I have been reading through the controller board threads and am very impressed by the level of sharing, cooperation and teamwork. I particularly enjoyed the thread about uli’s servo driver. I had thought about designing and building a similar controller, but this seems very done, but I had planned on starting with the Power Supply, and this doesn’t seem nearly so ‘done’. Obviously a line frequency transformer, a rectifier and some filter capacitors will do the job, but with the price of copper and the current focus on higher efficiency this solution is costly both initially and long term. A better solution would be a SMPS (Switch Mode Power Supply); however, SMPSs, while fairly simple in theory, present a great number of challenges to the average DIYer, and so the only available option is a commercial power supply to the tune of $200~$600.
Perhaps the greatest single challenge involves the transformer(s). Choice of wire size, core material, turns etc, etc can be a bit overwhelming. Even if all of these parameters are properly selected, obtaining transformers or even transformer cores in small quantities can be both difficult and expensive. I followed a thread in another section here where microwave transformers were used in an attempt to build a welding power source. It was fairly obvious to me that while a great deal of thought went into the project, the ultimate failure of the project hinged on an incomplete knowledge of the properties of the transformers.
With this in mind, I have been considering a completely different approach to a high VA supply. Unlike most SMPS designs this supply will need a great deal of flexibility with respect to the primary transformer. To achieve this flexibility I plan to use a micro-controller to optimize the switching elements based on feedback from the actual operating conditions. From a commercial SMPS design perspective this is an absurd expense; however, from a DIY’s point-of-view it offers tremendous flexibility at a considerable savings. That is, from a commercial design perspective a $50k investment in engineering R&D to minimize component cost is a trivial concern, but from a DIY point-of-view using a $10 component to allow flexible selection of other components is trivial.
Obviously a simple micro-controller is not going to take a $10 transformer from Radio Shack and magically provide us with an 800VA supply, but with some experimentation we might be able to come up with an easily fabricated transformer that will provide us with a 500VA to 1500VA supply. I have not begun experimenting with these transformer cores yet, and it may turn out that ‘off-the-shelf’ core materials simply cannot handle the flux densities requisite for a high-power supply and we have to fall back on a commercially available torrid core, but it is my hope this won’t be the case. Even if it were the case, having the ability for a single design to accept a range of available cores would allow DIYers a much better chance of success at a reasonable cost.
If anyone is interested in this project, either from a pure ‘user’ point-of-view or from a ‘wanting to help’ point-of-view chime in and let’s see if there is sufficient interest to move forward with this project.
Fish