Prius Multi Function Display repair

I'm going to start this page by giving credit where credit is due: Hobbit did the initial investigation of the 2004-2005 Prius Multi Function Display (MFD) communication failure issue, and everything I've done so far has been built on the work he has shared on his web site.

Please don't email or call to ask if you can send us your MFD for repair.
 Because:
1. We don't have a test bench set up to simulate a Prius, so we need to have the car here for testing.
2. We aren't set up as a mail order repair service, and dealing with emails & shipping would be a pain.
3. We have may not have perfected the repair procedure yet.

This is the story of an MFD that is in for it's second repair. We fixed it once, but it only worked for a month. Since the MFD is a $4000 part, it's probably worth spending some time to determine whether a reliable repair is possible, so we gave it another go. I attempted to repair a damaged solder connection using a reflow station on the first try. This time I've resoldered the damage using a soldering iron, and I also by-passed the suspect area with "green wire".

Warning: Even though there is only 12V going into the MFD, LCDs create very high voltage using a step up transformer. This is the story of what I did, not instructions for you to follow step by step. If you kill or maim yourself, don't blame me or Art's Automotive.

 

The symptoms for a failed Prius or Highlander MFD are:

An "audio off" screen, even when the radio is playing.
A "Check the connection of the air conditioner" message when the climate button is pushed.
And the Energy Monitor (Prius only) does not show any information about the regen status or consumption. There would be other symptoms if the car had NAV or other units on the AV bus.

When you enter the System Check Mode (press & hold the INFO button on the side of the MFD & turn the headlight switch to the parking light position three times in a row, starting & ending in the OFF position), you'll see a NCON (No Connection) warning for the AUDIO H/U as well as G/W.

The AUDIO H/U is the stereo.

G/W stands for Gateway. The Gateway ECU is the computer that allows the 3 busses (AVC, BEAN, & CAN) to talk together.

EMV is OK. What's the EMV? It's the MFD. Why not just call it MFD instead? I have no idea.

Basically, this means the MFD can talk to itself, but not to anyone else.
The LAN monitor says pretty much the same thing. The EMV (MFD) now says CHEK (which means there should be a code stored), but I couldn't get it to give up any codes while it was broken. Maybe I was just being too impatient. The MFD was moving very slowly. It seemed to be spending most of its cycles trying to establish communication, and was very slow to respond to user input.

In the Vehicle Signal Check Mode does not rely on information from the bus. (The bus is a two wire data path between control units). The MFD monitors these items directly.

IG is "Ready".
PKB is the parking brake.
TAIL is the headlight switch position.
ADIM/TCAN is the MFD dim/bright status (based on the headlight switch position it seems).

Since the headlight switch is used to enter diagnostic mode, its position must be monitored without using the bus.

The speed is likely used for function lockout based on vehicle speed. So if you are looking to hack this, so you can drive, dial, & crash, this is probably a good place to start.

Why the parking brake? I'm not sure. I do know that Toyota locks out some scanner utilities (like brake bleeding) if the parking brake is not set. So I'd guess that the PKB signal is responsible for some sort of lockout.

Anyway, all of the stand alone MFD functions seem fine. The problem is only with its connection to other control units.
Time to yank it out. The Prius has a well designed dash. My handy KTC plastic pry bar is all that's needed to remove all of the covers. You could use a sturdy tongue depressor, a taped screwdriver, or whatever.
Covers removed. Well, removed enough.
Two bolts hold the MFD in place.
First I check the resistance on the AVC bus. The terminating resistor for the AVC bus is in the stereo. An open would indicate a break in the wires to the stereo, or an open resistor in the stereo. Without the terminating resistor, the data signal gets big & sloppy. The control units will talk & listen, but have trouble hearing what's being said, and may not be able to communicate at all. The resistor should be about 60 ohms.
68.3 ohms. Close enough.
Next I hook up a scope. What will I be able to tell from this test? Not a whole lot. Mainly that data is being sent on the bus, but I can't make sense of the 1s & 0s, so I won't know what it means.
Yup, looks like data. 2.5 volts is about right. Looks like my B probe needs compensation though. Or maybe a loose ground clamp.
Time to tear it up. This is my third MFD disassembly, so it's pretty easy now. I know were every connector & screw goes. The first couple times my desk looked like an operating theater with bags & bins & labels. Now I'm getting comfortable and just lay it on top of all the work that has piled up on my desk while I've been working on the Prius. Hopefully OSHA doesn't notice my open can of V8 :)
The connectors are very small & delicate. If you just tug on the plastic connector, like you would in the normal sized world, you may break all of the solder joints where it mounts to the board. A better way is to gently insert a small screwdriver into the connector joint and twist. That way the force is applied to the plastic connectors, not the solder joints or wires. I found this out the hard way while fixing one of our LCD monitors here at the shop. I had to go back in to fix the problem I created after fixing the problem it originally had.

This is about as disassembled as it needs to get. Since this is our second attempt, I'm a lot more thorough in my checking this time. Last time I looked for a broken solder where Hobbit said it might be, found it, and fixed it.

This time I spent some time verifying where the bus goes with an ohm meter. This turns out to be a little frustrating. 1) because of disappearing traces, and 2) because the board is coated with something that makes is hard to get a good connection with the probe tip.

All of Hobbit's previous findings were verified.

I have a plan this time. Rather than assuming the surface mount connector is at fault, I'm going to solder test leads on either side of the connector so I can check the signal on either side of the connector, and know for sure the connector is where the break is.
I've been practicing soldering very tiny stuff, and my skills have improved. First I soldered two leads at TP47 & TP48 on the "small board", just like Hobbit did. Hobbit figured TP47 & TP48 were "Test Points" 47 & 48, and found that they were an easy place to access the AVC bus on the small board.
Next I start looking for a connection spot on the "big board". I could just backprobe the external connector, but I was hoping to find a location closer to the board to board connector. That way the connector would be better isolated during testing. Two of these tiny copper grommets go to a couple short traces the go right to the board to board connector on the other side. However, the grommets are so tiny 30 GA wire will not fit through and the coating on the board is making tinning them very difficult. After a while, I give up on this idea.

Thanks to the SMD soldering tutorials on Spark Fun and some practice, I decide that I'm not afraid to try soldering my test wires right to the tiny pins.

This picture is just to give you an idea of the scale. This is a standard 3mm soldering iron tip (cold). As you can see, it's over four pins wide.

A couple months ago, during my first repair attempt, I had used the reflow station because it was easy for me. It's a lot like plastic welding, and I'm pretty good at that. I had wanted to use the soldering station, but after some practice on an old motherboard, I realized I did not have the skill to work on a $4000 component yet.

This time I had more experience with a 1mm cone tip soldering iron & solder wick, so I was not worried.

 

The first wire went OK, but then I bridged two pins when trying to solder the second. A few months ago I'd be freaking out. Thanks to Spark Fun, I'm not even sweating. A little solder wick, and I give it another try.

Success! By soldering wires to the pins, I've also resoldered the suspect pins. This time with 60/40 solder instead of reflowing the lead free solder used in the original construction. If my test doesn't work out because the problem has been fixed, I won't be too upset :)

I figure I may as well do the small board side of the connector with 60/40 solder while I'm at it.

Now that I've had time to think about it a little, I wish I had soldered wires here as well. I think that having the wire soldered to the connector leg and the pad may add a bit of support to the joint. Hobbit pointed out that the leg contact area on the new improved MFDs were much larger. Soldering the wire to the side of the leg & the pad should bolster the joint and make it stronger. Oh well. Next time. (Hopefully on a different unit!)
Time to reassemble. I use pliers to reconnect the small connectors to avoid damage.
Now I've got one red wire & one white wire coming from pins 60 & 58 on the big board connector, and another red & white set coming from TP47 & TP48 on the little board. I'll be able to "see" data on either side of the connector, and I should be able to tell whether the problem is between the external connector and the big board connector, or between the big board connector and TP47 & TP48 (which are sort of near the little board connector). If the connector goes open, I should see data that looks "unterminated" on the small board side and normal data on the big board side.
The unit is back in the car with the rear cover off. The test wire sets are dangling out the back for easy testing at the first sign of trouble.
But there's nothing to test, the MFD is working perfectly. After a little shaking, smacking, vibrating, and connector tugging, I'm convinced it was probably fixed by the resoldering. I suspect reflowing the skimpy bit of existing solder does not create a lasting repair, and that adding 60/40 solder will be a better solution, but time will tell.
I figure I may as well solder my test wires together and tape them up. They will act as an alternate route around the connector while the unit is in service. And if the does break again, they will be a way to test the MFD quickly without much disassembly. I worried a little that there may be a shielding issue, but the scope pattern looked fine.
Now that the MFD can talk to the stereo, the system check now says there are some AUDIO H/U codes.
None of the codes are current though, so I don't bother to look them up. I imagine they have something to do with a loss of communication.
After clearing codes and a test drive down our roughest local street, everything is looking good.
The LAN monitor is happy too.

If you like the idea of fixing MFDs or electronic parts yourself, I'd recommend the following resources.

Hobbit's website (http://www.techno-fandom.org/~hobbit/) has a ton of Prius related stuff, including the original MFD repair page.

Spark Fun (http://www.sparkfun.com) They have SMD (Surface Mount Device) soldering tutorials, kits to build for fun & practice, You Tube soldering demonstrations, etc.. They're sort of like HeathKit from the 70s. Lot's of fun, if you like that sort of thing.

The CAN bus class taught by Edwin Alfaro & Harvey Chan was quite excellent. You may be thinking, "CAN bus, it's just a couple wires connecting control modules; what's to know?". If you are, you haven't taken their class! They usually post on IATN (http://www.iatn.net) to announce classes.

Toyota's information site (http://techinfo.toyota.com). Training courses and of course vehicle specific information.