Barn Door Star Tracker

I wanted to take some star pictures with my DSLR and decided a barn door style star tracker would be a fun project to build.

Early versions were nothing more than two boards connected by a hinge with a screw to adjust the angle between the board. One board was attached to a tripod and the other board held the camera. By periodically turning the screw a small amount the camera could be made to follow the stars. Manual adjustment could be eliminated by  using a 1 RPM motor to turn the screw. Using a straight screw leads to errors in the tracking angle for longer exposures. The Trott double arm drive was one effort to reduce the tracking error. The use of a curved screw eliminates the tracking error but requires the nut be turned rather than the screw. I chose to use the curved screw turned by a stepper motor. This way I can adjust the tracking speed in software and reset the tracker with a push of a button.

My design is a bit more sophisticated than two boards but still much less expensive than commercial camera trackers I have seen. I chose to make it out of metal.

Above is the bottom plate that attached to the tripod. It has a 1/4-20 hole in the middle for the tripod screw. Mounted on this plate is a microswitch assembly used in 3D printers. The switch signals the microcontroller when the drive is back to home position.

Above is the top plate that holds the camera. Also on this plate is a dovetail mount for a finder scope used to align the tracker with Polaris. These plates are 1/8″ thick and 3 by 12 inches.

Above is the stepper drive assembly. A stepper motor drives a timing belt pully which hold the brass nut. A ball bearing in the other side of the pully and a hub attach the pully to the plate.

Above is the curved screw. To curve the screw I bent a piece of all thread around a curved surface and matched it to a line drawn to the correct radius on a piece of cardboard. The nuts on the end attach the rod to the top plate. This rod is longer than I expect to need. I just used the full 12 inch piece I bought.

Above is the electronics assembly for the tracker. On the perf board is a Teensy 3.2 microcontroller by PJRC a Pololu stepper motor driver and a Pololu 5V switching regulator. Power from a 12 volt battery connects to the barrel connector and two pushbutton switches control the operation. I don’t have the buttons labeled yet. One is the track button and the other is the home button. Pushing the track button will start the unit tracking. Pushing it again will stop tracking. The same for the home button except homing will also stop when the limit switch is activated. The Teensy can be programmed with the Arduino IDE. I put a hole in the case to allow a USB cable to be plugged into the Teensy for changing the software and possibly controlling the tracker by USB.

Above is the completed tracker. The curved screw sticks out of a hole in the box. I added a wedge with a 42° angle for my latitude to allow the tripod head to stay mostly horizontal. A right angle finder scope help align the tracker. A ball tripod head lets me point the camera in most any direction. If the finder scope is in the way it can be easily removed after alignment.

Above is the tracker with a camera mounted. I got a F1.8  50mm fixed focal length lens for the camera. This is a way faster lens that the lens that came with the camera. Next is to try it out. I had to take this picture with my phone as my camera only camera was on the tracker.

Advertisements
Posted in Amateur Science, Astronomy, DIY, Electronics, Imbedded Software, Photography, Software | Tagged , , , , , | Leave a comment

Building A Sand Muller

To do sand casting a muller is almost a necessity. It is difficult to mix up foundry sand either green sand or oil bonded sand without a muller. I have been thinking about how to build a muller for some time and recently I started to build one. Basically a muller mixes and squeezes the sand a binder to get a uniform mixture with the binder coating the sand.

The basic design I chose uses a rotating cylinder with a plow and heavy cast iron wheels for mixing and squeezing fixed to a bar above the cylinder.  A transaxle from a garden tractor is used to reduce the speed of an electric motor and turn the cylinder. Besides the transaxle the cylinder is supported by four casters mounted to a frame under the cylinder.

I made a sketch using Fusion 360 to determine the size of the frame based on a drum diameter of 18 inches. I wanted the frame to be as small as possible but be able to mount  the drum support rollers on top of the frame on each side.

The first thing to build was the cylinder. It consists of a 18 inch diameter 1/4 inch thick steel plate and a 6 inch wide strip of 12 gauge steel. I had the disk plasma cut by the steel supplier and I should have had the strip rolled into a ring. Bending and wrapping the strip around the disk was a bit fiddly. I clamped the disk to my welding table and used a ratchet strap to tighten the strip around the disk.

The ratchet strap and the tack welding on the inside tended to make the rim lean inward. To fix this I used the jack to push it open before I welded the side seam.

The frame is build out of 1.5 inch square tubing. It took almost all of a 24 foot stick of tubing to build the frame. 

Here is the frame with the garden tractor transaxle installed for a test fit. The bottom of the legs have plates welded on that are tapped for casters so that I can easily roll the muller around the shop. The frame did not come out exactly square so it would rock when sitting on a flat floor. I fixed that by putting a washer on top of two of the casters. The two longer legs will support a crossbar on top of the muller that holds the roller and plows.

The pully that came with the transaxle was going to give a drum RPM that was higher than I wanted. This larger pully has been laying around and would give a suitable RPM. Problem is the original pully was attached with a special spline so I cut the hub with the spline out of the original pully trued it up in the lathe and welded it into the bigger pully.

The final part of the construction was building the plows and roller assembly. It took a bit of experimenting to get it right. One plow pushes the sand away from the center of the drum and the other plow pushes the sand away from the outside. This piles the sand right in the path of the roller. The roller is made of 4 five pound barbell weights glued together. After experimenting with this I may add two more weights to cover more of the sand with each revolution.

Above is the completed muller. It was necessary to clamp the bottom axle so that the drum would turn.

The first thing I tried with the muller was reconditioning some oil bonded sand. This worked well until the sand got well mixed. Then the oil bonded sand got sticky and built up on the roller until the roller stalled. At this point the sand was mulled well enough so it was not too big of a problem.

One of the reasons for building the muller was to be able to make and use water bonded sand (green sand). So I made a few batches of green sand. The bentonite clay that I used is oil adsorbent which is fairly larger granules. I probably could have just dumped it in with the sand and mulled away but I thought it would work better if I ground up the bentonite first. The muller did a good job of grinding up the clay and after about 30 minutes in the muller it was most a fine powder. To make the green sand I added the bentonite and fine silica sand to the muller in the correct proportions and let it run dry for a bit to mix it well. Then water was added in small amounts until the correct consistency was obtained.

Posted in Crafts, DIY, Machine Shop, Metal Working, welding | Tagged , , , , | Leave a comment

UVA UVB Sensor

I have a commercial UV Index sensor running but I found an IC sensor chip that would measure UVA and UVB separately and thought it would be interesting to see how the two UV bands varied with respect to each other.

The sensor IC is a Vishay VEML6075. This is a very tiny surface mount chip that would be very difficult for me to successfully mount on a board. Luckily I found a breakout board with the VEML6075 mounted. This board is the UV 2 Click by MikroElektronika a Serbian company. They make a variety of boards like this and microprocessor development systems.

Currently my favorite microprocessor board for this type of project the the Teensy 3.2 by PJRC. This board uses a NXP Cortex M4 processor chip and has plenty of power and IO including 12 bit A/D converter. This board communicates and is powered via a USB port. There is an add on for the Arduino software development system that allows code for the Teensy  to be written. It is really convenient that there are quit a number of companies  making modules like the Teensy and UV 2 Click. It makes projects like this possible without having to work with tiny surface mount parts. All these modules have holes on 0.1″ centers making it possible to plug them into a piece of perf board and wire them up.

The sensor IC is digital and uses I2C to communicate with the processor. Luckily I also found some code to control the sensor written for the Arduino environment. That saves a lot of time and head scratching getting up and running.

All of my environmental sensor are online. To do this for this sensor it is connected to a Win 10 PC out at the weather station tower via USB. A Python script running on the PC reads the serial data from the sensor and sends it to a MySQL server on the web.  A web page script gets the data from the MySQL server, plots the data and serves up a web page.

A sensor like this should accurately measure the energy falling on a horizontal surface. It should see the whole sky from horizon to horizon. It will have a cosine response to the angle of light from the sun. To do this a diffuser or cosine filter is used. This cosine filter uses a Teflon disk mounted at the surface of the housing so nothing blocks the light from the sun from horizon to horizon. The filter is a 3/8″ diameter disk of 1/32″ Teflon mounted in an aluminum holder. The holder is slightly cone shaped to help keep rain water from staying on the disk.

My favorite housing for outdoor sensors is plastic electrical junction boxes. For this sensor I used a 4″ X 4″ X 2″ box. These boxes are designed to be used outdoors making them water tight and UV resistant.

The cosine filter is attached to the junction box with some silicone sealant. The perf board with the sensor and processor attach to the inside of the box cover with plastic hex standoffs and screws.

The processor is mounted to the bottom of the perf board to allow the sensor board to be mounted close to the cosine filter and allow easy access to the USB cable and programming button. Unfortunately the button has to be pressed to put the processor in the program mode. So I can’t program it when it is mounted on the tower in its box.

The completed UVA/UVB sensor on the right mounted on the tower along with some other sensors that look straight up. I need to put a spacer under the sensor to bring it up to the same level as the others. Right now the taller sensor is blocking part of the sky.

 

References:

MikroElektronika https://www.mikroe.com/

Vishay https://www.vishay.com/

PJRC https://www.pjrc.com/

UVA/UVB sensor data http://www.ocrslc.net/sensors/uvab.php

Posted in Amateur Science, DIY, Electronics, environmental monotoring, Imbedded Software, Nature | Tagged , , , , , , , , , , , , , , , , | Leave a comment

IoT Rain Gauge

On thing that is missing from my weather station is automatic reporting of rain fall. I have a manually read CoCoRaHS rain gauge which is more accurate than say a tipping bucket gauge but I thought it would be nice to have a automatic gauge to compare with the CoCoRaHS readings and also to see the timing of the rain fall. The idea was to mount the tipping bucket sensor next to the CoCoRaHS gauge in the same Alter wind screen. It turns out there is not enough room in the wind screen for both gauges. So the tipping bucket gauge is on its own post near the CoCoRaHS gauge. The  rain gauge is not located near the weather sensor tower but it is close enough to the WiFi signal beam to use WiFi for communicating. The location also means that the sensor will have to be solar powered.

The tipping bucket sensor is from the weather meters sold by SparkFun electronics. It provides a reed switch closure each time the bucket tips. A cable with an RJ11 connector is provided.

For the electronics I chose a SparkFun ESP8266 DEV. This board has an ESP8266 WiFi chip along with a USB serial port for debugging and provisions to connect an external WiFi antenna if needed. I mounted this along with a tiny switching regulator on a piece of perf board.

For the solar power I used a small motorcycle size lead acid battery and started out with a 1.5 watt solar panel intended to trickle charge car batteries. I thought about using a rechargeable lithium battery but most of them will not charge when cold and I did not want the hassle of taking it inside during the winter. The 1.5 watt solar panel turned out to be inadequate and let the battery go dead after a week or so. A 5 watt panel would have worked but I could not find a suitable one so I settled on a 10 watt panel. Some software work on power saving might have allowed the smaller panel to work.

There were problems with the tipping bucket sensor. Spiders and earwigs got into the sensor by way of the drain holes in the bottom. The funnel has a grid to keep trash and insects out but the bottom had large holes. The earwigs would mess up the inside and the spiders would build webs that would prevent the tipping bucket from working. A piece of window screen and some duct seal solved that problem.

The battery and electronics are mounted in a plastic electrical junction box. The WiFi signals pass through the box easily so there is no need for an external antenna.

The software uses an interrupt to increment a count every time the bucket tips. For some reason I got 2 counts for each tip. Something to do with the way the interrupt works I suppose. There is timeout for contact bounce so that was not the cause of the 2 counts. No problem though just divide the count by 2.  Initially I was only going to send the tip count each time it incremented but there were problems with this approach. Without some activity the WiFi link would disconnect. I did not see anything in the software samples I looked at to deal with this problem. I settled on sending out the tip count every 5 minutes and whenever the bucket tips. This keeps the WiFi link connected and gives an assurance that the sensor is still working.

The tip count is sent as a HTTP get message to my web server. A PHP script on the web server then sends the tip count along with a time stamp to a MySQL server. Finally a PHP web page script accesses the SQL data, produces a plot and sends it to the user. The plot  shows the accumulated rain from midnight to the current time or the whole 24 hours if asking for data from a previous day.

There are a few enhancements I can think of that I will be working on.

  • Display the rainfall rate
  • Add a display of the rain for my CoCoRaHS reporting period 7 AM to 7 AM
  • Add some way to tell if the system is up and running

 

 

References:

CoCoRaHS Community Collaborative Rain, Hail and Snow Network https://www.cocorahs.org/

Rain sensor data http://www.ocrslc.net/sensors/rain.php

Posted in Alternate Energy, Amateur Science, DIY, Electronics, environmental monotoring, Nature, Software | Tagged , , , , , , , , , | 1 Comment

Laser Cut Gaskets

Now and then I need gaskets for one of my projects. Of course I could cut a gasket out of a sheet of gasket material with scissors or X-acto knife and use punches for the holes but using a laser cutter would make a nicer gasket and I would have a record of the design that I could re-cut anytime I needed another.

One of my current projects is building a filing machine from castings. It needs a gasket for the crankcase cover as the crank runs in an oil bath. It was easy to design a gasket for the filing machine as the necessary dimensions are available in the filing machine drawings. Using the dimensions I drew up the gasket in Fusion 360. There are CAM (Computer Aided Manufacturing) tools in Fusion 360 which even includes laser cutting. There is also available a post processor for Fusion 360 to output G-code for the Merlin software my laser cutter uses.

It only took a few minutes to draw up the gasket and output the G-code. I did a test cut on some printer paper to make sure the design was correct. I had to scale it up a bit to get it to fit just right. (The cutter probably isn’t calibrated quit right.)

The final gasket was cut out of Fel-Pro gasket material 1/32″ thick. I used three passes of the laser. Two probably would have done it but I wanted to make sure it was cut through. The gasket material comes in rolls and the biggest problem was getting it to lay flat in the cutter. From now on I am going to store the material flat.

Here is the completed gasket sitting on the cover it goes to.

Next I have to make a couple of gaskets for my lathe spindle bearings.

I haven’t tried it yet but another approach for a more complicated gasket like a small engine carburetor gasket would be to take a picture of the old gasket or mating surface then use Inkscape to  generate a path and then G-code for the cutter from the photo. I have done this with Inkscape but not for a gasket.

References:

https://jimhannon.wordpress.com/2017/01/10/cnc-laser-cutter-engraver-build/

Posted in CAD, CNC, DIY, Machine Shop | Tagged , , , | Leave a comment

CoCoRaHS sign

I have wanted a sign for my weather station for some time. Signs for CoCoRaHS has just become available so I ordered one. They made the sign with room at the bottom to put your station number. This article is a description of how I made the ID number and attached it to sign.

My home built laser cutter cuts vinyl nicely so I used it to cut the vinyl for the station ID number. Inkscape was used to draw up the lettering. Some test prints were made to get the letters the right size to fit on the sign and then a Inkscape extension was used to generate a gcode file for my laser.

Above is the laser cutting the letters. I used a Cricut mat to hold the vinyl. The speed of the cutting is adjusted so that it will not cut the back of the vinyl or harm the cutting mat.

Above is the cut out letters after they have been weeded. Also show is a piece of Peel and Stick which I use for transfer paper.

Above is the letters attached to the transfer paper and the backing removed.

Above I have applied the letters to the sign and have partially removed the transfer paper.

Above is the completed sign attached to my weather instrument tower.

Above a view of the entire tower. It was snowing when I put up the sign.

Issues:

I cleaned the sign with alcohol before applying the ID. The alcohol started to dissolve the blue. It would be better to use glass cleaner.

I was not careful and got the letters on a little crooked. Take your time and do it right.

I did not have any white vinyl to match the sign so I used yellow.

I am sure the sign would be destroyed by the wind around here if it were only attached using the two screw holes provided. Luckily where I mounted the sign I could use four screws near the corners.

I used Oracal 651 vinyl. They say it is good for 6 years. Time will tell.

References:

Inkscape https://inkscape.org

CoCoRaHS https://www.cocorahs.org/

CNC Laser Cutter https://jimhannon.wordpress.com/2017/01/10/cnc-laser-cutter-engraver-build/

Posted in CAD, Crafts, DIY, environmental monotoring | Tagged , , , | Leave a comment

Bandsaw Stand

I have had a 4X6 inch horizontal bandsaw from Harbor Freight for years. It works reasonably well. I have cut a lot of metal with it. The stand that came with it however leaves a lot to be desired. It is made of thin sheet metal and has become bent and broken. It is too short to use with my roller stands to support long work. It is also too short for me to use comfortably.  You are supposed to be able to lift the saw and pull it around on the wheels attached to the stand but it drags unless you lift it very high. I got busy building the new stand and forgot to take a picture of the old one. If you have one of these saws I am sure you know what it looks like.

Now that my new shop is operational it is a good time to build a better stand for the bandsaw. The legs of the stand are made with some 1 inch square steel tubing. I made a U shaped frame that fits up into the base casting of the saw and attaches to the saw with the same bolt holes as the old stand. Then the legs were welded to the frame at the same angle as the old stand.

Above is a picture of one of the leg assemblies clamped to the welding table. I made a wood pattern to align the legs to the correct angle.

Above is a picture of the leg assembly being test fit into the base of the saw.

Once the leg assemblies were built and temporarily attached to the saw I welded some 1 inch angle to the legs to serve as bracing and to support a pan to collect the chips. At the motor end of the saw a piece of 1/2 inch rod near the bottom serves as the bracing and an axle for the wheels. At the other end of the saw a piece of 1 inch square tubing serves as the bracing and a place to attach the handle for moving the saw. The old stand had a lift up handle but I made this one a pullout handle. Two 5 inch lawn mower wheels serve to make the saw portable.

Above is a picture of the completed saw stand. I can see from the picture I missed a couple of spots in the final painting. The handle is show in the pulled out position. It can be pushed in to get it out of the way. I used an old bread baking pan to collect the chips. It looks a little spindly but the wheels are mounted out and give it a wider stance. One thing I noticed after using the saw on the new stand, it is much quieter, the old sheet metal stand acted like a sound board for the motor noise. All in all a nice little welding project.

Posted in Machine Shop, Metal Working, welding | Tagged , , , | 1 Comment