SwedeSpeed - Volvo Performance Forum banner

1 - 14 of 14 Posts

·
Registered
Joined
·
502 Posts
Discussion Starter · #1 ·
Alright so I have been intrigued by the raspberry pi and adruino projects for a while now and just came up with an idea. Please feel free to shoot my idea down but be sure to give actual reasons lol.

So everyone knows that the magnetti marelli (sp?) Throttle body's are complete garbage and the newer style Bosch variants are much more reliable, correct?

That being said, from my understanding, the magnetti throttle body's include the module to run the actual throttle blade motor, interpret tps signals, etc etc. The Bosch TB's do not include a "module" and rely on the ecm to control throttle actions.

My idea is this: Use a dev board such as the arduino or rPi to act as the "module" for a newer style Bosch unit. Their are shields that have provisions to run DC/stepper motors with ease as well as handle signals (tps). It's all about how to deal with these inputs and how to send them back to the ecm seamlessly.

I understand that our magnetti TB's are "coded" to our cars and can be "recoded" if for instance we get one from another car of the same year in order to work correctly.

My main concern with this entire idea is the above mentioned hurdle. What would be stopping me from "downloading", in a sense, the magnetti module and re-implementing it into a dev board so that the car will play nice and be able to send/receive data nicely?

Like I said, feel free to crush my dreams but be sure to " learn me something" in the process.

Sent from my HTC One_M8 using Tapatalk
 

·
Registered
Joined
·
198 Posts
My main concern with this entire idea is the above mentioned hurdle. What would be stopping me from "downloading", in a sense, the magnetti module and re-implementing it into a dev board so that the car will play nice and be able to send/receive data nicely?
The "download" would be a code protected binary, so you wouldn't get any info from it.

About the idea in general: Yes it is theoretically possible, but no, I don't think you will achieve it. Programming a microcontroller to control some motors isn't a big problem itself. You wouldn't want to take the RPI though, because from what I know there still isn't any realtime OS for it. The problem is, normally any car ECU, and that is what you would be building, has to offer diagnostic functions an corresponding bus-messages. E.g. measuring how much power the motors use and therefore deciding if it is stuck etc. They are defined by the OEM and lead to check engine lights and low power mode if they are not handled correctly. I don't see how you would get the documentation needed and even if you have it they are a PITA to implement.
 

·
Registered
Joined
·
502 Posts
Discussion Starter · #3 ·
I would assume that anything vin specific or module would use some form of encryption but the question (or me thinking out loud) would be what form does it use.

As far as I know, any small-mid sized Linux platform runs on the RPI. There are many more shields for the adruino platform however from what I've been reading related to CAN.

I suppose the only way to figure out what these can messages would be as far as diag functions would be to reproduce certain occurrences such as stuck open/closed, etc.

Like you said though, handling one or two sensor signals and a motor is cake. It's making the vehicle communicate and function that would be the real endeavor. Connecting to the harness @ the throttle would be my best bet as far as sniffing the module I feel like. Since that would be right in-between the etm and the ECU so any communication between the two would be intercepted before being processed by any other module.

Sent from my HTC One_M8 using Tapatalk
 

·
Registered
Joined
·
198 Posts
I would assume that anything vin specific or module would use some form of encryption but the question (or me thinking out loud) would be what form does it use.
I don't see why there should be encryption on the communication of the mentioned module. The VIN is all over the place, so there is nothing to be protected there. But the software on the ecu will have some protection against copying it and using it on a copied PCB (because that is how cheap, unofficial replacement parts are made).
As far as I know, any small-mid sized Linux platform runs on the RPI. There are many more shields for the adruino platform however from what I've been reading related to CAN.
The last time I tried, there were no working RT-patches for the RPi Kernel tree, but you would need them to react timely on any input. I don't really understand the hype about "shields" soldering together a chip and some auxiliary parts isn't hard. The only important part are the drivers for set chips.
I suppose the only way to figure out what these can messages would be as far as diag functions would be to reproduce certain occurrences such as stuck open/closed, etc.

Like you said though, handling one or two sensor signals and a motor is cake. It's making the vehicle communicate and function that would be the real endeavor. Connecting to the harness @ the throttle would be my best bet as far as sniffing the module I feel like. Since that would be right in-between the etm and the ECU so any communication between the two would be intercepted before being processed by any other module.
On CAN you can connect in "listen only mode" and sniff all the messages, that's right.
To make sense of them would be a lot of work since you will not only receive the messages from the throttle but also from all other parts on the high speed CAN.

I did something like this once to emulate a CD changer on my old C70. There it was a direct connection and it was still a bit of work to get all messages right.
 

·
Registered
Joined
·
502 Posts
Discussion Starter · #5 ·
I don't see why there should be encryption on the communication of the mentioned module. The VIN is all over the place, so there is nothing to be protected there. But the software on the ecu will have some protection against copying it and using it on a copied PCB (because that is how cheap, unofficial replacement parts are made).

The last time I tried, there were no working RT-patches for the RPi Kernel tree, but you would need them to react timely on any input. I don't really understand the hype about "shields" soldering together a chip and some auxiliary parts isn't hard. The only important part are the drivers for set chips.

On CAN you can connect in "listen only mode" and sniff all the messages, that's right.
To make sense of them would be a lot of work since you will not only receive the messages from the throttle but also from all other parts on the high speed CAN.

I did something like this once to emulate a CD changer on my old C70. There it was a direct connection and it was still a bit of work to get all messages right.
Yea the encryption I was talking about would be when dumping (if at all possible) the etm module itself. I guess the hype with shields is just the user friendly aspect of it. Most if not all of the hardware config is already done for you and, like you said, the drivers are the deal breaker to get everything to work.

I would be interested in seeing how a dice communicates as far as software upgrades go but for that, I'd need a fully functional Vida subscription with the ability to do updates. Doubt a dealer is going to let me "sniff their units"

Dice has to send a certain message/code/signal through to the etm to enable it to be programmed. Figuring out what that particular signal would be, in my line of thinking, would be the cipher for said dump file. Reverse engineering that code would be the issue. In the end, all it does is send output signals to the DC motor and sends the tps signals to the ecm. The rest of the underlying code could be computed on a much faster processor using the adruino interface.

My main concern is I don't have the proper setup to sniff all of these scenarios that the ecm needs to see without throwing a limp mode of CEL like you mentioned earlier. Mainly a functional dice/Vida subscription.

Sent from my HTC One_M8 using Tapatalk
 

·
Registered
Joined
·
120 Posts
I work on the Magnetti Marelli throttles in my spare time and have rebuilt about 110 units over 4 years.

I can reprogram them with any program numbers that are viewable in Network in Vida. By now my programmer has over 40 ETM programs/firmware versions in his database. From my understanding the file is a 16KB file size for any given firmware version.

The engine's heat acting on a power supply and logic controller within the Magnetti ETM enclosure is, in my opinion the number 1 reason for failure followed closely by the carbon film & scrapers (which can easily be substituted by Sacer contactless)

Why not create a set of leads external to the throttle which handle all input/output of throttle signals to the board which can be housed in a finned circuit box away from the motor's heat?? The throttle body could stay bolted to the manifold and simply handle motor and throttle angle functions which could report to the vented board box. Just an idea and of course very time consuming to implement.

It wouldn't solve the problem of the expanding magnet problem which can be dealt with by removing the aluminium sheathing around the magnet in pretty much the same way that Maserati have dealt with this problem.
 

·
Registered
Joined
·
502 Posts
Discussion Starter · #7 ·
I work on the Magnetti Marelli throttles in my spare time and have rebuilt about 110 units over 4 years.

I can reprogram them with any program numbers that are viewable in Network in Vida. By now my programmer has over 40 ETM programs/firmware versions in his database. From my understanding the file is a 16KB file size for any given firmware version.

The engine's heat acting on a power supply and logic controller within the Magnetti ETM enclosure is, in my opinion the number 1 reason for failure followed closely by the carbon film & scrapers (which can easily be substituted by Sacer contactless)

Why not create a set of leads external to the throttle which handle all input/output of throttle signals to the board which can be housed in a finned circuit box away from the motor's heat?? The throttle body could stay bolted to the manifold and simply handle motor and throttle angle functions which could report to the vented board box. Just an idea and of course very time consuming to implement.

It wouldn't solve the problem of the expanding magnet problem which can be dealt with by removing the aluminium sheathing around the magnet in pretty much the same way that Maserati have dealt with this problem.
Hmmmm so what you are saying is basically retain the magnetti unit but re-locate the logic and controller side to somewhere else external of the TB housing. This would be much easier it seems like.

Since you work with these every day, do you believe the firmware that you get off of the units you work on could be controlled by something else with the correct outputs?

Your idea would definitely work and be, IMO, easier to accomplish but I'd still be curious if an alternative board could be used.

Sent from my HTC One_M8 using Tapatalk
 

·
Registered
Joined
·
120 Posts
Yes in a perfect world just site the board remotely and let the throttle act more like a Bosch unit. It's very doable. The benefit to this would be if you get another throttle board you would just plug it into the remotely sited box using pull-off connectors. I haven't seen a Sacer contactless sensor fail nor have I seen an armature/magnet setup fail once the aluminium sheathing's been removed.

Going back to duplication of the board:

The twisted pair of purple and pink is canbus which needs to report serial number of the throttle , OEM throttle number, diagnostic version number and finally the 2 x software part numbers. It also reports high current usage to the ECM when the flap gets stuck on carbon deposits or has a magnet to armature clearance problem.
This would be hard to duplicate on another board but I'm no programmer so I can't say.

The blue/red and green/yellow wires report both throttle angles (front and rear) to the ECM. The voltage is an analogue output ranging from 0.9V at idle to about 1.6V at full throttle. The curve for the throttle angle may vary depending on the aspiration and engine size (5 or 6 cylinders) It would need to be plotted with an oscilloscope in comparison to pedal angle so a correct throttle curve can be produced otherwise you would get throttle codes.

Finally thick green and brown are 12V+ and chassis respectively.
 

·
Registered
Joined
·
505 Posts
Why get into all this trouble when a xemodex unit will alleviate all the ETM headache?
 

·
Registered
Joined
·
502 Posts
Discussion Starter · #10 ·
Why get into all this trouble when a xemodex unit will alleviate all the ETM headache?
It will also alleviate my wallet. If I wanted to install a sacer type sensor I'd just buy one for 80 bucks on eBay and do it myself. I think all they do other than install the sacer is redo some solder joints on the board which I can do myself as well. Which I will more than likely do when I have a little bit of money to work with and the time. I try to be as resourceful as possible while not having a lot of resources ($$)

But the sacer combined with a remote mounted TB module would make this magnetti unit bulletproof in a sense. The potentiometers are usually the first things to go, followed by the motor or board due to heat. From what precopster said, the motor can be taken care of by taken out the sheath and there ya go.

I think remote mounting it would actually be a lot easier, cheaper, and safer than trying to reproduce the module itself. This may happen. I'm hoping that the added length of wire won't produce any voltage drop issues or "lag" time.

I'm thinking a good place for this remote mount would be underneath the airbox or in that area somewhere. That way, It's not too far from the factory harness so no crazy amount of wire would need to be routed, not a whole lot of heat around the airbox area, and lastly, has the most room for such a contraption.

Will take measurements sometime this week so I can figure out a project box to work with. Going to leave some room for a fan setup as well with a filter of some sort in-between the fan and project box so it doesn't blow crap into it.

Might also stop by the pull a part near me and yank the TB harness out of one of the cars and poasiboely the built in connector in the TB itself. Can't remember if it has the magnetti unit on it though. Will have to check.

Sent from my HTC One_M8 using Tapatalk
 

·
Registered
Joined
·
505 Posts
You'd be better off ditching the etm altogether, and going to a traditional throttle cable. Spend your resources figuring out how to disable torque demand via the etm coding and covert to actual throttle position sensor value.

Sent from my HTC6525LVW using Tapatalk
 

·
Registered
Joined
·
502 Posts
Discussion Starter · #12 ·
That would be a complete rework of the fuel requirement tables based on engine load and desired speed in relation to APP. Also would need to design a IAC so it would stay running under idle and during deceleration ("dashpot feature"). Basically "delete" all ETM functions, values, and tables from the ECM and start from scratch

I don't have the proper equipment to pick apart the ecm let alone program it :/ wish I did though. That'd make this infinite times easier

Sent from my HTC One_M8 using Tapatalk
 

·
Registered
Joined
·
505 Posts
There is a Russian R that did it, and didn't rewrite anything, just swapped to a standalone. Ecu.

If you can swap a contactless sensor then I'd do that and forget the rewire. I've had a xemodex for years and have zero heat issues.

Sent from my HTC6525LVW using Tapatalk
 

·
Registered
Joined
·
502 Posts
Discussion Starter · #14 ·
Oh believe me! If I could rip out this entire ecu, canbus H/L ckts, etc etc and program a megasquirt standalone EMS to run just the bare bones engine without all the failure prone electronics, that would be the first thing I would have ever done to this car. Unfortunately, I have emissions testing where I'm at so that's a no go, sadly.

Once I do the sacer install, I'll be able to take physical measurements of a project box I'd need if I remote mounted it since I'll have it apart already. I'll take some temp readings after a long drive and see what I'm working with as well

Sent from my HTC One_M8 using Tapatalk
 
1 - 14 of 14 Posts
Top