A Quiet Path

As with anything, real life can step in at any moment and wreck a lot of plans.

Read the rest of this entry »

My Diabolical Device uses stepper motors, which have drivers and a power supply. The drivers allow me to do things like set the motors to do steps of size 1 (meaning 200 turns per rotation), 1/2 (400 turns), 1/4 (800 turns), and so on. The power supply, naturally, powers the motors. Because the motors are powering electromagnets on and off internally, you can only send “step” commands to them up to a certain rate, otherwise they can’t power those magnets fast enough, and you start losing steps.

When I started building my device, I was using a 24V power supply that came with a 3 stepper motor kit I’d bought online. From reading various posts about these motors, I determined that a 36V supply might do better for me, since I use 4 stepper motors in the device, rather than the original 3.

However, after testing out the new power supply this morning, I’ve discovered that I can still only drive my motors at a rate of around 8000 Hz when set to 1/8 microstepping - that means I’m only getting about 1000 steps per second, or 5 rotations per second (note that changing the microstepping results in sending fewer commands - at 1/4 microstepping, I am only able to send 4000 commands per second, and so on, so the rate stays about the same, but 1/8 microstepping gives me finer control). The movement screws I use are 10 turns per inch, meaning that I’m theoretically getting around 30 inches per minute of movement along the screw. This sounds like a lot, until I compare it to other uses of these motors, where people effectively get 200 inches per minute of movement. I was hoping for a lot more out of this power supply, really.

I’m not sure where to go from here on this. I could accept the current rate, or I could try replacing the movement screws with something that would give me more than the 10 screw turns per inch of movement. Replacing the screws will cost around $120, and considering my recent and upcoming dental work, I think I’ll wait on purchasing anything else for the machine.

For now, I think, back to coding, to see if I can squeeze any more optimizations into that side of things - fewer movements will mean quicker run times, no matter what. And there is the possibility that the way I’m controlling these motors is incorrect or non-optimal in some way. Perhaps I should look into that code again, maybe I can get more speed there.

Yesterday morning, I got up to water my plants, as I’ve been doing every day since we’re in the middle of a heat wave out here. Surprisingly, I had a very tiny visitor, though - a ruby throated hummingbird was hovering over my patio, and apparently liked it there, since he started chirping at me, I assume to scare me off his new territory.

I guess I’ll have to get a hummingbird feeder now.

That night, I went in for a root canal. I remember reading that my chemo treatments could cause problems with my teeth, which I guess turned out to be true. One of mine completely broke, and needed the root canal and crown treatment. The crown will take another 10 days or so to get back from whatever facility they make such things at for my dentist, so for now I have a plastic replacement which I’m not supposed to chew anything sticky with. Unfortunately, they weren’t able to completely numb me as they expected, so I got a few jolts of major pain, but really it wasn’t as bad as I’d been led to expect. Afterwards, I was sore for about an hour, but I don’t really feel it at all any more.

Next week - wisdom teeth removal. I’m actually looking forward to that, though. The lower right one broke (much like the other tooth I just had crowned), and it’s poking at the side of my tongue.

Over the weekend, I was trying out some new (to me) materials for costuming. I want to carve designs using a Dremel power tool, to create various armor and other props. So, for my tests, I attempted:
* Wood (Poplar)
* Wax (block candle wax)
* Expanding Foam (from Tap Plastics)
* Plastic (1/16″ thick)

Wax carved the easiest, as one might expect. Unfortunately, the shavings from the wax were slightly warmed, and would stick back together and to the source block of wax I was working from. Still, it was fairly quiet, and can probably work, given time and limited situations.

Next easiest was the expanding foam. Once hard, the foam is quite rigid, and easy to carve away, holding a fairly good surface. Slightly porous, but that can be solved with some paint, I think. However, carving deep into the foam was rather noisy.

Plastic.. didn’t work out so well. My Dremel is fairly weak, and needs to be spun up to a high speed to carve plastic well. High speed == loud, and I dislike possibly annoying my neighbors. They haven’t complained about my antics yet, but I do worry.

The wood also had this problem. As soon as I’m carving any depth into the wood, the noise level increases dramatically. I want to try a few other flavors of wood, like pine and balsa, but I’m dubious of the possibility of carving wood indoors in an apartment.

Of the materials I tested, I’m most hopeful for the foam. It holds a surface better than wax, and isn’t as loud as wood. It’s also fairly cheap, though I wonder if I could find a supplier of pre-expanded foam sheeting - I’ve heard that places like Home Depot might sell the stuff as insulation by thick sheets, but I haven’t had a chance to check that out yet. The noise level bothers me, though. I really need a garage. Or a house that doesn’t share walls with neighbors.

Creating a diabolical device in an apartment is, perhaps, not the easiest thing in the world. One problem I didn’t count on at the start was the level of noise created by the machine during testing. It’s not horribly loud (especially as diabolical devices go - for that matter, it’s quieter than my vacuum cleaner), but it is noisier than I’d like. It may be diabolical, but I’d still like to be considerate to my neighbors.

So, the first stage of getting it quiet was to install rubber foot pads along the base, rather than relying on the direct metal frame. This helped a bit with the noise level, and also dampened down some of the vibrations that were being transmitted to the floor as the device was in motion. Unfortunately, doing this required that I disassemble nearly half of the machine in order to install the new parts, then put it back together. I took advantage of that, though, in order to re-design a few parts while I was in there.

The next stage, oddly enough, was giving the motors more power and running them faster. It’s somewhat counter-intuitive at first, but here’s how it works. I’m using stepper motors in this machine - they require an on and an off pulse to be sent in order to move each ’step’, which is about 1.8 degrees. If you are running the motor slowly, it has to counter inertia in order to start moving, then again to stop, for each and every step. If you get the motor moving fast enough, though, inertia starts to work for you instead. I had tuned my interface program to run as fast as the motors would accept, which was around 70-100 pulses per second. After reading up on the motor system I’m using, though, I discovered that if I increased the voltage running to the motors, I could run them faster - to around 500 pulses per second with my current power supply.

Thankfully, this not only quieted down the machine, but also sped it up significantly. It’s still not as fast as I’d like for a diabolical device, but to get any faster I believe I’ll have to rework the motion systems, or get a new, higher voltage power supply for the motors.

I’m still not sure the machine is quiet enough for apartment use. I may have to make a soundproofing box to go around it. We’ll see.

I need to learn to keep a better eye on the machine as I’m using it, however. It started to go out of control tonight, which wasn’t dangerous to me at all, but it started to bend some of its motion control screws, bowing them outwards. I reversed the motion and the bowing mostly went away, but they are still somewhat curved. It doesn’t seem to be affecting motion (yet), and fortunately those screws are easy to replace as needed.

Yesterday, as I was playing hacky sack with my coworkers, I decided to jump a low bush as a shortcut. Now, I’ve done this tons of times in the past - in fact, I’d done it twice yesterday before my fateful jump.

On the third, and final, jump, I landed wrong on my right foot, and felt a series of four pops inside of my ankle as it twisted inwards. That… was not good. Turns out that, as the x-ray technician put it, “your superpowers appear to have saved you this time” - it’s just a sprain, not a fracture or serious break. They gave me a nice ankle brace, and told me to keep it elevated and iced for the next few days, and I should make a complete recovery in 3-6 weeks. Guess I won’t be doing any Parkour, climbing, or even normal jogging for a while.

On the plus side, I spent the morning working with my freshly compiling, ReplicaNet-enabled, Diabolical Device code, and got the network client interface working like a charm. Had to add a few new network messages to support my newest features, but things are now going along smoothly.

RE - Yesterday’s Post:
Yes, I know odds don’t work that way. I was just having fun with their numbers ;)

Today I picked up my mail, and saw one of those “special invitation to win prize” mailings, which has some interesting numbers running across it:
Prize 1: Odds are 1 in 14,140
Prize 2: Odds are 1 in 14,140
Prize 3: Odds are 1 in 14,140
Prize 4: Odds are 14,136 in 14,140
Prize 5: Odds are 1 in 14,140

Soo.. this totals to odds are 14,140 in 14,140 that I’ll win something. Big surprise. And look, the scratch-off number that has to match the revealed number on the back… matches. Another big surprise.

But, the place where the math gets very odd is the numbers of available prizes. We have one each of the big prizes (1, 2, 3, and 5), which makes sense, but only 300 of the high-chance prize. So there are 304 prizes available for 14,140 entries sent out.

I really hope only 304 people show up, otherwise there are lots of Prize 4 winners who are going to go home sad.

And I hope that one of those 304 actually gets the grand prize, rather than the 13,836 entrants who should probably stay home, since they won’t get anything.

This weekend was exciting and slow all at the same time.

On the exciting front - I have completed initial construction on my Diabolical Device! It’s somewhat on the loud side, though far quieter than my vacuum cleaner, and since it has its own room for the time being, I can just close the door and more or less ignore it. There are still several upgrades I want to perform on the machine, but it can move, and I was able to begin testing its functionality.

On the slow front - bugs, bugs, bugs. Having both software and hardware bugs is a new experience for me. The hardware bugs are mostly simple ones, such as certain clamps on the device need to be replaced with sturdier versions, and one moving section was placed too high, so it wasn’t able to interface with its matching side.

Software bugs, however, are going to take longer. I finally got the networking client and host to connect, but at this point they refuse to talk to each other. Even using just the local version of the control program currently has bugs in some of the movement code, which should be easy to fix, but again, take time away from using my machine for its intended purpose.

Since it didn’t happen this weekend, I hope to run my operational test program this coming weekend. But we’ll see. I’d much prefer to get this done right than to get it done fast, especially with everything that could break on the hardware side if I do something wrong.

I’ve decided to upgrade this blog to the latest version of WordPress. That means possible downtime while I work out the kinks, and possible lost features (like pictures) until I can get everything smoothed out again.

Progress on my Diabolical Device has ground to a halt. As I mentioned last week, it moves! However, it’s only around 95% constructed, and has been stuck there ever since while I’ve been waiting on parts to arrive. I got a box yesterday… only to discover that the wrong parts were sent. Which means another week of waiting before I can continue assembly. If I’m lucky, the parts will arrive on Friday, but knowing my luck lately, it’ll be next week before I can finally assemble the last of the machine.

The coding side is at a functional stage. It runs, but only in the most basic mode. I have further refinements in mind, mostly adding features - like the ability to control the device from a remote computer, which I’m working on now. Unfortunately, the networking library I’d used in the past was on an older version of Visual Studio, and so far I’ve been having trouble getting it to compile properly under the newer version I’m now using. Worst case, I’ll fall back on doing my own networking support (since I won’t need much in the way of networking messages), but I’d like to get the ReplicaNet library working again, since I could also use it for a few games I’d like to create.