MS2/extra beta 6 (20070722)

Post by **muythaibxr** » Tue Jul 31, 2007 9:52 pm

Ballistic wrote:Today I ran 30 minutes or so at the speed that was causing problems earlier and everything seemed OK. My runtime today was about 90 minutes.

I need to recheck my O2 sensor calibration, and may reenter my settings from scratch because my VE table is rich by ~20%.

Yeah, I've definitely not done anything there... the only thing I can come up with is EAE settings... have you tried without EAE just to rule that out? EAE is the only thing I've changed that could potentially have the effect you're seeing. I don't think it should cause you to be 20% off across the board though.

Liike I said though, I have a few different ways I can make EAE dual-channel capable, and you're using one of the ways. There are 2 other ways that I can think of that I could have you take a shot at if you want. That is once you've tried without EAE and made sure that's the difference.

Ken

Ballistic · Post by **Ballistic** » Wed Aug 01, 2007 3:59 am

I'll see if I can determine what's happening today.

With my VE table seemingly off, I didn't make changes to EAE. With the same settings as I used previously, acceleration transient AFRs were fairly flat -- probably better than before, but I'd need to make a side by side comparison with an old datalog to be certain. Deceleration resulted in a slight rich dip and then a large lean plateau.

The lean condition seems to result from the EGO correction on initial decel quickly adjusting to remove fuel, but then when pegged lean at 22.30 AFR it is slow to restore it.

Is there some boundary condition in the PID algorithm that prevents the full proportional change from being applied when full-scale lean?

I'm willing to try other codes if you'd like.

Post by **muythaibxr** » Wed Aug 01, 2007 7:14 am

Ballistic wrote: Is there some boundary condition in the PID algorithm that prevents the full proportional change from being applied when full-scale lean?

Not that I know of... At some point or another I mean to dig into that algorithm and see what I can find out, mainly because it seems to me that I should be able to make it react faster but without overshooting.

I don't know if I'll get to it for 2.0 though, as there are some other things I was thinking about doing to get better quick-stab response that I've been thinking about for about a week or two now.

Ken

cmcook · Post by **cmcook** » Thu Aug 02, 2007 3:24 pm

I've been out running with beta6 and noticed a problem with the output ports. I am using the IAC outputs to control what will become shift lights, but for now I am using them as a crude AFR gauge. I have a small diode and transistor circuit on the back of my dash to allow me to use the two outputs to control 3 LEDs. After I've been out for a while they all come on - I will have to check what this means in terms of what the IAC output pins are doing. If I stop and restart they work again. The megatune AFR gauge still appears to read properly which the outputs are stuck.

I'll dig up my MSQ but I'm not sure I'll be able find the bit in the log where it happened.

I'm also finding that my engine will die sometimes when slowing back down to idle revs with the clutch open. I have not tried much to tune it out which I guess will sort it but the VE table originated from my standard MS2 setup and it didn't it with that code. This I can find a log for but its getting late and I have to go to work early tomorrow.

Chris

Post by **muythaibxr** » Thu Aug 02, 2007 3:41 pm

Yeah, I wouldn't expect the tune to remain the same from ms2 standard to ms2/extra... we've redone a lot of the fuel calcs to help make them more accurate, but as a result, a moderate retune is necessary...

As far as the output ports doing weird things goes, we are using the exact same code that the ms2-standard 2.6xx series was using, and we have not changed it at all....

I've also used those ports myself (the IAC ones) to do fan control and VVT control with no problems like this... so if there's a problem, it's probably either hardware, or specific to the exact way you have the ports configured, so an msq will be good in helping us diagnose/fix the problem.

Ken

Post by **muythaibxr** » Sat Aug 04, 2007 9:06 am

Ballistic:

Have you had a chance to drive around with EAE turned off and on again to see if that's what was affecting your tune?

Glad to hear AFRs were pretty flat even with the new EAE code you're using.

Ken

Ballistic · Post by **Ballistic** » Sat Aug 04, 2007 7:38 pm

muythaibxr wrote:Ballistic:

Have you had a chance to drive around with EAE turned off and on again to see if that's what was affecting your tune?

Glad to hear AFRs were pretty flat even with the new EAE code you're using.

Ken

I haven't, but will check that Monday.

cmcook · Post by **cmcook** » Sun Aug 05, 2007 2:04 am

My IAC outputs did it again yesterday. Working fine then locked up - cycling the ignition made them work straight away. It could be a problem with the wiring of circuit for my light because its rather ropey but I wouldn't expect a reset of the MS to fix that. It happens in the log linked to below but I'm not sure exactly where - not much help I know but please keep going.

In looking through the log I noticed at at one of the points I thought it might have happened that the rpm drops to 0 for one point. I think the second time this happened could be when my lights lock up. The times when this happens are approx 592,1279,1456,1688,2357 and maybe more after this time.

Curiously my coolant temp started reading about 4-5 C below my instruments when it has alway been within about 1C previously (I checked it the day before and it was ok). After reflashing the table with the same (GM) values it read ok again.

I tried EAE with a later msq and got a very oscillatory idle. When I opened the tables they are very stepy which seems strange to me. I have not changed them from the defaults. Is this how they should look?

Some of this could be wiring and will check them out but reseting MS seems to sort them out.....

Link to msq and log: http://www.cmcook.f2s.com/cmcook.zip

Thanks,
Chris

MS2E Beta6, with EDIS ign

UnaClocker · Post by **UnaClocker** » Sun Aug 05, 2007 7:33 am

My IAC control seems to come and go as well. One day it wants to idle at 3,000, the next it's idling at 600.. I keep meaning to jumper around the resistors to see if that will improve my consistancy.

Post by **muythaibxr** » Sun Aug 05, 2007 8:24 am

cmcook wrote:My IAC outputs did it again yesterday. Working fine then locked up - cycling the ignition made them work straight away. It could be a problem with the wiring of circuit for my light because its rather ropey but I wouldn't expect a reset of the MS to fix that. It happens in the log linked to below but I'm not sure exactly where - not much help I know but please keep going.

Alright, I'll have to check into that.

I tried EAE with a later msq and got a very oscillatory idle. When I opened the tables they are very stepy which seems strange to me. I have not changed them from the defaults. Is this how they should look?

The default settings may look weird, but they work for a lot of people. If they make your idle oscillate, then they're wrong for your engine, and you should change them. I'd start out by making the two rpm graphs all 100, then adjust the main graphs at cruise until your AFR stays stable during transients... then use the RPM graphs to adjust idle and high rpm.

Ken

Ballistic · Post by **Ballistic** » Mon Aug 06, 2007 2:42 pm

Ken, turning EAE off did not negate the large change in my VE table as cruise AFRs stayed the same either way. Deceleration was richer with EAE off, but I attribute this to EAE needing adjustment which I haven't done since the VE table isn't dialed in.

Something strange is happening in the ~1300 to ~1800 rpm range under light acceleration. It's very slight and is almost like a lean surge but AFRs are OK when it happens. It is a cyclic oscillation that happens ~2x/sec. that can be so slight that it can barely be felt but may be heard as a pitch change. Timing is stable when checked at that RPM and I can't find anything in the datalogs other than a slight RPM ripple.

With my earlier testing of this s19 I was focused on the higher RPM ignition glitches and thought maybe I was sensing something but wasn't sure.

My next step is to recreate the msq from scratch and see if anything changes.

Post by **muythaibxr** » Mon Aug 06, 2007 6:18 pm

OK, well I haven't changed anything that should affect the VE table, nor have I changed anything that I can think of that would cause that cyclic effect you're noticing... so let me know how that recreation of the msq goes.

Ken

blk 02 · Post by **blk 02** » Mon Aug 06, 2007 10:26 pm

I'm Having a small issue with the tach out. Its hooked up to the fidle output pin. As rpm increases the signal begins to fluctate. At 6000 rpm the signal fluxuates as much as .5msec. Its fairly stable around 2k. This is tested on the bench with stim hooked up and the tach out hooked to my o-scope and my tci transmission computer. Megatune showes a steady 6000 rpm and the tci computer will fluxuate from 5850 to 6300.

If this helps I tried this circuit hooked to my other MS1 loaded with 29l and had the same issue but worse but when i flashed to 29v it was rock solid.

Thats for all the hard work guys
Graham

Ballistic · Post by **Ballistic** » Wed Aug 08, 2007 3:55 am

muythaibxr wrote: ... so let me know how that recreation of the msq goes.

The recreated msq exhibits the same behavior.

mattd860 · Post by **mattd860** » Wed Aug 08, 2007 7:12 am

Just a quick question. When tach out is set to ON and on Normal speed, does the megasquirt output a voltage between 0 and 5 volts, or between 0 and 12 volts? Thanks

Keithg · Post by **Keithg** » Wed Aug 08, 2007 8:19 am

MS puts out a pulse train 0-5V. You will need a simple 12V circuit to drive the tach. This circuit is shown in the MS1/E docs.

KeithG

mattd860 wrote:Just a quick question. When tach out is set to ON and on Normal speed, does the megasquirt output a voltage between 0 and 5 volts, or between 0 and 12 volts? Thanks

mattd860 · Post by **mattd860** » Wed Aug 08, 2007 8:26 am

Keithg wrote:MS puts out a pulse train 0-5V. You will need a simple 12V circuit to drive the tach. This circuit is shown in the MS1/E docs.

KeithG

mattd860 wrote:Just a quick question. When tach out is set to ON and on Normal speed, does the megasquirt output a voltage between 0 and 5 volts, or between 0 and 12 volts? Thanks

Actually my stock tach just was a wire that goes right to the megasquirt output pin. When tach out is active it controls my stock tach perfectly. The RPMS reported by that tach gauge perfectly match the megatune rpm gauge. So I don't and never had a need for that circuit. However, the other day my tach just suddenly stopped working after I replaced the alternator. I figured it was the MS so I reflashed the firmware and reloaded the msq but this didn't change anything. So with the car running, I checked the wire coming from the tach output (JS11) to the tach gauge and it read about 1.03 volts or something like that. I'm just trying to rule out possible problems. I also want to apply max volts to the tach to see if I can get it to move. I don't want to apply 12v is max is only 5volts. Thanks.

Post by **muythaibxr** » Wed Aug 08, 2007 9:08 am

Ballistic wrote:
muythaibxr wrote: ... so let me know how that recreation of the msq goes.
The recreated msq exhibits the same behavior.

OK, I'll see if I can find anything with your last msq that you gave me that would explain what you're feeling. Any idea what load you're at when you see the problem?

Ken

mops · Post by **mops** » Wed Aug 08, 2007 2:37 pm

mattd860 wrote:
Keithg wrote:MS puts out a pulse train 0-5V. You will need a simple 12V circuit to drive the tach. This circuit is shown in the MS1/E docs.

KeithG

mattd860 wrote:Just a quick question. When tach out is set to ON and on Normal speed, does the megasquirt output a voltage between 0 and 5 volts, or between 0 and 12 volts? Thanks
Actually my stock tach just was a wire that goes right to the megasquirt output pin. When tach out is active it controls my stock tach perfectly. The RPMS reported by that tach gauge perfectly match the megatune rpm gauge. So I don't and never had a need for that circuit. However, the other day my tach just suddenly stopped working after I replaced the alternator. I figured it was the MS so I reflashed the firmware and reloaded the msq but this didn't change anything. So with the car running, I checked the wire coming from the tach output (JS11) to the tach gauge and it read about 1.03 volts or something like that. I'm just trying to rule out possible problems. I also want to apply max volts to the tach to see if I can get it to move. I don't want to apply 12v is max is only 5volts. Thanks.

i run my tach straight of coil negative, so dont worry applying 12V ;)

Ballistic · Post by **Ballistic** » Wed Aug 08, 2007 2:42 pm

Ken, in looking up an answer to your last question I checked a burst datalog I had made but forgotten to view. Check out the AFR.