Tims Logbook

Sunday, 28 September 2008

Just hit 45mpg.

I've had a few runs between work and home that are getting 45mph *average*. Generally, it's over 40 for the commute. Not too shabby. My volvo averaged 26mpg.

Diesel is only $4.00 right now. So it's still amazingly cost effective compared to gasoline. I'm getting about 500 miles between fill ups.

I still hope to get the bio-diesel reactor going, but now I've got too many projects going. I need to just finish these things, not to mention save some money for equipment.

Posted by tim at 7:00 PM in General Stuff

Yet another dream project: my own TARPS.

I have baggage pods for my Long-EZ. Plenty of room in just one of these, I'm thinking I could make a 'recon package' for the backseater to play with. Think about it... a gimbal mount could be made with some stepper motors, a DSLR camera could be stripped and mounted on it, strap on a decent telephoto lens, and the whole thing be controlled via USB. For a few more.. uh, thousand dollars, a uncooled thermal imager could be on the same platform. Overkill perhaps, but so...very...cool. I could imagine a backseater having great fun snapping 12(20?) megapixel 20x zoom photos of anything remotely interesting on the ground during a flight.

I'm not sure how to get the camera to 'track', might be an image processing problem, or maybe it could guess using current GPS position, airspeed, and gimbal orientation.

It's probably rife with issues, camera shake, targeting, user interface, and so on. Still, seems like it'd be fun to work on.

Ah well, maybe after I've close all my other open projects down...

Posted by tim at 6:56 PM in News about my Long-EZ or other noteworthy aviation news.

Tuesday, 2 September 2008

New Car

I just bought a 2009 VW Jetta TDI. Already getting 45mpg in it. Awesome stuff. It's quiet, smooth, and powerful. Why doesn't everyone drive one of these?

Posted by tim at 2:16 PM in General Stuff

Tuesday, 19 August 2008

New elements installed

I picked up new elements from Kiln Elements. Stan was very helpful and even made up custom coils for my FX23 that were 7 ohms instead of 8's and 9's. Because my SSR is rated for 50 amp, going to 7 ohms nudges up the amp draw a tad, which should help be reach cone 10 easier (or at least with a margin)

Posted by tim at 2:13 PM in My adventures in throwing clay, slip casting, and firing.

Friday, 8 August 2008

I need more lock washers and K-Locs

So my starter was getting flakey. I was fearing the worst, might need a new starter, or at least a new relay.

Nope. Turns out the signal wire (the one that tells it to "start") was just dangling on the bolt, with no nut! The thing had backed off nearly an entire inch in one of my flights.

I'm always amazed at how screw + nut + vibration = lost nut. I would think it would 'vibrate' randomly, but for some reason, it tends to favor 'unscrew and fall off'. Which is why, I presume, they invented lock washers, or nuts like the K-Loc.

I think I need to make a sweep of all the nuts on my plane, and make sure they are k-locs or have lock washers.

Posted by tim at 12:38 PM in News about my Long-EZ or other noteworthy aviation news.

Maybe next time, I'll read the circuit diagram.

My kiln kept popping the breaker. a 50 amp breaker, and it was supposed to only be a 25 amp kiln. This puzzled me for a long time, since the mechanical timers worked fine, and never popped the fuse. So I started to poke around and noticed the resistance in my setup. It was 3 ohms. That means 240V /3 Ohms = nearly 80 amps! I then remembered I could get the original circuit diagram from Cress, and looking at that, I noticed the coils were wired as 2 series resistors in parallel (4 coils total). After rewiring and checking resistance... 10 Ohms. Ah, much better, 240/10 = 24 amps.

I have no idea why my SSR didn't melt down or asplode(other than the active fan on it), or my circuit breaker didn't pop immediately. I'm thankful it didn't cost me another SSR. at $35 each, they are sorta pricey.

Moral of the story: check your wiring diagrams kids.

In other news, the kiln is working! The software needs a tad more automation to be useful. That'll be the next step.

Posted by tim at 12:33 PM in Posts about making small computers do my bidding.

Tuesday, 29 July 2008

New SSR, now with cooling.

I strapped a CPU fan to my SSR. I ran some low duty cycle tests on it for about an hour last night, and it is still ticking. The first one without heatsinks died in about 20-30 minutes.

Hopefully it now has enough heat dissapation.

I'm now wondering if I should make a *new* controller with a fan speed input and a 2nd thermocouple just to monitor the SSR health. The last one failed 'open' (i.e. power shut off), but if it fails closed, that could be very bad.

Ah, the trials and tribulations of automation.

Posted by tim at 11:28 AM in General Stuff

Saturday, 26 July 2008

New EFIS computer booting from Compact Flash

In the end, it was simple, just let the installer do it's thing, including NTFS file system. What I thought needed to be done was stick with FAT32, but that was nothing but problems. So in that regard the advice of the masses of internet posts was amazingly wrong. I did learn a bit about MBRs and boot records, but I would have prefered just to get things working. I currently have a 1Ghz fanless "Nano-ITX" board booting off compact flash in about 45 seconds. I think I can reduce that further by trimming some services, and using Minlogon, which supposedly speeds up logins. As it stands though, I do have a working computer that would probably hold up until SSDs get cheaper (and in stock, can't find a OCZ core anywhere)

Also, there is a tool to try called EWF, or EFW.. er, anyway, it makes the drive read only, which means the CF will last just that much longer. The main task now however is to mount this thing back in the plane.

Posted by tim at 12:51 PM in News about my Long-EZ or other noteworthy aviation news.

Tuesday, 22 July 2008

Kiln controller now two-way, but eats relays

I've done some low duty cycle tests, and was able to fry my SSR (apparently, it needs a heat sink beyond 5-8amps), while I was also able to melt and/or boil off solider from the coil-relay replacement. That means it got to 400 degreesF at the contact point! Insane. I've tried beefing this up (more conductor should reduce resistance), but I've still had it fail. Might have to make a new board with large area for heat sinking. Or just wait for the replacement SSR to show up (which I will attach to an old CPU heatsink+fan).

The control software is coming along slowly. I'm probably going to make it more of a follow-the-curve sort of system instead of a true feedback or PID. Right now I'm doing everything I can to increase safety, like auto shutoffs and timeouts.

Posted by tim at 2:27 PM in Posts about making small computers do my bidding.

Thursday, 17 July 2008

Kiln controller almost works....

So very close. I have the kiln controller reading ambient temp, reading the type-s thermocouple, and sending UDP packets with its updates. I need to get it receiving packets, to control its duty cycle. The solid state relay already can be switched. Once this is done, I can then have most of the controll logic in a simple .NET app. Lessons learned (and applied) in this build:

Voltage reference is on board.

Hooked up all pins properly, for example, RB2 for the interrupt.

Properly managed the 5V and 3.3V portions of the board with no ground loops.

Checked all pins before plugging in expensive part.

I screwed up the TC amp again, but I use SPICE to design a new Layout, which is now on a tiny dauther board, preventing another pcb.

I bought at least two of each required part. this came in handy.

I had stocked up on basic opmps and a wide variety of resistors/caps/etc Prior to the build, this too came in handy.

I discovered that opamps have an input offset voltage. The TLC272 ended up having about 1.1mV. Which ends up being quite a few degrees C if I didn't compensate for this in the software. Ill have to watch out for this in the future.

Grounding as much of the copper as I can seems to help tremendously. This board has yet to crash or reboot. (cross fingers)

Another efis update

I ve installed the efis in my plane, even flew around with it. The biggest problem is the hard drive. The thing cuts out when I take off. Perhaps its the change in pressure. I've ordered a nano-itx mobo with a CF slot. So ill be running all solid state, which should cure my problem. The software has come along pretty far. I made a pipeline that consumes and compresses airport data directly from the FAA. The touch screen works well. Im able to scroll the map with just a finger swipe. Tapping an airport shows its data, including a runway diagram. Once I get set up with new computer and fly a bit, ill be adding airspaces. Then the next thing is to redo the engine monitor.

Wednesday, 11 June 2008

So amazingly cool, I can't stand it.

A hexapod robot with a router motor for a 'head'. It's only 7-8bits of precision with 0.5mm repeatability, but wow, this blows my mind.

This kinda thing can only get better with time. I'm boggled by what the software must look like to run this thing.

Posted by tim at 12:00 PM in Posts about making small computers do my bidding.

Monday, 12 May 2008

Finally, the touchscreen works

In the end, i had to resort to buying a pre built controller. My board almost worked, but i didnt take into account the current drain and inevitible voltage drift. If i would have read the ads7843 datasheet a little more closely, i could have duped its functionality (especially the differential mode). the next thing i tried was to use a actual 7843, though i never could get it to do anything. I did learn about the SPI HW on the PIC18F though. The docs are kinda wrong. Note to readers, check the spp bit in the interrupt flags, do *not* ready the BF flag in sspstat. Doing so will halt the transfer! ( the interrup flag triggers weather or not interrupts are enabled for the ssp). Thankfully i have a logic analyzer. I would hyave been screwed without it. the other thing i learned was about USB. I brought up the mouse example and sucessfully modified it to report absolute mouse positions. Note, this was all firmware wrangling, the built in HID mouse driver in winxp/vista took care of the rest. pretty sweet actually. Having delved into usb, i feel more confident about rewriting microchip's usb "framework". which is a bit of a mess. next challenge, complete the kiln controller. all the parts are finally here. lessons learned:

expect the first board or design to fail.
always include a good voltage reference if using an a/d.
Make sure not to tax the output pins with devices that draw excessive current.
Always build in sampling pins to hook up the logic analyzer.

Tuesday, 11 March 2008

My mogul just got better.

FINALLY, HTC/Sprint released a new rom update that brings Rev A. EVDO to my phone, as well as the software to run the GPS. I thought it was "AGPS", i.e. cellphone tower triangulation, though in reality, it's a real GPS receiver! I've been wanting a GPS in my phone for a long time, I was using a bluetooth GPS till now. I can now use LiveSearch map function and have it center on the GPS. GPS assisted 911 calls are also enabled.

Posted by tim at 9:25 AM in General Stuff

Thursday, 7 February 2008

Kiln controller update

My first trial run discovered that relay 2 had a fault. I traced it to the pullup airwire, so after a quick solder fix, all relays work again. I also improved the connectors. In the end, I was able to sucessfully switch 240V without anything melting down.

The control board still suffers a bit from instability. Even though I grounded the large copper pour area, (which helped tremendously), it still can flake out sometimes and end up rebooting. This causes the USB connection to drop, confusing the driver.

I've been unable to figure out a combination of calling Reset() and have the C# code attempt reconnects on the USB com port to restablish a connection. I was hoping that if the C# app 'backed off' from the port, and the board reset itself, I might get it to reconnect. But so far, no dice.

I might have to decouple it from USB entirely, and instead, go to a wireless connection. That way, there is no problem with 'loosing' a usb com port. Maybe I could try a C++ app, so I really know when the app has 'let go' of the com port. In C#, there isn't any way of really knowing when it's done. And it's wrapped up in all sorts of nasty exceptions (which probably keeps the port open)

Posted by tim at 1:29 PM in General Stuff