And now a warning:

I THINK this is a likely fix for many ARC problems. I THINK that even if it doesn't fix the issues, it can't make things worse. I THINK everything I am saying here makes sense, and it definitely worked great for me. However:

1) I've only tried this on one board
2) I only did this yesterday and haven't subjected it to extensive testing yet
3) It's quite possible to blow things up if you mess something up. Case in point: while testing in the car, I slipped with my voltmeter probes on the display board and ended up smoking one of my ARC computers. Fortunately it ended up just being a burned PCB trace that I was able to bypass and fix, but it could have been much worse.

I suggest you proceed at your own risk until more data comes in. My hope in posting this is that someone else out there (likely with a disconnected display board who has nothing to lose) will do some work to validate my findings.

I will keep you posted of any updates in my testing.

 

Procedure

1) Remove the ashtray

2) Remove the two screws on the ARC display and remove the display from the dash. Disconnect the three connectors.

3) Unscrew the two screws on the back and remove the clock.

4) Unscrew the three screws on the back holding the display panel together.

5) Gently separate the white back of the display panel from the black front. You can see the catches around the edges - pry these gently and separate. When you do this, there are two lenses over the master warning light that will probably fall out. Don't lose them.

6) Remove the black piece with the graphics from the white back of the display panel to reveal the circuit boards.

7) There are two boards: the front board with the LEDs and the smaller rear board with the plug connector. These are joined by a short flexible cable. Gently slide them out.

8) The voltage regulator is on the rear board under the big heat sink. The heat sink is riveted to the regulator and then attaches via a post to the circuit board. To remove it, I heated the underside of the board to melt the solder where the post goes through the board. Then I gently pried the post off of the board using a small screwdriver. Once I did this, I desoldered the regulator/heatsink and removed them.

9) I re-used the heat sink by drilling out the rivet. I then used a pop-rivet to attach the heatsink to the new regulator. BE SURE TO MOUNT THE REGULATOR IN THE SAME ORIENTATION or else you'll get the pins backwards. The pinout of the new regulator is exactly the same as the original one.

Edit: I suggest you use thermal grease in step 9 between the heatsink and the regulator

10) Bend the pins and solder the new regulator into place. the heatsink is grounded, so make sure it will not touch the other board when you reassemble everything. The post was there to keep the heatsink from vibrating and cracking the solder, so make sure you use some method to mechanically fasten the heatsink to the board to prevent this. I used tape but will probably try to come up with something better.

11) Reassembly is the reverse of assembly.

 

Someone should sticky this. I'll test my ARC display that went haywire when I go back home for it.
Thanks for the writeup!

 

Hmmm, I'll go to Radio Shack and get a voltage regulator. I've noticed that in changing the displays that some are more erratic than others.

Keep it up Gubi! I'll report back if it works or doesn't.

 

Oh Tom! This is awsome news! (but why couldn't you have done this a few months ago )
If this works, and continues to work, you may have found the cure for one of the most irritating gremlins that these cars suffer from
(in case you are wondering why i said about doing earlier, ive removed mine, and its wiring completely )

 

G'Day Tom,

Sounds like a possible fix, I had assumed that the 5V regulator was a 7805 but have now opened one up to find a TEA7605SP rated at 500mA as you have stated.

The main differences between the two are output current (500mA v 1A) and input voltage (the TEA7605SP only requires an input voltage greater than 5.6V to give 5V out whereas the 7805 requires greater than 7V in to give 5V out).

So using the 7805 I think the only downside (apart from stuffing up the changeover) is that there could be funny things happening with your ARC during cranking with a less than fully charged battery. This may not be a problem at all.

 

Sounds very interesting. I think I may wait for the testing to occur, but it'll be easy enough to get a regulator from work.

On a side note - Over the past two months or so, my ARC Trip Computer has been freezing. I'd come out to the car and find the clock had stopped and nothing would work, the three buttons would make no difference. I'd have to disconnect the battery and reconnect it for it to work again. It's done it three times now. Has anyone else ever experienced this before? I'm yet to get a chance to switch it over to the other trip computer we have, but it's incredibly bizarre.

ROCK ON
R~R

 

That's what I figured from the specs, Craig. Interestingly I did notice it going a bit wacky today during cranking (car took a little bit to start) but it settled down once cranking stopped. Beyond that it's been working utterly flawlessly, which has been sweet.

I'm pretty certain they used different components depending on when these boards were built - I know that the ICs on the two boards I have are from different manufacturers. I don't remember the part number, but I know the regulator I pulled out was ST manufactured and rated at 500mA.

[Edit: it's an ST L4705CV]

Fry's Electronics around here carries NTE components and the NTE equivalent part was a NTE1951 (1A). Unfortunately this isn't one of the parts Fry's stocks on the shelves, so out of desperation I went to Radio Shack and was pleasantly surprised to find they had an equivalent in stock.

For the record, the web page for the Radio Shack part says 1A, but I'm 95% certain the packaging on the one I bought said 2A. Likely 1A is enough, but it's voltage regulated so no risk of blowing something up with a higher amperage version.

[Edit: I'm wrong, the Radio Shack one is a 1A 7805 as it says on the web page]

 

G'Day Tom,

Yes the NTE1951 would be a better choice with the lower drop out voltage but the 7805 should suffice and is as common as dirt. The 78xx series in the TO220 package is normally rated at 1A but since the theory here is that 500mA is good enough most of the time (but only just), 1A should be more than enough.

I guess we need more people to try this on known dud ARC's and report their findings.

 

I just finished trying to repair an ARC using the GUBI mod. The display I used was totally dead and was completely erratic. IT WORKED!. The display worked perfectlly from cold. No Milano Christmas tree. I couldn't believe my eyes. I swapped back to the one that I used before and sure enough, dancing lights. On the repaired one, the connection between the cable and the display was a little dirty so I used CAIG contact enhancer and it was working 100%.

I did this one fast. I think I'll use thermal compound between the heatsink and the voltage regalutor. I used an aluminum rivot to hold them together.

I'll report back after testing! If it works without fail I'm buying GUBI pancakes!

 

Pancakes? Dude, if this works, every Milano owner in the world owes me a beer

Glad to hear of some confirmation. I wonder, perhaps, if one reason these got worse with time was that the regulator gets hot and degrades over time? I haven't checked how hot that heat sink gets in normal operation.

I suppose if that's a contributing issue it's easy enough to add some vent holes to the housing.

 

Nice work

Great job. I don't know why I didn't check that.
One word of caution though, NTE doesn't have the best QC, but since that
component is obsolete & the markets flooded with knock offs & remarks, you have to get what you can.

Regards,
John

 

hmm I took out the part above the glove box and it seemed to be fine I even changed out all the caps.
I did not know there was anything on the other part I will take mine out and take a look.
I was starting to think that the flash in the MCU was starting to lose the program. but I never went back to look at it anymore. I was going to go log all the voltages and see if they were in the right range but never got to it. I was thinking I might have to make a new box from scratch. but all this time I was thinking the display PCB had no active parts on it.

 

I pulled it out and swaped in a normal 7805 I have a box of them.
I did not reuse the riveted thing I just put in a standoff with some locktight
I think the whole thing took like 15min. start to finish. And it looks better allready. the break light went off as soon as the ABS pump pumped up. and the hand break light went on/off with the hand break.

I have not driven yet. but it looks like it is working.....

 

Used a LM340t5, sub for NJM7805a or AN7805f. Went through several hot/ cold cycles &
viola, no more ARC disco shows. I feel so ****y now I might even try to fix the oil level
sensor. Thanks a million Tom.

John

 

Tom - Roger B. & I swapped out the stock voltage regulator with a LM7805 ($0.69 at Halted), reused the original heat sink and ... IT WORKS

Yes, we do owe you a beer.

 

Sweet. Sounds like we've got this mo-fo licked

 

How do we pin this or make this part of the FAQ?

 

No more use for Louis' GIF then. He should still post it here anyways..

 

Hahaha. Just a picture of a blank ARC maybe?

I meant to pick up one from work, but forgot.

It's amazing though, it's taken 20 years for someone to finally figure out the real problem! A $1 problem.

ROCK ON
R~R

 

Yeah, a one-buck part and an incredible number of man hours to figure out the problem. Somewhere there is an Italian electrical engineer who made my s--t list! I mean, seriously, Alfa sold these cars for how many years and couldn't fix this?

I'd been working on this quite a while, actually, swapping parts, checking voltages, checking grounds, checking current flows to ground, and just generally trying to figure out what was going on. Once I had enough parts stockpiled, I was able via swapping to nail down that the ARC computer didn't make a lick of difference but the display did.

After that, it took my buddy and I a couple of hours to figure out what the board was supposed to do. It's a pretty simple circuit, but we had a hell of a time figuring out the details for the driver chips so it took a while - just couldn't find them anywhere. Eventually we found a similar VFD driver chip with similar specs.

Once I knew how the circuit was supposed to work I spent an afternoon just looking at voltages on the board as various things happened, and even with a DVM I could see the voltage get noisy as things turned on and off. Using the freeze spray to further isolate the problem helped a lot. Hell, if I'd started with the freeze spray I probably would've figured the whole thing out a lot faster.

I actually bought an HP 1631D logic analyzer and oscilloscope to check for electrical noise and read the digital data stream but ended up not needing it. No big - been looking for an excuse to buy an oscilloscope for a long time.

So now I have a working ARC and a cool oscilloscope. Life is good

 

You should make a roll call of all Bay Area Milano Mafioso (BAMM!) to fix their ARC's and come down to IHOP to show off their (working) ARC's

 

I bought one of the VR's from Radio Shack this past weekend as I wanted to make the fix on one of my spares. I'm going nuts now trying to find even one of the two ARC's I know I have buried in the many totes of Milano and Spider parts I have in the basement! I feel like I'm on a Snipe hunt.

Thanks for posting this fix!

Kevin

 

Wow

Gubi what you have done is the Milano electrical gremlin equivalent of inventing penicillin!
Congratulations and thank you. My Verde display works well but no doubt some day it will need this fix.

 

Okay, maybe not *completely* fixed. I took a long drive on Saturday and by the end of it I was getting an occasional blink of the master warning light. As I mentioned earlier, I'm guessing the issue here is probably heat since this didn't happen on shorter trips.

Since I wasn't really happy with the tape method I'd used to attach the regulator anyway, I took it back apart and re-soldered it, this time properly attaching it to the board with a stand-off. I also took the opportunity to dab some thermal grease between the heat sink and the regulator to ensure good thermal contact (it was a little loose with my original rivet, and with no thermal grease the heat transfer may have been poor. This is what I get for rushing through a job)

We'll see if this shows further improvement. If not, I'm going to drill some holes in the plastic case to improve air circulation.

 

Tom - when I reassembled mine, I found a standoff from my old PC that fit, along with using a generous amount of thermal grease to the (reused) heat sink. So far so good. I actually have to pay attention to it now because it actually does tell you what's wrong