First Week of May 1999 - Reservoir Tank
We have found a suitable container for a replacement reservoir tank so we have spent some time fitting the new tank. It is much better, it does not leak!, and we do not get the frothing of the oil now. The down side is we are now using more oil and the tank is a slightly heaver than the original.
We have also started on the nitty-gritty programming for the walking software. We are using the backbone of the Jim Struts software, but have stripped out the majority of the walking routines and started again. The reasoning behind this was the Jim Struts software was structured around left and right routines, this became a handicap in controlling all four legs at the same time. We tried during January with limited success. The other main difference is we are using precalculated look-up tables for the loci of the feet instead of tables calculated on the fly. In theory we should be able to have a more sophisticated step patterns.... in theory.
We spent the weekend working on the software. On testing, when requested to walk forward, the legs would moved to their maximum or minimum extents, and then stop moving. Not good... Then spent hours looking through the code to no avail. Then spent the rest of the weekend making up and hooking into Miss Struts computers a Liquid Crystal Display(LCD) to help debug the walking program. This has proved a very worthwhile task as within five minutes of getting the LCD running it showed that the Analog to Digital(A>D) convertors for the leg sensors where incorrectly assigned.
We gripped the A>D allocations and the walking cycle now operates correctly. It is quite entertaining because the hydraulic solenoids are switched on and off so quick you see the rams juddering and making banging noises which you can just hear over the engine noise..... like a growl.
Did Miss Struts walk, well the answer to this is no! As each leg of Miss Struts is lifted off the ground that corner of her dips down. We made some minor modifications to her feet and we have seen her walk about a foot but only in one direction. Investigations continue..........
Well after having a think all week, finally realized that the walking sequence was phased incorrectly and the exactly opposite foot was been lifted off the ground as compared to what was required. The correct sequence was reprogrammed and Miss Struts stepped out about six foot forward and about the same back. We are all smiles now. Now we just need to solve how to get her to turn around corners. On checking over the legs, we found, that where the legs rub on the chassis there is excessive wear. The plastic legs were not going to survive basic walking, no chance in a battle. So then spent a lot of the weekend rebuilding the legs. The down side of this is it has added 1.6 kgms to her weight. While rebuilding the legs we did some minor modifications to the software, and during one of the programing cycles managed to plug the CPU incorrectly on to the mother board. This had the result of burning out one of its A>D ports and one of the outgoing data bus ports. The smiles have now disappeared. Oh dear.................
It has been a very busy weekend which despite some set backs ended on a high. During the week we worked furiously at the walking software to get Miss Struts to turn, and took delivery of a replacement CPU chip. During the weekend there was many visits to the EPROM eraser, and program compiler, staring at feet and LCD displays, with the net result that Miss Struts turned mid-Sunday afternoon. We tried two turning strategies of which one worked really well and the other was a waste of programming time. We then spent the next hour letting our driver practice maneuvering, until one of the legs jammed, then spent the next couple of hours with a big hammer and file fixing it. Getting Miss Struts to turn was anticipated to be the most difficult task in the programming and so the smiles are back and are huge. We are now making preparations for the auditions this coming weekend.