PDA

View Full Version : Food for thought, camfire ignition system?



GTIspirit
10-28-2009, 04:08 PM
I've been trying to figure out if there is some way to install an aftermarket/standalone/programmable ECU on say a VW with CIS-E, aka, KE-Jetronic. The limiting factor seems to be the lack of a meaningful timing signal. I'm sure someone could configure MS to use the camshaft signal in the distributor, but that signal only comes from a four window hall effect sensor, greatly hindering accuracy of the ignition system. The rules expressly prohibit crankfire ignition systems, but what about "camfire" ignition systems? :D

So I'm wondering, would it be allowed to replace the four window hall effect sensor with say a mini 36tooth wheel? And replace the hall effect sensor with a sensor that would read the toothed wheel? The rules say this on the subject:

"Any ignition system which utilizes the original distributor for

spark timing and distribution is permitted. Internal distributor
components and distributor cap may be substituted."

I suppose the intent of the rules was that the distributor could be replaced with one of similar function but from a different supplier, like Meyle and Febi parts for example. But the rule does expressly say the "Internal distributor components.... may be substituted." The function of say a mini 36tooth trigger wheel is similar to the function of the four-window vane inside the distributor. And the function of an inductive sensor is equivalent to that of a hall effect sensor. The signals are functionally equivalent, it's just that one has a lot more resolution and a gap at TDC, but otherwise the signals are the same in that they are just pulses read by the ECU.

I realize that the camshaft rotates at half speed, so a 36 tooth camshaft wheel is equivalent to an 18tooth crankshaft wheel, and there are physical limits to how small a toothed wheel could be made and still have the inductive sensor read the gaps.

I'm sure it is prohibited due to some flaw in my logic because I'm prone to creative interpretation of the rules. But I'm wondering why not so I'm throwing this out there for the discussion.

joeg
10-28-2009, 04:19 PM
I think what you propose is legal since distributors are free but the accuracy (non-crank sensor) is a big technological hurdle.

shwah
10-28-2009, 06:06 PM
There are a lot of people that run VWs on Megasquirt using the 4 window hall signal.

It will be no less accurate than the current ignition, but you can map it.
Yes the timing will 'walk' a little bit with rpm, but again, you can map it to compensate.

The bigger issue has been controlling the fuel side of CIS-E in a more sophisticated manner than WOT enrichment. Been working on it for a while, and IMO there is a solution, but I am not convinced that it is not better to just use Digifant hardware and make it simple to integrate an aftermarket ECU. I am converting my car with DIYAutotune's help, and will have some data on area under the curve for both setups to compare.

jlinfert
10-29-2009, 06:32 AM
I've been trying to figure out if there is some way to install an aftermarket/standalone/programmable ECU on say a VW with CIS-E, aka, KE-Jetronic. The limiting factor seems to be the lack of a meaningful timing signal. I'm sure someone could configure MS to use the camshaft signal in the distributor, but that signal only comes from a four window hall effect sensor, greatly hindering accuracy of the ignition system. The rules expressly prohibit crankfire ignition systems, but what about "camfire" ignition systems? :D

So I'm wondering, would it be allowed to replace the four window hall effect sensor with say a mini 36tooth wheel? And replace the hall effect sensor with a sensor that would read the toothed wheel? The rules say this on the subject:

"Any ignition system which utilizes the original distributor for

spark timing and distribution is permitted. Internal distributor

components and distributor cap may be substituted.

"



I suppose the intent of the rules was that the distributor could be replaced with one of similar function but from a different supplier, like Meyle and Febi parts for example. But the rule does expressly say the "Internal distributor components.... may be substituted." The function of say a mini 36tooth trigger wheel is similar to the function of the four-window vane inside the distributor. And the function of an inductive sensor is equivalent to that of a hall effect sensor. The signals are functionally equivalent, it's just that one has a lot more resolution and a gap at TDC, but otherwise the signals are the same in that they are just pulses read by the ECU.


I realize that the camshaft rotates at half speed, so a 36 tooth camshaft wheel is equivalent to an 18tooth crankshaft wheel, and there are physical limits to how small a toothed wheel could be made and still have the inductive sensor read the gaps.



I'm sure it is prohibited due to some flaw in my logic because I'm prone to creative interpretation of the rules. But I'm wondering why not so I'm throwing this out there for the discussion.

I agree with your rules interpretation but a 36-1 wheel that small would be iffy as MS seems to have a hard time whe the teeth get really close. We had problems losing rpm signal with a 5" 60-2 tooth crank wheel above 6000 rpm on our GT cars but a 4 1/2" 36-1 works fine. HOWEVER, MS2/Extra code supports a 12-1 tooth wheel and lets you choose if it is cam or crank speed so that could be made to work

jlinfert
10-29-2009, 06:36 AM
There are a lot of people that run VWs on Megasquirt using the 4 window hall signal.

It will be no less accurate than the current ignition, but you can map it.
Yes the timing will 'walk' a little bit with rpm, but again, you can map it to compensate.

The bigger issue has been controlling the fuel side of CIS-E in a more sophisticated manner than WOT enrichment. Been working on it for a while, and IMO there is a solution, but I am not convinced that it is not better to just use Digifant hardware and make it simple to integrate an aftermarket ECU. I am converting my car with DIYAutotune's help, and will have some data on area under the curve for both setups to compare.
I think I've gotta argee with you Chris, given my choice I'd just do a digi conversion.

Flyinglizard
10-30-2009, 09:26 PM
Seems like you can run a trigger on the cam wheel itself.
IMHO. MM

shwah
10-31-2009, 10:00 AM
Do some research into motorcycle stuff. IIRC some of them run trigger wheels in small spaces. Not huge tooth counts, but better resolution than the hall sensor wheel.

jlinfert
11-01-2009, 05:40 PM
Do some research into motorcycle stuff. IIRC some of them run trigger wheels in small spaces. Not huge tooth counts, but better resolution than the hall sensor wheel.
Actually we are already looking into have some small diameter 12-1 tooth wheels made. Cad file was sent to the mfr, we are just waiting for the protos.

GTIspirit
11-01-2009, 07:21 PM
Seems like you can run a trigger on the cam wheel itself.
IMHO. MM

But then how to get the sensor? Someone previously noted that there is no allowance to add sensors, hence me wondering if the wheel could be put inside the distributor and the hall effect sensor on the distributor changed out for an equivalent sensor that can read the pulses from a toothed wheel. But if the wheel could go on the camshaft and and the addition of a sensor was allowed then said wheel could have more teeth.

jlinfert
11-02-2009, 06:38 AM
But then how to get the sensor? Someone previously noted that there is no allowance to add sensors, hence me wondering if the wheel could be put inside the distributor and the hall effect sensor on the distributor changed out for an equivalent sensor that can read the pulses from a toothed wheel. But if the wheel could go on the camshaft and and the addition of a sensor was allowed then said wheel could have more teeth.

Our idea is to replace the stock distributor pickup sensor with one that can read the toothed wheel. Using a wheel on the cam is a no go in IT unless one is already there.

shwah
11-03-2009, 09:41 PM
Jeff - I would be interested in that part.

Flyinglizard
11-03-2009, 10:32 PM
I use a digi dist, with only the module, for a fixed timing on our ministocker. The timing is set at 31 * advance. The timing does not move over 1 degree, ever. Less than 1 degree really. I have used this for 2000 laps or more, never had any problems or failures. I have two modules mounted, but never had to use the other one.
The engine is 12 to 1, runs from 4800 low end to 7300 upper. 100 octane fuel.
What could be improved upon? You are still limited by the dist/belt/shaft twist. Enlighten me. Do I pick up some torque with less lead at 4800?
TIA.

shwah
11-04-2009, 09:50 AM
I run fixed timing on my car as well, but with the 4 window wheel, and have not had issues. I would still be interested in the 12 tooth wheel to test and compare. Not convinced that I need it at all, but don't know unless I test it.

GTIspirit
11-04-2009, 10:47 AM
I don't think the stock system has fixed timing, at least not on my '87 16V, unless we mean something different by fixed timing. I believe the Triumph Adler knock box has some kind of electronic ignition advance map. So while it's "fixed" in that I don't believe it can be changed via calibration or EEPROM or chip-tuning, it's not fixed in the sense that there is the same advance at idle as at 7000rpm. The only change that can be made is to rotate the distributor, which will make a global offset to the ignition map.

Correct me if I'm wrong, but the way I think the ignition system works on the VW's with KE-Jetronic, (KE-Jetronic is only for fueling control) is that the knock box contains some kind of ignition advance map. And said map can be retarded in case knock is detected, but said map can never be advanced electronically, only by rotating the distributor. I believe what happens is that the hall effect sensor is in effect a trigger wheel. A pulse from the hall effect sensor goes to the knock box. The knock box knows the engine rpm, and depending on the engine rpm, sends a trigger to the TCI-h ignition coil, which in turn sparks the correct cylinder through the distributor. So there is a delay time between the signal from the hall effect sensor and to the ignition coil. This delay time would be greater at lower engine rpms, and at higher engine rpm's, with more advance, the delay time would be reduced, according to the map.

Now I'm not an engine calibration engineer, but I'm told by someone who is, that there will be a lot of scatter in the spark when the input signal from a four vane hall effect sensor. The four window hall effect sensor simply doesn't provide enough resolution to accurately control the spark timing, and there will be a lot of variance relative to say a 36 or 60 tooth crank trigger. But even a 12 tooth camshaft trigger would be better than a four window/tooth. I'm also told that the spark advance map has a greater effect on engine output than the fueling, that 1-2* difference in timing can make a difference, especially in the all important area under the curve.

So, the way I understand the rules is that whatever will fit inside the stock distributor is fair game. But the stock method of spark distribution must be retained, so the cap and rotor must still be used. There will be some hardware limitations to advance based on the length of the rotor edge, but I would think the four window hall effect sensor is more of a handicap. So if someone has designed a 12tooth wheel to fit the stock 16V distributor they should offer it for sale. ;)

Now about the comment about belt stretch/variance between the cam and crank. I suppose this is why all the OEM's use a crank trigger wheel for timing, the camshaft is only used for synchronization. But still, a 12 tooth trigger wheel on the camshaft is better than a four window vane sensor, and maybe 12 teeth is the resolution limit, above which the effects of the timing belt are noticeable. I don't know, it's a good question.

shwah
11-04-2009, 11:57 AM
I know that the knock box has ignition mapping in it. I run fixed timing on my car. I found that the knock box was altering the timing when I did not want it to, and felt that it was an additional failure point that I did not want to leave in the system. My car is wired per the CIS-E Golf without knock sensing diagram in the Bentley shop manual.

The 4 window sensor provides enough resolution to control the spark timing for every single CIS-E/KE-Jetronic car that VW sold. I found that the 'scatter' of the spark timing reduced dramatically when I removed the knock box from the system.

I'll define 'better' between 4 window and 12 tooth setups with empirical testing for my car, as the better that I care about is one that impacts the performance/reliability of the system, and we may just not need the higher resolution.

Flyinglizard
11-05-2009, 10:49 AM
Chris and I both just run the module, no knock box.
There does not appear to be any spark timing varience. If any, it is less than 1 degree. The belt/ shaft/drive has that much. I doubt that any 12 tooth setup will reduce the 4 door hall wheel spark bounce, (that really doesnt exist).
If the dyno shows that a nice advance curve will make more torque @ 4700rpm, than I am interested.
The stock setup takes care of phasing, by having the timing and the roter move at the same time. If you fire it from a wheel, with much of a curve, you may run into phasing issues.IMHO.

pfcs
11-05-2009, 09:45 PM
The CIS E GTI ignition box and the Digi ECU both have an inbuilt processing delay of about 60*. (If you bypass them, your timing will be over-advanced by that amount and will need resetting to fixed total timing of your choice (31-33* generally for 8v)
With a new distributor there is very little ignition scatter at any rpm. Crank-fire is nice if you build a hand-grenade 14/1+ GT motor which is very sensitive to over-advance. One degree of scatter/timing off by 2* in this app=virtually no power difference around proper setting.
If you set your timing with locked/no advance setup and see several degrees of retard @6k, it's probably your light and not the ignition that is late. (search MSD ignition info which put me present to this and their recommendation for a specific Sears Penske light which made me a believer when I saw my Volvo motor was overadvanced 4* @6500 if I timed it with $$ Snap-On light!)
And, yes-for IT app, advance curve is unnecessary/no help to power in usable rev band.

jlinfert
11-06-2009, 07:14 AM
The CIS E GTI ignition box and the Digi ECU both have an inbuilt processing delay of about 60*. (If you bypass them, your timing will be over-advanced by that amount and will need resetting to fixed total timing of your choice (31-33* generally for 8v)
With a new distributor there is very little ignition scatter at any rpm. Crank-fire is nice if you build a hand-grenade 14/1+ GT motor which is very sensitive to over-advance. One degree of scatter/timing off by 2* in this app=virtually no power difference around proper setting.
If you set your timing with locked/no advance setup and see several degrees of retard @6k, it's probably your light and not the ignition that is late. (search MSD ignition info which put me present to this and their recommendation for a specific Sears Penske light which made me a believer when I saw my Volvo motor was overadvanced 4* @6500 if I timed it with $$ Snap-On light!)
And, yes-for IT app, advance curve is unnecessary/no help to power in usable rev band.
Since when? I just had an ITB mk3 Jetta on the dyno Wednesday,(one of our megasquirt customers) and we saw gains of 4whp and 6 wtq to the AVERAGE power across a 4000 to 6500rpm powerband. The peak numbers were only improved by 1 or 2 but the curve flattened out nicely and wide power curves make your car very easy and fun to drive.

Flyinglizard
11-09-2009, 11:33 PM
Jeff, are you a diplomat in training?? Phil has shown to know of what he speaks. I have known Phil for about 17yrs, from a distance, but at the same tracks, mostly Watkins Glen. He has shown to know the higher tech areas of making slow shit go fast. like Volvos and VWs.
What are you saying in your post?? It is not clear, not informative. Please elaborate your found data.
I have found no advantage to more lead than about 32 degrees on this ( Non IT legal ) engine. at any rpm that this car runs at( 4700- 7300). More than 35* seems to shorten head gasket life, and often distorts the chamber squish area. I run fixed timing to deal with the vacuum rules at my local circle track. I have found no timing varience, ever.
The car pulls the same with the advance dist also, just idles poorly.
The IT cars run the advance curve for now , but If i find any bounce , will go fixed on them also.
Thanks, MM

jlinfert
11-11-2009, 08:02 AM
Jeff, are you a diplomat in training?? Phil has shown to know of what he speaks. I have known Phil for about 17yrs, from a distance, but at the same tracks, mostly Watkins Glen. He has shown to know the higher tech areas of making slow shit go fast. like Volvos and VWs.
What are you saying in your post?? It is not clear, not informative. Please elaborate your found data.
I have found no advantage to more lead than about 32 degrees on this ( Non IT legal ) engine. at any rpm that this car runs at( 4700- 7300). More than 35* seems to shorten head gasket life, and often distorts the chamber squish area. I run fixed timing to deal with the vacuum rules at my local circle track. I have found no timing varience, ever.
The car pulls the same with the advance dist also, just idles poorly.
The IT cars run the advance curve for now , but If i find any bounce , will go fixed on them also.
Thanks, MM
I'm not suggesting that more than 32 to 34 degrees total advance is going to make more power. However with a standalone ecu that allows you to control timing precisely and customise your actual curve, a fair amount of power can be found, typically "under the curve" ie below torque peak. This effectively widens the powerband of the engine making it easier and more fun to drive. Example playing with the timing curve on a GTL motor found 15ft/lb wtq at 5000 rpm (numbers went up 15 ft/lb) and 10whp at 7000 rpm. the power curve and torque curve used to be bell shaped and now they are both pretty darn flat. The same thing can be done with any motor when you can truly adjust your timing curve. Of course the amount of gain will vary but the average gains tend to be decent.

pfcs
11-11-2009, 09:11 AM
Hogwash. To see those kinds of changes your fueling curve would have to be terrible.

jlinfert
11-11-2009, 01:28 PM
Hogwash. To see those kinds of changes your fueling curve would have to be terrible.
1: I have the dyno sheets to prove it.
2: Never touched fuel, afr was hovering around 13:1
3: Just because say 32 deg timing works best for peak power (at say 6700rpm) it doesn't mean it's best at 5000rpm or at 7200rpm. If you've ever monitored timing with a scan tool on a modern efi car you'll see that the timing is very dynamic. Even above 4000 rpms. The manufacturer does it for economy and emissions but it can be tuned the same way for power.
4:A year ago I would have said the same thing. but after 9 megasquirt installs and hours of tuning them on the dyno I know differently.

pfcs
11-11-2009, 03:10 PM
"found 15ft/lb wtq at 5000 rpm (numbers went up 15 ft/lb) and 10whp at 7000 rpm. the power curve and torque curve used to be bell shaped and now they are both pretty darn flat"
Since HP equals torque @5250rpm, sounds like you gained about 15hp@5000! and 10@7000-- and that seriously flattened the power curve? Only if you started with a disaster (quite possible w/DIY TuneItYourself systems-where do I start?) But then you say a/f was 13/1 thruout. I hear that you have "data". And I know what I know. Somethings wrong with this picture. For these values to flatten the curve, you would have to start with a huge hump in the middle; if timing was locked at some sensible value thru-out, you wouldn't get this result-what was your timing control and what were your tq and hp peak values and rpms? What dyno?

jlinfert
11-12-2009, 01:40 PM
Bell shaped is a bit of an exaggeration but the torque rose rapidly and steeply to peak at 5300 or so and the hp dropped rapidly after peak at 6800 or so. Torque at 5000 was 115 it's now 130. Hp at 7000 was 130 it's now 140.

Timing has never been outrageous 30-34 degrees max but small changes of 1 or 2 degrees above where a tradional curve would be constant seems to work well.

This is an SIR motor so that may have something to do with the unusual gains. But I still have seen gains on IT motors of 2 to 3% by messing with the timing above 4000rpm.

As far as timing control, the megasquirt is directly driving an Electromotive 4 cylinder DFU (waste spark coil pack). It's using a 36-1 tooth crank trigger for rpm/tdc. Timing is very accurate and contollable to .5 degrees.

All of our testing was done on a portable dynojet, outside air and it's VERY consistant.

While I'm well aware of the fact that standalone systems can create a disaster, I have a lot of experience with megasquirt (as I said 9 installs with tuning in the last year) and once you know your way around it it's very friendly.