M939 CTIS Reprogram, Anybody Done It?

[therooster2001]

One more. The DANA manual doesn't specify what configuration options 0-8 are. Does anyone know what they are (specifically), are they exposed in the DANA tool? I am assuming they might be user configurable settings so if you set those options they may be different from default, or its hard coded. This is probably how the wrecker gets its default increase in pressure as the truck harness is probably jumpered. I wondered what else lurks in those banks...

 

[aleigh]

What I was suggesting is building a hardware replacement for the ECU. I think you took me to mean building a hardware code scanner (which is also a laudable and practical project). What I am proposing is building a new combined ECU+interface that plugs into the factory harness and deletes the existing Spicer/Eaton/Dana unit. In the case of the scanner I sure wish I had one here, but at the same time, I think it's just as easy to build one that works with a laptop. Everyone has a laptop by now, right.

There's no reason to deal with the J busses (either to consume or provide) with a project of this nature. There are no other electronics in the system, everything is present on the harness at the ECU. It's just a simple process control / SCADA exercise. It's straight-forward to program. If you have expertise with enclosures though that would be super helpful as it's certainly not mine, I'm a software guy.

I have two real motivations for proposing an ECU project; the first is to build a unit that provides fault codes without a scanner while giving better information about what is going on with the system (e.g. which valves are open when, transducer reading, etc).The second is that a lot of times I just want to bump the PSI a little. Like maybe I took on a heavy load. A related solution might be to source one of the civilian CTIS controllers in the secondary market because they have LCDs and are driver programmable.

My personal take is that the CTIS systems are reliable when they are maintained well (and all the rubber bits are new), but, when they break on our older trucks that have sit for many years and are in who knows what condition, diagnosing is a lot more trouble than it has to be. The fixes mostly seem simple - a seal here, an o-ring there, bad sensor, once the mechanic gets to the bottom of the issue.

I looked around for information on the configuration jumpers and didn't find anything. In the PDF they are just kind of listed as configuration 1 through n. A thought I had is that only maybe a certain configuration enables the diagnostics and that is why he's having trouble reading the bus, but, without the terminators, I know that's a non-starter anyways, so I'm eager to see what he comes up with.

 

[therooster2001]

Building a new ECU is laudable, but yes, my vision was an augmentation that covers what you stated, using the existing ECU, that allowed user "changes" that were nothing more than level changes, with real time display of diagnostic things we all want to know (current pressure, target pressure, faults, and location). Maybe we splinter this dev into two projects, we'd both learn something from the existing system, as you will still have to contend with the 20-some odd inputs that the current ECU handles, unless you have other plans. I would love to geek on it.

 

[aleigh]

20? There's only 2 data inputs and 5 relay outputs (3 if you have single-channel CTIS) that matter. You can add one data and one relay if you want to preserve the over-speed indicator, but I don't see that being a pressing need. As far as I can tell from reviewing their material the entire (necessary) ECU functionality is replaceable with an arduino, a relay shield, a lcd, a few buttons, less than a thousand lines of code, and the amphenol jack.

Everything is present at the harness already. It's practically a high school electronics project. If anything the biggest (and most expensive) part of the project would be having the enclosure fabbed. Preferably the right dimensions to bolt into the same position as the ECU we have today so that it's just a drop-in. I doubt it would even be a PCB, we're just talking about an arduino and a shield in a box and some jumpers.

You have to go to all most of that trouble to build the scanner anyways, and if the intent is to leave it in the truck, and you have put all that money and effort into it (implementing the J busses is going to be way more complicated than just the simple logic for the ctis), why not just replace the ECU? It's not like there is anything special about the one they provide, just the opposite. They are (relatively) expensive and one more thing that can break. I'd just as soon be rid of it.

I do think a scanner would be valuable if you didn't intend to leave it in the truck (ie, use it as a traditional diagnostic tool). Guys would be borrowing that all the time. But for all that, why not just use a windows laptop and a RS-485 interface. Just about everybody can do that already without the expense.

 

[74M35A2]

Hold tight. I hope to try the termination resistors today. For some reason, I always pick the coldest days to do this stuff outside. Here's a little love for the warmer of folks this morning on the board:

 

[JohnnyBM931A2]

Hold tight. I hope to try the termination resistors today. For some reason, I always pick the coldest days to do this stuff outside. Here's a little love for the warmer of folks this morning on the board:

lol.. 19 degrees warmer by you than it is by me Low here later today is -6, -20 if you include the wind. Fun stuff.

 

[therooster2001]

20? There's only 2 data inputs and 5outputs .

Yup, I meant 20 connections, not inputs, and some optional. Agree the Arduino could replace it. Besides potential failure and lack of visibility, what don't you like about the current ECU?

Edit: you also said expensive. I sold several for $40 and the going price for take off seems to be $60, more than an Arduino setup and fab...

 

[aleigh]

First and foremost I just like simple. It seems simpler to replace the controller than graft another system onto it which talks to it over a data bus via a reverse engineered protocol. If you go the scanner route you wind up with two ECUs, a harness (which leaves you soldering up more wires than my ECU proposal anyways), and you have to find some place to stick the scanner. Which granted is not exactly a difficult challenge in a military truck, but now how long is the harness wiring?

The buttons all seem to wear out on the stock controller. Mine's certainly cracked and that seems pretty par for the course in the used ones I have seen.

Then the things already mentioned; no feedback about current status / valve configuration on the stock unit, no ability to manually actuate valves for testing, no ability to read the PSI off the pressure sensor (which required coordination with the valves), no ability to set the desire to fine-tune the desired PSI, human-readable errors limited to blink codes too granular to be useful (5 leds blink - thanks ecu!), no ability to configure quick set-points (sand,xc,hwy) without a diagnostic tool.

I've never seen a CTIS controller for $60 that wasn't all banged up with cracked plastic on the buttons. Look at prices for NOS replacements. And every day there are less of these, not more. What I propose would be a brand new condition solution, so I think it's probably most fair to compare it to NOS prices. First to agree that I don't think a kit could be put together for a replacement ECU at the $60 price point - the plug alone is what 40 bucks in low quantity - but - that isn't what you are proposing either if we are being fair. You have to provide nearly all of this stuff (minus the relay board if you weren't going to provide valve override) for a handheld diagnostic tool that grafts on the functionality you have mentioned. I think the costs between the two solutions would be pretty similar.

I'm sure Dana did a nice job all ISO9001 six sigma kaizan wazoo accredited to a huge pile of military standards. It's probably literally EMP hardened and I'm sure they sent test units through paint shakers. Skipped the button pressing tests I am guessing, but I digress. I'm also certain Uncle Sugar paid handsomely for them. And that's nice - certainly desirable if you are one of the guys who wants a 100% stock show/parade truck. But honestly, just speaking for me here, if I am traveling with it I would find it a lot easier to service a box that just has a handful of common commercial off the shelf boards in it running open-source software.

But, at the end of the day - we shouldn't overstate the logic involved in CTIS. It's not complicated. You charge the system every so often, read the pressure, and fiddle some valves. You look at what the pressure sensor does over time and draw some conclusions and then give the operator feedback.

I've got the parts coming to build a prototype ECU in any event, so we'll see first hand if I am being foolish.

 

[therooster2001]

Buttons. Yep. That's a good one. Removal of the extra protocols is a plus, as it's the only thing on the bus that talks those I think. And I don't think anyone's debating simplistic operational design of it. I'm all about the electronics and redesign of it. I'm a stock, but not parade guy, so prettiness doesn't matter to me. Happy to work on it with you, and a diagnostic tool will help us get there.

You should have snagged my $40 ECU shipped of FeeBay, I had to pay $12 shipping and $6 of fees, I practically gave it away. Side track:Stop shopping and selling on FeeBay, the change in fees is killing sellers. Stay in here!

One other question is whether the Eaton controllers were the same design and wiring, that's all I have... Meaning if they even present the protocol, or went cheap and didn't use it (or need it)

 

[aleigh]

I totally get where you are coming from on the stock angle. I wonder if a new ECU could be crammed into the stock housing, nappy buttons and all. Could blink more intelligent codes with the LEDs already there, multi button presses for features. It'd be cryptic but hey least then you'd retain the authentic stock appearance, although that might still offend philosophically.

You are right it's the only thing on that bus. Really the bus isn't a bus, it doesn't go anywhere (at least not in my LMTV). The allison MD3700 on my truck also speaks J and that is wired to the under-dash diagnostics port. Why they didn't wire the CTIS and the transmission together on the same bus which you can get to from the port - no idea. Don't get me started on the state of engine diagnostic tools in the industry... It pains me because implemented well electronics can make maintenance of trucks /easier/ not harder, yet look at where we are...

I think the biggest mystery right now - besides him trying to get the J bus to work with the NexIQ - is what is the sensor data coming back from the valve module. IIRC it's a 3-wire setup and I am not sure what it is for (closed loop on the valve status? something else?). The pressure switch is elsewhere but maybe that is just a low pressure cut-off (the 60PSI to save the brakes) and the working pressures come from the valve controller. Any light that could be shed on that would be helpful.

On my end of things my CTIS is broken - that's how I would up here in the first place, the 5 lights blinking - so I may start by building the diagnostic end of the ECU first to allow me to open and close the valves manually, then work up to reading back the pressures, and then finally the reguar logic last. Obviously it'll be really hard to QC/test the ECU software with broken mechanicals.

On the wiring end, if we stick with DIN headers or screw terminals, it could be reconfigured easily for different trucks wiring schemes. I am not sure that they are different between the trucks though. I think we'll have to find that out. I'll probably take my ECU apart in the next few days and photograph it all just to see what is going on and what chips they are using. In turn maybe those reference photos will help other guys if you want to compare your ECUs.

 

[Coffey1]

You will have to scrape all jelly or whatever it is off it.
That looks to be fun

 

[The HUlk]

Oh I should say - failing all this I'll just build a new CTIS ECU from scratch. With an integrated LCD display it could show the PSI and allow operator control of the pressure (raise/lower manually) as well as end-user programming. Then it can just display actual human readable faults. Probably would end up $150ish in parts, I'd give the controller software away for free (I have no interest in making money here). It's not a difficult exercise with something like an Arduino. If a solution like this is of interest to people, please let me know.

I'd be interested just to see the source code as I've never worked with an Arduino. Motorola 68HC11 is the last embedded platform I worked with years ago.

 

[aleigh]

OK I tore into the controller. Jeez coffey was not kidding about the jelly. Pics are here, they seemed to be too big to want to upload:

https://www.flickr.com/photos/adventioneering/albums/72157661316728564

The Amphenol jack connects to a PCB which then connects to the mainboard with a pair of 10-trace ribbon cables. On mine, the A/D pins are NOT wired. This is supposed to be J1939. Little "a" and N ARE wired. This is supposed to be J1708.

They are using power transistors, 6 of them, along one of the edges of the board. Probably makes sense instead of relays since the unit is obviously not end serviceable.

For trivia's sake the microcontroller is an Intel [SIZE=-1]AN87C196KR.

The bubble switch panel interfaces with a thru mounted ribbon cable, and then the LEDs are (of all things) through mounted. They are dialite packages mounted directly onto the PCB.

Also - I'm pretty sure my perfectly working ECU probably does not work anymore

[/SIZE]

 

[SLOrazorsedge]

Boy sure enjoy following this thread, don't understand a dang thing you guys are saying. Just hoping you solve the "5 flashing lights" gremlin for us lessor poor souls!

 

[74M35A2]

What a mess. The donor module I disassembled had zero potting in it.

 

[aleigh]

Maybe you lucked out! How does that module compare to the one I posted (make/model/date/etc)?

 

[JohnnyBM931A2]

While we're on the subject of disassembling CTIS components, has anyone taken apart a CTIS proportioning valve? I'm curious what it is that actually "fails" inside there..

 

[aleigh]

I dunno but 25 bucks will get us a broken one on that well known auction site...

 

[74M35A2]

Maybe you lucked out! How does that module compare to the one I posted (make/model/date/etc)?

Not much ID on it (no labels like yours). Also I didn't take any other pics than what I posted because I was just removing the internals to have the housing and connector. Sorry, I would have paid more attention otherwise.

 

[aleigh]

c'est la vie. I can scrape everything out to potentially re-use this enclosure, but, I will destroy the mainboard removing it for sure. The jelly is on the bottom underneath the board and is essentially gluing it down.

One thing I'll say is the thing seems to be well thought out other than the bubble buttons (I still wonder who thought that was a good idea. Do they last, ever, in any application?). They went to a lot of trouble to put protection circuits and diodes on the inputs from the look of it, little touches like that.

The enclosure is fairly nice but not up to the quality of some of the radio gear I've taken apart. Interestingly my bottom plate (the one you remove to get to the PCB) was not gasketed and had a couple holes in it. So any moisture resistance is coming from that jelly which certainly would do the job.

I may try baking some of the jelly to see if it melts. That might be a simpler way of getting it out of an enclosure that is going to be re-purposed or repaired.