Idle PWM at 16ms?

[Blazer79]

I have Megasquirt 2 v3.0 mainboard
Firmware V2.90500
Bought assembled from DIYAutoTune
I also have the USB to Serial adapter bought from them.

Engine is a Chevy 350 TPI with a HEI distributor with module removed, just the VR signal going directly into MS, then into an Summit Racing MSD clone.
According to Desktop Dyno, this engine should make 406ft/lbs@4000rpm and 357hp@5500rpm. It has a Crane 216/228 hydraulic compu cam, and ProTopline 180cc heads.

I tried generating VE table from http://www.megamanual.com/v22manual/vetable.htm but apparently that page was disabled because it was generating dangerously wrong tables.

So I started with the VE table that comes in the fimware file. After a short while of trying to start the engine I had fouled my spark plugs. I just changed to new spark plugs and noticed that my pulse width at idle is around 16-18ms, which I believe is extremely high. What I did (and I'm sure I'm gonna be laughed at for doing it) is took the whole VE table and multiplied it by 0.20. Now the idle PWM is about 1.8-2.0ms. But during cranking is still around 8.0ms. Of course the car hasn't been able to run for more than 2 seconds before dying. Just checked a spark plug and still very wet. Not fouled though (yet)

I double checked the injector size and engine size on my MS setup. I also put the correct size on the Required Fuel Calculator, which gives me 19.30ms. I think this contributes to the extra long PWMs. I lowered it to 14.00ms and combined with the lowered VE table, appears to give me the more civilized arount 2.0ms. What else could be causing such high PWMs? Attached my current MSQ.

Any help greatly appreciated.

2014-09-01_00.56.26.msq

[Bernard Fife]

Blazer,

An MSQ is only half of what we need. We need an datalog of the problem too - having both is the best way to see what might be happening.

Lance.

[Blazer79]

OK here's the datalog and a new msq file. Sorry it took so long. I can only work on this car a few weekends per year.

I decided to start over. Here's what I did:

-Flashed the latest firmware to my Megasquirt 2.920
-Created a new project and made sure it's using the freshly downloaded .ini file
-Entered the initial constants again just to make sure I had no mistakes:
-->My injectors flow 24#@39PSI, which means they flow 25.185891168# @ 45PSI. I converted this using the online calculator http://www.megamanual.com/mt29.htm#introadvice which gave me 3.174g/s
-->My engine is a Chevy 350 0.020" over which gives me 353 cu. in.
-->I have a TPI setup, so under Injector Control I typed 8 injectors and 2 squirts per cycle, Alternating, Four Stroke, then proceeded to Calculate the Required Fuel. It calculates 18.4ms
-->I disabled ASE since I'm psychologically traumatized about flooding my engine and fouling my plugs
-->VE table has not been touched and was left as it came with the firmware. Actually all tables are still untouched.

So I set my Megastim to about 215°F, IAT to 95°F, and proceeded to vitually crank the engine by moving the RPM pot slowly. When I'm at cranking RPMs, I get a PWM around 5.249ms, but as soon as the engine is considered as not cranking by the MS, the PWM goes up to around 14.5. I checked a datalog from my old GM 730 ECM and it had about 2ms PWM when idling.

One thing to note is that the wall wart I'm using for megastim only outputs about 7V instead of 12V, so I'm getting a Battery Low signal on TunerStudio, and I'm sure It's still getting richer by this, but not to the extent of throwing 7x as much fuel as I need.

So is there a parameter I'm missing or misunderstanding that I'm setting wrong? Any help will be greatly appreciated.

2014-09-15_11.44.46.msl

2014-09-15_11.44.36.msq

[Bernard Fife]

the wall wart I'm using for megastim only outputs about 7V instead of 12V

Blazer79,

There's your problem. 7 volts not only triggers very long pulse widths due to the voltage correction, it is also borderline for the processor to operate, and it does very funny things at that point (such as getting the ADC's all wrong, including reporting the supply voltage incorrectly - usually higher than it actually is).

When I use your MSQ on my bench (with 98.9 kPa, 12.20 indicated volts (12.5 from the supply), 95ºF air temp. and 215ºF coolant temp.) I get a cranking PW of 4.078 and a running PW of 14.310 msec. Since your req_fuel is 18.4, this seems about right.

Note that in the real world your idle should be about 60 kPa or less in most cases, and this will reduce the PW to about 6.55 msec (both because of the lower MAP and the lower VE cell in the table).

Once you get your engine started and warmed up, the first thing to do is to work on the VE table to reduce the MAP as much as possible, see: http://www.megamanual.com/ms2/tune.htm#howto

Lance.

[Blazer79]

Thanks for your quick reply Lance.

I still think I must have a parameter wrong somewhere. Even with your corrected voltage, 14.310 is still extremely high for my application. That's about what I was getting when I fouled my spark plugs. Even with the 6.55ms you mention I'll still be too rich. If you think about it, if TunerPro is calculating 18ms for 100%VE, there is no way I need 14.310ms or even 6.55ms at idle.

I was looking into getting somewhere around 2ms. I played with req_fuel and it must be turned down to 1ms to get the PWM to less than 3ms, which doesn't make sense either. That's why I was hoping someone would point out some stupid mistake like for example converting the injector size to g/s incorrectly. I also tried multiplying the whole VE table * 0.25 in order to lower it. I'm sure that's not the right way either.

A friend showed me some datalogs of some corvettes and camaros and they are idling around 2ms as well, and get around 13ms at WOT and 6000rpms, but they are using HP tuner and the stock ECU.

Is there some other parameter I should look at? Anyone with a 350 TPI can give me their req_fuel and inj size so I can check if I'm in the ballpark?

Thanks again.

[trakkies]

I'm slightly confused here. Are you attempting to drive all the injectors using a wall wart? That won't work. You need something like a 10 amp supply to do this adequately.

[Blazer79]

Trakkies

No, I'm not trying to drive my injectors with a wall wart. I'm using a Megastim to test my MS II before fouling another set of spark plugs. The generic wall wart I bought for the Megastim has a voltage selector, but at 12V it only outputs 7V.

My problem is I'm getting way too high PWMs at idle. My req_fuel has been calculated at 18ms which seems logical, but idles at 16ms. Even on Lance's test bench (with correct voltage) it idles way too rich. I'm sure I'm missing some parameter somewhere that's causing this.

I did this test: with the Megastim idling at 800rpms and 85kpa vacuum, on the VE table, I copied the VE value in the 800rpms and 30kpa cell (the lowest possible kPa value) to the 800rpms and 85kpa cell. PWM went down to 9.5ms. Then I changed te battery voltage correction to 0, just to simulate a good battery. PWM went down to 8.3ms. Still too rich.

If I try to lower the value of this cell(800rpm and 85kpa) to try to lower my PWM, I go all the way to 1, and still PWM is 4.94ms. Still too rich. That's why I'm sure I'm doing something else wrong.

Please help

[kjones6039]

A bit off topic, but I feel compelled to correct some terminology being used here.

PW relates to injector Pulse Width..........
PWM refers to Pulse Width Modulation........

In the context of this discussion, we are talking about Pulse Width (PW) not PWM........



Ken