Pinball Rehab

pinball repair and restoration

Repairing Alkaline Battery Damage Repairing Alkaline Battery Damage Hot video

I see a lot of CPU boards where battery damage is not being treated properly and as a result the alkali continues to eat away at the traces on the board.  Eventually it eats through a trace, the board fails and it's sent out for repair (often after being damaged further in an attempt to repair it).  This is a preventable issue which has been turned into a major issue.  It is important to note that a lot of board repair guys won't take boards with major alkaline damage since it takes more time to repair than the board is typically worth.

Friendly Reminder: Get the batteries off the CPU board!  

Note: While no boards were damaged in the making of this article, stunt doubles where used in some cases.  In other words, there are several different WPC CPU boards used as examples.

Potassium Hydroxide

The stuff (technical term) that leaks from alkaline batteries is potassium hydroxide, which is a very strong base with a pH of 12.  It absorbs carbon dioxide from the air to form a feathery crystalline structure of potassium carbonate (11.5 pH).  Before it transitions to potassium carbonate it is clear, so just because you don't see any white crystals doesn't mean there isn't potassium hydroxide on the board.  

Potassium hydroxide/carbonate is like a cancer in that it will expand and migrate across the board where it oxidizes copper traces and components leading to permanent circuit damage.  In Image 1 you can see where either through capillary action or electrochemical migration it has gotten under the solder mask (the dark areas indicate corrosion) and is eating away at the copper.

Image Gallery

Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage
Repairing Alkaline Battery Damage

In Image 2 it has completely eaten through the last 1/4" of the ground plane and the next portion is falling apart.  You can also see where the trace has lifted from the board due to the alkali destroying the adhesive holding it to the board.  Also note the corrosion on the inductor.

Evaluating Alkaline Damage

Noble metals are those that are resistant to corrosion and the less noble a metal is, the more susceptible it is to corrosion.  Lead-Tin solder is less noble than copper so as you look for alkaline damage on a circuit board the solder joints are often a leading indicator of problems.

The board in Image 3 has minor alkaline damage. It is limited to the one small area, and below the batteries the solder joints are still shiny and there are no dark spots showing through the solder mask. In this case removing the battery holder and properly flushing the board (vinegar/water/alcohol) is sufficient.

The board in Image 4 has severe alkaline damage. While some areas are rather obvious (the damaged components and the solder mask at the bottom of the board) other parts, like the dull solder joints and the dark spots showing through the solder mask, are not. All of that damage needs to be removed and simply flushing the board is not sufficient.

The board in Image 5 has alkali that was hidden until J212 was removed. It is unlikely it would have been neutralized by flushing the board with the connector installed.  This is why it is a common for components (and solder mask) to require removal during the evaluation and neutralization process.

Treatment and Repair

If a board has minor leakage then removal of the battery holder and a flush with vinegar/water/alcohol is sufficient. If a board has moderate to severe damage, it requires an involved process. The key is knowing what to look for and how to properly treat and repair the board.

Moderate to severe alkaline damage is best left to the pro's. Soldering alkaline damaged boards is extremely difficult since in addition to eating away at the copper the alkali will get under the copper and destroy the adhesive bond.  Even for an experienced solderer it is difficult to work on an alkaline damaged board without lifting a trace or pad, and it's more difficult yet if previous work was done on the board.

I use the following process to treat and repair boards with moderate to severe battery leakage.

  1. Inspection and Preparation.
  2. Chemical neutralization.
  3. Abrasive treatment.
  4. Chemical neutralization.
  5. Circuit board repair.
  6. Testing.
  7. Final cleaning and sealing.

Inspection and Preparation

Any socketed chip's that will get wet during the flushing process need to be removed. In the case of a WPC CPU board you can usually get away with leaving any chips above the batteries in place if you're careful while cleaning. While you've got the chips out inspect the lower part of the component lead for corrosion. If there is any corrosion you need to clean or toss the chip and replace the socket since it is likely corroded internally.

Note: When desoldering a board with alkaline damage you always want to clean the joints with a fiberglass pen and add flux. It is also common for the solder joints to have been eaten away and be low on solder. In that case you will typically need to add some solder prior to desoldering.  I'm not going to go in-depth on the soldering techniques, but I do suggest you review my soldering guide (see references), which also covers advanced repairs like jumpers, stitches and eyelets--which will all likely be needed.

So the big question at this point is which components should be removed. The majority of the damage will progress down from the point of leakage and spread out in an upside down V.  It is also not uncommon to see alkaline damage to the left and right of the battery holder, and in rare cases you will find corrosion on the upper portion of the board.

The battery holder needs to be removed along with any sockets, chips or connectors with visible corrosion or nearby corrosion under the solder mask. While we could do this after flushing the board, I prefer to remove these types of components first.

I always remove U18, U19 and U20 on a WPC board.  If there is corrosion on the dual-lead components below these IC's you can almost be guaranteed there is damage to the traces under these chips.  This is due to the fact that the corrosion migrates from top to bottom slowly over time and therefore has had a longer period to cause damage to the traces that are closer to the battery holder.

In Image 5 the left battery leaked and the alkali traveled down the left side of the board. So I removed J205-J207, J212  and the chips between them and the battery holder (there was no corrosion under J206 and J207 so they were re-installed prior to the photo).

In Image 6 I've already done some cleaning and you can see where the right battery leaked and how the damage spread as it moved down.  I did remove all of the connectors but there was no alkali damage under the left and center ones.

One important thing to note is the damage to this board didn't happen overnight. The longer you let a board sit with alkali on it, the more the damage spreads and the bigger the repair process. So if you swap out a CPU board with battery damage and throw the old one on a shelf, at least take the time to flush the board properly even if you do nothing else. 

It is important to note though that neutralizing the board will not remove any alkali that has gotten under the solder mask.  If you store the board in that condition the alkali will continue to eat away at the copper traces.  The board in Image 7 has been partially blasted and you can see where the alkali has started to damage the ground plane.

Chemical Neutralization

Everyone has their own method of neutralizing alkaline damage, but the goal is the same: neutralize the alkali with a mild acid, flush the acid from the board and remove any moisture from the board.  I use alcohol and compressed air while others may finish with water and heat.  Some will also finish with water rather than alcohol, which is fine as long as all moisture is removed.  Personally I think using the alcohol step adds an extra margin of safety.

The most important thing to note is that splashing some vinegar and water on the board is not sufficient.  You really need to spend the time to do this properly.

  1. Neutralize alkali with vinegar.
  2. Flush with water.
  3. Flush with alcohol.
  4. Blow off excess moisture.

I use a baking pan (wive's love when you use their kitchen utensils for your projects) to catch the fluids and place a piece of wood under the end I want to keep dry so the board is at a slight angle. A fairly stiff ESD safe brush and a syringe or squeeze bulb will also come in handy.

While rinsing with vinegar (a mild acid with a pH of 2.4) you want to give it enough time to completely neutralize the alkali. I will spend 5-10 minutes cleaning both sides of the board (most of the time on the component side). Use a syringe or bulb and make sure you get vinegar under any chips and sockets. While the board is soaking you want to scrub the board as this helps loosen and remove the alkali. Keep the board wet as you are working so the vinegar has time to complete the neutralization process.

For those who like to see things bubbling the video below shows the interaction between vinegar and the alkali. 

After neutralizing the Potassium Hydroxide we are left with water and a salt, which needs to be removed and any remaining vinegar flushed off.  Potassium salts are water soluble, and are not alcohol soluble, so water provides the added benefit of dissolving them.

While I prefer distilled water (potable water contains chlorine, fluorine and halides), unless your area has hard water then tap water is acceptable for the second step.  Since all we're doing in this phase is diluting and flushing off the vinegar you can move along more quickly. Although you do want to use plenty of water and a syringe or bulb to get under sockets and chips.

The reason our last flush is with alcohol (Isopropyl, 91% or higher) is that it will displace water. Again, make sure you get under any components.

Once you're done use some canned air or an air compressor to remove any remaining liquid from the board. You will want to do this from a couple of different angles and make sure you get under the components. We could let the alcohol evaporate, but this is faster.

In a production environment neutralization is typically done with full immersion in an ultrasonic cleaner.  Obviously this is not feasible for most hobbyists.  

Abrasive Treatment

Now that we have the surface alkali neutralized we need to expose any areas where it has gotten under the solder mask. If we don't it will continue to eat away at the copper.   I also remove any oxidation on the solder joints, but you can consider this an optional step if you're working on your own board.  If the oxidation is not removed though it will be difficult to reflow or solder those joints in the future.

If you do want to remove oxidation on the solder joints micro-blasting is the best way to get around the components and leads (this will be discussed more later).  You could do this with a fiberglass pen, but that approach is a lot more difficult and time consuming.

The following two images are after chemically cleaning the board.  In Image 8 you can see where the alkali has eaten into the board at U20. There are also several damaged traces and through-holes. The trace circled in red was intact but lifted from the board, and when I prodded it with a solder pick it broke. In addition the area above circled in red has cold solder joints from the alkaline damage.

I also circled some of the areas where the alkali has gotten under the solder mask.

In Image 9 you can see damage to the ground plane and corroded pads where the connector was removed. If these pads are not properly cleaned up solder will never properly flow through the hole and the result will be a cold solder joint.

In addition to exposing any alkaline damage under the solder mask during physical cleaning we also need to expose any traces/pads we will be working on or that are suspect.

One non-abrasive option I should mention is chemically stripping the solder mask.  The most common chemical strippers will contain methlyene chloride, a strong solvent. Not only will methlyene chloride based strippers quickly remove the solder mask, they'll deteriorate the base material if exposed to it for a prolonged period.  I would advise against this approach.

There are several options available for abrasive cleaning: Dremel tool with emery paper or abrasive buff, electric eraser, Scotch-Brite pad, fiberglass penCratex abrasive bullets or micro-blasting.  If you only need to touch-up a small area, or remove solder mask, any of these techniques will work OK and are easier to implement than micro-blasting.  On the other hand, if you want to remove corrosion from solder joints then micro-blasting is the best way to go since you can get in and around component leads.

Image 10 is before cleaning with a Dremel tool and Cratex abrasive bullets and Image 11 after cleaning (thanks to johnwartjr for the photos).

In Image 12 you can see where I have blasted the ground plane and lower part of the board (the pads for the connector will be finished up with some sandpaper), and in Image 13 the upper section has been blasted.  Note how much of the corrosion on the solder joints has been removed compared to previous photos.

In a production environment micro-blasting is typically the chosen approach for removing large areas of solder mask or corrosion on solder joints.  A micro-abrasive blaster has a small diameter tip (think of an airbrush) which allows us to surgically remove solder mask and corrosion.  It uses an ionizer and a conductive work chamber in order to minimize the risk of ESD damage and the small tip allows for surgical removal of the solder mask and corrosion.  

Since most people are not going to use micro-blasting at home I will not cover it in this article.  If you are interested in learning more there is a link to my micro-blasting article in the references (all it requires is an air compressor, $28 air eraser from Harbor Freight and the media).  If you are considering using a Dremel tool you will also want to read that article since you will need to take ESD precautions with that method as well.

I do though want to talk briefly about about ESD since many people ignore proper ESD practices.  It is estimated that up to 35% of total IC field failures are ESD induced.  Before you start thinking that you've never blown a chip due to ESD, the truth is you really don't know.  A large percentage of ESD damage does not immediately blow the chip, but rather causes internal damage which shortens its life.

Images 14 and 15 are microscopic views of components with ESD damage.  In Image 12 you can see in the upper circle where the circuit was damaged rather than completely destroyed as the others where.  In Image 13 the drain has been damaged on an NMOS gate and while still functional, it won't last as long as its buddies.

Chemical Neutralization 2

The reason we do a second neutralization is because alkali that was previously under the solder mask or copper foil and therefore not treated has now been exposed.  In a production environment we used Bromothymol Blue to test our process for the removal of all alkali.  What we found is that untreated alkali was exposed during the abrasive cleaning phase thus requiring a second neutralization step (not every time, but a reasonable percentage of the time).

When looked at under magnification the board surface and traces are not as smooth as they appear to the naked eye.  The solder joints in Image 16 are not uncommon and prime candidates for trapping alkali particles.  It is impossible using abrasives to get into these microscopic deformities and joints.

The other benefit of a second flush is the vinegar will lower the pH of the copper traces and reduce the chances of oxidation in the future.

Just follow the same procedure as we used the first time, although you can work a little faster with the exposed copper.

Board Repair

It is important that you understand what you're getting into before attempting to repair an alkaline damaged board.  I won't go into details on the specific techniques since they are covered in my soldering guide, but this is more akin to surgery than soldering.  The examples are very typical of a board with moderate to severe alkaline damage.

Here's the process I use.

  1. Visual inspection.
  2. Continuity test.
  3. Through-holes/pads.
  4. Traces
  5. Dual-lead components

I first do a visual inspection and continuity test on the circuits (typically switch matrix and/or dedicated switches).  It is important to do both since a trace can be partially eaten through and pass the continuity test but not work under load.  Obviously you need to be able to read a schematic for the continuity tests.

In Image 17 you can see the extent of the damage to the through-holes and traces.  The solder joints on the resistors above the chips are so corroded that the resistors will need to be pulled, the pads cleaned and new resistors installed.

Upon closer inspection some of the through-holes were able to be saved since the majority of the copper liner was intact.  I used a toothpick to push the liner back into place and will install a socket so no future soldering will be required on those connections.  I installed 5 eyelets on through-holes that were not salvageable (see Image 18).

I also pulled D7, D8 and C8 since the component's packages were damaged.

In Image 19 I'm starting to install jumpers and am beginning to pull the resistors above the chips and clean up the pads (I was out of some resistors and had to order them so I temporarily forged ahead with the existing resistors so I could do testing).  I use toothpicks to hold the jumpers in the through-holes while soldering.  For jumpers that are going from trace to trace I use Bondic (UV activated glue) to hold the jumper in place while soldering.

In Image 20 I've installed one socket and working on more jumpers, and in Image 21 I've got all of the sockets installed.  I ran out of machine sockets so one of them got a dual-wipe type socket.  I'm of the opinion that either one is fine.  While the machine sockets do have some technical advantages it's really not that important in a pinball environment.

In Image 22 you can see my "surgical" tool lineup.

If there is corrosion around the dual-lead components it is not uncommon to find one or more where the connection between the lead and the package has been weakened.  While these may be working at the moment, they won't be for long so each one is tested by gently prying up with a solder pick (see Image 23).  In fact one popped off on this board but I forgot to take a photo.

 Finally I reflow the solder on all of the dual-lead components.  I add flux to both sides of the board to help in cleaning up any corrosion and heat each lead (adding solder if needed).  I will admit this isn't absolutely necessary, but it helps narrow down any dual-lead components that need to be replaced.

In Image 23 I've circled two components where the solder did not flow properly (these are just examples and not all inclusive).  I will pull the resistors, spot neutralize the area using syringes with my liquids, mechanically clean the pads and install new components (the leads on the old ones are corroded and not worth trying to save).  Also any dual-lead components with damaged packages will be replaced.

Final Cleaning

Since you have been soldering on the board it is important to remove any flux residue.  This step should be done whether you used standard flux or no-clean flux.

Any remaining flux residue will impede proper adhesion of the sealant to the board.  This results in a minuscule gap between the board surface and the sealant (see Image 24).  The result over time will be corrosion just as we saw previously under the solder mask.

I typically use a toothbrush and Super HFE Electronics Cleaner from MG Chemicals.  Solvents such as Alcohol or Naptha can also be used.  Be aware though your results with different solvents will vary based on the type of flux used. 

If you have an ultrasonic cleaner large enough to fit a circuit board the UltraDust sold by Comet Pinball is an excellent solution and is similar to the type of solution used in a production environment.

Sealing the Board

The final step is to seal all exposed copper.  Some people prefer to use a clear coating so that any future alkaline damage can be spotted.  Clear conformal coating (MG Chemicals 419Cis a good choice for this since you can apply it to leads/pads and the board can still be reworked in the future since the coating will melt at soldering temperatures.

Another choice for small areas is a green conformal pen, although I am not thrilled with the color match or flow of the paint.

Of course we could use green solder mask, but it is difficult to work with and almost impossible to match the original finish.  From a repair perspective we don't really care about the chemical properties of the solder mask as it relates to the manufacturing process, only that it will protect the traces from oxidation and have an original look. 

I prefer to camouflage the repair using Pebeo Vitrio 160, a transparent paint used on glass which can be airbrushed or applied with a brush.  The Mint 37 is a perfect match on WPC boards and if you need a tad darker try the Tea Green 15.  For WPC boards the Mint is the best match.  Several coats are required since it is transparent.

When using paint do not cover leads or pads since it makes them impossible to solder without removal of the paint.  You will also want to leave a bare area on the corners where the board mounting screws go through.  This provides a path to ground, and I usually just add a little solder to the area to protect the copper trace.