Exhaust ECU aka SDECU

Jim,

Do you think you could build a circuit to output like this graph? Maybe a new business for you. I would be glad to beta test it. We would want to use any K temp probe plug and play to the processor and to the Motronic and end our pain with the Ferrari SDECU. Dr. Bob came up with a sort of close off the shelf exhaust monitor but still gets CEL's until warm because one part of that circuit does not output what the motronic wants to see until it warms up.
Image Unavailable, Please Login

 

FBB,

What I've got rigged up now for my 328 doesn't output a voltage, but instead just completes the SD light circuit (blinking or solid-lit) if the threshold temperatures are reached. (Actually I've got it set to blink the SD light in a way that tells me what the cat temperature is under normal operating conditions, and it also logs the temperature to an SD card, both of which are interesting for diagnostic purposes.) I think that adding a digital-to-analog converter to my current set-up to output the required range of voltages would be feasible, as long as the required voltage is in the 0-5V range, which the plot suggests is the case.

Is supplying the plotted output voltage as a function of cat temperature the only functionality that the TCU needs to have, or does it also need to take care of switching the slow-down light on? I thought I'd read somewhere in this or another thread that the Motronic itself deals with the SD light (based on the voltage it's getting from the TCU), whereas for the older cars (328s, etc.), the TCU performs that role by completing the SD light circuit, as I note above.

 

All the 5.2 motronic needs is the voltage output and the light is regulated by the motronic. So the SDECU just needs to process probe signal to output a corresponding voltage.

 

I assume that the white "trigger" wire carries the voltage output from the SDECU to the motronic?

Do you have a source for the SDECU connectors you show in the photos in this first post in the thread? The two-wire connector is for the thermocouple, right?

Once I finish my initial work on my 328's SDECU substitute, I'd be interested in making up a voltage-output-to-motronic version that I could send to you for testing.

I haven't been able to find a source/sources for the connectors on the older-style SDECUs, so if anyone's got a lead on that as well, I'd welcome the info.

 

Yes that is what white does.

I will hunt for connectors and get back to you. It would be great to make a plug and play. But if an aftermarket systems works better than Ferrari which is easily done I imagine then we could just cut off the old connector and do a nice weatherpak one.

 

I have a source for the 4 port plug on the SDECU and the 2 port as well. The 2 port is the thermocouple and 3 of the 4 ports are used on the 4 port connector. ( ground , DC voltage and signal wire). I am speaking of the viscovini SDECU in 355/360. I not familiar with 328 but if they are identical I can help you source the connectors/pins , etc.

 

Thanks. I'd be interested in pricing a set of the 4-port and 2-port plugs that correspond to the 355/360 SDECU. Looks like it's the male plugs that are needed for the ECU end.

The ECU/TCUs on 328s and (I think) 308s have different style connectors, which varied a bit over time. Here's a pic of my '89 328's ECU:
Image Unavailable, Please Login

 

Great Bob you were 1st on the list to contact. Can you post a link to the female connectors like on the SDECU in my 1st post?

 

This thread has some nice photos of the 355 SDECU connectors:

http://www.ferrarichat.com/forum/348-355-sponsored-bradan/378429-thermocouples-slow-down-lights-related-cel%92s-how-conquer-them.html

 

Will do when I get to the office today Carl.

 

Here are the 4 port connectors. These are the females, the link for the mating ones (males) is on this page as well. You need to order the housings, the pins and the seals.

https://www.eficonnection.com/eficonnection/ConnectorDetails.aspx?ItemId=435

Here is the link for the 2 port connectors for the thermocouples. The guy who runs this site happens to be a Jersey guy. No affiliation but a nice guy who just bought a red NSX. His name is Joe and you will be very pleased with his site and service. Hè has a nice selection of crimp tools and extractors , etc. and infirmation on which to use. Just ask him.

http://www.cycleterminal.com/amp-superseal-connectors.html

 

Awesome DrBob! Now whatever Jim builds can be plug and play. Jim I'm pretty sure this is the unit DrBob experimented with that works but not completely. It might be possible to just crack one of these open and design around existing circuitry. We need to make a robust unit that will take say 350*F temps and it should last a long time.
Image Unavailable, Please Login

 

Yes that is it and I have three of them installed in my car and they work fine less the blinking sdl at cold start up due to the .2v initial output.

I have had many conversations with EGT about thiß and most recently they told me they are working on a unit that will be able to be user programmed for variables like the initial voltage. They said they will get it to me to test as soon as it is ready. Stay tuned.

I have a unit on my work bench from another outfit that does the same thing as the EGT unit does. Our cars want .5v on cold start.

 

Right. That all sounds good especially user programable. I don't think we even need an SDECU to log or output JohnC's graph. We only care about 4 things:

1.5V keeps the MIL off
3.7V Flashes a MIL warning 1650*F
3.91V MIL solid lit and Motronic ECU goes into limpmode 1720*
5+V destroys the MOtronic ECU

So if we could have a device that just outputs those voltages when those temps are constant and drops those voltages as temps fall we would be golden. Not being the EE guy I have no clue what it takes to make either circuit but it can't be that hard.

 

Thanks, DrBob.

FBB - I've just ordered a digital-to-analog converter chip that ought to allow me to program the microprocessor I'm using to output the required voltages you list, at the threshold temperatures for those voltages. (That may be the same basic approach that EGT is taking, re. programmable parameters.) The circuitry for that is actually simpler than what I presently have built for the older 328-style SDECU which does the switching on of the SD light rather than that being handled by the motronic.

But you hit on a big concern I have regarding using the kinds of standard electronics bits I'm working with - the ability to withstand high temperatures. Everything I'm using (basic off-the-shelf hobbyist/tinkerer hardware) is rated at, at best, 125 deg. C operating temperature. Potting the electronics might help insulate things against transient heat spikes, but not sustained high temps. There's specialized high-temp hardware out there, but it would mean bringing more expertise to bear on the problem than I have, since I likely couldn't use the pre built sub-components I'm relying on at this point.

If I can make a prototype that's worth testing (and hopefully you're willing to give that a try), then a possible way forward re. engineering for high temperature operations might be for me to stop by Dave Helm's shop to talk to him about possible ways forward.

 

Vescovini encases their PCB in potting compound. They switched from a black to a green material on the units I am familiar with. I have some dead ECUs here and was going to try and determine what typically fails on them Revs did a nice write up on the solder joints on the connector pins but some tried redoing those joints and that wasn't the issue.

At $400 each, the real issue with the Vescovini units is the price. We are fine replacing 02 sensors that can be had for under $100. If a EG temp module could be made for less than half of the OEM cost, then replacing them when they die would just be a cost of doing business in my mind.

I have sourced and make up complete thermocouples that I sell for around half the OEM replacement price. Having an alternative to the Ferrari OEM pricing for this stuff will allow these systems to be maintained and repaired much like changing 02 sensors.

Vescovini has the connectors mounted on the PCB which makes for a nice package that can be considered after the board is proven. The thing that baffles me is all of the stuff on the Vescovini board to do this job. The EGT board is about 1"x1" and has minimal stuff on it. I asked Dave Helms if he had any clue what all that electronics was doing and he said that he didn't really. In the end all it does is transmit a voltage related to the EGT down the signal wire to the car ECU. The simple EGT unit that I have installed in my car works fine except for that initial voltage issue.



I think encasing the finished boards in a good, high temp, potting compound would be good enough and adequate.

 

The Vescovini units for the older cars (potted with black material) are even more expensive - about $1000 each, last time I checked. If it turns out that standard electronics can be used rather than needing special high temperature hardware, then building a replacement that could sell for $200 sounds feasible. As you say, at that price then there's no need for a super-engineered version that could last for 20 years.

In any case, the hardware I'm finding online that's advertised for "high temperature" use is only rated up to 175 or 200 deg. C or so.

 

Well for the sake of easy to buy easy to sell plug and play units we need to survive high heat or whatever kills electrical components. We need Japanese OEM robustness to really stamp out the pain that is the Ferrari SDECU. So what does it take to achieve that I do not know. It is possible to relocate the SCEDU in the passenger compartment with just longer wires. However, the footwell of a 550 will get near 200*F, I have measured it, and I do not know how hot a black car gets inside on a 100*F ambient arizona day. Engine bay temps are also like a 200-300*F oven on a 550.

 

From what I can find regarding availability of specialized hardware able to withstand really high heat, an ECU built from such parts would probably be expensive. The table below is from a Toyota design document for 2005 vehicles. It suggests that Toyota ECUs perhaps are designed to handle up to 110 deg. C (230 deg. F.). The components I'm using for testing are rated to 85 deg. C (185 F) but it might be possible to build the same package using off-the-shelf components rated to 125 C (257 F). Anything rated higher than that would probably beyond my tinkerer/hobbyist level of expertise, and maybe not worth the cost anyway.

What I could try to do is make the 85 deg. C rated version as a prototype for initial testing to see if it's worth pursuing further. Even if it's not suitable for hot locations in 355s or 550s, it might still be acceptable for the placement locations in the older cars (in a compartment inside the luggage space in 328s, for example). If nothing else, it's almost certainly a better/safer option than folks driving around with cats but with disconnected SDECUs.
Image Unavailable, Please Login

 

Sounds good! The most important thing in your prototype is to make sure the output line voltage can never exceed 5V. John fried his ECU by accidentally letting the trigger see 12V battery power.

 

When you say older cars, what models are you talking about? These models have EGT modules? Do they have SDL systems? What does the car do with the output of the thermocouples/modules?

 

I'm referring to 308s, 328s, and TRs (I think). Not sure about others of that era. I've mainly just paid attention to 328s since I'm trying to make a substitute SDECU for my car. These cars indeed have dashboard slow-down lights tied to single or multiple thermocouples in the cats, linked to single or multiple SD/Cat ECUs. The control unit reads the thermocouple voltage and, based on voltage, grounds a wire that's connected to the dashboard slow-down light, which then blinks the light or turns it on continuously, depending on cat temperature.

There's also another output connection that apparently plays a role in some versions of the K-Jetronic, but according to Steve Magnusson, that output doesn't actually control anything on our cars:

"Question posted by jmaz:
Can anyone confirm that, for the 308/328 series (particularly my '89),the only role of the catalytic converter electronic control unit is to trigger the slow-down light, and that it doesn't interact with the fuel injection system ECU?..."

"Reply from Steve M.: That is correct. On the K/KE-Jet F (and carbed F), there is no intelligent connection between CC ECU and the injection/ignition ECUs -- the CC ECU just runs the overtemp warning light."

So, these SDECU/CC ECUs (as far as I understand them) only control the SD light and don't do anything else such as shutting down an engine bank or affecting fuel mixture. It would be interesting to know what that is sent through that additional wire though.

 

Ok thank you. I have seen the thermocouples for those cars but wasn't sure what they are controlling. So it is just the lights.

In 355/550 it shuts down engine banks at a higher voltage.

I'm sure I could make up thermocouples s for those cars for around $160 also if I have a demo.

 

What are the implications of there being a delay of a second or two, after the ignition is turned on, before the microcontroller-based ECU would start supplying 1.5V? During that delay (the time needed for the microprocessor to start), the voltage output from the ECU would be 0V. This isn't a problem for the SD light/Cat ECU version I've made for my 328 but it sounds like it might be for a version that would work on a 355 et al.

 

The microcontroller-based SD/CAT ECU version I've now put together for my 328 (shown below) is built with "automotive-rated" parts (AEC-Q100), which are typically rated to 125 deg. C operating temperature, along with some other parts also rated to 125 C but not with a specific AEC-Q10* rating.

I suspect that's good enough, but I'll do some testing using a high-tech temperature chamber sitting on my kitchen counter (a.k.a. my toaster oven). But I have to wait until my wife is out of the house.
Image Unavailable, Please Login