Moisture-Proofing the Griso Digitek Dashboard
by Moto
This is the first of a set of three (now four) posts that were intended to be parts of a coherent discussion of moisture-proofing from both theoretical and applied points of view. I probably should have posted all three as a single document. Since they've now been folded into a thread with the resulting discussion, I have put links to each of the other parts at their ends for convenient access.
The three parts are:
1. Theory and Dashboard Design
2. Adding a Desiccating Breather (http://wildguzzi.com/forum/index.php?topic=84968.msg1340261#msg1340261)
3. Applying a Conformal Coating (http://wildguzzi.com/forum/index.php?topic=84968.msg1340261#msg1340339)
4. Roy-Type Breather (http://wildguzzi.com/forum/index.php?topic=84968.msg1354221#msg1354221)
Part 1. Theory and Dashboard Design
Griso dashboards have a poor reputation for reliability and cost over $500 to replace. This has gotten many of us interested in how to prevent failures. Since I haven't been convinced we have attained an understanding of how to do this, I decided to take a fresh look. I'll report my findings and two practical preventative treatments here, hoping this will be of interest and/or benefit to others. I'll start with discussing principles and theories of dash failure, and the particular construction of the Griso dashboard. Followup posts will cover the two preventative treatments I have implemented, one involving complete disassembly of the dash and the other leaving it intact.
Possible causes of failures
I tried to understand what the cause(s) of the Griso dash failures might be in theory, the main possibilities being vibration, external wiring failures, and moisture.
Examining the dashboard (see photos below) and its flexible mounting, I quickly discounted vibration as a likely cause of failures. Besides its apparent high-quality build, I considered the background of the builder, Digitek SpA. Digitek was founded in 1983 in Modena, Italy, to provide onboard electronics (including telemetry) for Formula 1 cars. It subsequently expanded to supply dashboards for Ferrari, Lamborghini, etc., as well as mainly Italian motorcycle makers. I decided that if this firm didn't know how to deal with vibration, nobody would. (Digitek was acquired by Italian electrical equipment maker MTA Group in 2008, and still exists as an arm of what must be its later descendant, Microtec, also based in Italy. There have been some statements on the web that the Digitek that manufactured the Griso dash is the Australian electronics manufacturer of that name, but this seems to be wrong; I see the Italian firm's pre-acquisition logo imprinted on the dash circuit board, and this logo does not appear on the Australian company's websites that I have found; nor do any claims of constructing our dash.)
I don't discount the likelihood that external wiring problems account for some incorrect diagnoses of dash failure, but I decided only to pursue the moisture angle, since it seems pretty well attested, and because I felt I could make more of a contribution there. That being said, it is very important to investigate grounds and electrical connections outside the dash when a problem arises. My goal is to keep some actual dash failures from occurring.
Effects of Liquid Water
The big difference between the Griso dash and that old Marantz amplifier in your basement that's been running perfectly for 40 years is its operating environment. Being exposed to temperature swings, and also open to entry of moisture, the conditions in the dash often produce condensation, which occurs when the air inside drops to the dew point temperature. Condensation can affect the electronics -- on a printed circuit board (PCB) in this case -- in more than one way.
(http://i127.photobucket.com/albums/p135/motocomo/fogged%20griso%20dash.png)
As water vapor condenses, it forms droplets that can bridge circuits on the board. The condensate is pure, distilled water, which is a very poor conductor of electricity. But water on a PCB immediately begins to dissolve components, like glues, used in constructing the board, forming ions that increase its conductivity, making an electrolytic solution. If enough current goes through a droplet between circuit traces or terminals, the temporary short circuit can disable the board until the droplet has the chance to evaporate away. Worse, it is possible that the short circuit is serious enough to permanently damage the board, meaning the problem will not go away after evaporation.
Another possible outcome is the formation of dendrites -- metallic, crystalline, fern-like growths that are deposited molecule-by-molecule on the board between charged conductors. As they grow, dendrites make temporary short circuits increasingly likely when condensation forms, since the distance between the circuits to be bridged decreases. Ultimately, assuming the PCB continues to function long enough, the circuits will be physically connected by the metallic bridge of the dendrite, causing a permanent malfunction. Here is a photo of such a short-circuiting dendrite on a PCB (source: www.multisorb.com/files/9113/4315/2180/article-moisture-in-electronics.pdf (http://www.multisorb.com/files/9113/4315/2180/article-moisture-in-electronics.pdf)).
(http://i127.photobucket.com/albums/p135/motocomo/Dendrite.png)
The Griso Digitek dash is particularly vulnerable to short circuiting of all types because of the very small pitch of the layout: the conductors are very close to each other. (The small pitch was remarked upon by a web guru called Techrat1, who will be introduced more thoroughly later.) The small pitch means greater amperage short circuits through the electrolyte can occur at lower voltages, and that dendrites have shorter distances to bridge.
Layout of the Griso PCB
Here are photos of the upper and lower surfaces of the PCB in my dash.
(http://i127.photobucket.com/albums/p135/motocomo/IMG_20160628_222648758.jpg)
(http://i127.photobucket.com/albums/p135/motocomo/bf0bde32-23c6-44c8-96df-7a577c238252.jpg)
Most of the interesting stuff is in the lower photo. The white circle (3) is the tachometer needle stepper motor. Above it is the pin connector (4) to the Griso wiring circuits. The large gray rectangle toward the lower left (2) is the Fujitsu CPU. At the left edge of the board is the ribbon connector terminal (1) for the LCD display. To its right, the small rectangle (5) is the serial EEPROM, a memory unit that holds the parameters we set on the dash and also stores the odometer mileage record. At about the four o'clock heading from the stepper motor, right on the edge of the board, is an array of ten unused connector terminals (not numbered). These are a programming interface to the CPU.
I didn't figure all this out on my own. The information comes from an exploration of the Aprilia SR50 scooter's dashboard, done by user Techrat1 on www.apriliaforum.co m. (Here is the thread: http://www.apriliaforum.com/forums/showthread.php?180601-Digi-Dash-modding-or-repair/ (http://www.apriliaforum.com/forums/showthread.php?180601-Digi-Dash-modding-or-repair/).) That dashboard is a simpler version of the Griso's, lacking the CAN bus, and also the air pressure sensor that will be an important part of my tale, visible as a twelve-sided shape just below and to the right of the stepper motor (6, above). Here is a photo lifted from his thread, showing the similarity to our dash:
(http://i127.photobucket.com/albums/p135/motocomo/SR50%20dash%20by%20Techrat1.png)
Techrat1 was originally responding to a request about changing the color of the LCD display, but became more deeply interested in the dashboard. Aside from identifying components, he hooked up a terminal to the ten-pin connector and considered reprogramming the dash for various functions, including a Game Boy-like interface. (I think the CPU might be shared with the Game Boy, but didn't find out; the model number of the CPU seems to be different from ours, judging by the fuzzy lettering in the photo.) He found that the odometer reading in the external EEPROM could be read, and could be reset to zero but not to another value, because of a check sum that he could not crack. Along the way he demonstrated a deep understanding of what he was looking at, and made several comments useful to me, including the one about the very small pitch of the PCB, which is related to the condensation threat. (He was not concerned about condensation.) Hats off to you, Techrat1! (Let me know if you object to my pirating your photo.)
Sealing Against Water Vapor Entry
Though all of the damage will be done by contaminated, liquid water, the root of all this evil is water vapor -- disassociated H2O molecules forming a gaseous component of air. Water vapor is every bit a gas, as much as the oxygen (O2) and nitrogen (N2) components of air, and completely distinct from fogs or mists made up of fine droplets. Its molecules (H2O) are smaller than the oxygen and nitrogen molecules (about 2.3 angstroms). Once it is inside the dash enclosure it is free to condense into harmful droplets when the dew point temperature is reached.
At first I supposed the answer to the dash's condensation problem would be to seal it hermetically -- against all entry of gasses. But there are two things preventing this.
First, water vapor freely penetrates, or permeates, most if not all plastics. The molecules are small enough to travel between the polymer strands making up plastics, and the result is the equalization of the partial pressure of the water vapor inside and outside the enclosure. (They boldly go where O2 and N2 cannot.) This takes time to occur, but the result is that the water vapor in the enclosure reflects a long-run average of the environmental conditions. Even if the environment is only moderately humid, the water vapor in the enclosure will condense when the temperature drops far enough. Because of this problem, tables of Water Vapor Transmission Rate (WVTR) values are published. With the data in them engineers can model the permeation of vapor during the expected lifetime of an electronic component.
But there is a bigger problem that means the dash should not be hermetically sealed, even if it were practical. As mentioned before, mounted directly on the PCB is an air pressure sensor that is used to adjust the fuel/air mixture. This sensor is about 1/4 inch across, as seen in the two close-up photos below, before and after I removed the plastic cap that covers the transparent sensing membrane visible at the right. Successfully sealing this sensor from the atmosphere as part of a moisture-protection scheme would cripple the engine.
(http://i127.photobucket.com/albums/p135/motocomo/IMG_20160403_095235047_HDR.jpg) (http://i127.photobucket.com/albums/p135/motocomo/IMG_20160403_094606337_1.jpg)
The Dashboard Vents
The dashboard has two small vents at the back of its lower surface, looking like two posts protruding about a quarter of inch:
(http://i127.photobucket.com/albums/p135/motocomo/IMG_20160305_221740886.jpg)
As can be seen, the two vents are provided with baffles looking like half moons. Because no light is visible through them, they have often been mistaken for casting channels (I suppose) and ignored. (Blowing through them confirms they are vents.) Some have overlooked their presence at first, and thought they were added later in the production runs, but they are present in the drawings of the dash in the original workshop documentation. (And they are also there on the SR50's dash.) The Griso dash's vents should never be sealed up, because of the PCB's air pressure sensor.
What's Next?
In the final two posts of this set I will describe two methods of combatting the moisture problem. The first, non-invasive procedure adds an external desiccating breather to admit air from which water has (largely) been removed. The second, quite invasive approach is to disassemble the dash completely and treat it with a "conformal coating" that controls the approach of water vapor molecules to the PCB. I'll post pictures of what's involved.
I've read quite a bit about this stuff, and want to avoid going on at too much length. I can provide references to some sources I recorded, all from the internet, but including corporate product pitches, university-posted material for students, and scientific/technical reports in addition to forum exchanges of dubious provenance, such as this one.
Skip to:
Part 2. Adding a Desiccating Breather (http://wildguzzi.com/forum/index.php?topic=84968.msg1340261#msg1340261)
Part 3. Applying a Conformal Coating (http://wildguzzi.com/forum/index.php?topic=84968.msg1340261#msg1340339)
Part 4. Roy-Type Breather (http://wildguzzi.com/forum/index.php?topic=84968.msg1354221#msg1354221)