Accelerator Rebuild - lessons learned

Often called the "big hand" syndrome, when the NHW11 accelerator gets electronically noisy the car goes into "safe home" mode. It is as if a big, invisible hand reaches down and slows the car. If one parks, turns off the car, and rapidly cycles the accelerator a half-dozen or so times, one can restart the car and keep going ... until it happens again.

Inside the accelerator encoder are two, independent, variable resistors:


Taking a Dremil tool, Doug Schaffer pioneered removing the thin plastic lips that hold the cover sides and rounded end so the cover can be removed showing the internal wiring:

The six pins on the right contact the cover and the little fingers on the rotating disk are the wipers. Notice only the lip of the straight section on the right remains as the side and rounded area lips have been removed WITHOUT going so deep as to remove the recess that aligns and holds the cover.

The cover has the carbon traces that provide the variable resistance when the disk rotates:

You'll notice there is a rubber seal that keeps out dust and dirt.

The past rebuilding practice has been:

• remove the cover carefully just taking off the thin retention layer
• using contact cleaner, wipe the cover, fingers, and re-tension them with a small flat screw-driver
• clamp the cover on and apply a JB Weld epoxy bead to reseal the cover, takes hours until properly cured

Doug used a battery and speaker so when the accelerator was pressed and released, he could hear the electronic 'noise.' I mounted the accelerator on a board and using jumpers and a turnbuckle, measured the resistance at different turns:

This shows the accelerator on the test jig and the six, exposed pins.

Here is the turnbuckle configuration:

With this configuration, jumpers and a volt ohm meter, we could observe and plot the resistance changes before and after repair. Then I would swap it with the accelerator in our 2003 Prius and ship that one to someone needing a rebuild.

This is a before and after plots of accelerator resistance:

Now what you can't see in this graph is how the resistance would vary with each turnbuckle rotation. One turn and the resistance would spike up and then wander down yet still rest at a higher than the cleaned value. It was as if some random semi-conductor were in the encoder wandering all over the place. Disassembly and cleaning 'cured' the problem yet unmagnified examination of wipes showed no evidence of any significant debris.

Last year, I learned about the NASA tin whisker investigators at Goddard Space Flight Center:

In fact, they had looked at a Camry accelerator with similar problems to the NHW11. Of the lessons learned:

• modern VOMs may or may not detect tin whiskers
• older VOMs can 'burn out' the whiskers when trying to measure the resistance
• tin whiskers can form a thin oxide layer and have semi-conductor, variable resistance
• less than one volt can 'burn out' existing tin-whiskers

So as an experiment, I used a 'wall wart' power supply and jumpers on known, bad accelerators. First I confirmed they had the problem; then applied the 12V 'wall wart' power to the different encoder pins and cycled the accelerator. When I then went back to check, the resistances were solid and did not wander. In effect, the 'wall wart' burned out the tin whiskers.

Now I had been rebuilding accelerators until my bank started charging commercial accounts with too low of a balance a fee. I did not want to keep $2,000 in an interest-free, bank checking account needed for PayPal payments. So I closed down my rebuilding business and sent my inventory to the GSFC tin whisker team.

Today, my recommendation is to take a 9V battery, jumpers and apply it against the wiper pin and fixed pins and cycle the accelerator. Do both potentiometers and this will 'burn out' any tin whiskers. This will not prevent re-growth as the GSFC team has documented but it is a much easier, faster, and less invasive procedure than having to remove the cover, clean, and JB Weld the cover back on.

The other approach is to buy a replacement accelerator either new from Toyota, ~$500, or salvage (if you can find one.) Experimentally, it looks like Scion encoders are in the general resistance range and might swap for the NHW11 encoder. I have one but not yet 'done the experiment.' But at least you now have options.

GOOD LUCK!
Bob Wilson

 

Checking with the schematic:

• 1-4 fixed - 2 variable
• 3-6 fixed - 5 variable

Connector A13 has two rows:

• 1,2,3 - latch adjacent row
• 4,5,6 - second row

Without a VOM, I would put put the jumper on one of the center pins and then visit the outside four pins and cycle the accelerator with each pair. Then move the jumper to the other center pin and do the same. Of course if you don't have a VOM, you have to touch the battery to your tongue . . . which will remind you to buy a VOM.

Thanks,
Bob Wilson

 

This is a hard problem. I know there are folks who are advocating a general recall and the only way that makes sense would be to recall and replace all similar design accelerators including Scion and our NHW11. But what to put in its place?

The dual-pots connect to hybrid and/or engine control computers that are expecting given range of values. But the dual-pot, resistance range varies as I found testing a similar scion encoder.

Conversations with the GSFC team suggests 'tin whisker' growth is not predictable. Some metals are more prone to growth (tin and zinc) yet when and how whiskers form remains an area of active investigation. The whiskers look like 'metal crystals' but no one knows, yet, what causes the nucleus to form or how adjacent atoms get sucked from the metal matrix and added to the base of the whisker.

The most likely failure is the control computer detects the problem and puts the car in 'safe home' mode. Doug developed the rework process in response to the $500 part + labor cost of a latent design defect. My thinking is the fair answer is an 'extended warranty'.

As for unintended accelerations, this remains a risk. The first time a defense lawyer gets a professional engineer to analyze and find tin-whiskers in an accelerator . . . add to that the GSFC documented case . . .

Bob Wilson

 

Some of the earliest reports were from tube radio, tuning caps. Still Toyota went with Hall effect sensors in 2004 and that pretty well eliminated the problem.

Bob Wilson