Cutting metal always works better when you apply cutting fluid (except cast iron, which you cut dry). It lubricates as well as cools the cutting tool. Serious machining usually uses flood coolant, which is very messy. A better solution for the small machine shop is a fogless coolant sprayer.
These basically mix air flow with a trickle of fluid. The better ones pressurize the fluid as well as the air. The design is very simple, 2 holes to let air and fluid into the mixing block, a .040 hole lets fluid into the air stream. As .035 hole in the end of the nozzle and you are done.
A visit to eBay for parts, Home Hardware for a 1/4-18 NPT tap and off to the workshop.
I made a hose barb fitting from some scrap brass to match the one for the coolant intake filter.
The various parts, the block is not finished yet.
The block is now cleaned up (edges chamfered to 45°) and 2 mounting holes counter-bored, using a 5/16 end mill.
After more than a year away from hobbies, I am back! In meantime we moved into a fabulous house, with 2 workshops! One for wood, in the garage, the other in the basement for electronics, models and machining. Pretty close to utopia for me.
First some catch up from July 2014. Having made the cylinders and pistons, the next item is the cylinder block. A nice chunk of brass was squared up with a 3/8” end mill and then marked up for the center of the combustion chamber. The block was then centered in the 4 jaw chuck and drilled out. I started boring the hole larger but the boring bar has the wrong back rake for brass, resulting in a mess! I left things here as we were packing up the condo in preparation for our selling it.
Rolling forward to the present:
Having access to all my tools is so empowering. I made a new HSS boring tool with geometry to suit machining brass. This solved the issues I was having while boring out the cylinder block to fit the glass cylinders. They are a somewhat lose fit to allow the Loctite 603 to do its job, and to prevent the block from cracking the glass once it heats up during operation.
I also bought the rest of materials that I needed for this project. The supports were not the same width so I fly cut them to be the same. The usual mess of chips flying everywhere but they are now within .002” end to end. Due to the small work envelope I had to shift the pieces 3 times to machine the entire 6.2” length. Thanks to the very solid milling vise this works out well.
We now have a working 3D router built up from a set of 3 Velmex positioners from the 1990’s. The built in limit switches are also connected. With Sonny Jeon’s excellent work on Grbl over the past year this is now a fabulous combination.
A short video of the first real cut. There is still some packaging work to do and it needs a control panel, esp. an E-Stop panic button. Router bits are on order from Richon Tools. These are inexpensive, a factor because I expect to break a few while figuring out the optimum cutting speeds, esp. for aluminum.
Grbl runs on an Arduino Uno but has no user interface. Instead you need to use a sender program. I tested out a bunch of different ones but the best for our purposes is Universal Code Sender. This program runs on various Windows, Mac and Linux. It provides various controls such as jogging and a file loader to send the actual Gcode cut program to Grbl. And it provides a cut visualizer.
The Grbl site has a copy of UCS but it is out of sync with the .9G version. So I installed NetBeans, , replaced the out-of-date RXTX library that is in the git source and compiled UCS 1.0.8. It is rock solid when run by double clicking the .jar file, no communication errors as before. (I do not use the start-windows.bat file, an open Command window seems to affect behaviour.) All testing is done on 64 bit Windows 8.1
I am presently blocked on the Glass 2 Cylinder project while I decide what grinder tool rest I am going to build. Boring the cylinder head, which is brass, is showing me that my hand grinding of cutting tools is not adequate. This requires a trip to a metal supplier and so need to fill up the hobby cash pile first.
In the meantime, I decided to revisit my previous CNC efforts. For the Unimat I used Mach3 (free edition) on a dedicated small PC. This worked well enough but it seemed overkill plus the learning curve for Mach3 is steep. After a bunch of browsing I came across Grbl and SmoothieBoard. SmoothieBoard requires a processor that I don’t have (it is partly based on Grbl though). For Grbl I only need an Arduino Uno, which I already have. It took some investigating to see if the 2.5D milling/routing that we want to do is supported by the limited G Code subset of Grbl. There are enough free tools to generate the code we need. Even helical threading is possible!
Now that I have 2 cylinders, it is time to make the pistons. These are supposed be graphite, which is easily obtained. But that only works if the cylinders are perfectly round inside. And regular borosilicate glass isn’t perfectly round. Close but not close enough. Option B is to make pistons out of brass and add a groove for an o-ring.
I’m a big fan of Plan B anything. First up, cut some hex brass to just over the final length of 2 pistons plus room for finishing the ends and the .05 wide cutoff blade. My 4×6 band-saw takes no time at all. Some time spent aligning its blade means that cuts are close to perpendicular. Less post processing that way. At same time I also cut the cylinder block. It amazes me sometimes just how much scrap material my dad collected from various places he worked. The waste was incredible. After 40 years I am still using the scrap leftovers.