M35A2 Turn Signal Flasher Repair...

[stumps]

Hi,

I just finished repairing a couple of M35A2 Flasher modules, and thought some might be interested. The problem is the circuit board is single sided, and the manufacturer drilled the holes too big. That means that if the component leads are not bent over onto the board, the potting compound will stress and crack the solder that bridges the gaps (over the too large holes).

The challenge in repairing potted blinker modules is to get the potting compound off of the circuit board. That is fairly easily done, if you have a source of hot air (some would say that I AM a source of hot air, but I digress...) I use a hot air work station for repairing surface mount circuit boards as that source... but you could use an oven set for 250F.

The potting compound gets soft as eraser rubber if it is warmed up, and it can be picked away carefully with a small bladed screwdriver, or an xacto knife.

Once the potting is off, you will be able to easily see the broken solder joints, and resolder them up using plenty of solder. Use 60/40 solder, not 63/37. 60/40 builds up better than 63/37.

The picture is the board after I have repaired the bad joints. It is now blinking along merrily.

-Chuck

My next project is to fix an alternator regulator that has the same failure.

 

[leepotter]

What will you use to cover the board back up?

 

[DieselBob]

Thanks for the FYI.

 

[stumps]

What will you use to cover the board back up?

I plan to use clear GE Silicone II caulking compound. I have found that it is non corrosive, and does a pretty good job if you are neat about it. If the new Loctite silicone windshield repair compound is non acid curing, it would work very well... it pours like liquid...I haven't tried it lately, so I don't know if it is the right stuff. A pourable 2 part RTV compound would be best, but it is too expensive for such a small job.

Note, you don't want to use any of the RTV silicones that smell like strong vinegar. And you definitely don't want to use epoxy, unless it is loaded with a rubbery filler like the original.

-Chuck

 

[John S-B]

Would a heat gun (for removing paint) work, or is that too much?

 

[Nonotagain]

If the sealant was being used as a vibration reduction compound, then some good choices would be DC 732 or 3145. Momentive RTV 102 (two part) would be a very good choice.
If it was used as a dielectric compound I'd suggest using an acrylic conformal coating over an epoxy or silicone coating as the latter are difficult to remove once cured.

 

[stumps]

Would a heat gun (for removing paint) work, or is that too much?

A heat gun is fine, but remember the case that holds the potting is plastic too, and will melt if you let it get too hot.... 250F is about as hot as you should let it get. Apply a blast of heat, and then pick at the potting... more heat, and then more picking...

-Chuck

 

[stumps]

If the sealant was being used as a vibration reduction compound, then some good choices would be DC 732 or 3145. Momentive RTV 102 (two part) would be a very good choice.
If it was used as a dielectric compound I'd suggest using an acrylic conformal coating over an epoxy or silicone coating as the latter are difficult to remove once cured.

The sealant is there mostly to keep moisture out. Remember that the deuce can be submerged.

-Chuck

 

[cranetruck]

While it's exposed, trace out the circuit. Did that on several units for my MVM article ("Solid state devices for the deuce", Feb 2004 issue #101) and found some do not have overload protection, the power transistor simply "blows" like a fuse.
There are also special chemicals available to remove the casting epoxy, but I like your method since it doesn't effect the components so much or at all. I simply chipped away the epoxy at room temp...

 

[cranetruck]

There are small kits of casting epoxy formulated for electronic circuitry available from places like Alllied Electronics, Newark etc. Try MG Chemicals Part# 832B-375ml for example.

 

[crazywelder72]

don't they have to be cured with heat?

 

[cranetruck]

don't they have to be cured with heat?

Depends on the product, this one cures at room temp in a couple of hours...checked the package, actually fully cured in 24 hours...

 

[stumps]

While it's exposed, trace out the circuit. Did that on several units for my MVM article ("Solid state devices for the deuce", Feb 2004 issue #101) and found some do not have overload protection, the power transistor simply "blows" like a fuse.
There are also special chemicals available to remove the casting epoxy, but I like your method since it doesn't effect the components so much or at all. I simply chipped away the epoxy at room temp...

It is a lot harder to get the board out the rest of the way. Since I was 99% sure that I would find broken solder joints (differential expansion of the epoxy potting compound, and the copper wire on the component leads), I only exposed the bottoms of the boards. I have another bad blinker module, to work on... it is actually cracked, so I will remove the board and trace it out.

-Chuck

 

[stumps]

I did some further checking around... MSDS's, etc... and the Permatex flowable silicone windshield sealant should be perfect for potting the board back up. Basically, you don't want to use any of the silicones that make acetic acid (vinegar) on curing. Those that shouldn't be used will list Ethyl triacet oxysilane, and Methyl triacet oxysilane in the ingredients list.

-Chuck

 

[Wildchild467]

So do you think that mounting the flasher on the firewall vs. the fender is a better idea? it seems like there is less vibration on the firewall. i mounted mine on the firewall and had no problems yet. I was also thinking i could mount it using the rubber isolaters that are used on the deuce headlights. just dont forget about the ground wire and i think i would be all set.

 

[stumps]

Here is a picture of the potted unit, and what I used for the potting compound.. Plain old Permatex Flowable Windshield & Glass Sealer. Make sure that you read the label, though... The old version of this sealer was the type that made acetic acid while it cured, and the new is safe for electronics.

As to the vibration, that isn't what is killing these flashers. The problem is the dissimilar expansion of the potting compound, and the components on the board. The potting compound pushes and pulls on the solder joints every thermal cycle. Eventually it will break again. The best way of protecting against that problem is to crimp the leads that pass through the printed circuit board so that there is a good mechanical connection... barring that, use plenty of solder.

-Chuck

 

[cranetruck]

Hmm...I think your broken solder joint is a very rare occurrence. These boards are wave soldered to exact specifications, the potting compound formulated to match the electronics thermally and electrically. Billions of applications in the automotive, military and aerospace fields can not be wrong.
The potting of the boards prevent damage from vibration etc, but also hides design compromises by the manufacturer. The early solid state flasher modules (=from the sixties) were designed to handle short circuit overloads until the CB would trip. Expensive, with SCR and associated electronics was later easily replaced with a cheap flip flop circuit and a power transistor using commercial components...apparently in some cases without short circuit protection...the potting compound hides it all.

 

[stumps]

Hmm...I think your broken solder joint is a very rare occurrence. These boards are wave soldered to exact specifications, the potting compound formulated to match the electronics thermally and electrically. Billions of applications in the automotive, military and aerospace fields can not be wrong.

Where to start? I work in this industry as an engineer. There are very exacting specifications, as you say, but this manufacturer didn't follow them. The broken solder joints I found were on the IC, where the holes were easily 50% too large, and the lead ends were not crimped over to make a strong mechanical joint. They were destined to fail from the day they were made. There is a second area where these blinkers will fail, the leads on the TO3 packaged transistor that does the switching are similarly wrong. Mine were showing cracks in the solder fillet, and would have soon failed.

The potting of the boards prevent damage from vibration etc, but also hides design compromises by the manufacturer. The early solid state flasher modules (=from the sixties) were designed to handle short circuit overloads until the CB would trip. Expensive, with SCR and associated electronics was later easily replaced with a cheap flip flop circuit and a power transistor using commercial components...apparently in some cases without short circuit protection...the potting compound hides it all.

On the contracts where I have made this type of device for the US Army, I have been required to provide all design materials necessary to produce the device: schematics, pcb layout, drawings for molds and cabinet parts, etc.. The potting is there purely to protect the electronics from vibration and the environment. Anybody within the government that needs to know what was done can look at the schematic. Anybody outside of the government that wants to know what was done can do what I do and remove the potting.

I have done repairs and post mortem analysis on dozens of such devices. The most common failure I have found is broken solder joints.

-Chuck

 

[steelandcanvas]

Very interesting thread. What I've gained here wll allow me to fix the module myself, instead of tossing it in the trash and buying another unit. Thanks Chuck!

 

[TIGERFANS2]

anyone have the PN for the power transistor, I think a friend's might have gone south when he was trying to wire in a trailer connector