Need someone to look at my log...

[Sanjuro]

After getting through setting up my trigger, O2 sensor, fixing 1 fried FET and 2 fried coilpacks, I'm finally able to start the engine!  well somewhat - here is the issue:

It is very hard to start unless I give it 10% throttle.  Then it proceeds to cough and puff until it makes it through the afterstart duration.  It would then run smooth for a little while, then proceed to die.  This is all done with at least 10% throttle.

That is really disturbing is that occasionally MAP, RPM, TPS and EGT would spike through the roof (I don't even have EGT!), then the VEMS would proceed to reset itself.  I haven't caught anything suspicious on the scope for the VR and MAP sensors, but it might be something that is very small.

Here is my log and my msq:
http://www.mediafire.com/file/zymjnlvnayz/datalog200906101806.xls
http://www.mediafire.com/file/onnm4mnyidc/megasquirt200906101733.msq

What could this be?  The gizmo says I have trigger error, but I'm not sure if it is caused by low RPMs during cranking or wiring.  Any ideas?

[dnb]

Sounds like you might have a sensor ground problem from your description - it's worth double checking it.

I will have a look at the log files tomorrow evening - I have a little experience of the afterstart problem you describe, so may be able to suggest some config and table changes.

[Sanjuro]

thanks dnb, I'm about to head down to the dungeon to look over my wiring and do the InputTrigger/TriggerLog test.

[Sanjuro]

ok, so I did the following, and it is still broken:

InputTrigger/TriggerLog - Wheel Error Counter
did mlp01/mll.  3 sets of 3second crank, w/ injector/coil fuse unplugged.  No trigger error during cranking, pulls 1 trigger error when I stop cranking (I think this is normal)


InputTrigger/TriggerLog - Serial Logging
Results show good 1:3 tooth times between the actual tooth and the missing tooth of the 60-2 wheel.  VEMS correctly identifies all "MISSINGTOOTH 58":

Code: [Select]

276 24081 0.82
277 4269 3.1 MISSINGTOOTH 58
278 44293 1.83 MISSINGTOOTH 22
279 1351 0.31
280 18181 0.41
281 1388 1.02
282 27654 1.52 MISSINGTOOTH 2
283 1377 0.99
284 24837 0.89
285 1404 1.01
286 31750 1.27
287 1380 0.98
288 25605 0.8
289 1373 0.99
290 23813 0.93
291 1378 1
292 25094 1.05
293 1372 0.99
294 23557 0.93
295 1389 1.01
296 27909 1.18
297 1379 0.99
298 25350 0.9
299 1367 0.99
300 22277 0.87
301 1370 1
302 23046 1.03
303 1365 0.99
304 21765 0.94
305 1382 1.01
306 26117 1.19
307 1368 0.98
308 22534 0.86
309 1369 1
310 22789 1.01
311 1364 0.99
312 21509 0.94
313 1341 0.98
314 15621 0.72
315 1362 1.01
316 20998 1.34
317 1366 1
318 22021 1.04
319 1351 0.98
320 18181 0.82
321 1346 0.99
322 16902 0.92
323 1338 0.99
324 14853 0.87
325 1341 1
326 15621 1.05
327 1349 1
328 17669 1.13
329 1330 0.98
330 12806 0.72
331 1345 1.01
332 16645 1.29
333 1325 0.98
334 11525 0.69
335 1321 0.99
336 10501 0.91
337 1320 0.99
338 10245 0.97
339 1331 1
340 13061 1.27
341 1321 0.99
342 10502 0.8
343 1330 1
344 12805 1.21
345 1308 0.98
346 7173 0.56
347 1325 1.01
348 11525 1.6 MISSINGTOOTH 33
349 1310 0.98
350 7685 0.66
351 1310 1
352 7685 1
353 1320 1
354 10245 1.33
355 1303 0.98
356 5893 0.57
357 1315 1
358 8966 1.52 MISSINGTOOTH 5
359 1306 0.99
360 6405 0.71
361 1312 1
362 8197 1.27
363 1302 0.99
364 5637 0.68
365 1302 1
366 5637 1
367 1313 1
368 8453 1.49
369 1294 0.98
370 3589 0.42
371 1305 1
372 6405 1.78 MISSINGTOOTH 7
373 1320 1.01
374 10245 1.59 MISSINGTOOTH 1
375 1292 0.97
376 3077 0.3
377 1302 1
378 5638 1.83 MISSINGTOOTH 2
379 1304 1
380 6149 1.09
381 1301 0.99
382 5381 0.87
383 1310 1
384 7685 1.42
385 1306 0.99
386 6661 0.86
387 1303 0.99
388 5893 0.88
389 1288 0.98
390 2053 0.34
391 1276 0.99
392 64527 31.43 MISSINGTOOTH 7
393 3963 3.1 MISSINGTOOTH 58
394 31493 0.48
395 1249 0.31
396 57605 1.82 MISSINGTOOTH 2
397 1304 1.04
398 6149 0.1
399 1300 0.99
400 5126 0.83
401 1288 0.99
402 2053 0.4
403 1289 1
404 2309 1.12
405 1300 1
406 5125 2.21 MISSINGTOOTH 5
407 1271 0.97
408 63237 12.33 MISSINGTOOTH 1
409 1270 0.99
410 62981 0.99
411 1305 1.02
412 6405 0.1
413 1286 0.98
414 1541 0.24
415 1288 1
416 2053 1.33
417 1257 0.97
418 59653 29.05 MISSINGTOOTH 5
419 1300 1.03
420 5125 0.08
421 1260 0.96
422 60421 11.78 MISSINGTOOTH 2
423 1304 1.03
424 6149 0.1
425 1272 0.97
426 63493 10.32 MISSINGTOOTH 2
427 1278 1
428 65029 1.02
429 1257 0.98
430 59653 0.91
431 1278 1.01
432 65029 1.09
433 1284 1
434 1029 0.01
435 1271 0.98
436 63237 61.45 MISSINGTOOTH 5
437 1290 1.01
438 2565 0.04
439 1258 0.97
440 59908 23.35 MISSINGTOOTH 2
441 1271 1.01
442 63238 1.05
443 1288 1.01
444 2052 0.03
445 1279 0.99
446 65285 31.81 MISSINGTOOTH 3
447 1276 0.99
448 64518 0.98
449 1290 1.01
450 2564 0.03
451 1270 0.98
452 62981 24.56 MISSINGTOOTH 3
453 1272 1
454 63493 1
455 1282 1
456 517 0
457 1273 0.99
458 63493 122.81 MISSINGTOOTH 3
459 1286 1.01
460 1541 0.02
461 1282 0.99
462 517 0.33
463 1287 1
464 1797 3.47 MISSINGTOOTH 3
465 1278 0.99
466 65030 36.18 MISSINGTOOTH 1
467 1291 1.01
468 2820 0.04
469 1269 0.98
470 62725 22.24 MISSINGTOOTH 2
471 1277 1
472 64774 1.03
473 1297 1.01
474 4356 0.06
475 1263 0.97
476 61190 14.04 MISSINGTOOTH 3
477 1297 1.02
478 4356 0.07
479 1265 0.97
480 61701 14.16 MISSINGTOOTH 2
481 1280 1.01
482 5 0
483 1267 0.98
484 62213 12442.6 MISSINGTOOTH 2
485 1273 1
486 63749 1.02
487 1288 1.01
488 2053 0.03
489 1263 0.98
490 61189 29.8 MISSINGTOOTH 3
491 1276 1.01
492 64517 1.05
493 1261 0.98
494 60677 0.94
495 1286 1.01
496 1541 0.02
497 1270 0.98
498 62981 40.87 MISSINGTOOTH 4
499 1274 1
500 64005 1.01
501 1254 0.98
502 58885 0.92
503 1286 1.02
504 1541 0.02
505 1266 0.98
506 61957 40.2 MISSINGTOOTH 4
507 1241 0.98
508 55567 0.89
509 3857 3.1 MISSINGTOOTH 58
510 4357 0.07
Check Ground
TPS/MAP/VR sensor ground/cable shield join with VEMS sensor ground 10cm from where all other VEMS grounds converge, and a single thick wire runs to the cylinder head (away from starter and alt).  Coil pack ground separately to chassis.

Cranking shows the usual small noise in signal, but nothing dramatic.

Running Log
So far I've seen either MAP spike/TP drop or MAP/TP/RPM spike, followed by reset:






  So what gives?  Backfire?  Freakishly hard to track down ground loop?

[rob@vems.co.uk]

How is your alternator behaving?  Is Vbat nice and constant?

[Sanjuro]

How is your alternator behaving?  Is Vbat nice and constant?

I don't currently have the alternator hooked up, should it be?

Vbatt seems to spike or drop with the RPM/MAP/TP spike.  I have a racing battery relocated to the trunk powering the system/starter solenoid, and a standard automotive battery just for the starter. 

What I don't understand is which factor is causing the other.  Vbatt -> RPM? RPM -> other?  MAP -> other?  Is there some way to isolate?  The log doesn't really tell me the cause, but that they are happening

[Bat]

Hi,
How is the racing battery isolated from the starter?
I'm just thinking that if the starting battery has an intermittent connection whilst cranking the starter may then try to get its current from the racing battery which would pull the voltage down very quickly causing the reset to occur?
Cheers,
Gavin

[dnb]

Several things immediately spring to mind with your MSQ:

disable MAP multiplication because you don't need it for MAP based operation.

disable overrun fuel cut by setting both parameters to 9000 rpm or so.  (You can turn it back on later, but right now it will be in the way)

Change the afterstart added value to something like 80 and the 77 degree scaling factor to somewhere between 15 and 20.  You might need to fiddle the VE table to make this work, but it should be a closer match to what you want.

Drop the first column of your VE table to around 40.  Right now you're flooding the engine with far more fuel than it needs - big VE, multiplied by big cold start factor etc...

I have no idea about the resetting, other than the RPM reading indicating the engine has stopped.  This could be due to losing sync when the VR sensor voltage is very low at extremely low RPM.   When the car is running right they might not reappear.

[Sanjuro]

Hi,
How is the racing battery isolated from the starter?
I'm just thinking that if the starting battery has an intermittent connection whilst cranking the starter may then try to get its current from the racing battery which would pull the voltage down very quickly causing the reset to occur?
Cheers,
Gavin

The racing battery is on a completely different circuit from the starting battery, the only thing they share is the chassis ground.  what I might try is to run everything except for the vems on the starter battery, and do the vems on the racing battery to rule out any issues.

Several things immediately spring to mind with your MSQ:

disable MAP multiplication because you don't need it for MAP based operation.

disable overrun fuel cut by setting both parameters to 9000 rpm or so.  (You can turn it back on later, but right now it will be in the way)

Change the afterstart added value to something like 80 and the 77 degree scaling factor to somewhere between 15 and 20.  You might need to fiddle the VE table to make this work, but it should be a closer match to what you want.

Drop the first column of your VE table to around 40.  Right now you're flooding the engine with far more fuel than it needs - big VE, multiplied by big cold start factor etc...

I have no idea about the resetting, other than the RPM reading indicating the engine has stopped.  This could be due to losing sync when the VR sensor voltage is very low at extremely low RPM.   When the car is running right they might not reappear.

Thanks for the headsup,  I will make the changes to see if things improve.  I don't know about the low RPM though, because the trigger wheel performed without issues at ~200rpm cranking speed.   a smoother running engine never hurts though.

does vems use the o2 sensor in afterstart?  I checked the plugs and it is running very rich.

Current Status
Moving in a few days, won't have much time to mess with engine.  Hopefully I can try dnb's suggestion tomorrow before I pack everything away to see how things go.  Currently this is the latest datalog highlighting the problems:

[dnb]

No, the o2 sensor is only used after it has warmed up.  This can take 30 seconds or more depending on the ambiant temperature.  After this time, the o2 sensor readings are available for display, but they are not used in feedback control until the coolant temperature has reached the threshold set in Megatune - in your case this is 60 degrees.

[Sanjuro]

No, the o2 sensor is only used after it has warmed up.  This can take 30 seconds or more depending on the ambiant temperature.  After this time, the o2 sensor readings are available for display, but they are not used in feedback control until the coolant temperature has reached the threshold set in Megatune - in your case this is 60 degrees.

Sounds good, I'm hoping to see lambda that is not 0.6 or 0.84 all the time

Have to put off testing for now, need to pack up and move.  I'll be following up with this problem from California in a week or so

[Sanjuro]

Finally mustered enough courage to attack the problem again.  I set up the vems on the bench, hooked up to the audio trigger generator.  Did 180rpm for 60s, then 1280rpm for 60s.  Behold!  The problem did not go away:

Here are the changes I've made since the last graph:

• Changed R57 to 75k per Samspade's post
• disabled MAP multiplication
• Disabled overrun fuelcut

What gives?  it seems that the frequency of feakouts are proportional to the input RPM.  Rob I once again submit myself to your infinite wisdom...  Also if you don't mind I think this thread belongs in Triggering now.



The LM1815 out looks good too...

[rob@vems.co.uk]

So let me confirm this... you're getting resets/trigger errors with the Audio trigger?!?

[Sanjuro]

That is a positive sir.  Power comes from a switched 12V computer supply, nothing else hooked up except for the audio trigger. 

I'm working on getting a serial communication module for the scope, so I can log all 60s worth of LM1815 output and compare it to the VR (or audio) trigger input.  I figure this would really help me identify whether or not the finicky bits of the hardware are working. 

Is there anyone local here is southern California that wouldn't mind letting me test their board with my configs?

[SamSpade]

Can't really comment until we see the LM1815 input (pin 3) and output (pin 12) waveforms together, and on a shorter time base like 50us/div.