Lorenfb

"It would lose track of this as soon as you turn off the ignition though, wouldn't it? Maybe the sensor was used so that it can use the established map at startup instead of having to figure it out from scratch?"

That's true. I don't know whether this is saved (CMOS static RAM) in the EZK.
I haven't done that much reverse engineering on the EZK (they never fail).
Maybe John knows about this.

Louie928

I don't think the individual cylinder knock information is saved. If no knocks, it reverts to the mapped value within ten seconds. The knock counts are saved on later EZ-Ks for diagnostic use. It seems as if they could have made the EZ-K so it would do individual cylinder retard without the Hall sensor if they had wanted to do it. The Hall sensor is really needed for sequential injection. It doesn't go to the LH for that function though. Speculation... I wonder if sequential injection was planned at some point and a bean counter said "Nein". Could be Porsche used an off the shelf ignition controller that was designed to be used with a sequential fuel controller. Guess it doesn't matter. It's what we have.

I had an EZ-K function fail on my '87. The engine still ran normally. It couldn't pass information from the EZ-K to the ST was the problem. John fixed it. I think he fixes quite a few EZ-*

Lorenfb

Most ECUs in the 928 are very reliable with exception of the LH which uses a hybrid circuit versus
all monolithic ones as in the EZK. Monolithic circuits for the most part are more reliable than hybrids,
e.g. sealed package. The LH hybrid exhibits a higher than normal failure rate mainly because
of it not being fully protected from overvoltage conditions, e.g. a charger used on a fully
dead battery. The same type of LH failure occurs on Volvos.

Many ECU failures occur mostly because of; wiring shorts, poor grounds, overvoltage, or poor
troubleshooting techniques.

John Speake

Hi
Hall and sensor faults are stored in the EZK SRAM for viewing later with a diagnostic tester. As louie says, the dynamic knock retard info is not stored in any way.

Hammers and Spanners can count the number of knocks in a 10,000 duration period, and give a knock count total. According to Porsche, 50 or more knock registered indicate a fault, apparaently less than 50 is considred "normal".

So when I remap the EZK I try to set them up for less than 10 knocks or so for a quick 1500-6500rpm run through a suitable gear.

The EZK are pretty relaible, conpared to the dreaded LH. Although I had a little pile waiting for me to repair when I finished my EZK test jig earlier this year. The faults were relatively easy to fix.

1slo928

Hi Jhon I was wondering what you found when going over the failed ezks is there a repair that can be done and how can you test the Ezk to find out if it has failed
Thank You Chris
PS I will read more but I am trying to purcase a car that needs to be shipped and it died before being loaded.

John Speake

Hello Chris,
They had various small failures, random. No real pattern. I wouldn't recomend a DIY attempt at repair.

Try and swap the EZK into someone elses car, or their good one into yours, assuming you have already swapped out the EZK relay

H2

During testing, is there any chance than swapping a defective EZK into a car with no problems could cause any kind of damage to the properly functioning car?

Harvey

Lorenfb

'During testing, is there any chance than swapping a defective EZK into a car with no problems could cause any kind of damage to the properly functioning car?'

As I said before, the EZKs are VERY reliable and DON'T really fail unless someone damages one, e.g.
by jumpering around with ignition modules.

"Many ECU failures occur mostly because of; wiring shorts, poor grounds, overvoltage, or poor
troubleshooting techniques."

H2

Yes...but what I mean is, if your buddy's car that's not running has a wiring short or overvoltage problem that burned out his EZK in the first place...and you're helping him troubleshoot by loaning him your good EZK to eliminate that as the problem...it would seem that the wiring short, etc. in his car could also burn your unit out? So, perhaps it would be better to take his potentially defective EZK and try it in your known good running car? What I'm understanding is that the former situation could be hazardous to you EZK but the latter situation may not be? Am I right?

Harvey

Lorenfb

"it would seem that the wiring short, etc. in his car could also burn your unit out?"

That correct! So the better approach is to place the bad unit in the running car, always.
This is case unless it's some type of high power unit which could damage the good car's
wiring. Also, if ECUs communication with other ECUs, a bad ECU in a running car could
damage the running car's other ECUs. This, though in earlier cars (928s), is of lower risk.

Most "no spark" conditions in 928s result from a +12 source (relays), a pickup
sensor, coils, or spark modules. Remember, the EZK must function properly first
before the LH will produce injector pulses.

BrianG

Excellent idea! Certainly the dead car would have nothing to lose, and the functional once should remain functional. I have never heard of a dead EZK or LH causing any issues with an otherwise healthy car....

Tony

All i have to say is .... to John!

Cool shtuff!
Thanks for all your work.

drnick

i can add an endorsement as to how much fun it is to drive and play with all this but the real value is being able to instantly see what the engine management system is doing either in real time or via datalogging. very informative to see which cylinders were knocking and under what conditions, heat, rpm and load etc this is an excellent tool.

i quizzed john about a further adaptation of his sharktuner, one that was permanently interfaced with the LH and EZK with a handheld PC driving it, that would be much more fun than the "retard" light. at the moment it seems you need to sharktune after every mod if you want to optimise running, but with the system installed permanently you would be sharktuning continuously.

Vilhuer

If handheld can run normal full scale PC 32 bit Windows there shouldn't be any problems in running it right now. I'm not sure if current software works in Windows CE or what ever these latest handheld operating systems from Microsoft are called. In any case it should be possible to port Visual Basic program like ST into Microsoft operating systems.

John Speake

Thanks Tony, but my partner in this development - Niklas in Finland - has made it possible.

He's a software genius ! Hats off to Niklas !

Tony, are you going to use a SharkTuner soon ?