I just bought a VacuTrace from Jim and really liked this device.
The first thing I wanted to do was to automate the collection of valve data. My thought was that if I could collect the digitized data, the curves could be fit to an equation which would then make calculation of tube parameters a breeze.
Neophytes, like myself, wouldn't have to struggle over basic tube equations to generate and design circuits--the math could be all part of an Excel spreadsheet. Thus leaving me free to tweak with the fun part.
I have a Velleman PCS 500 Scope, which can sample at up to 1gigHz, and connects directly to my PC for it's display. IT has an X-Y mode, so I thought I was home free.
If you look at my web site
www.entreprenertia.net (UNDER CONSTRUCTION PLEASE!!), and click on the "Hagerman VacuTrace work in progress" you will see two graphs. The top graph is the scatterplot from the DSO (digital sampling ocilloscope) as captured to the PC.
THis is pretty much useless for automated math since the traces are not separated. So I spent a little time, and used a physicist's standard toolbox--which is to cheat if you know a little about the hardware--and came up with a program that separated the traces in bins related to each grid step. This is the second picture with different colors for each trace. I threw away the bottom 5% and top 2% of the graph to make my job easier, but as you can see--
Now there is a plot that can be fitted to a curve and automated. The missing parts aren't going to affect design issues or curve fitting.
I'm posting this to get some feedback. Is there interest in continuing this path? If I post the Excel spreadsheet (all the math is in VBA for now) would it be of interest? Would others take the next step and create curve fitting programs, and put together the math? I'd be willing to post curve data for vitrually all tubes, or any tube sent to me, if someone doesn't have a DSO.
The math is tuned for my DSO, but the technique is simple and I think it could be modified to work with any DSO. THe key is that the horizontal scan is pretty much identical for each line, so 2-D binning and rejection is easy. (That is part of the cheating--a real mathematician would generalize the solution to work with any 2 dimensional space. I don't have the luxury of that much time.)
Feedback please??
Also--note what I'm trying to do on my web site. Does that make sense?
