Can't get an easy cold start #2

[billjam]

Bill,

Just for kicks, could you post a composite tooth log of the engine cranking with no spark/fuel?

Would that be for your kicks or mine!

I can hardly wait to try it tonight.

[billjam]

Have you actually tried r56 and r52 at full CCW?

I'm just curious as it has been mentioned and you haven't responded.

David

Yes David, been there, done that.
Log 1 in one of the posts above is with R56 and R52 at full CCW.
Logs 2 and 3 (not posted) are at same settings but with fuel and spark on. They look identical to log 1, so no point in uploading them.

Log 8 (attached to a post above) is a log that was run with R56 starting at 0 and gradually being increased to about 8 turns CW while the engine was being cranked with compression for about ten seconds or more.
The shape of the bars gradually changes and after a while, the missing bars disappear. I don't know at what number of turns on R56 these changes occurred because it wasn't possible to note that during the test. The purpose of the test was just to see what effect, if any, R56 was having.
Basically, it made some minor changes to the shape of the log, but had no effect on the main problem (which is the size and frequency of the humps in the log at every second compression event).

I have asked a few times what effect R52 has and if it should be adjusted, but no response yet on that. I did try a log with it at 1 turn CW. It didn't make any difference and I was concentrating on other issues, so I didn't persevere with any further adjustments.

[billjam]

Bill,

Just for kicks, could you post a composite tooth log of the engine cranking with no spark/fuel?

Dennis, for your kicks I have attached not one, but two composite logs! One is without compression and one is with compression.
When I ran the logs, I forgot to turn off (pull relays) for fuel and spark. Sorry about that, however, having these two functions on or off hasn't made any difference to previous tooth logs, so hopefully they show you something useful.
Fuel pump was switched off so no likelihood of firing or starting. I can re-run them tomorrow without fuel and spark if you think it will make a difference.
R56 and R52 are set fully CCW.

I should point out that these logs are meaningless to me, so I can't tell whether they are good or bad.

[billr]

"log1" has no compression, though. I'm waiting to see a tooth-log with compression, pots full CCW, and no spark (and no fuel, if you prefer). Two of them, actually: with the engine both cold and hot.

[billjam]

"log1" has no compression, though. I'm waiting to see a tooth-log with compression, pots full CCW, and no spark (and no fuel, if you prefer). Two of them, actually: with the engine both cold and hot.

Bill, please check log #8 that I uploaded on Sept 25.
Log 8 is continuous cranking for 10 seconds or so as I slowly turned R56 from 0 to about 8 turns CW, so I don't believe that R56 is the problem. It helps clean up the signal marginally, but it isn't the problem. The first few seconds of that log are with R56 fully CCW. The position of R56 has almost no effect on the logs.

I would love to be able to show you a log from a warm engine, but need to get it started first.

BTW, I have found a guy at the local Porsche garage who knows how to use a scope, so I am trying to arrange a time for him to come over and help me with it.
If I connect the scope to the disconnected sensor and crank the engine, I presume we would get some kind of trace that has 58 bumps and a gap? Yes, no?
If the sensor is connected, there is nowhere to connect the scope without cutting into the harness or dismantling the DB37 connector.
Even if I can pickup the sensor wires somewhere near the MS3, I don't see how the scope trace would look any different because its signals are still being measured before its signals get into the MS3 box.
Do steer me in the right direction if I am not understanding the basics here.

[billr]

That "log8" was taken while you were turning R56? It's tough for me to tell what I am seeing there, to separate the real problem areas from what is caused by fussing with R56.

I thought your only problem was getting and easy cold-start, will it no longer start at all?

For the scope there are two fairly easy ways to connect the scope while the sensor is connected. On the mainboard there is jumper to select either VR or Opto as the input, hook the scope probe there. Or, take a sewing needle and pierce the input wire kind of like when they draw blood from you arm; then connect the probe to the needle. It will make a very small hole in the insulation that will probably close and "self heal" fine, but it you do it near the MS connector then chances are that area doesn't get flooded with water often and the little hole doesn't really matter. You can also use that needle to "back probe" into connectors, reaching in until the needle tip hits the connector contact; but that can be a bit trickier to do.

Whether the sensor is connected or not, you are correct: you should see a wave that is trying to be a sine wave with two cycles missing. My experience is that problems with the wave are right after the "missing tooth" gap; the first cycle after that tends to get amplitude variations. Syncing can usually be done with careful adjustment of the trigger level, as there will usually be one cycle (hump) that is a bit higher than the others. If that doesn't work, then trigger off the second channel, connecting it to the cam input. That way the CKP channel will reliably show you two crank revolutions, one full one with gap for sure.

[Dennis930]

Bill,

I looked at the composite log with compression. It basically reflects the tooth log: one hump on one revolution and two humps on the next revolution. I do not understand that. Also, are you using the cam sensor? If so, it is not showing in the composite log. It looks like the engine sync's up and then loses sync as you crank.

I have posted below a tooth log and a composite log for my engine that I did today. My engine is low compression compared to yours but three humps are visible for every engine revolution. In the composite log, the engine sync's up after passing the missing tooth, it does not lose sync after that and you can see where (at the end of the log) the cam signal is polled by MS.

77TurboToothLog.jpg (125.51 KiB) Viewed 790 times

77TurboCompositeLog.jpg (145.4 KiB) Viewed 790 times

For the o-scope, I think the jumper billr is taking about is shown below. Then you will need to find a spot to attach the negative lead.

MS3XV357VRjumper.jpg (29.29 KiB) Viewed 790 times

[billjam]

Thanks for your comments Dennis.
I have just compared our logs and I agree, something doesn't look right with mine, but why?
My tooth log clearly shows only three compressions for each two revs, not six. What can possibly cause this?
I just can't imagine any amount of fiddling with R56 or R52 will fix this issue. But then, as long as the -2 spikes are in the right place, the compression humps really shouldn't matter, should they?
The -2 pulse is regular at 58, so that seems OK, but how can there possibly be only half the number of compression events on an engine that has six good compressions and which runs perfectly once it starts?

Log 7a.jpg (60.81 KiB) Viewed 773 times

As far as the composite log is concerned, I have no idea what it should look like, mine doesn't look anything like yours and it just doesn't make sense to me.
Is there some other way I can check the camshaft pulse?
I'll attach my tune file again just in case something is screwed up with the ignition settings, but to me, it looks OK for 60-2 crank wheel and a one tooth cam wheel.

Did I mention frustration!

[billr]

Frustrated? Indeed, I know the feeling. Unless you will try things I have suggested, post the logs (not screen-shots), and answer all questions asked, my efforts are futile; and it is painful to hear of your wanderings "in the weeds". I will wish you the best of luck but am inclined to leave this thread.

[billjam]

Bill, sorry to hear that your enthusiasm is waning.
I am not ignoring your suggestions, just struggling to get time to work through this and figure it out. I usually work a ten or eleven hour day and can't always get time to spend on the car, however, tomorrow I am planning to try the oscilloscope test of the crank sensor. I still have no real idea what I am going to see and haven't figured out exactly how I'm going to do it, or record it, but I will give it a try.
As far as posting logs are concerned, I just figured it was easier for everyone to look at a picture than dick around downloading files and reading them in MLV, especially when they are basically all the same. I thought the log that I attached where I adjusted R56 from zero to about eight turns CW would have been quite useful to show the effect that R56 was having (or not).

I am really struggling to see how tooth logs are going to solve this problem. As I discussed in my last post, there is definitely something screwy with the logs with regard to the compression humps and clearly, R56 has nothing to do with this. What about R52? Should I be adjusting this too?
The composite logs I posted make no sense to me. I was kinda hoping that they might shed some light.

Unfortunately, I can't get the engine to run now, so hot logs are out of the question. Keep in mind that this difficult starting problem has been with the engine from the beginning. When it does run, it is 100% OK. When it is warm, it starts OK. This probably means that the cranking speed is a key issue.
At least summer is not far away.

[billr]

This is kind of a long-shot, it's not clear to me if it is even possible for the engine to run except in "wasted COP", but I gotta ask...

Is it possible this engine has non-stock cams that have it firing two cylinders simultaneously; like a "flat crank" V8 does? That would certainly account for the hard cranking and missing humps.

[Dennis930]

Bill,

Have you adjusted the valves lately? I am just wondering if one of the cams is mistimed or slipped adjustment. Even that does not explain the good compression reading though. Maybe you could turn the engine maually at the fan housing to see if you are getting compression at every 120* marked of the crank pulley.

You should have a cam signal on the composite log. The only thing I see different in your MSQ is "Second Trigger Active On" is set to Falling edge and the manual says Rising Edge, but it also says to verify it with a composite log and you are not getting any signal. I would check all the connections and make it is getting 5 or 12VDC reference voltage. Also, I assume the Clewett hall sensor needs a pullup resistor, so JP7 on the MS3X board needs jumpered; both R11 and R32 fully turned anti-clockwise (7 turns), then turn R11 three turns CW. One other thing, do you know if the cam pulse will come before the #1 tooth angle(87*)? I am sure it probably does.

As Billr suggested you could switch to Wasted Cop in the "Number of Coils" to see if that would help it start, it would take the cam signal out of the picture.

[billjam]

Thanks for the tips guys. I have a few hours to spend on this today, so I will work through your suggestions.

[billjam]

To start with today, I tried wasted spark. This has been suggested before, but I couldn’t find where the setting was until Dennis led me to it. I tried to start, but no joy.
Even ran a few logs, but they didn't look any different to previous ones.

Next, I did a few runs in test mode just to be sure that all coils are firing in the correct sequence. I can easily see when the spark pattern is good from the LEDs I have on the spark relay box.

Spark box 13c.jpg (30.86 KiB) Viewed 762 times

This video is with the test mode running at 240rpm, so it is easy to follow the firing order 1-6-2-4-3-5.
https://photos.app.goo.gl/v0qMXX59sGtJghgG3
It is also easy to confirm that the coils are actually firing by pulling the coils out of the engine and plugging in a spare spark plug grounded to the engine. So no problems there.

During cranking to start, you can see in this video how the spark pattern becomes random. This is how I have easily been able to identify that the spark timing was being screwed up by something.
https://photos.app.goo.gl/rEIyk7HLaRzjGqq22

Next step was to look into the strange compression pattern on the tooth logs.
Hand cranking the engine as suggested confirmed what the logs have been telling us all along … compression is occurring every 240*, not 120*. That is three compressions per two revs. This implies that two cylinders are firing at once.
When the engine comes up to compression, it is seriously hard to turn over. = two cylinders on compression?
Knowing something about the internals of these engines, I think you’ll agree Dennis that there is no way a cam can just slip and be out of timing by a whole cycle. Until I enquire further (on PP forum), I just can’t get my head around how the cams could be assembled with two cylinders having compression at the same time. It kinda sounds possible, but I can’t imagine it would run very smoothly.
If I was running wasted spark, I could understand how the engine might run, but it has been using sequential spark from day one, so it is a bit of a mystery that it runs so well.
Anyhow, at this point, I am not going to continue chasing my tail with sensor testing and pot twisting until I confirm that the cam timing is correct. Sh!t, if this is the problem, I can’t even blame someone else because I built the engine! (one of many, I might add)

All along, I have had the gut feeling that sync loss was the result of a problem, not the cause. My theory has always been that the sync loss was due to the erratic cranking speed, and even though it might mean an engine teardown to rectify, I am much happier doing that than learning how to use an oscilloscope.
I will check with the PP forum before going further, but at last it seems there is light at the end of the tunnel.

Thanks Bill and Dennis for steering me in the right direction. I will post more details here later when/if I confirm that cam timing is the problem. I need to figure out the easiest way to check for two cylinders at compression, probably by checking rocker positions.

[Marty Wagner]

On a flat 6 engine with overhead cams, if the cams in one bank are 180 cam degrees out of phase you will get 2 cylinders compressing at once. Is it possible to install Porsche cams this way?

[billr]

How about taking tooth-logs with only one cylinder at a time with the plugs in it? That way there is no way there can be more than one piston on compression, and I think the starter will turn the engine OK. We should see a nice single hump, and should see a "phase shift" relative to the missing-tooth gap that identifies each cylinder if six separate logs are taken like that.

[billjam]

Bill, I know you love logs, but we are at the point where the logs have done their job and identified the problem.
I haven't had time to do a physical check of the cam timing, but it is pretty clear that one cam is 360 crank degrees out of position. The engine will run OK in this condition as long as it has wasted spark. That is confusing because it hasn't been set that way, however Dennis has suggested that maybe the ignition has defaulted to this mode in the apparent absence of a cam signal.

[billr]

"Postby billjam » Sun Oct 01, 2017 1:54 am
To start with today, I tried wasted spark. This has been suggested before, but I couldn’t find where the setting was until Dennis led me to it. I tried to start, but no joy."

Crap, I'm getting confused again! That post quoted above indicated to me that wasted-spark did no good and the engine wouldn't fire or run; the "no joy"...

Now you are saying it will run on wasted spark???

Your choice about looking at more tooth logs, but the purpose was to help you determine if it was the first three in the firing order coming up on compression at the wrong time or the last three.

[billjam]

Bill,
I haven't been able to start the engine for weeks, probably because the weather is colder and the starter can't crank the engine fast enough.
The primary factor is that the cam timing is wrong and two compressions are coming up together. This almost stops the cranking every 2/3 of a revolution and it loses sync.

"Now you are saying it will run on wasted spark???" ... The only way this engine could have been running for the past four years is if it was in wasted spark mode all along, however, I didn't know that was the case until after that test. My ignition setting for Number of Coils has always been "Coil on plug".
As I suggested above, maybe in the absence of a cam signal, the system defaults to wasted spark. Is that what happens if there is no cam signal? I don't know, but it would explain how both cylinders on compression were getting a spark at the same time.

"That post quoted above indicated to me that wasted-spark did no good and the engine wouldn't fire or run" ... exactly, it didn't make any difference, it still wouldn't start and the log I ran at the same time was exactly the same as before. I am sure that the engine won't start regardless of spark setting because it is getting sync loss due to the erratic cranking speed.

Whether or not I run more logs isn't relevant any more because regardless of the logs, I need to physically check the cam timing to be 100% sure of the problem before I drop the engine out.