Stokes' ProjBlog

Come up and see my etchings December 9, 2007

I finally got around to doing some PCB etching. For my first experimental attempt, I made some little carrier boards for an SMD IC: a M62364FP digital/analog converter, intended for the analog meter project.

It’s been more than a month since I did the toner-transfer with a clothes iron to the copper-clad board (some incredibly thin stuff I bought at the MIT Swapfest a while back), but I’ve been reluctant to break out the ferric chloride and do the actual etching. The transfer was only moderately successful; there was a small degree of smudging. I don’t think the paper moved; I may have used too much pressure and/or heat. The smudging would have been acceptable for a board with larger traces, but it was enough to short some of the board’s traces. I had tried to scrape away some toner with an X-Acto knife before the etching, but that was only moderately (if at all) successful. I may try scraping away the tiny shorts with a diamond-tipped stylus; I think I have one around, somewhere.

The etching took a remarkably long time. I’d pre-heated the ferric chloride by placing the bottle in a tub of warm water, and the ambient temperature was probably 65°F, but it took nearly 30 minutes to complete etching. I etched two, one after the other, and each took about the same length of time. I should have probably left it in even longer; it could have removed the shorts.

While this wasn’t the most successful attempt, it was my first try. I’m optimistic. Once I get some heavier copper-clad board, I’m going to try making the motor controller board. More board, along with some 3A diodes, is in the mail.

Images added to the original Meter Thing post. September 30, 2007

Having gotten my camera out to take reference photos during a work field trip to the zoo, I took the opportunity to snap some pictures of the TINI and the meter collection. These have been added to the original Meter Thing project post.

‘The Meter Thing Project’ September 4, 2007

The ‘Meter Thing’ — to be granted a better name when one occurs to me — is a project to display on a set of vintage analog meters several numeric data (or data that can be turned into numbers) scraped from the Internet.

The heart of the beast is an older model TINI, a little Java computer with Ethernet and loads of I/O, all in the shape of a 72-pin SIMM — roughly the height, width and maybe a third the thickness of a pack of gum (although newer models use the DIMM form factor). This sits in an extremely simple breakout board, which basically just provides a RJ45 network jack and a breadboard-compatible set of DIP pins for all 72 conductors on the SIMM. The meters I bought in a lot of ten on eBay. They are all from different sources, are different sizes, and have different units and scales displayed.

The first challenge is to generate the analog signal for the meters. My original idea was to use 1-Wire digital potentiometers. The TINI supports 1-Wire exceptionally well, which is not surprising, as it has invented by the same people. 1-Wire would let me add meters as needed, just attaching them to the 1-Wire bus; I could then direct data to them by the unique ID of each 1-Wire potentiometer.

The problem with the digital potentiometers is the lack of range. Using the potentiometers to drive the analog meters directly did not give me the even coverage I need. The meters, regardless of their displayed units, are ammeters. Adjusting the resistance gives me a reciprocal curve (if I’m using the term correctly): x/1, x/2, x/3, et cetera.

I need some sort of ‘real’ digital/analog conversion; either PWM or some more sophisticated DAC chip. 8 bits of resolution are probably enough; if a meter has an arc of 60 degrees, that gives me increments less than half a percent. I have a couple of candidates in mind.

More background on this project and its current status to come.