Hi guys, I have opened my dead sdecu to see the internal circuit. It was a very hard job. The main IC is an LM124. I am searching to recreate the circuit with an ad595. Fabry Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login
Please explain. An LM124 is a quad OP amp. Must be a simple circuit to recreate. An AD595 is a Type K Thermocouple ic. Are you trying to fix/reverse engineer the circuit or are you trying to recreate/redesign the SD ECU with a thermocouple connected to the cat? As you redraw the schematic, please post it. You appear to be up to something big here. Keep us posted. Kudos!!!
Well done, thanks for the photos. I'd like to mirror Richard's comments... are you designing a replacement for this? I'm guessing that the 324 is acting as a thermocouple conditioning amp in this circuit, so I'm also guessing that you're looking into a complete redesign with the AD chip. Please let us know... this is very interesting. I have 30+ years of analogue circuit design experience and I'll happily contribute what I can. Thanks, Charles
When my company says "Diversity is our life blood" I say yeah yeah yeah, whatever. When I see someone take on a task that would make my eyes blur, with knowledge I don't have and have little interest in obtaining, I will shout DIVERSITY OF INTEREST IS GREAT!!!!!!! Kudos for digging into this. We all add something different to this great place and end up with a very comprehensive list of answers to those who make it to our doors. Thanks for taking this on. I'll be watching your work with keen interest even if I don't understand it.
Looks like the one diode next to the supply wires in the last pic is broke, besides the one capacitor. Might be an easy fix. What is that goo on everything?
The whole thing is potted in a solid brick of epoxy -- it's a minor miracle that he was able to get this much exposed without more damage.
1. R (red) = +12V power to run the thing. 2. N (black) = Ground to run the thing. 3. G (yellow) = A wire that is switched to ground to flash, or solidly illuminate, the warning light in the dash (the other side of the warning light filament is always at +12V) when it thinks the thermocouple is reporting a really hot (flashing) or really, really hot (solid illumination) temperature. 4. B (white) = This one I'm not sure about, but, I believe that it is a wire that does something (not sure what) when it thinks that the thermocouple is below ~300 deg C and does something else when it thinks the thermocouple is above ~300 deg C. It seems to only be used on the "carb with cats" (to fiddle with control of the air diverter valve) and "K-Jet with Lambda with cats" (to "tell" the injection ECU something) models, and is not used on the "K-Jet without Lambda with cats", "KE-Jet with Lambda with cats", or "Motronic 2.5 with cats" models. One of the tricky issues of making such a replacement for others (if one was to do that) is that the connectors (both the 4-pin and the 2-pin for the thermocouple) are not the same on all models -- even though the functionality is the same.
I didn't know the plugs were different. I suppose it could be done with 4 bare wires, maybe solder ended and 4 pins, or with pins attached if they are all the same and the end user could reuse their connector box (for lack of better term) For reference these would be before pics. That has been part of the challenge of these, they are in a brick of solid rather hard goo. epoxy. Image Unavailable, Please Login Image Unavailable, Please Login
Yes, not an insumountable problem, and many models might share the same one as yours, but I know that the later KE-Jet and Motronic 2.5 (that use 141037) have a different body form and connectors (even though the internals have the same functions), but the wires may even be the same colors so could be spliced: Image Unavailable, Please Login
Fabry: I think the transistors you've identified as 2N292 are something else... 2N292 is a germanium part from the 1950's. Could these be 2N2925? That is a popular Si NPN general purpose transistor. There is a discussion about the SDECUs at http://www.ferrarichat.com/forum/showthread.php?t=328364 There are also discussions on another forum where the linear output voltage versus thermocouple temperature is graphed. Use the search function to see what Cribbj (John) has written on the subject here on F-chat. He has done the research on the SDECU, graphed the performance by using a thermocouple simulator to drive it, and analyzed the function. I think his car is a 550, which might explain the linear voltage output he speaks of (I presume a more modern unit might provide an analog signal). He is presently having a Canadian manufacturer of aerospace instruments build dual EGT gauges which also operate the slow down lights. With regard to using the AD595 in a new circuit: I think it is a good choice for this application. It requires a minimum of external parts. Have a look at the application note if you haven't already http://www.analog.com/static/imported-files/application_notes/AN-274.pdf The app. note describes how to scale the output voltage range, improve noise immunity at the inputs, and so forth. If you need a linear output, the AD595 normally produces 10mV/Deg C ... and judging from the graphs mentioned on that other site, the existing unit is roughly 5mV/deg C with an approx. 0.5 volt offset. This would be easy to replicate with the AD595 if needed. With regard to the output that drives the slow down lights: You could easily implement this with, say, a LM339 quad comparator which would compare the output of the AD595 to a set of voltages established by a resistive voltage divider, with the outputs of the comparators driving transistors which pull the lights to ground when you wish to illuminate them. Blinking could be achieved with the venerable 555 timer chip. A more modern and probably cheaper approach would be do use your favorite microcontroller and do all of your signal conditioning and alarm generation in software. I'd think that an AD595 and an AVR would work very well in this application. Your total parts count would be these two chips, a three terminal voltage regulator, some transistors for the open-collector outputs used to drive the lights, and a handful of passive components. BTW, the Mondial T with 2.7 Motronic the specification is given as: slow down light blinking = 900 +/- 20C slow down light constantly on = 940 +/- 20C cut off = 980 +/- 20C I assume it is something similar for the older cars.
Hi Mulo Rampante (bel nick), I think that the component is a germanium component, on it is wrote " Ge2n 295". In this period I don't have so much time because I'm in the process of renewing my actual visa. I'm trying to write down the ecu schematic when I have some time free in the evening. Your help writing down the schematic with the ad595 will be very appreciated. Ciao Fabry
Hi Fabry: In spite of what it says on that transistor, I find it very hard to believe it is a 2N295. (I know the writing on the package seems very compelling). Those markings may be a "house number" (when transistors are purchased in large quantities or to a customer requirement, they often get a "house number" -- only the manufacturer and the customer know what it is.) My reason for saying this is that the 2N295 was completely obsolete back when I was a little boy and Ferrari was making the 250GTO! It is an alloy junction PNP transistor, an example of the very first type of transistor made. Also, I have never seen a germanium transistor in an epoxy package; all the ones I've seen are in metal cases, in very obsolete styles. On the other hand, a 2N2905 is a very common PNP Si transistor... maybe its one of those. That part would be more in keeping with the other components on the board. I would recommend you reverse engineer the board just to understand its function. But I think there is little point in trying to duplicate it, time has moved on. Your choice of a replacement IC is a good one. I will try to sketch up an idea and post it in the next few days... rather busy here myself. In the meantime read the pdf file in the link in my last post. It will give you a clear idea of what's involved in setting up a circuit with the AD595. I'll try to explain some of the other ideas I mentioned in the next few days. I mentioned using an LM339 to compare the output voltage of the AD595 to some fixed reference voltages in order to create the different alarm outputs. This page http://home.cogeco.ca/~rpaisley4/Comparators.html has everything you need to know. Scroll down to the part entitled "4 Level - voltage Detector" to get an idea of what I was thinking of. If you, or someone you know is comfortable writing software for a microcontroller, then you can just sample the voltage output of the AD595 and do all of you alarm generation in the microcontroller, using some of its digital outputs to drive transistors for controlling the lamps and cut-out circuit. This also lets you do data logging, drive external displays, etc. Charles