Scamby Posted May 23, 2020 Report Share Posted May 23, 2020 Greetings, I am experimenting possible misfires issues with my car. It happens after the car is running after a while driving which means like when the engine bay is really hot. The car starts to misfires even at little load. I checked everything and is good in terms of fuel (pump and injectors which by the way are new from DW) base pressure looks good also. When I checked the last time the coil was burned (picture attached) which tells me that there is a problem with the dwell. The coil was brand new. I changed the coil with another one that I have spare and the car starts to work fine until I am starting to have the same issue again. The dwell time for this car (Mazda 323 GTX Turbo BP 1.8) is 3.5ms which is right. Because is a distributor as the rev increase the ECU should decrease the dwell but not much as this will create a weak spark and under boost is not good. But when I checked the logs I see spikes of dwell times of 3.8 and up which is not right. I think of the last burn coil and because now I am starting experimenting the same issue with the spare coil and what I am seeing in the logs with the dwell, I am pretty sure I have something wrong with the dwell configuration that could be frying the coil. For what I understand the ECU must not send more than 3.5 ms of dwell at any time. Logs and .pcl file attached. Please advise. Regards. Mazda 323 GTX Map Dwell.pclr Log 2020-05-23-10-59-45-am.llg Quote Link to comment Share on other sites More sharing options...
Adamw Posted May 23, 2020 Report Share Posted May 23, 2020 Dwell scatter is just a result of a poor trigger, the ecu is getting unreliable and erratic information on engine position so it cant reliably predict when the next firing event will be. In order to get the spark at the right commanded angle then it has to shorten or extend the dwell. Those engines only have 4 teeth in the distributor - so the ecu only gets an update on engine position effectively once every 180 degs of creank revolution - effectively like a 2 tooth crank wheel. On top of that it is made worse by the fact that it is mounted to the end of the camshaft with a coupling that has significant backlash, which is exposed to all the valvetrain harmonics since the drive end of the camshaft is at the other end of the engine, which is then made worse by being coupled to the crankshaft with a rubber belt. If you want good dwell control then you will need a good crank trigger. Quote Link to comment Share on other sites More sharing options...
Scamby Posted May 24, 2020 Author Report Share Posted May 24, 2020 Greetings Adam, Thanks for the feedback. However it could be possible that the CAS has issues? I ask because this is a complete stock engine the only mod is the ECU swap and 430cc DW injectors. No cams or other big horsepower mods that this could happen (ignition scatter). Looking at the logs I don't see any trigger error issues which at first looks like there is no issues with the CAS and the ECU is getting a clean signal. The CAS that this car use (and Miatas) from the factory is the Mitsubishi 4G63 which is the same CAS that the Eclipse and other DSM cars use and also driven by the exhaust cam. I use the same factory CAS with other Mazdas that I had in the past with the same engine pushing 500+ whp without any single issue with other ECUs. Also I tune cars with the factory ignition pushing more that 300whp with the same CAS and again no issues. Now those car was running in sequential mode and the trigger was setup for 4G63. This car is batch firing injection because I am using the stock OEM GTX wiring. Which means I think the ECU does not need the home signal from the CAS for the calculations. The trigger pattern that my G4+ ECU setup is for MX5 which is the one that comes with the base map. I believe NB Miatas runs sequential injection. Based on that can you suggest if changing the trigger pattern in the ECU can resolve this issue? Please advise. Regards. Quote Link to comment Share on other sites More sharing options...
Adamw Posted May 25, 2020 Report Share Posted May 25, 2020 The dwell scatter you see in your log is pretty typical for a low tooth count trigger connected to a camshaft. Early evo's and MX5's are just as bad. If you want less dwell then you can command less in your dwell table but there is always going to be a lot of scatter either side of whatever you command. The only way to improve that is more trigger teeth, and less mechanical backlash in the drive system. Quote Link to comment Share on other sites More sharing options...
Scamby Posted May 25, 2020 Author Report Share Posted May 25, 2020 Greetings Adams, Thanks for the feedback. Link tech support advise to me to switch to a 3D table and setup the dwell that I want the ECU to command to the module. With a 3D table, the ECU will command whatever I setup in the dwell table no matter what? I apologize but I want to fix this issue as again with the stock ECU for example I dont have this problem whatsoever. Please advise. Regards. Quote Link to comment Share on other sites More sharing options...
Adamw Posted May 26, 2020 Report Share Posted May 26, 2020 20 hours ago, Scamby said: Link tech support advise to me to switch to a 3D table and setup the dwell that I want the ECU to command to the module. With a 3D table, the ECU will command whatever I setup in the dwell table no matter what? No, it is not going to make any difference. The only thing a 3D table will allow you to do is run less dwell at low loads to allow the coil to cool down when high energy is not needed and more dwell at high loads when it is needed. You will still get the same amount of dwell scatter. 20 hours ago, Scamby said: I want to fix this issue as again with the stock ECU for example I dont have this problem whatsoever. How did you log the dwell with the factory ECU? Many ecu's dont even have a variable to let you see the actual final dwell that was used - they only show the value that was commanded from the table. Quote Link to comment Share on other sites More sharing options...
Scamby Posted May 26, 2020 Author Report Share Posted May 26, 2020 Greetings Adam, About the stock ECU is that when I plug in the stock ECU back I don't have any misfires issues at all. Probably the stock ECU is sending less dwell that I setup. I will setup the 3D table for other things that tech support advise me. Thanks for your support. Regards. Quote Link to comment Share on other sites More sharing options...
Scamby Posted June 2, 2020 Author Report Share Posted June 2, 2020 Greetings Adam, I just checked the trigger wheel on my car and for my surprise it has one big slot for cam and not 2 like the Miatas.(Picture attached of the trigger disc and trigger scope). So in this case is not a MX5 trigger. That MX5 trigger comes with the base map supplied by Link for the 323 GTX. Based on the post that you helped the other guy with the same GTX engine, I setup the same trigger parameters that you advised and now the dwell scatter issue is not present anymore and the car runs a lot better. Based on the trigger scope, those trigger paramaters are the correct ones for my setup? Also I have doubt about the trigger filtering parameter for both triggers. Right now is setup to Level 1. With the MX5 trigger it was set to Level 2. Because is the same optical sensor that the Miata trigger use, do I have to set that filtering to Level 2 also? Please advise. Regards. Log 4-1 Trigger.llg Mazda Protege Map Trigger 4-1.pclr Quote Link to comment Share on other sites More sharing options...
Adamw Posted June 2, 2020 Report Share Posted June 2, 2020 Yes, your settings should be fine. Since you only have 4 teeth you can run the filter quite high with no negative effect. Level 2 would be suggested. Quote Link to comment Share on other sites More sharing options...
Scamby Posted June 2, 2020 Author Report Share Posted June 2, 2020 Greetings Adam, Will do. Thanks for your support. Regards. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.