It is currently Fri Aug 17, 2018 3:31 pm



Forum locked This topic is locked, you cannot edit posts or make further replies.  [ 8 posts ] 
Author Message
 Post subject: Timing precision
PostPosted: Thu Feb 26, 2015 8:03 pm 
Offline
User avatar

Joined: Wed Jan 21, 2015 1:53 pm
Posts: 49
Location: WxSW U.S.
Might seem a little late in the game to ask a question like this but it is one of both theoretical and practical interest. Assuming a high-performance engine, if there were a way to continuously track crank position in terms of angle either using some sort of hardware assist or (incl.) sophisticated software methods so that timing events such as fuel injection and ignition are angle based, what would be: 1) The minimum acceptable resolution of the estimated crank angle value in degrees or fractions of a degree and, 2) The nice-to-have resolution in degrees or fractions of a degree. This question is mostly directed at Mark since I seem to recall from his profile on the old site or somewhere that he has a lot of experience setting up and tuning EFI systems on high performance engines.


Top
 Profile  
 
 Post subject: Re: Timing precision
PostPosted: Fri Feb 27, 2015 9:09 am 
Offline
User avatar

Joined: Sat May 11, 2013 9:45 am
Posts: 729
Location: PA, USA
The eTPU crank position code DOES continuously track crank position ;)

What's needed depends on the engine. on most engines if spark timing moves 0.5 degrees you will see a result on the dyno. On most engines you will not see a 0.25 degree change. You won't see 0.1 degree on any engine I don't think and that it just so happens is the spec most OEMs use to maintain emissions which turns out is more sensitive than the dyno.

To maintain 0.1 idle to redline you generally need at least 36 crank teeth or you run into a problem where if you run the tooth to tooth clock fast enough to get resolution at high rpm then it times out at idle or cranking speeds...at least that is true for the 24 bit clock the 5xxx family uses, a 32 bit clock could probably deal with a 24 tooth wheel and hold 0.1 degree.

So, a true 0.25 accuracy is generally ok, a true 0.1 degree accuracy is all you'll ever want I think. ALL oem setup give you thins, o5e has it, motec, haltech the 32bit stuff), Bosch for for have it. 16 bit products are going to struggle.....you'd need a 60 tooth wheel I think which is probably why electromotive going back 20 years was only supported 60-2 wheels.


Top
 Profile  
 
 Post subject: Re: Timing precision
PostPosted: Sun Mar 01, 2015 2:02 pm 
Offline
User avatar

Joined: Wed Jan 21, 2015 1:53 pm
Posts: 49
Location: WxSW U.S.
Really appreciate the answer. That's exactly the info sought.
mk e wrote:
The eTPU crank position code DOES continuously track crank position

I attempted not to pose the question in terms of a particular implementation since someone we know is promoting an alternative to the eTPU.
mk e wrote:
You won't see 0.1 degree on any engine I don't think and that it just so happens is the spec most OEMs use to maintain emissions which turns out is more sensitive than the dyno.

Hadn't considered emissions even though I live in the state with the strictest emissions enforcement in the U.S. (the biennial PITA).
mk e wrote:
To maintain 0.1 idle to redline you generally need at least 36 crank teeth or you run into a problem where if you run the tooth to tooth clock fast enough to get resolution at high rpm then it times out at idle or cranking speeds...at least that is true for the 24 bit clock the 5xxx family uses

Yeah, looks like a careful selection of tooth count and timer tick resolution will work for everything but cranking and initial startup. Since the ECU "knows" when the engine is cranking, timer overflow during cranking and startup could be handled as a special case.
mk e wrote:
So, a true 0.25 accuracy is generally ok, a true 0.1 degree accuracy is all you'll ever want I think.

So 0.1 deg. should be the target both for performance and emissions with possibly some allowance for exceeding 0.1 deg. at very high RPM.

What prompted the question was looking at the way the TMS470 handles the angle conversion (no documentation yet for the equivalent TMS570 functions) and the opportunities for loss of precision to creep into the signal processing chain from the crank trigger wheel and VRS sensor through the VRS signal conditioner (MAX992x or equiv.) and then processed by the HET (predececessor of TMS570 N2HET). What jumps out though is that a lot of attention should be given to the trigger wheel and sensor to ensure that the signal output from the signal conditioner is consistent tooth for tooth and gap for gap at a given steady RPM. That would imply a certain level of attention to the material and fabrication tolerances of the trigger wheel and how well it works with a particular sensor.


Top
 Profile  
 
 Post subject: Re: Timing precision
PostPosted: Sun Mar 01, 2015 7:11 pm 
Offline
User avatar

Joined: Sat May 11, 2013 9:52 am
Posts: 304
Location: Over here, doing 'over here' things.
US56 wrote:
...What prompted the question was looking at the way the TMS470 handles the angle conversion (no documentation yet for the equivalent TMS570 functions)
There exists much documentation regarding N2HET and NHET... both of which exist on 470 and 570 chips.

_________________
/me goes off to the corner feeling like Jerry Springer with a mullet.

My O5E candidate: 1982 Honda CX500TC motorcycle.


Top
 Profile  
 
 Post subject: Re: Timing precision
PostPosted: Mon Mar 02, 2015 9:00 am 
Offline
User avatar

Joined: Sat May 11, 2013 9:45 am
Posts: 729
Location: PA, USA
US56 wrote:
So 0.1 deg. should be the target both for performance and emissions with possibly some allowance for exceeding 0.1 deg. at very high RPM.

What prompted the question was looking at the way the TMS470 handles the angle conversion (no documentation yet for the equivalent TMS570 functions) and the opportunities for loss of precision to creep into the signal processing chain from the crank trigger wheel and VRS sensor through the VRS signal conditioner (MAX992x or equiv.) and then processed by the HET (predececessor of TMS570 N2HET). What jumps out though is that a lot of attention should be given to the trigger wheel and sensor to ensure that the signal output from the signal conditioner is consistent tooth for tooth and gap for gap at a given steady RPM. That would imply a certain level of attention to the material and fabrication tolerances of the trigger wheel and how well it works with a particular sensor.


This is an important point for all systems I think. i was ...ummm chatting I'll say wiht someone recently about MS3 and resolution and the fact that just because you see 0.1 degree "precision" on the tuner screen doesn't mean you get it at the engine.

On this project in early on engine testing I couldn't get the ECU to sync....there main issue was the hall sensor I bought from DIYautotune had something like +/-20% time between teeth signal at steady rpm. I sub'd in a high quality industrial sensor I had on hand and got +/-10% which fitted the sync problem...but it was still +/- 10% of +/- 1 degree on the 36-1 wheel I was using. I tested a couple VR sensors and they were well under 1% error.

At this point I finally understood why VR sensors were so popular with OEMs, but it also finally occurred to me that the trigger wheel you pick matters and if the MS ecus have no trouble reading the crap DIY crap sensor then there is simply no way they are error checking the crank signal which is really scary.

To your point, it's the system as a whole that matters. All systems have a pretty large number of processing steps to get from sensor to spark plug...so the only way to know how well they work is to test the system.


Top
 Profile  
 
 Post subject: Re: Timing precision
PostPosted: Sun Aug 09, 2015 12:30 pm 
Offline
User avatar

Joined: Sun Aug 09, 2015 12:09 pm
Posts: 9
mk e wrote:
On this project in early on engine testing I couldn't get the ECU to sync....there main issue was the hall sensor I bought from DIYautotune had something like +/-20% time between teeth signal at steady rpm. I sub'd in a high quality industrial sensor I had on hand and got +/-10% which fitted the sync problem...but it was still +/- 10% of +/- 1 degree on the 36-1 wheel I was using. I tested a couple VR sensors and they were well under 1% error.


How are you calculating tooth-tooth variance? One tooth to the next or one tooth to the average over the last full cycle?

Depending on how many cylinders your engine has, the tooth variance could be real. For 4 cylinder engines, the variance of one tooth to the cycle average can be fairly large, as the engine accelerates and decelerates due to compression/expansion, as well as manufacturing tolerances in the wheel itself. I wouldn't be surprised if a lot of your 10% variance is mechanical.


As to using 0.1 degree precision, 0.1 degree (or even 0.5 degree) won't make a difference in torque if you're at MBT spark timing. In fact, it takes about 3 degrees to see a change of about 3% in the indicated torque. If you're not at MBT, it makes a much bigger difference, up to a few % per degree. As with everything though, each item in the chain (sensor readings, sensor circuits, software, actuator control precision, actuator precision, ...) makes a difference, so targeting 0.1 degree is probably good.


Top
 Profile  
 
 Post subject: Re: Timing precision
PostPosted: Mon Aug 10, 2015 1:32 pm 
Offline
User avatar

Joined: Sat May 11, 2013 9:52 am
Posts: 304
Location: Over here, doing 'over here' things.
Once crankshaft position has been established, be it by missing tooth/teeth or cam trigger, the eTPU program should be tracking position on a tooth-tooth basis, and as such variance is limited to tooth-tooth.

If I remember correctly, the eTPU uses the period between the previous tooth signals to anticipate position between the most recent tooth and the next, and self-adjusts after the next tooth if that period varies from the previous.

_________________
/me goes off to the corner feeling like Jerry Springer with a mullet.

My O5E candidate: 1982 Honda CX500TC motorcycle.


Top
 Profile  
 
 Post subject: Re: Timing precision
PostPosted: Thu Aug 13, 2015 7:39 am 
Offline
User avatar

Joined: Sat May 11, 2013 9:45 am
Posts: 729
Location: PA, USA
abecedarian wrote:
Once crankshaft position has been established, be it by missing tooth/teeth or cam trigger, the eTPU program should be tracking position on a tooth-tooth basis, and as such variance is limited to tooth-tooth.

If I remember correctly, the eTPU uses the period between the previous tooth signals to anticipate position between the most recent tooth and the next, and self-adjusts after the next tooth if that period varies from the previous.


Been on vacation sorry.

Correct, it's tooth by tooth.

apalrd wrote:
Depending on how many cylinders your engine has, the tooth variance could be real. For 4 cylinder engines, the variance of one tooth to the cycle average can be fairly large, as the engine accelerates and decelerates due to compression/expansion, as well as manufacturing tolerances in the wheel itself. I wouldn't be surprised if a lot of your 10% variance is mechanical.


I was working with a 1 cylinder thinking is was both worst case for variation and cheap if something went wrong.

The biggest issue though was the random +/-% part because with +/-20% error going on even at a smooth steady rpm on an electric motor (which is how I took the % error readings) there was no way to maintain sync, at least not on a 36-1 wheel because if I opened the error window large enough to find the teeth it couldn't find the gap. I know the eTPU code ran on F1 cars and is the basis for several OEM and aftermarket ECU applications (Motec M1 and holley).....it works fine. That was just an issue with a total crap sensor.

I also know MS has no issue with that same sensor so they must be averaging several teeth which OEMs can't do because you can't detect mis-fire without a tooth by tooth position.


apalrd wrote:
As to using 0.1 degree precision, 0.1 degree (or even 0.5 degree) won't make a difference in torque if you're at MBT spark timing. In fact, it takes about 3 degrees to see a change of about 3% in the indicated torque. If you're not at MBT, it makes a much bigger difference, up to a few % per degree. As with everything though, each item in the chain (sensor readings, sensor circuits, software, actuator control precision, actuator precision, ...) makes a difference, so targeting 0.1 degree is probably good.




Turbo?

My experience on NA engines is they are quite sensitive and anything more than 0.5 degree shows up easily on the dyno....but there's always a detonation dance going too so thinking about it now I may have never truly been at MBT.


In other news Marcus is still chugging away on the HW....almost there.....


Top
 Profile  
 
Display posts from previous:  Sort by  
Forum locked This topic is locked, you cannot edit posts or make further replies.  [ 8 posts ] 

Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Theme designed by stylerbb.net © 2008
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
All times are UTC - 5 hours [ DST ]