Speeduino is now on Github Sponsors (Rather than Patreon): https://github.com/sponsors/noisymime
Help with building your Speeduino, installing it, getting it to run etc.
By JoostXT
#44051
Good morning guys/girls,

After trying my hardest for quite some time, I've finally given in and humbly ask you support in getting my car running right.

My last hope was an firmware issue, (see viewtopic.php?f=12&t=4048&p=43776#p43776) but as it turns out, this was not the case. Which is good, in a way.

Okay, so some background information first. I've had this car for over 17 years now, driven it a lot, fixed it even more, usual story I guess. After some time the original ECU went bad and I had to switch to something modern. Enter speeduino. At first I wanted the ECU's to be hotswappable so I modified a spare original ECU housing to fit the Speeduino into and re-used all wiring and sensors. This got the car running, but poorly. In an attempt to get it running right I've slowly but surely replaced all the ancient sensors to modern standards. I am however still having issues with it!

Technical rundown;
EA82T engine, 1800cc, 100kw, batch fired, port injected, 8,7:1 compression, turbo, no intercooler, flat four, 180k km's
CA18DET low Z with resistors 380cc injectors
Stock coil with ignitor and cap, rotor, wires, plugs (all new, or swapped)
upgraded fuel pump, walbro 255 lookalike
aftermarket boost referenced fuel regulator, set at 30PSI base
36-1 crank trigger wheel, VR sensor, shielded wire to VR conditioner
MAP sensor mounted in engine bay, fitted to intake plenum.
VW IAC valve, only for warmup
14point7 spartan2 WBO2
Tried loads of different base maps and guidelines, non seemed to work really well.
runs on euro95/10% ethanol

Issues;
- very high fuel consumption, combined driving; 6km/liter, granny like highway speeds; 10km/liter
- low on power
- injectors too large so won't idle below AFR 13 (hard to get through emissions!) (this is probably just my fault by fitting these)
- will not run smoothly or strongly with AFR above 13
- with lower AFRs it's still rumbly/rough but slightly better
- anything above AFR 14,5 will result in rough running, misfires.
- needs 1 sqrts/alternating to get PW high enough, but only runs right/strong with 2 sqrts/alternating (but won't idle)
- occasional hick-ups (every couple of minutes, have not found any logic to this), sometimes misfire (AFR high spike), sometimes something else (AFR low spike)
- very hard to map, fuel demands seems to be continuously changing, ignition changes don't have a serious effect on power/fuel consumption. Very high advance results in knock under boost. Very low advance makes for a very slow, but pretty smooth drive.
- runs pretty OK on boost, but low power.

TL;DR car runs rough, needs loads of fuel to run OKísh, fuel consumption through the roof, intermittent misses/hickups

edit; forgot the most important part! MSQ and log added. The log is from my way home from work, I was running VEAL at the time, so keep that in mind.
Attachments
(2.41 MiB) Downloaded 35 times
(58.23 KiB) Downloaded 37 times
By Vyizis
#44056
I would give it a try with the 'MAP Sample method' changed from 'Cycle Average' to 'Event Average'.

I would also increase the fuel pressure up to 2.5 bar as per the CA18DET factory pressure, that should give better atomisation and help with idle.

The general recommendation at the moment is also to not use the 'Multiply by ratio of AFR to Target AFR' setting
By JoostXT
#44058
Vyizis wrote:
Fri Jul 03, 2020 12:30 pm
I would give it a try with the 'MAP Sample method' changed from 'Cycle Average' to 'Event Average'.

I would also increase the fuel pressure up to 2.5 bar as per the CA18DET factory pressure, that should give better atomisation and help with idle.

The general recommendation at the moment is also to not use the 'Multiply by ratio of AFR to Target AFR' setting
Can you explain why and how changing this MAP setting would benefit me? To be honest I've never touched or even looked at this setting so I immediately went out and test. It does have effect, I always felt the ECU was lagging behind on the engine and this seems to be a a lot better now. It's running smoother and VEAL was immediately changing stuff all over the map, so it definitely did something! AE also seems to be working a lot better. Improvements all over the place!

The hiccups and low power are still there though. I checked the spark latency by fixing the timing and looking at the trigger wheel with a strobo and found about 5 degrees retardation at 4k rpm, is there a way to offset this besides just upping the spark map?
The general recommendation at the moment is also to not use the 'Multiply by ratio of AFR to Target AFR' setting


Again, can you explain (or point me to a topic) how and why? Switching this off would mean re-mapping the VE table right?

I have included another log, this is my test run with the MAP sampling changed, I also tried to mark the hiccups/stumbles but my reaction times are only so-so so you'll need to look a bit ahead of the marks ;)
Attachments
(1.9 MiB) Downloaded 30 times
By Vyizis
#44059
It is changing the filtering in the map reading the ecu uses for fuel calculations, instantaneous means no filtering and it will take the sensor value at the time it needs it, Event is over a single ignition event so 180deg and Cycle will be full 720deg of engine rotation. Essentially the ECU is reacting quicker to what the engine is doing so your ' I always felt the ECU was lagging behind on the engine' is quite true!

Spark latency is more that I would expect, there may be something in your trigger setup causing this but I don't have any experience in your setup so can't really help. EDIT: Is it possible that IAT retard is kicking in when you hold 4k rpm?

There is quite a bit of info on the AFR ratio feature here, it probably will mean having to redo some of your VE table unfortunately. I would also turn off ego correction until you have a better running engine. viewtopic.php?f=21&t=2510&hilit=Multipl ... Target+AFR

Are your temperature sensors calibrated correctly? Looking at that log your seeing very high intake (103C) and water (110C) temperatures after what is a fairly short pull.
By JoostXT
#44060
Vyizis wrote:
Fri Jul 03, 2020 2:28 pm
It is changing the filtering in the map reading the ecu uses for fuel calculations, instantaneous means no filtering and it will take the sensor value at the time it needs it, Event is over a single ignition event so 180deg and Cycle will be full 720deg of engine rotation. Essentially the ECU is reacting quicker to what the engine is doing so your ' I always felt the ECU was lagging behind on the engine' is quite true!

Spark latency is more that I would expect, there may be something in your trigger setup causing this but I don't have any experience in your setup so can't really help. EDIT: Is it possible that IAT retard is kicking in when you hold 4k rpm?

There is quite a bit of info on the AFR ratio feature here, it probably will mean having to redo some of your VE table unfortunately. I would also turn off ego correction until you have a better running engine. viewtopic.php?f=21&t=2510&hilit=Multipl ... Target+AFR

Are your temperature sensors calibrated correctly? Looking at that log your seeing very high intake (103C) and water (110C) temperatures after what is a fairly short pull.
Why would you set the MAP filtering at a higher or lower filtering level? I feel kind of stupid not looking into this :lol:

Spark latency is a bit high, I agree, IAT retard kicks in at 90c but I doubt it would hit that free-revving, I will check though.

EGO has 0% authority, physically turning the EGO control on/off requires a ECU restart so this is easier ;)

Heat is an ongoing issue, IAT might not be perfectly calibrated, but ambient air temperatures are always correct when engine is cold so it should be somewhat right. It's an open type VW sensor I believe. CLT's are always high, the radiator is small and inefficient so heat has always been my enemy. But, I don't really believe the CLT values, the rad fan kicks on and off at idle at quite a low interval and IR thermometer and external thermocouples put the coolant right around at 90c (fan switches at 88c).

The short pull with 100 degrees IAT did trigger the IAT retard though putting Advance at 17 degrees, but with the spark latency of 5 degrees it might explain the lack of power...
By ric355
#44061
JoostXT wrote:
Fri Jul 03, 2020 2:54 pm

The short pull with 100 degrees IAT did trigger the IAT retard though putting Advance at 17 degrees, but with the spark latency of 5 degrees it might explain the lack of power...
If by spark latency you mean the difference between the advance as measured with a timing light and the advance commanded by the ECU, this should only ever be at most 1 degree out. I think you said you are using a VR sensor? Check it is wired with the correct polarity and associated setting in speeduino; one of the effects of incorrect polarity is a significant variance in commanded versus actual advance since the crossover point of the signal is not steep enough for the conditioner to detect it at the same point every time with incorrect polarity.
User avatar
By PSIG
#44063
I'll add that not only must VR polarity be correct, but your Trigger Edge setting must be selected based on the conditioner module you are using, in order that the correct edge and missing teeth are properly read, e.g., Rising for MAX or LM conditioners, and Falling for most others. Don't guess or assume.

In looking at your tune and log, it would appear your biggest battle is incorrect settings, which are fighting you, and often I'll bet you don't even realize it. The weird or incorrect changes and guesses accumulate to make a confusing mess for Speeduino to sort. For example, you have no Barometric Fuel Corrections set. This means every time you start your engine, and the weather is a bit different, fueling will be wrong for those conditions. ;) Why it was changed I don't know, but it needs to be re-set where it was, along with a number of other settings that are not good or not-quite-right.

So first, I suggest you save your current tune as a new tune, perhaps "Subaru XT turbo MOD" or something that indicates changes to your original tune. This way the tune you have (Subaru XT turbo.msq) will not be altered and may be used for reference. Try stuff on your new tune. That's how you learn what the real-world effects are. Once you know what it does and if it helps, you can save it, or reload the MOD tune to go right-back where you were. :)

On the new tune, begin looking at every setting, in order. For each setting, consider if it should be set as it is, or if you know why it is set that way. Beginning at Settings tab and selecting each sub-menu item, verify why the settings are there and set however they are for a specific purpose. In your settings tab, I found at least one poor setting for your engine, fuel and application under each item.

Unfortunately, while we could suggest settings to you, tuning requires you understand what each setting is for, if you need it, and if so what it should be at any point. Only in this way can you make changes that will benefit your tune as you progress, so that's up to you. If you are not sure about something, then certainly ask, and we will help you to understand if we can. Begin with your settings and let's see how things improve, as it is difficult to find solutions to real issues when they are masked by settings issues.

David

As you practice your tuning skills, you can "see" weird stuff, and ponder or search why it is happening. Here is a shot of MegaLogViewer (my workhorse for tuning and data), showing at about 422 seconds your timing is commanded at 16°BTC. Your Spark Table is calling for 17-18°, and your IAT Retard is not doing anything yet… so what's going on? ;)
JoostXT_log_anomalies.jpg
JoostXT_log_anomalies.jpg (359.74 KiB) Viewed 989 times
By JoostXT
#44064
PSIG wrote:
Fri Jul 03, 2020 6:52 pm
I'll add that not only must VR polarity be correct, but your Trigger Edge setting must be selected based on the conditioner module you are using, in order that the correct edge and missing teeth are properly read, e.g., Rising for MAX or LM conditioners, and Falling for most others. Don't guess or assume.

In looking at your tune and log, it would appear your biggest battle is incorrect settings, which are fighting you, and often I'll bet you don't even realize it. The weird or incorrect changes and guesses accumulate to make a confusing mess for Speeduino to sort. For example, you have no Barometric Fuel Corrections set. This means every time you start your engine, and the weather is a bit different, fueling will be wrong for those conditions. ;) Why it was changed I don't know, but it needs to be re-set where it was, along with a number of other settings that are not good or not-quite-right.

So first, I suggest you save your current tune as a new tune, perhaps "Subaru XT turbo MOD" or something that indicates changes to your original tune. This way the tune you have (Subaru XT turbo.msq) will not be altered and may be used for reference. Try stuff on your new tune. That's how you learn what the real-world effects are. Once you know what it does and if it helps, you can save it, or reload the MOD tune to go right-back where you were. :)

On the new tune, begin looking at every setting, in order. For each setting, consider if it should be set as it is, or if you know why it is set that way. Beginning at Settings tab and selecting each sub-menu item, verify why the settings are there and set however they are for a specific purpose. In your settings tab, I found at least one poor setting for your engine, fuel and application under each item.

Unfortunately, while we could suggest settings to you, tuning requires you understand what each setting is for, if you need it, and if so what it should be at any point. Only in this way can you make changes that will benefit your tune as you progress, so that's up to you. If you are not sure about something, then certainly ask, and we will help you to understand if we can. Begin with your settings and let's see how things improve, as it is difficult to find solutions to real issues when they are masked by settings issues.

David

As you practice your tuning skills, you can "see" weird stuff, and ponder or search why it is happening. Here is a shot of MegaLogViewer (my workhorse for tuning and data), showing at about 422 seconds your timing is commanded at 16°BTC. Your Spark Table is calling for 17-18°, and your IAT Retard is not doing anything yet… so what's going on? ;)
JoostXT_log_anomalies.jpg
Thank you for your extensive response, let me answer some parts and I'll check the rest!

The trigger edge is set to falling as indicated by the supplier (diy-efi), I thought I had checked the VR polarity but I will double check tomorrow, it should be correct...

You are totally correct that I am fighting incorrect settings! This stems from a corrupted tune somewhere in the past, maybe I should start a new project and import the maps. I have been methodically setting stuff in a way it seemed logic to me, but some things were apparently wrong (like the MAP sampling or "incorporate AFR target" settings which seemed logical but generated issues!). The baro settings never occurred to me to be of much importance especially since the greatest elevation I can get over here are speedbumps (I live in the Netherlands) ;)

I will let you know how I get on!
By JoostXT
#44066
I've been looking into the barometric corrections, but all base maps supplied have a straight line, ie no correction!

Also, the barometric pressure varies around 99 to 102kpa where I live with a yearly average of 101kpa, is this really that much of an impact that it should correct for this? I really thought barometric corrections were only of use at elevation?

I looked at the speeduino base map and mx5 base maps to get a feel for the different correction tables but could not find anything way off, I changed some things but nothing that really impact performance? I changed some stuff in ASE and priming which may help starting, but that was never really an issue.

EDIT: @PSIG, maybe you can tell me where this 1 degree of timing went, because I for sure cannot find it. I though it *may* have something to do with increased CLT but I couldn't find any reference to this. To be able to find this missing degree I need to know exactly what speeduino does and need to calculate spark timing and check those parameters.
By LPG2CV
#44076
You may want to try changing the o2 corrections to pid to help with afr. Not sure about the injection timing being at 180*. Perhaps revert back to 355. I would imagine that being a flat 4 is why you changed it, however, im not convinced its any different from a 'standard' engine, and its possible with the turbo, if you have valve overlap, it may just blow fuel out of the exhaust. And this could affect the afrs the wb is producing. And low mpg.

Something David (PSIG) has suggested a number times before, is to wedge the waste gate open and tune the NA first. Then bring the boost in after that area is tuned.

I dont have any experience with boost, however I suspect you require more advance when over 3500k rpm and below 100 kpa. There is an ignition table generator, and a link to it somewhere on the forum. Perhaps do a comparIson with a generated one.

just re eead ... Download the zip from github. ba[…]

have you calibrated the temp sensors?

Modified 3 cylinder

I have a 3 cylinder engine that I have running on […]

You can delete the maf d speedy doesn't need it. Y[…]

Still can't find what you're looking for?