It may be working on the scope for you, but I've never tested the settings the way you have them. The way I told you is the way I designed the code to run, and is the way that pretty much everyone using the code uses it. Its well-tested in that configuration. Your way may work, but I've never tried it, and as far as I know, neither has anyone else.

Technically the rotary is a 4-stroke anyway, and setting it up as a 4-stroke 4 cylinder gives you the correct spacing between ignition events and the correct RPM.

If you have the CAS stabbed for the 1-3-7-9 ms1 setting, then the tooth #1 angle of 60 deg is correct, but other than that, you should use the settings I listed.

Ken

 

If you had to use anything other than rising edge for the secondary trigger, then you have the G sensor wired backwards.

For the Ne sensor, if you have TSEL wired to VROUTINV, then you'll want falling edge for the primary trigger input capture, otherwise you'll want rising edge.

It's important that you get the correct sensor input set to the correct edge. It may appear to work with 2nd trig at falling and primary at rising (assuming VROUTINV is wired to TSEL like it would've been for ms1), but the timing will be WAY off.

You most likely did not have to change your lm1815 circuit from what you used with ms1 as long as the edges are set right, and you set your settings correctly.

I have 2 MSes here, 1 is an ms1-v3 with lm1815 2nd trig, and the other is ms2/extra 2.1.0 beta (compiled out of my CVS tree, the latest dev code really) with lm1815 2nd trig. My two lm1815 boards are exactly the same, including the 1-shot output. I have driven the car about 20 miles or so on 2.1.0 with no problems with the settings and circuit I suggested.

Ken

 

next time I have the MS inside, I'll see how it works that way (although i'm pretty sure i tried that just on the bench and was having sync errors... but maybe thats just b/c i always had the 2nd input source set to cam instead of crank).

I have the primary wheel set to falling edge since its jumpered to VROUTINV, and the secondary set to rising. I've verified both of these at the 40 pin chip holder and compared them to the actual VR signal, so I know they're correct for my setup.

 

YES! She RUNS!

Took a while to get workable PID settings for EGO and Idle valve control down (and they still dont seem to be reacting quite right, but that'll come later), and I definitely need to revisit the VE table hardcore, but she's running!

Whenever the secondary injectors would stage on, it would go very lean and buck like hell (until i let off the gas and pushed the clutch in as quick as I could anyway); I'm assuming that they're set up the same way, so I dont know why it was happening. I've also been hearing that the time delay between staging the secondaries up and primaries down has been implemented in 2.1, but I dont see a control for it? Is it still in alpha?

This is using the settings Ken posted up earlier, and the original LM1815 circuit from MS1-e

Thanks for everyones help thus far!

 

The time delay for staging is implemented but not in a released code.

It should be in the next beta.

Ken

 

Ok, I'll be looking forward to that. In the meantime can you email me the bit of code you and Jobro were trying out (I PMed you on msextra)

 

Ok, i uploaded the file you sent me, and have noticed a few things. First, sync errors seem to have gone down a bit, which is a definite plus.

Second, when using closed-loop PWM idle, if I even touch the gas pedal, it will die on me. The duty cycle drops by about half (without increasing, and with decel enrich set to 100% using accel wizard). However, I'm just using PWM warmup and the same acceleration enrichment settings, its fine (accel settings at bottom of post).

Third, when the secondary injectors come online, a duty cycle is registering and showing up in the datalog, but physically the engine just goes lean, bogs, and refuses to keep accelerating. In injector characteristics, does Bank 2 need to be on for this to work properly, even though a duty cycle is still being recorded?

Thanks a lot!

TPS-Based

Rate (%/s)___Value (ms)
20.0________10.0
10.0________7.0
5.0_________4.0
0.5_________0.5

TPSdot Threshold = 1 %/s
Accel Time = 0.4 s
Accel Taper Time = 0.4 s
End Pulsewidth = 0.5 ms

 

I forgot to add to the last post that there hadnt been any problem before loading the new firmware up with closed-loop PWM idle control (so with 2.1.0 beta it was ok)

 

well, after further testing, I think I burned out Q5, which is why the secondary injectors arent working. I've ordered another one from mouser, as well as a few other components for the secondary injection circuit, just in case. The conflict with closed-loop PWM idle and touching the throttle is still present though.

 

I changed the closed-loop algorithm since the 2.1.0 beta.

You will probably have to retune the PID settings to get the same behavior back. My main tester for that code tells me that it works a LOT better than it did last beta.

Also, make sure you calibrated your TPS properly.

The PID closed-loop code closes the idle valve when you step on the throttle, which could also be the problem you saw... If it's a small throttle movement and the valve closes, MAP might actually decrease. You can make that close take longer which should help get rid of this problem. I'm adding an option at some point that makes it so the close is optional.

I fixed the TPS accel enrichments so that TPSdot would be more accurate, so that may have affected you as well. Before, accel may have been triggering when it should not have, causing the close of the BAC valve to be masked.

I've run that code on the bench and in the car without any issues. The only major change to staging was the addition of a duty-based staging mode, where you can stage based on primary duty cycle instead of rpm,tps,map, etc... If duty-cycle is being recorded, then the secondaries are squirting.

You should verify that none of the settings were corrupted during your upgrade.

EDIT: Just saw your Q5 comment. I was assuming you had already checked for hardware failure.

Ken

 

Please send a datalog of the "step on the throttle" problem.

Ken

 

I had been playing with the accel enrichments at the same time, but regardless of what I set them to, or how far/quickly i opened the throttle, it would still do it. In the datalog I was also fooling around with the idle PID settings, delays and whatnot, and nothing seemed to make a difference. After about 2/3 of the way through, i turned off PWM closed loop and proceeded to tune the accel enrich stuff and some of the PID controls for EGO correction, but even after doing this, (after 3/4 or so) touching the throttle still kills it.

I emailed you the datalog since it was a bit too big to attach

 

Alright, I tested the exact code I sent you in my rx7 before I sent it and didn't have the issues you're having, so there must be a corrupt setting somewhere. Send me your msq in addition to a datalog if you didn't already.

Thanks,

Ken

 

Ok, i sent you the msq.

 

I don't think google let me get that xls... We might have to figure out another way to get it to me.

Maybe just cut down the datalog to 1 instance of the issue happening?

Ken

 

I got your datalog and msq.

I loaded the msq and could not reproduce the issue.

The code I sent may have had some experimentation I was doing for O2 in it so I sent a new .s19 and ini to you.

I also sent you several suggestions to get better behavior out of the various closed-loop algorithms.

Ken

 

Ok, the secondaries are up and working again, so I can continue to play with it this weekend. It had been a joint problem of a trace in the PCB burning itself up and then the LED for injector 2 in the JimStim mysteriously dying and not providing the pull-up that I was seeing on the injector 1 channel, affecting what was showing up on the o-scope and completely confounding me for 5 hours today.

 

Try giving the duty-cycle based staging with the new delay setting a shot too, this should completely remove any lean spot on staging without having to add rich spots to the VE table.

The duty-based staging is very simple and should replace rpm/tps/map based staging if you use it... basically if you go over a certain (user-specified) primary duty, it stages. I'd set this to 80%. Then you're all but certain that it'll only ever stage when it's a good time to do so.

Ken

 

Yea, I'll continue to play with it, but that sounds like a very good way of controlling it.

 

What happens if the TPSdot AE fires when you are at something like 70% duty, at say 4000rpm. Is there time for an additional 5-7mS of fuel inside the window?

Does the code instantly snap to use both injectors when the duty should exceed 80% in a single cycle?

Or are you suggesting using a really low number for gradual staging, and a low number for reduce primary pw delay?

 

btw if you gave topless and me the same code, then I think he just needs to tune is throttle pump. It is working fine for me roughly my settings


TPS-Based

Rate (%/s)___Value (ms)
20.0________7.0
10.0________5.0
5.0_________3.5
2.0_________1.5

TPSdot Threshold = 1 %/s
Accel Time = 0.2 s
Accel Taper Time = 0.1 s
End Pulsewidth = 1.5 ms

I will double check these values later, but remember I am using a twin 50mm 48IDA style throttle body setup

This is the most solid piece of code I have used in a while

ALSO!

Megalowviewer v2.7.4 (as in diyautotune software package, I assume because it doesn't annoy you for registration every 15 seconds like the more current versions) does not seem to be able to save your 2.1.0 files after using the VE analyser, MLV says it saves the changes but the changes don't make their way into the new output file. Does the new version work?

 

the problem I had been having was something unrelated to accel enrichments, but its gone now with the newest file that Ken sent, and replaced with something more sinister (but unrelated to the code version, as its happening with MS1 too).

http://www.msextra.com/viewtopic.php?f=89&t=29908

 

Well, that problem is solved, but I need to revert back to MS1-e for a week or so since I'm taking her about 1500 miles in the next week and don't have time to tune or play with her beforehand, so efforts on MS2-e will continue in early Dec.

 

It'll switch on the secondaries in that situation. However, I may make a mod to that code so that the gradual stage can be made shorter during a quick stab. You can also set a second condition with this turned on to try to make the accel stuff behave.

Ken

 

I believe it works, but it's been a while since I messed with it. I'll have to double check.

Ken