jleidel

This post is pure greatness.

Alan

Bravo for modelling in a simulator - there are small problems with simulator accuracy depending on the models used - however those are usually manageable - the bigger issue is with accurately modeling the environment in which the modeled circuit will live.

The real challenge in this case is knowing what parameters to use, what's the correct supply range, what are the likely extremes of input signalling, how might the input slew rates vary across conditions, and what's the relative signalling timing and how does that vary across other conditions, for the output loads what are the reasonable load ranges and how are they modeled to look like a realistic motor load.

For the non-standard cases - (e.g. safety engineering) does the circuit handle spikes on the supply or various inputs, what happens if the input signalling is erroneous, what happens if the outputs get shorted, what if various components fail in various ways (e.g. how do you guard against catastrophic thermal runaways in the extremes). Tons of considerations for failure modes... very hard to model - even if you know all the answers going in ... which I'm pretty sure we don't.

So after simulation - breadboard to test in the real environment AND do major component failure evaluation... Auto companies do this quite well - their stuff may not be super reliable but it usually fails safely...

If you analyse Porsche circuits much you will see that they do some very weird things that make absolutely no sense for normal operation - and its almost always to take care of possible failure modes so that it fails gracefully & safely.

e.g. If your HVAC head unit or vacuum system (or both) fail totally - you will still get some airflow to the windshield and some heat... and the blower is separate enough that it will probably still work (that one wasn't quite as well done...) important stuff.

So the cooling fan controller should know if the fans are running normally, shorting the fan output should be detected and be a recoverable fault, OC fan circuits should be detected - Ideally if the controller itself has any problems - the fans should default to run at full speed, if one fan dies the other should run at full speed...

Alan

fraggle

For manual full power fans I was going to hack the controller that's next to the passenger seat. There's a fail mode signal that kicks the fans to high, but I can't remember which input it used.

Lorenfb

"Our lab guys built a software model of a 27-layer PCB w/ both through-hole and surface mounted components."

Hardly what's needed for a very very simple 928 fan controller.
Using SPICE or any computer aided design app for this is a real joke!
The circuit could easily be designed and tested (a bread-boarded version)
in the car in about 2-3 hours.

Bottom line: Where are the results, it's been over 5 months?
Maybe, stop running the SPICE app on your old 8088 PC.

SeanR

What's stopping you from doing it if it is that easy?

928 at last

TOP LINE: SOME OF US HAVE JOBS OR CAREERS
That tend to command our attention for a considerable number of hours each day......
Also, if I understand the project correctly, John's not looking at replicating the existing technology, but improving on it.
I'm illiterate in this field, but I'm not retarded, (excuse me, intellectually impaired), and I'm willing to bet that even I could replicate the exisitng circuitry with a little effort and research.

It's always easy to be critical. Bit more difficult to provide either technological insight or constructive criticism.

To paraphrase Jon Bon Jovi's lyrics...."Have a nice day!"......

928 at last

What Sean said......

jleidel

Loren, wow. What got your vagina all twisted? If you don't like the design methods, you don't have to utilize the end result. Indeed, we could source the original parts and/or solder on a new set of fets. The whole point of this is to improve the overall design and potentially allow for the future addition of more cooling fan options.

...and its not an 8088. Its an 8087. We need float.

fraggle

whoa life got in the way of a project? say it isn't so! that never happens!

dr bob

There are a few "ism's" that hang on the office wall. Maybe more than a few. Prominent is one shared by my mommy: "Anything you can do better than I can do, you can do." For a long time I'd suffer through potshotting and sniping, second-guessing and less-than-constructive criticism. Then I started assigning the 'better way' task to the person who thought that they had a 'better way'. Show me, don't tell me. Life got easier, I was able to sleep bettter at night, more frequent calls to the doctor for e*******ns lasting longer than 4 hours, etc.

--------

There are only a couple people on my "ignore" list here. Lorenfb heads that list. FWIW. YMMV and all that.

Lorenfb

"Also, if I understand the project correctly, John's not looking at replicating the existing technology, but improving on it."

All negatives aside, actually that original 928 ECU works just fine. Of all the
ECUs I've tested from shops that claim they're bad, most all (99%) are O.K.
and the assumed failures result from inadequate troubleshooting. With the
exception of the LH ECM, most all the 928 electrical system elements are quite
reliable and don't really need a redesign.

dprantl

Someone wants to use their personal time and expertise to provide a modern equivalent to the S4 fan control unit, possibly with more features than the original. They are not doing it for personal gain, but for the good of the 928 community and because they can. You are criticizing this because... ?

Dan
'91 928GT S/C 475hp/460lb.ft

17prospective buyer

Put me down for one of these Leidel.

rgs944

Keep us updated, I need to get my fans working properly on both cars.

rgs944

Is this project still in the works or did it hit a snag?