Crummy servo tester

The prototype viscometer board

• First 2 bytes are settings, things like block size, filename size and special behaviors like for EEPROMS that can only be written to with blocks.also defined is the address size 8b, 16b or 32b
• Each file in the FAT consists of a name of either 2,4 or 8 bytes. this is determined by the setting.
• Following the name is one byte detailing the flags such as read-only, archive, and whether it’s open or not.
• After the filename and attribute byte, are the addresses. These addresses can be 8, 16 or 32 bits, however, an 8 byte address can address something larger, say 2k rather than 256B if you set the block size appropriately. This of course lowers the efficiency of storage space but allows for some growth in the file. The file can have any number of addresses in order to remove time wasted moving stuff around.

Anyways, that’s what I’ve been working on as of late. Fun!

The weight in grams is held on a pulley and pulls on a rotating pulley that is 1.125″ in diameter. That pulley rotates some gears or belts at a ratio of 11:1 (1 rotation of the pulley = 11 rotations of the spindle). Taking the torque applied on the main pulley and dividing it by 11 results in the actual torque to the spindle.

Simply because I use it in these cases, here is the conversion ratio for grams hung on the instrument to oz-inches. Also, one should keep in mind that there is some loss of torque due to mechanical limitations.

oz-inches == .00367056 * grams

therefore, using this formula, a KU meter ranging from 32 grams to 1099 grams ranges from .1174 oz-in to 4.0339 oz-in.

yay! Hopefully someone finds this useful as well.

Anyways, here are the links to the drawings in PDF format.

stormer paddle drawing PDF
qy-2004a drawing in PDF

I’m going to draw up some of the rarer parts I have, I may need them later and somebody else might be able to make use of them. Keep in mind, these are rough drawings taken from measured features. Don’t assume that the measurements are the same for all units except perhaps the stormer paddle.

The DIY Measuring arm
This idea came from the need to build a measuring arm at a very low cost. I simply used some wood and some regular linear potentiometers.

The cheap and quick measuring arm, this image doesn't show the wiring or the board behind it.

Anyways, things are going well but it’s a lot of birds in the bush and none in the hand. I’m just writing to keep a log of what’s happening

Chinese 9G RC Servo Photo

Chinese 9G RC Servo drawing

[caption id="attachment_548" align="aligncenter" width="373" caption="Chinese 9G RC Servo Solid"][/caption]

As the project progresses I’ll review the functionality of the servos. After I had ordered them I had read that they are susceptible to outside interference, especially from the likes of human contact around the housing. From what I read, this made the servo “Freak out”. We’ll see I guess, I’m going to make a standard testbed circuit for testing anyways.

So, anyways, for anyone who wants it the drawing, here’s the PDF

Chinese 9G RC Servo PDF

Hopefully someone will find it useful.

Originally, my father and I spent a day working on both the frame and the electronic portions of the unit. I quickly whipped one up with a protoboard I had laying around and a PIC16F690. I used an SN754410NE H-Bridge driver for this design as well. Unfortunately we didn’t finish the project that day, and I wasn’t keen on programming the PIC having to pull it out of the IC socket every time I wanted to test it.

The original defunct board. This board was omitted due to the fact that it had no ICSP provisions.

First, I took some Stripboard and planned out the pinouts and connections. Stripboard (or veroboard) is my favorite since it’s so damned easy to plan. Despite there being a great many other protoboards out there, veroboard has been the most useful, for me anyways.

Planning it out

Starting the board

Halfway done

Board pretty much complete

Astrophotography electrics working with keypad, LCD and stepper motor

Electrics in Enclosure

Anyways, I’ll report further progress, as usual, on my blog here.

Here are some interesting links
http://www.keteu.org/~haunma/proj/barndoor/
http://www.cs.uiowa.edu/~jones/step/types.html

Comment if you wish.

Maybe I’ll make a full map one of these days, though, looking at some of the elaborate maps that the community is putting out, I’m not sure I’d have the time to compete with such standards. It’s fun nonetheless.

Astrosmasher is a partial clone of an older colecovision game bit I got bored with it pretty quickly, I think I’d like to write games that are a little more fun despite the fact that I have no artistic ability.

Anyways, here are some screen shots and a like to the game. The game requires windows 2000+.

Consider this game, for the most part, incomplete. I got bored with the idea and made it workable. It is a hard game though. My high score thus far is 35100. Give ‘er a shot.

The living room temperature as seen over 6 days

Sorry about the small scale of the temperature graph, the graphing program is still a work in progress.

1. Prototyping.

The initial version of the sensor, using older stuff from viscometer project.

Cup o' snow to check and see if the Opamp is biased correctly

Temperature check with the IR gage

The voltages swung just fine with a bit of extra range. I have about 50C to work with in range, good enough.

3. Making the board. For this I used a board from measurexplorer. I have tons of these but haven’t had much luck using them. The only ones that have worked well for me are the ones with 3 holes per pad. Anyways, here are some pictures.

Start of prototyping the board

[caption id="attachment_397" align="aligncenter" width="400" caption="Finished Sensor board with both light sensing and temperature sensing"][/caption]

Anyways, This board took me a couple hours to make but it works well and required no rework, thankfully. You can see both the LM335X (TO-92) and the CDS for sensing the light. This board interfaces to an RS232 board that I’ve made and that I use for some of my other projects.

4. Getting ‘er running. While the unit itself is already programmed in terms of the MCU. it needs some adjusting for voltage on the pot and that’s about it. now to affix it to something so it doesn’t move around.

RS232 and Sensors glued to a board

[caption id="attachment_400" align="aligncenter" width="400" caption="Tracking it on the computer"][/caption]

As you can see, I simply used hot glue to affix both the RS232 board and the sensor board to the block of wood. Its a temporary arrangement while I come up with a good enclosure for outside. I brought my old laptop out into service for this project, works well just for collecting data.

Anyways, here are some images of some collected data.

Living room temperatures (click to see description)

Basement (click to see description)

Well, so far so good. Now I’ll make the enclosure for outside and improve the sampling. hopefully I can leave it out all spring/summer and see the patterns.

I didn’t tag them or describe them. There is a seat, a guard and a valve plate. I omitted the dowel pin, bolt, center nub and lock nut. I just wanted to see what it looked like in solids.

That brings me to my next topic. What I should do.

Firstly, I think I’m going to start taking more pictures of Edmonton again and post them on Panoramio again. Frankly, this is the reason I started taking pictures in the first place was to provide images of Edmonton for whoever wanted to look and explore both the nice parts and boring parts of our city.

Secondly, I’ll finish that fucking clock. Really not much more to do it it, just make it so you can set the time.

Thirdly, focus on building a working vibratory viscometer. It’s the ideal way to measure Krebs units in an environment as harsh as paints and it will be substantially cheaper.

Anyways, End of Line.

First, I took one of those dollar store cookie sheets and drilled it through for 40 LEDs. 12 for the hours, 12 for the minutes, 5 for the seconds, 5 more for the div/5 minutes and five more for the outside and the PM light.

Drilling it wasn’t much fun but, when it was finished, I started to insert the LEDs with the grounds all towards the outside of their respective circle. Then I mixed up some epoxy and drizzled it over the backs of the LED’s. Unfortunately, the epoxy didn’t hold very well on the other epoxy surfaces so I used hot glue to tie them down.

As you can see, I labeled it in reverse of the other side so that I could get ‘er working. Thus started the long job of soldering the whole thing.

The image show isn’t the complete one, but it took me several hours to solder it all together. I used 7 groups of 6 LEDs in order to display the image. In this case I used the 7 groups for the duty cycle, thus a 1/7 duty cycle was used. Each of the 7 groups is fed ground through an MPSA13 Darlington transistor. Here’s a picture of the board that controls it.

Pretty fucking ugly, I know. But it works and I used an MCU (PIC18F4685) that was a little overkill but that’s fine, I’ve got quite a few of those. Anyways, so far so good, though it seems a little impractical as a clock and I think I set the rings of the minutes and hours too close together. Oh well.

Here’s a blow by blow of how it was built.

First I took the thing apart. The wire itself is pretty tough to chew on so I kept it in there so I can save the lamp itself. The lamp was bought from Wal-mart for about $18.

Here’s a blurry pic of the control board I designed. It’s generic in that it can take both digital and switch input and output 200ma per channel on 3 outputs. It’s nothing special but it’s small enough and it works. The MCU on it is a PIC16F505, not a great MCU but it works.

And here’s the casing on the base. Inside was some sort of bizarre weight made of something I did not want to cut.  I took the guts out and popped two 5/8 holes for the buttons. I didn’t have a proper drill so I used a forstner bit, not ideal bit it worked surprisingly well.

I inserted the guts of the machine and had to follow the existing wire in and keep the controller towards the front. As you can see, it’s pretty ugly, as well I used hot glue to affix the boards to some wood which was then affixed with hot glue to the inner casing. It’s not an ideal solution but it relieves me of having to use bolts and it was fast.

And there you have it, it seems to work. No fire, or smoke and it works as programmed (kind of). Now I’ll have to go to work and build a new base, one to keep the unit steady, also I think I’ll add a little piezoelectric tweeter for audible confirmation of time selected.

Anyways, I’m glad I made it this far on the project in such a short amount of time.

Electronic Side

Magnet Drive (click image to see full size)

This unit is the main magnet drive. It is responsible for the pulse width modulation control of the four H-Bridges below it. Each H-Bridge is fully optically isolated, can happily push 15A at ~3KHz with any duty cycle. The H-Bridges were also designed in such a way that they don’t produce much power supply noise and very little inductive hiccup. The Board on the top has a microcontroller that controls the four H-Bridges and accepts input from the main control board and the dip switch set on-board for debug purposes.

The unit has been fully tested (waveform, heat and max current) and works, no more modification or re-design is required.

Input Totem (click image to see full size)

This unit was built to stem problems arising from EMF and Static discharge. Earlier versions of the project would suffer from random shutdowns, errant signals and noise. I wasn’t entirely sure for some time why certain events would occur for no reason but even pressing a switch sometimes was enough to roast a microcontroller. I recently tracked it down to long input lines picking up noise and also generating power dips when, say a switch, is activated. This unit isolates the inputs from the faceplate and debounces the signal in order to remove potentially hazardous voltages from making it to the mainboard. This unit also dissipates large transient voltages through the use two voltage suppressors per input line.

Currently this unit works and has been tested. It does require some additional programmatic tweaking.

Main Motherboard (Click on image to see full size)

This unit comprises of a larger PIC18F microcontroller with a 4 digit LED display, this is the primary control board. This board is responsible for controlling the Magnet drive, The LCD Front Panel Display, The Debug Display (The one on the board), calculating the results of data, outputting said result and it also handles the filtered input. This is basically the ‘brain’ of the machine and this particular part of the project has gone through the greatest number of iterations, probably 8 or so.

This unit, circuit wise, works and has been tested. It still requires programming as the entire unit has been redesigned and functions in a different fashion than the previous versions.

General overview (Click on image to see full size)

This is a general overview of what the internals are going to look like when assembled. Of course, these are not the final positions of each unit and a couple of units are missing from this picture (namely the +5v isolated supply and the Serial Out board)

Mechanical Side

The Current Design (click on image to see full size)

This is the latest design of the unit. Several design changes have been made in relation to the previous 6 designs (3 were made). Following is a list of changes.

• Made the magnet units smaller but longer, opting for size in heat-sink area.
• Magnets are closer together towards the paddle
• Sensor unit no longer consists of two reed switches and now consists of one hall effect sensor.
• Timing is found by center point activation rather than activation at extents of travel.
• Wiring is changed from Single solid-core Teflon coated wire to shielded, multi-lead stranded wire for easier assembly and maintenance. Also the shielding provides protection for outside interference and electrical problems.
• Top and bottom plates are solid instead of multi-piece in order to simplify manufacturing
• Paddle is shorter and has less area, PWM will ensure proper shear rates on paddle.

The previous designs had the following flaws:

• Distance to paddle from magnets was too great, creating a ‘dead’ spot in the center of travel
• Making a magnet larger does not mean greater field intensity at larger ranges. Each previous iteration was based on the false assumption of larger=stronger.
• Reed switches were problematic. Either moisture or improper magnetic application proved to create intermittent contact on the switches making for a difficult to track problem.
• Previous units were susceptible to heat problems. Smaller magnets, lower current and warm-up cycle solved the issue. Plastic sometimes melted under large loads
• Units were difficult to assemble, thus, difficult to maintain in an industry setting.
• Units had lots of nooks and cranny s making it difficult to clean.
• Previous paddles were either too larger or were of the incorrect length at both the signal and Thrust sides.

As far as this part of the unit is concerned, I’m not far from completion, aside from machining.

What is left to do?

So far, not much is left to do on a surface level. I have the following items to complete.

• Finish design for unit (Decide on arrangement and potting of sensor)
• Machine parts for prototype
• Finish some programming on the input totem.
• Start testing with dummy board
• Once tests on mechanical portion are done, main board needs programming
• Program main board (in no particular order)
  • Confirm and program communication scheme
  • Derive algorithms for appropriate sampling and error correction
  • Program debug display
  • Program for parallel LCD display
  • Create initial setup and calibration routines
  • Create run-time user interface and system.
• Test, test, test.

That’s the list thus far, and completed as per my knowledge.

It should also be noted that I’ve built an automated rotary paddle type viscometer (rudimentary prototype) that is functional and can easily be adapted, with appropriate components, for in-line use. I’ve also produced a working vibratory viscometer, it however doesn’t follow non-Newtonian fluids well.

Hopefully this was informative, thanks for reading.