Thread: Scion xB '06, in-dash Atom 330, Lilliput 889GL; details, pictures, links. Index: pg 1

rdholtz i lost a 3.5 ide drive to a pothole after having my pc under the drivers seat so hopefully the bracket you made our of pipe will insulate better than how i had the case bolted to the underside of the drivers seat. the shocks suck on the scion its no Lexus of course but i figure i was going like 4o and hit the pothole so it did take a pretty hard hit. im thinking of placing the computer in the rear storage bin instead now
ill try to put up a work log but im not as organized as you

BTW how much for you to make a second one for me $$$

Vibration-Control Mounting

I learned an important thing on the 11-day road trip: I want more vibration control in the PC mount. I covered just under 3,000 miles, part of it over some less than pleasant roads, like some jarring, potholed streets in Washington, D.C. I also ran through construction on I-75, some beat-up two-lane roads in the mountains, and a couple of passes over railroad crossings that I thought would loosen my teeth. I had the laptop on this trip, and the laptop holder (see this thread) did its job perfectly; the laptop was stable, even over the railroad tracks. But, on the rough stuff, I could still feel plenty of vibration in the laptop when I put my hand on it. So my next effort will be to add another layer of insulating foam to the laptop holder. If that's successful, I'll add that same material to the PC mount, which is constructed similarly (see post #2 of this thread).

My biggest concern is the life of the hard drive. Small cars aren't noted for soft ride, anyway; they get bounced around pretty well by crummy roads. In the Scion, the seats take pretty good care of the humans, but the PC and the screen won't ride on cushioned seats -- they'll mount to the car. The screen will be bolted in, and will just have to tough it out. Since it has no moving parts, it should be okay.

But the hard drive is a whole different matter, because it's a moving, gyroscopic part. Although it's a laptop drive, added loads from vibration and suspension impacts will still wear the drive faster. Good data backups will help keep the data safe, but if a little extra padding will keep the hard drive alive somewhat longer, then it's a good investment of time and money.

I'll also have to rethink the hard drive mounts. I used simple metal brackets, but I think I'd better consider going for a softer suspension system -- as used in some laptops -- to isolate it better.

And then here you are, LostReception, with your story:

Originally Posted by lostreception

rdholtz i lost a 3.5 ide drive to a pothole after having my pc under the drivers seat so hopefully the bracket you made out of pipe will insulate better than how i had the case bolted to the underside of the drivers seat. the shocks suck on the scion its no Lexus of course but i figure i was going like 4o and hit the pothole so it did take a pretty hard hit. im thinking of placing the computer in the rear storage bin instead now

BTW how much for you to make a second one for me $$$

I thought about installing the PC in the rear storage bin, but decided I wanted easier access to it, less cable to run, and better airflow for cooling, so that compartment was a no-go. That brought me to the position under the passenger seat, and the tubing mount.

Before we talk about making you another mount similar to mine, I'd better get it -- or the replacement for it -- fully tested. I think more padding would be a good idea, but I'm seriously considering tubes across the top, as well, to keep it from bouncing up if I find more rugged railroad tracks. I commented about my original design that it's not in an off-road vehicle, but the railroad tracks launched my wheels into the air, and that produced the same effect as off-roading: impact. I think I can improve on the old design, possibly by just adding to it, but I believe a complete new design could protect the computer from backseat passenger feet, as well.

There's another consideration: my insurance agent -- from State Farm -- says that any device that's permanently installed -- they say "bolted down" -- is insured as a part of the vehicle. I'd probably be wise to devise a system for at least nominally fastening the case in place to meet that requirement.

One final thing -- it may be possible to make this new one much easier to produce by using 1/2" PVC tubing and the matching size of foam. I think that'll be the next trick. Some reasons:

• Tubing has some really good structural qualities; that's why race cars use tubular space frames, though those are steel tubing.
• Almost anyone who can use glue can make something out of PVC tubing.
• It's low-cost; heck, the cleaner and glue cost as much as all the tubing and fittings.
• Tube and fittings are available at any hardware store.
• PVC tubing is easy to cut -- if you don't have a $10 or $15 tubing cutter, a saw works just fine.
• PVC tubing is easy to glue.
• PVC takes paint well, so color is no problem.

So there's a new section of this project on the horizon. Watch this space for the revised edition . . .

Trapster

If you have or will have GPS in your car, or even in your cell phone, here's a thread you'll love. It talks about Trapster.com, a (free) service that uses GPS to locate you and to inform you of speed traps in the area. It's a user-supported system, so I recommend signing up and actively supporting it with notices of speed traps. I'm signed up.

Check it out; you can join the PC version beta starting next week. Information on the beta test signup is in Ozzy71's thread above. You can sign up for the service for some cellphones at Trapster.

12V Adapter

Some time back, I bought one of the Coleman thermoelectric heating/refrigeration units that keeps food hot or cold, depending on which way you run the power. I hardly used it, and then switched to the Scion xB. In this smaller vehicle, the big cooler took up half the back seat, and I went for a smaller one that fit the smaller car. I'd about decided to give the original one away, and right about then it came time to test the CarPC on 12V. That's when I remembered that the hot/cold unit came with a 120V to 12V adapter with a socket for a standard lighter plug. Very useful, and it'll run a load up to 5A. Here's the way it looks:

Click image to enlarge.

This is the unit I mentioned briefly in post #64 of this thread. It's so handy, I thought I should add a little more information here. Coleman sells the adapter alone for about $40 on their site, but they're available elsewhere for $25. Lower-amperage units are lower cost.

Screen Test

I took the whole CarPC system on the road trip with me so I could test it, and the adapter made it possible to do the testing away from the car. In the hotel room, I got the PC up and running just fine under 12V power, and did the initial RoadRunner installation. But I'd somehow forgotten to take the button bar for the Lilliput, and, without that, I couldn't test the screen under 12V. Bummer. I had an LCD panel with me as well, so I could still work on the CarPC, but I had to wait 'til I got home to test the screen.

Today was the day for that. I did the first testing of the complete system, and it wasn't pretty. I connected everything except the Lilliput, flipped the power switch, and it all booted fine. So I shut down, plugged the power into the Lilliput, flipped the power switch, and the system fan started and stopped immediately. No joy. Tried it three times, with the same result. Something was making the system very unhappy. I unplugged the power from the Lilliput, flipped the switch, and the PC ran fine. So we know it's the Lilliput screen that's the problem.

This is ugly. Now I have to figure out where the failure is in the screen, preferably without burning it out, if it isn't already toasted. This is one of those times when the leading edge is the bleeding edge.

it's very possible that when you add the screen there is not enough power anymore... try just the screen by itself to see if it's a power issue

Thanks -- I'll give that a shot. I'm running both PC and screen off the M2-ATX, and it should have enough juice to do it, since there's not a lot of power grabbed by the PC, with a 25W processor and a single SATA drive. Still, there's no sense in ignoring an obvious solution. I'll test that tonight.

More Screen Testing

Separate power didn't change anything, nor did using a powered hub to run the USB connection. Looks like there's something fundamentally wrong here. Time to recheck all the connections, make sure the polarity is correct, and make sure everything is plugged in the way it's supposed to be. If all that fails, I'll set her aside until I can get to a repair shop, and invest my time in more bezel construction. It probably wouldn't hurt to start laying out the head unit mount, too.

I wonder if there's any such thing as a repair manual for the Lilliputs, or if this is all done by the seat of our pants.

Even More Screen Troubleshooting

So what's wrong with the touchscreen? Actually, what's wrong, not just with the touchscreen, but with the whole display? I let her sit for a few days to ponder, and I come to one conclusion: a dead short. This idea makes sense, because when I put power to the screen, it kills the whole system, and that indicates a dead short. In thinking about this, it occurred to me that the Lilliput case is made of plastic, and the printed circuit board is isolated from the screen itself except for the wires that connect them. I've taken everything out of the case, and have both screen and PCB mounted to a steel support frame I built, and so they're grounded together. Maybe that's not the way it's supposed to be.

There's a question that comes up: after carefully checking the way I wired the power cable to the screen, did I somehow do it exactly wrong, and feed power the wrong way through the screen? Would that produce this result? Hmmmm . . .

In the final analysis, either the monitor isn't getting what it wants from the PC power system, or it's toasted, and I'm not going to get what I want from it. Either way is a bummer, but the "toasted" scenario is a bigger bummer.

Step One in the troubleshooting process was to plug everything together and apply power. That was a miserable failure. So Step Two was to take the system off the mount and see if I got any results. Now, I know that by running this with a dead short, it's possible that I had fried the system. Oh, that would be very ugly, indeed. I was hoping it was still alive . . .

Having taken it all apart, I plugged her in. Here's how it sounded:

"Okay, she's all connected. We flip the switch and . . . aaauuuggghhh, horse puckey."

Nothin'. Nothin' at all. I tried switching some connections around. Still no joy. But, on the other hand, there was no big streamer of smoke, either.

I guess the next trick is to do some multimeter checks and see what we have. That'll wait 'til I go back up to Atlanta where Ozzy71, the electronics guru, lives. He knows bunches more about this stuff than I do. He informed me once that electronics devices run on smoke. "You let the smoke out of 'em, they don't run any more." And I've lived by that wisdom ever since: don't make 'em smoke.

What's Next if it's Dead?

If she's dead -- smoked -- then I'm going to take a hard look at two ideas: a regular, non-touchscreen LCD panel (with a touchpad or the already-planned joystick/trackball), or a nice, bright transflective (transmissive-reflective) touchscreen display. The plain panel is lower cost, but the transflective display is much better in the Ooh and Aah department.

I may consider pulling back to a 9-inch monitor, as well; that wouldn't change the cost very much, wouldn't change usability a whole lot, but it might make the bezel production somewhat easier. I guess I'd have to ask the mods to change the name of the thread, wouldn't I?

Rather than letting this be a complete roadblock and putting the project on Hold for a couple of weeks, I think it's time to make a mount for the head unit. Whatever I do, the HU will be moved, and it has to attach to something. That mount is where we're headed next.

Moving the Head Unit

To make space for the screen -- whatever screen I end up with -- the head unit has to move. It goes to the bottom of the dash, in the space previously occupied by the cubbyhole, ashtray and lighter. There's just room for it. Here's the original dash, with the HU up on top, and then a closer shot showing the holes made by removing the HU above and the cubbyhole/ashtray/lighter unit below::

Click images to enlarge.

Finally, here's the dash with the HU moved down -- but not mounted to anything -- and the screen mockup in place, first from the driver's viewpoint, then looking in the passenger door:



Head Unit Mount

Now the challenge is to mount it this head unit. I want to use a system similar to the OEM support: two sheet metal brackets that are bolted into the sides of the HU, and then are bolted into the vehicle.

In the space where I want to put it, there's a solid fastening point on the left side, where the Phillips-head bolt attaches the dash to the dash sub-frame:



(That cable bundle is well behind the bolt, although this camera angle makes it look like it attaches there.)

That fastener only makes the left side work. Almost everything else in the area is part of the dash, which means it's only plastic. That's not the best support material for a relatively heavy unit. By pulling back the carpet on the right side, I see that, down at the floorboard level, there's a steel bracket held on by two Torx T40 socket-head fasteners:



If I can build a bracket that attaches to those two points, I'll have a total of three points to mount on, and that should be solid.

(If you're wondering what those black plastics things are in the last two pictures, they're ducts sending heat and a/c to the rear seats.)

Space in this location is relatively tight, and the bracket down the right side is going to be a long one. It may be wise to make the bracket that bolts into the right side of the HU separate from the part that extends down to the floor. That way, I can put the long lower right bracket in first, mount the side brackets on the HU, and bolt the HU in place.

That's the plan so far. Now I'll build some mockup brackets from cardboard, because it's so easy to work with. Then, when I have something that looks like it will support the HU -- and that I can get in and out of the dash -- I'll build them in light sheet metal, then 22-gauge steel. The light sheet metal isn't a necessary step, but it lets me test the theory in a material strong enough to do a test installation. If it works in that material, or if it works after some refining, the test piece is a rough template for the final steel part. Doing it in the light material lets me make mistakes that are easy to correct, and means the final part won't have suffered through extra bends and unbends, which weaken the material, and it won't look cobbled together. Even if it's hiding behind the bezel, I want it to look OEM and be at least OEM quality.