CAT 3208 monitoring

[nedb]

Sorry but I don't know of a monitor for the 3208. Unlike more modern engines, my '91 300HP doesn't have a "computer" to manage the engine, and therefore has no data bus to provide trouble codes, engine data, and the other monitor functions that the DDEC computer and data bus collect and provide to the dashboard area.

Still, my WL has plenty of sensors and interesting places to collect information, including the usual things like coolant levels, temps, pressures, and so forth. It would be useful to break out and monitor the individual sensors that are currently aggregated into the "Check Engine" warning light. I know of a sensor on the leveling jacks, a sensor on the coolant tank on top of the engine radiator, and I assume there are more. Any of these will turn on the generic "Check engine" warning. I hate it when that happens, because I want it tell me which system the problem is in.

So I think about lashing up a short custom wiring harness to feed the data from sensors to an Arduino inside the coach. They have plenty enough processing power to log, filter, transform, and transmit a handful of raw data inputs for display. I'd probably try to send it wirelessly to dashboard area for display.

I've never used Arduino for wireless data transmission. I picked up an XBee shield thinking it was what I needed, but it needs the actual XBEE wireless radio component as an add-on. The XBee web site says this shield model is deprecated! So I am back at square one on that aspect of the project.

David mentioned once that he was thinking about an engine fan speed monitor, and I thought this sounded like a cool idea so I put together a prototype based on a bicycle speedometer and a few lines of code modified from a sample code I found on the web. It works perfectly as a counter--every time I wave the magnet near the sensor, the arduino reads it as a high value. When the magnet is out of range, the arduino reads it as "0". All that remains is finding a mounting place close enough to the Horton fan to pick up the magnet as it goes by, attached to a fan blade. Heh--an Arduino fan speed sensor!

I get carried away with the possibiiities. I know better, but under this hardcore veneer of 3208 "KISS" discipline, I am at heart a fully-instrumented kind of guy. I don't mind the uncertainty of taking the bus on the road with the prospect of mechanical problems at some point before returning home. But I would love to be able to see problems developing before they occur by keeping an eye on more aspects of critical systems and engine operating parameters. There's no particular reason why I couldn't monitor many more things than my coach in factory trim is equipped to do.

Is anyone else hacking the 3208/Arduino frontier? Care to share?

--Ned

Ned Bedinger
'91 SP36

[Arcticdude]

So how will you know if its actually a problem arising , or a sensor malfunction or wiring issue? With a mechanical engine, there's very little to go wrong. While I understand the frustration of a light and what is it exactly trying to tell you, but it's ONLY capable of telling you a couple of things so why over think it? Aren't those things easy to check? A quick look at your jacks, coolant level, oil pressure, maybe a fuel sample (if its capable of water in fuel detection). All good? Then keep going. Why, if you're running down the road and the engine light comes on, what's the coolant temp?, what's the oil pressure?, engine running smooth with good power?, any unusual vibrations/noises? If no, then why fuss over it? I've got all the sensors and my cel will light on occassion. Yep, some stupid sensor is failing. All other gauges are normal. Eventually, I'll spend the $3 large and buy the diagnostic computer to trace the issue down. I'll end up replacing a $10 sensor or repairing some chaffed wiring. What have I gained that a non-ecm engine didn't have? If I could take all my ecm stuff off, I'd gladly do so on EVERY vehicle I've got. My last non-ecm vehicle had a 351 2 bbl in it, weighed 3500#, had the aerodynamics of a brick, yet got 22mpg on the highway, could run a high 13sec quarter and lift the front wheels when it was feeling good. Lets see, I can't even buy a big 2 door anymore, especially one that can handle 4 people. I'm lucky to even find a mid sized V8 and it sure won't break the 20mpg mark. Ok the Vette can, has a true V8, but it won't seat 4.

Yes, we've gained vehicles that can tune in 300 radio stations, warm or cool our butts and now even heat our hands. But we can't fix them anymore, because you need a $10k computer to talk to the $3500 computer to see if it can tell the bigger computer what's wrong.

Sensors, schmensors!! You can keep them!!







Ok, it still would be nice to see what the cel is trying to tell you!

[davidbrady]

(02-27-2013 03:46)nedb Wrote:  Is anyone else hacking the 3208/Arduino frontier? Care to share?
Ned Bedinger
'91 SP36

This is great Ned. I've been wanting to do this for some time but lately I've gotten myself sidetracked...

I have an Arduino development platform loaded on my lenovo Thinkpad running slackware 14.0, and I've been writing really basic software for my Arduino Uno, simple stuff just to get the configuration and makefiles sorted out. I haven't yet bought any shields. I'd like to build a sensor area network for my bus to collect parameters that the DDEC doesn't monitor, so heck yeah, let's put our heads together to see what we can come up with. I agree, stretching wires is the last thing I want to do.

Of course my bus is J1587 equipped and I have Silverleaf's VMSpc up front. So the natural thing for me to do is take my fan speed data and push it onto my J1587 bus as a diagnostic PID, but I haven't found a gateway (or shield) to help me do that. Because the J1587 bus is mission critical for the engine, transmission, and ABS brakes I want to be careful about interfacing to it, and I may abandon it all together in favor of a separate network. So maybe we can find a solution that'll work for both of us, maybe an 802.11 Wi-Fi shield: https://www.sparkfun.com/products/11287

VMSpc gives me a fairly complete view of what's happening at the engine, but there are still a bunch of items I'd like to sense: fans speed, engine room temperature, engine vibration, etc. There are a host of sensors available to the hobbyist, https://www.sparkfun.com/products/11283 Some of these will work for a while, but they're not exactly engine room hardened.

I've never written software for Microsoft Windows, so I'm a little wet when it comes to putting together a GUI to display gauges. My easy way out is to use VMSpc, but that's an expensive option if you don't already have it.

So yeah Ned, let's roll!

(02-27-2013 10:17)Arcticdude Wrote:  But we can't fix them anymore, because you need a $10k computer to talk to the $3500 computer to see if it can tell the bigger computer what's wrong.

Sensors, schmensors!! You can keep them!!

John, I think that's the rub, "we can't fix them anymore". Why is that? Because we don't have the source code, the development environment, nor a way to connect to the proprietary systems in our vehicles - we don't have the tools. That's the frustration. If we had those things then debugging electronic engines is no different from debugging a mechanical engine's overheating, or vibration, or smoke out the stack. In the mechanical engine we break out our wrenches and we pride ourselves on fixing it. Well with our homebrew sensor area network it's the same deal. We break out our software tools and we fix it. As long as we have the tools, and we do cause we wrote the stuff, it's a non issue. The frustration with today's "can't fix 'em anymore" vehicles is really a frustration over a lack of tools.

Now, tell me you wouldn't like to know when your fan is on or off, and whether it's operating within the correct rpm range?

[Arcticdude]

David,

But the sensors and electronics and a whole layer of complexity that's primarily unnecessary for adequate engine operation. So not only do you have to worry about fixing the original problem, but you've got to fix the secondary issues as well.

I remember a friend working on a Chevy several years ago that had some electrical issues. After disassembling the car, a pinched wire in the factory harness was found. Truly, the fender was off the car, the dash was mostly out and the harness had to be unwrapped to find the issue.

Ok, so that was only one out of millions on the road. But the owner of the car certainly didn't care! It was a 100% failure rate for them!

While it might be "cool" to know what my exact fan speed is, my coolant gauge tells me it's working or not.

[davidbrady]

John, the solution here is clearly sensors for the sensors! ROTFL

You're never gonna get the geek out of me John. This is all cool because we can. And, it's cheap - an Arduino Uno is twenty-five bucks. And, it's intellectually challenging. And, it's fun! None of it is mission critical; it's just a layer of toys. Besides, these wanderlodges are so reliable I need something to work on. I need to create my own Spaghetti code so I have something to do!

[Arcticdude]

[Itchintogo]

Seems like you guys are going to take these mechanical engines and bring them up to current diagnostics. Won't it be interesting to see an early eighties FC roll into the campground and the owner knows already what he is going to do to fix it because he has a screen to tell him where the problem is with out unhinging the front bumper and taking all those fasteners off and draining the radiator a bit and swinging it open.

The Arduino is a fun little item. I did some reading lately thanks to David's mention of it here on WaGu. I have to admit it is all geek to me. But they look and sound wonderful for their size.

My Arduino comes out of a green bottle from Italy and looks red once carefully placed in a long stemmed glass frequently stored in the front passenger side glass cabinet of an Lxi.

[pgchin]

(02-27-2013 14:23)Itchintogo Wrote:  Seems like you guys are going to take these mechanical engines and bring them up to current diagnostics. Won't it be interesting to see an early eighties FC roll into the campground and the owner knows already what he is going to do to fix it because he has a screen to tell him where the problem is with out unhinging the front bumper and taking all those fasteners off and draining the radiator a bit and swinging it open.

The Arduino is a fun little item. I did some reading lately thanks to David's mention of it here on WaGu. I have to admit it is all geek to me. But they look and sound wonderful for their size.

Ditto! Yeah if Ned and David get something going, there are a TON of 3208's in birds and most everyone carries a puter of some sort so they will have really helped those guys and gals out!!!!!! Go for it!

[nedb]

Hey 'Dude--

(02-27-2013 10:17)Arcticdude Wrote:  So how will you know if its actually a problem arising , or a sensor malfunction or wiring issue? With a mechanical engine, there's very little to go wrong.

*cough*

I've seen more than a few threads about Facit pumps failing on the road. I'd bet I could catch it before it failed away from home. Fuel pressure monitor and maybe monitor resistance of the pump wiring.

(02-27-2013 10:17)Arcticdude Wrote:  While I understand the frustration of a light and what is it exactly trying to tell you, but it's ONLY capable of telling you a couple of things so why over think it? Aren't those things easy to check? A quick look at your jacks, coolant level, oil pressure, maybe a fuel sample (if its capable of water in fuel detection). All good? Then keep going. Why, if you're running down the road and the engine light comes on, what's the coolant temp?, what's the oil pressure?, engine running smooth with good power?, any unusual vibrations/noises? If no, then why fuss over it?

I've got to hand it to David here, when he admitted that it is a geek thing. And you're right too--the manufacturers gave me gauges as well as an idiot light. So technically it isn't an idiot light. I guess.

Anyway, picture me, a noobie in my first diesel electro pneumo mechanical vehicles. I'm well into the first thousand miles of ownership, breezing down the expressway when I get the warning. Check the gauges, everything is OK. Pull off, shut down, realize I don't know what all is capable of triggering the alarm! I've read one thread about coolant level sensors failing, mimic'ing a loss of coolant. But in this case, the engine temp is, as I said, OK. So long story short, it turned out to be the coolant level sensor after all--broke a fanbelt, and the fanbelt whipped around and parted the sensor wire. I've had only this one CEL incident, and the system gauges couldn't lead me in the right direction. As a noob with foreknowledge that a new installed engine runs $20K+, I'm just bound to get anxious and want better information on where the problem is. Keep on going with the CEL on? Not on a dare! Can I get along without the elaborated monitors? Well, ... well yeah sure. But then I don't get a merit badge for proactive management of my vehicle

(02-27-2013 10:17)Arcticdude Wrote:  I've got all the sensors and my cel will light on occassion. Yep, some stupid sensor is failing. All other gauges are normal. Eventually, I'll spend the $3 large and buy the diagnostic computer to trace the issue down. I'll end up replacing a $10 sensor or repairing some chaffed wiring. What have I gained that a non-ecm engine didn't have? If I could take all my ecm stuff off, I'd gladly do so on EVERY vehicle I've got.

Absolute agreement with you on this, for all the reasons given by you and David. I currently or recently have had two big block chevy 454 gassers in and a small block Ford 351. Two are carburetted with Holly 800 cfm 4bbl. The other has a Quadrajet. The boat with the 351 is the extreme example--if I have a problem on the water, I want the fuel/spark/air mantra to guide me so I can power outta there. Northwest waters, with vessel traffic and some wild weather, are something to avoid after dark in a dead boat! Spare coil, distributor cap, spark plugs, and multimeter, that's my carb'd engine kit. Right on.

But the WL and diagnostic computers, whoa. This one vehicle is many many times more complicated than any of my gassers. When you consider the long learning curve for where everything is, how it works, and how to work on it, I think that looking at it as a monitoring problem is a high-yield proposition. If I took your advice and didn't worry about it 'cos I've got gauges, I would likely never need to know what sensors can trigger it. But devising monitors, I get the hands-on familiarity with it, in nice manageable one-system-at-a-time chunks. Give me a couple of years and I will build a diagnostic computer out of Arduinos

(02-27-2013 10:17)Arcticdude Wrote:  My last non-ecm vehicle had a 351 2 bbl in it, weighed 3500#, had the aerodynamics of a brick, yet got 22mpg on the highway, could run a high 13sec quarter and lift the front wheels when it was feeling good. Lets see, I can't even buy a big 2 door anymore, especially one that can handle 4 people. I'm lucky to even find a mid sized V8 and it sure won't break the 20mpg mark. Ok the Vette can, has a true V8, but it won't seat 4.

Corvette did something radical a few years ago to give really great mileage. Or else I'm a big sucker for whopper lies. Talking to a vette driver in a gas stop, he said he was getting 30-something MPG. Well OK, he was driving like he had a bowl of goldfish on the seat instead of a sports model gal. But 30 mpg in a hot V8? What did they do to accomplish that? Compression releases on 4 cylinders when cruising?

(02-27-2013 10:17)Arcticdude Wrote:  Yes, we've gained vehicles that can tune in 300 radio stations, warm or cool our butts and now even heat our hands. But we can't fix them anymore, because you need a $10k computer to talk to the $3500 computer to see if it can tell the bigger computer what's wrong.

Sensors, schmensors!! You can keep them!!

Luddite! Go geek, they don't leak.

(02-27-2013 10:17)Arcticdude Wrote:  Ok, it still would be nice to see what the cel is trying to tell you!

That's the stuff!

And thanks for the comments, I'll try to keep it real.

--Ned

Ned Bedinger
'91 SP36

[Arcticdude]

Ned,

As far as the Vette is concerned, higher horsepower, lower 1st, 2nd and 3rd tranny gears, but high OD 6th coupled with a low ratio (~2.41 iirc) gives you strong acceleration and let's the engine idle at typical highway speeds. Any of the newer factory hotrods will essentially do it, if driven in moderation on the highway. Even a Viper! Of course, having low cd's also helps.

You guys keep on trekking! I obviously like playing the devils advocate. It's all very interesting to consider. And enjoyable!