I have a vintage lathe that has few features. I've decided to add a quick change gear box (QCGB.) A QCGB is a gear-filled gizmo that connects a lathe's spindle to its leadscrew using any of dozens of ratios. It is is very much like a car transmission. The result is that the carriage moves a fixed amount for each revolution of the spindle. This is how threads on bolts are made.
I checked out buying a vintage QCGB but they are hundreds of dollars. My lathe may not even accept one unmodified. So I considered building one. The cost of the gears was prohibitive. So I considered making the gears and decided it wasn't happening - making gears is a lot of work and I don't have the tooling. In short, I ran into a dead end.
The other option to was to computerize the lathe. I don't want a computer in the shop as I program for a living - I don't want to do it at home, too. But perhaps I could sneak in some electronics.
And thus begins the story of my virtual QCGB. The basic idea is instead of having the spindle drive the leadscrew through a train of gears and a 'trasmission', I'll collect signals from the leadscrew and use these to drive a stepper motor which turns the leadscrew appropriately. Such conversions are actually pretty common. Most solutions, however, involve the use of a computer in the shop and are basically low-end CNC. My goal is a very spartan implementation so the computer isn't needed.
The only thing that is holding me back is that I know zip about electronics. Other than that, and my lack of knowledge of mechanics, I'm good. But my pal Marco is basically a mad scientist. When I mentioned my idea, he gave me a box of electronics parts and supplies. Then I bought a bunch of stuff he told me I needed including some common IC chips, proto boards, and selections of resistors and capacitors.
With some guidance from Marco and the web I made my first circuit. It just lit an LED. My next circuit used the ubiquitous 555 timer chip to make the LED pulse about once a second. No great shakes - examples are all over the web. I did the math just the same.
I liked that circuit, so I set it aside and pulled out another small protoboard. I added a 4040 counter chip. I salvaged some LEDs from a death ray or something Marco built back in the late '40s when he lived in Roswell. The end result is that I could use the 555 to drive the 4040 which in turn made some LEDs flash. Oooo pretty. Here's an early version:
Later on I pulled to 4040 (the larger IC to the left) off and moved it to the same board as the LEDs. Now I have a self-contained counter module to use for my next experiments.
The reason I started with the 555 and 4040 is that, when creating an electronic QCGB, you have to add an encoder wheel to the spindle. As the spindle turns, it generates a series of pulses. While I'll eventually have to work up the full pulse generation circuit, for now the 555 can simulate this. The 4040 is interesting to me not because it counts or makes LEDs light up, but because half of the real QCGB deals with dividing spindle speeds by 2, 4, 8, etc. The pins on the counter do exactly the same thing. Thus I have already done a proof-of-concept on an important part of my electronic QCGB. Referring to the picture above, the rightmost LED is on. This LED is connected to the 'divide by 2' pin on the 4040. Visually, that LED lights on every 2nd pulse from the 555. So if I wanted to slow the leadscrew down by half, I would tap the output from this pin. Similarly, by tapping the pin associated with the leftmost LED, I'd slow the leadscrew down by a factor of 128! I might use that to make a nice finishing pass, after all the heavy material removal has happened.
Lessons learned so far -
1. If you don't protect LEDs with a resistor, they die.
2. If you leave a 4040's reset pin floating (not connected to a ground) you'll get funky results.
3. Wire kits for prototype boards just rock.
My next series of experiments will likely be a proof-of-concept of an encoder wheel for the spindle. This is an opaque disk with equally spaced holes in it near the rim. As the wheel turns on the spindle, a stationary LED shines through the holes producing sort of a strobe effect. The light flashes fall on a phototransistor which generates a pulse we can use to drive the counter previously built. Will a good pulse be generated? How small can the hole be? How quickly can the holes pass in front of the LED and still let enough light through to produce a usable pulse?
I haven't a clue. Yet.
No comments:
Post a Comment