outathebox

Building a 3L turbo for track use and although lots of guys have made it work, I'm personally not really enthralled with a crank mount 60-2 trigger wheel. The good kits are $500 and IMO it still seems like a vulnerable mechanical kludge. I'm a retired EE with lots of time but limited $$ and i'm thinking how hard can it be to converts the stock 132/tdc ref mark sensors to 60-2? ... the answer was a bit harder than i originally anticipated, but it's working reliably on on the bench with a purposely "jittery" 132 tooth signal .from 100 rpm to 7K. Here's a shot from logic analyzer::

SirLapsalot

Agreed regarding the front mount 60-2 options, none of them are fantastic options. A 60-2 flywheel is the best option but also is quite expensive, and maybe harder to come by.
So what is your intention with that logic analyzer?? What SW/ECU are you planning on using for your engine management? I do know of at least one ECU supplier that has the 944 132tooth trigger option available in their SW by the way.

944M3

I’ll bite, how did you do it?

Really interested in this. I’m still on the fence whether to build 3.0 turbo but this might be one check mark towards going for it. I have an S2 block and head.

Droops83

@outathebox I'm definitely following this with interest. I have the VEMS ECU in my '86 944 Turbo, winch is one of the few that plays nice with the stock 132+1 arrangement, and even then some VEMS users have had trigger issues. Mine has always worked very well, however.

From what I have read, the chip in the original Bosch DME that converts the analog speed signal to a digital square wave is the most expensive and sophisticated item within----but surely something like that must be cheap and easy in 2022, at least for an EE like yourself?

outathebox

@SirLapsalot The proto is buith and running on an ESP32 , written in c using the ESP-IDF APIs . The logic analyzer is showing the input and output of the prototype. Top two traces are the stock 132 lywheel and ref sensor (after converting from raw sensor signal) The bottom trace shows the 60-2 output.... the position of the blank teeth is totally adjustable. I'm planning on using Haltech..not becuase they are the best, but because it's adequate to do the job (goal is about 400 reliable track hp) and their end user support is way better than most others .AFAIK the only ECU suppliers that are doing 132 are not excatly mainstream suppliers and IMO Haltech is a better overall value even though there is a bit of a premium...and likely more reliable. Having a 132 to 60-2 converter means you can use any engine management ECU out there. It's not at all difficult (now that code is "done" ) to convert from any pattern to any other pattern so i expect this could be useful for any engine that has or can be fitted with flywheel sensors that aren't a 60-2 pattern. .

@944M3 short answer "Magic"
The slightly longer, more truthful answer - basically a software based phase lock loop. Not really sure where i'm going with this ... guess it depends on the interest and wheter or not i have any time to spend outside of my own 3.0L build this winter. One should also keep in mind the 90-10 rule in engineering - you get 90% of the way in 10% of the time..... "The devil is in the details" . Buildiing one of anything is a piece of cake....building N of them is a bit more challenging. . .

outathebox

@Droops83 "The original Bosch DME that converts the analog speed signal to a digital square wave is the most expensive and sophisticated item within----but surely something like that must be cheap and easy in 2022"

yes in the last 35+ years it's gotten waaay easier. it 's a $5 off the shelf chip that's more reliable when coupled with some decent firmware. . I don't count on it , but what also helps is that most mainstream ECUs are tolerant of less than perfect trigger signals, becuase if you have a crank mounted trigger wheel with a less than solid mounting you need it.

SirLapsalot

Interesting, was expecting this to be for a mega/microsquirt or VEMS application. You may want to do a bit more research on current standalone ECU's though, as your assumptions are bit off. Cool project nonetheless!

So it sounds like you'll be simulating a hall sensor by outputting your systems 62-2 digital signal to the "trigger 1" input pin on the Haltech? And also the TDC signal you'll output to the Haltech as a hall sensor for position/home reference? Or is that not correct? How confident are you that you'll still get an accurate reading on the Haltech at higher rpm? Not an EE but seems like you could be introducing possibilities for offset/error between the physical VR sensor reading on your flywheel, your systems conversion, and the haltech's calculation of engine speed. I'm sure your system can read accurately to 7krpm but what kind of time delay will there be between your output and Haltech's processing of the signal?

944M3

Maybe I’m missing something but it was my understanding that EMU ECU and Link support 132-2 setup. This is not enough for individual plug firing but fine for batch firing, which is what stock is. Someone correct me if I’m wrong.

Im personally leaning towards EMU ECU Classic for my build and keeping batch firing.

outathebox

@SirLapsalot - when you say my assumptions about ECUs are off ...which of my many assumptions?

The ref signal from the flywheel is not passed along it's merely a reference for crank position relative to TDC. The home sensor on ECUs like Haltech is to figure out where you are in the engine cycle since there are two revs between an individual cylinder firings. I beleive the home signal is only required for sequential injection and if required, it usually comes offf the cam shaft. Howver it's timing is not even remotely critical since all it determines is wheter you are in the 1st or second revolution of the cycle It's not used to calculate ignition timing

The delay between "input" and output * is very very small, a couple of microseconds (1/8 of a degree at 7K ). Not that significant.IMO especially since most of that differential is consistent and you would account for it in your ignition map when you did your tuning. . It's hard to see, but on the LA picture every 5th 60-2 pulse aligns with every 11th pulse of the 132 signal. What might matter more is that any differential between the ref sensor pulse and the 60-2 blank tooth gap is consistent.because then the ECU can calibrate this out. This is similar to the calibartion you would do with a 60-2 crank sensor..

SirLapsalot

Your comments about suppliers that offer 132 tooth options not being "mainstream" and overall values/reliability. ECU Master and Link are two that both offer 132 tooth options and are both "mainstream." Link's G4X line has a faster and more capable processor than Haltech's Elite series and I'd argue Link has better support and also overall makes a better ECU. Haltech obviously makes a good product but if one wanted to use the OEM 132 trigger without any additional processing of the trigger signal they could just install a Link ECU. No experience with ECU Master's EMU but their other products I've used are really good and are widely used/respected in the industry.

Sorry, I misused terminology there with Haltech. They do use home as their engine position reference (yes from the cam typically) for sequential injection. I saw you included it in your screen shot so thought maybe you'd be passing it along as well, but there would be no need as you said.

Not trying to downplay or argue against your solution at all. It's sounds pretty damn cool.

fast951

Nice fun project. Converting the 951 132 crank pattern to a different more common trigger pattern is doable. I would have chosen a different output pattern than the 60-2.

Each tooth on the 132 crank wheel is 2.727 degrees, while on the 60-2 it's 6 degrees.
Every 2.2 teeth of the 132 wheel equates to 1 tooth of the 60-2. How do you handle the delta of 0.2 teeth (or 0.54 degrees)?

When you start testing on the car, you may need additional input signal filtering (additional HW). Lots of "little things" to worry about, detailed fault handling will consume time.

Keep us posted.

Nowanker

If you have an ECU that'll support sequential, why not take advantage of that...
Seems like it would be mostly irrelevant where the cam trigger pulse falls... it's either on power or exhaust, one or the other.
Counter intuitive, but the location of the cam trigger signal IS critical on on AEM Infinity, and apparently not user-adjustable.
Haltech says it can be easily configured in their stuff.

Nice work on your converter!
Might have some of interest within the LS engine world.
GM switched from 24 tooth wheels to 58 tooth (60-2?). Lingenfelter already sells a converter box, but from the feedback it seems a little sketchy.

outathebox

@944M3 i completely understand your choice but not supporting sequential operation doesn't make sense for me personally . AEM and Haltech both told me the number of teeth including "missing teeth, had to be a submultiple of 360 ie, 60, 36 etc. Anyway the ECU is purchased and project is in motion so no going back now

@SirLapsalot yes i agree they are mainstream, i stand corrected and no offense taken. To be honest i kinda cut some cormers on the reseach of ECU mfg, ... but as i mentioned above the cake is about to go in the oven.

@fast951 as i mentioned above with the 132 input and the sixty tooth output they actually align 12 times a revolution (30 degrees) - . 11 teetrh*2.72727 = 5 teeth *6 degrees. it also aligns every 1/4rev (33/15 teeth) and every 1/6 rev. (22/10 teeth)

what trigger pattern would you have chosen and why?

60-2 just appeared to be universally acceptable, however any output pattern is just a small matter of firmware tweaks. BTW pretty sure the pattern shown is not what the output from a hall sensor would look like. it's closer to what a VR sensor does.depending on the profile of the blank tooth portion of the wheel.

You are spot on about testing that's where the 90-10 rule rears it's ugly head. I plan to have the firmware do some internal consistency checks and aesssment of incoming signal quality. and there is always the trusty mixed signal scope with deep memory and the ability to record signals for a very long time. From what i understand the Haltech has "tooth" diagnostics and has a limited degree of forgiveness with tooth tming as poorly mounted hall sensors don't provide the best signal quality.

@Nowanker - agree with you on sequential, , the alternative is more boost..not a hard decision.. Was not aware that cam sync was critical on AEM. AFAIK ift's not that hard to get accuracy within 5 or so degrees with cam sensor. I beleive a lot of the issue issue is the slop/jitter in the cam belt drive - Thanks for the heads up on the "older" 24 tooth LS1 crowd

944M3

“i completely understand your choice but not supporting sequential operation doesn't make sense for me personally . AEM and Haltech both told me the number of teeth including "missing teeth, had to be a submultiple of 360 ie, 60, 36 etc. Anyway the ECU is purchased and project is in motion so no going back now”

Got it, you get to keep the stock flywheel and have the ability for sequential with your method.

Since we’re learning, what benefits do you expect going sequential over batch?

outathebox

@944M3 I am totally unqualified to answer that question

I'm not club racing any more, but for better or worse some mindsets just stick with you . Being able to adjust mixture/injector timing and ignition timing on a per cylnder basis was just considered a "best practice"...you can't do that unless the ecu knows which cylinder is next to fire. ..All the ecu's i was considering support it and i didn't see a big initial cost differential in my case. That being said the biiger investment is probably in spending the time/money tuning to get the advantages of "sequential" . if you are tuning foir the street might be a very different answer. i just googled around and found what looks like a good dscussion on injection timing supported with evidence from dyno results as oppossed to hearsay. https://www.hpacademy.com/previous-w...haltech-elite/ Per cylinder ignition timing is another related subject.