to 0V when button pressed ? => if so, a clear indication of HW (wiring) issue
- can you measure DC voltage on the ALS button ?
- if it ever goes above 5V or below -0.5V, that can kill the AVR (so it shouldn't happen).
- megatune communication goes away ?NO (not 100%sure)
- When button pressed, engine stopped immedately.
- But does it start immediately even while button is pressed ?
- Or it stays in "blackout" and only starts when button released ?
- fuel pump relay output ?
- other outputs ?
suspect:
- 5V is pulled to GND too strong (a short between 5V and GND ?)
- what is the pullup resistor value ?10k
- what is the GND return point of the switch ?Sensor GND
- cable-length ? approx. 1m
- any noisy lines nearby, like ignition ?NO
If the problem is GND noise, a 5V1 zener (red-black SOD80 small cylinder with endcaps in rescue2 kit) parallel with the switch for "transient protection" purpose, between GND and switched input could help:
- marked (black circle) goes to switched input
- the other (non-marked) cap goes to GND, EC36pin26
Closer to the EC18pin.. (ALS input) should be preferred.
However, if wired right, this "transient protection" zener shouldn't be necessary.
---
During the Boostcontrol runs i experienced some weird injector correction behaviour.Forunately we made logs so furher investigation is easy.
here the log:
see it at 9037.410 sec.
Here you can see that PW is immediatly dropping from 11.7 to 7.4
WHILE ENGINE IS AT FULL LOAD!!!!
This is completely not acceptable.
EGO limits were plus/minus 3%,and EGO was switched off at full load,with RPM and TPS 'switches'.
I experienced it constantly right after starting the engine,so it might be associated with EGO swithing on.