Saturday, January 9, 2010

Configuration Control

Those of you who work with software know what SVN and CVS are. And you've probably found yourself in the position of having code that used to work several versions ago, but you've messed it up and you pull your hair out trying to figure out what went wrong. So, what you do is go back to the last working version and start over, making smaller changes.

It sure would be nice to have SVN for hardware.

I posted a video of my torch working well. I then welded it onto the chamber for testing and found that every time the spark triggered, the microcontroller would reset. In the standalone configuration, I had never experienced a reset. Unfortunately, I had made several changes at once that contributed to the resets.

First, I had made a new custom PCB to replace the protoboard I was using previously. Because the board was new, I focused on it as the source of the reset issue. I did eventually find an issue with the ground plane design which was contributing to the resets (It is the first PCB i had ever designed). However, the resets continued but now they only occurred when propellants were flowing in the igniter!

At this point I began to suspect that my ignition box was bad. I have been using a CDI box designed for RC engines for some time now and thought perhaps it had been damaged. After perusing this page: http://rocketworkbench.sourceforge.net/igniter/igniter_blog.html and talking with some of my fellow Arocket-eers, I thought I would try a simple inductive coil setup. I bought via mail several different hardware options including a replacement for the CDI I had been using (I'll post a follow on about all the different ignition systems).

The various setups still suffered from similar problems - at this point I was baffled! Eventually, I made an Edisonian discovery on the bench - if the spark plug electrode was shorted to ground when the coil attempted to fire, it would cause a reset every time. Placing a bead of water over the electrode and touching the ground had the same effect.

As it turns out, when I welded the torch to the chamber, I had welded a new spark plug bung onto the igniter. This was done as an afterthought - I wanted to clean up the multiple bungs I had welded onto the igniter when trying different threads for different plugs. My intention was to maintain the exact spacing, but I must have made the new bung slightly shorter than the original.

I added a spacer to the bung to move the tip back just a little --- and everything worked just as it had before. I inject both propellants as liquids and apparently the spark plug tip was seeing too many liquid droplets, causing the spark to not fire and the resets. All of the ignition systems work just fine in this configuration, including the original CDI box.

So, SVN for hardware would have saved me tons of time, headache, and grey hairs.

Stay tuned for a post about ignition options.

No comments: