Sunday, when I did my last update on getting the printer going, I concluded with a test plot that showed some oddities. I spent some time on Sunday troubleshooting that by the seat of my pants. I certainly made it better and ended up with a higher quality version of that test print.
Armed with that, I decided to make the first "real" part that I had designed. It was for an off the wall (and not very well designed) experiment to substitute a microswitch for the ignition points in my model engine. The ones I bought last month from NAPA have crapped out and I ordered replacements last week. The engine's mounting bracket has two holes for the two screws in the point. Those are 1-1/16" apart and threaded 10-32. The microswitch has two mounting holes that a 2-56 screw goes through. It's smaller than the 1-1/16, so maybe I could mount a switch on a little adapter board that I make and mount that on the engine? The adapter would need to have the spacing for both sets of holes and one of the big holes would have to be an arc to allow some tweaking of position.
I made two different versions of this adapter, one using the default settings for the Cura Slicer program and one where I attempted to make a better part. The attempt at better first:
There are some obvious places where I think the plastic didn't adhere well to the glass plate and other errors on the bottom layer. Notably, it isn't that flat.
Now it was time to go over to the one with the default Cura profile settings for my printer.
What can I say? This one is better all around. Better adhesion, more accurate features, flatter back, everything. Look at the two of them standing on their sides on the printer bed. The first one I printed just gets mysteriously thin on it's right edge.
The bottom line here is that I have two good prints. The first might be usable with that oddity to its shape if the part was usable at all. The second, even though it came out better, isn't usable because the design is faulty. This is the second one with the microswitch mounted on it sitting below and in front of the points it would replace. The geometry is wrong. It might work if I modified things more, but I'd need to make a 3D model of the switch and everything there. I'd probably need to change the timing on the engine (position of the flat on that cam at the top that drives the points).
At this point, I think it's safe to move on to more prints. There will be more learning curve and more supplies to get, but I think I can say it's up and running.
Armed with that, I decided to make the first "real" part that I had designed. It was for an off the wall (and not very well designed) experiment to substitute a microswitch for the ignition points in my model engine. The ones I bought last month from NAPA have crapped out and I ordered replacements last week. The engine's mounting bracket has two holes for the two screws in the point. Those are 1-1/16" apart and threaded 10-32. The microswitch has two mounting holes that a 2-56 screw goes through. It's smaller than the 1-1/16, so maybe I could mount a switch on a little adapter board that I make and mount that on the engine? The adapter would need to have the spacing for both sets of holes and one of the big holes would have to be an arc to allow some tweaking of position.
I made two different versions of this adapter, one using the default settings for the Cura Slicer program and one where I attempted to make a better part. The attempt at better first:
There are some obvious places where I think the plastic didn't adhere well to the glass plate and other errors on the bottom layer. Notably, it isn't that flat.
Now it was time to go over to the one with the default Cura profile settings for my printer.
What can I say? This one is better all around. Better adhesion, more accurate features, flatter back, everything. Look at the two of them standing on their sides on the printer bed. The first one I printed just gets mysteriously thin on it's right edge.
The bottom line here is that I have two good prints. The first might be usable with that oddity to its shape if the part was usable at all. The second, even though it came out better, isn't usable because the design is faulty. This is the second one with the microswitch mounted on it sitting below and in front of the points it would replace. The geometry is wrong. It might work if I modified things more, but I'd need to make a 3D model of the switch and everything there. I'd probably need to change the timing on the engine (position of the flat on that cam at the top that drives the points).
At this point, I think it's safe to move on to more prints. There will be more learning curve and more supplies to get, but I think I can say it's up and running.
Looks like your set screw hole ate up the rubbing block on the points. I think the best way to alleviate that would be to place the set screw into the center of the timing cam flat. That way when the points close, the rubbing block would be held above the set screw hole as it rotates past. Unfortunately, it would mean having to manufacture a new cam.
ReplyDeletePolishing the timing cam and using a points specific grease would mitigate wear on the rubbing block as well.
The printed mount looks nice and would function well with a low speed cam timer system. However, I think that the weight of the micro switch lever / roller assembly would have too much inertia for the internal spring to overcome at operational speeds. There is a reason that points have such a high spring rate. It would be my assumption that the micro switch would fire the coil, at low speed, then "float" over the cam flat at operational speed.
Not to cast a pall on your design, SiG. I don't believe that the micro is up to the challenge. If you want to get away from the points, might I suggest a hall effect type ignition.
Just my $.02 .
Leigh
Whitehall, NY
I don't know if the spring strength of the points matter. The spring on a set of points is stiff so it can follow a 4, 6, or 8 lobed cam at speeds approaching 6000 RPM. SiG isn't spinning his engine anywhere near that speed, and he only has one lobe on the cam.
ReplyDeleteGood catch on the rubbing block getting eaten by the setscrew hole. I was about to ask how they failed, and now I see it.
That should be "3000 RPM". I forgot most engines spin the distributor shaft at half crankshaft speed.
DeleteI have dealt with points in my dirt race car, and micro switches in my profession as a maintenance mechanic. The comparison between the two spring rates is miles apart.
DeleteMy premise is that the abrupt transition between the cam diameter and the flat will cause the roller to jump off of the diameter and land part way into the flat. Then the transition from the flat to the roller would launch the roller off of the cam to land beyond the transition. Much the same way a valve can get launched by the valve train when the opposing force of the spring is insufficient to maintain contact between the lifter and cam lobe. Hence, "valve float". As the rpm increases, the distance the roller "jumps" off of the close cam runs into the opening cam. At that point, the switch no longer cycles.
You also have to consider that the points cam is running at crank speed as well. Also, SiG's cam is essentially a flat milled into a round disc, with a slight radius on the transitions. The heavy spring pressure of the points would be able to overcome the sudden transition and close long enough to complete the circuit. Even if they didn't exactly follow the close/open ramps exactly. That would be evident on a swell meter hooked to the ignition circuit, as the rpm varied.
I don't want to appear argumentative, just explaining my thought processes and reasoning therein. And by no means are my ideas definitive. This is just my SWAG analysis. If it works as intended, I will be the first to raise my hand and declare myself wrong. I've been wrong before, just ask my wife. ;-)
Leigh
Whitehall, NY
"transition from the flat to the roller" - that is supposed to read, transition from the flat to the diameter.
DeleteSwell meter??? That is supposed to be DWELL Meter.
Grrr - I proof read it at least twice and those still slipped by me.
Leigh
Whitehall, NY
Good stuff, Leigh. The failure mode of the old points was different than the setscrew hole tearing up the rubbing block. I think I posted before that when I got these points I went to try to to start it one day and the spring wouldn't close them. I could take the points off, push them wide open and they'd just sit there open, in my hand.
DeleteThe temporary fix was to oil the place where the spring goes around that mounting hole on the right side. Before I'd try to start the engine, I'd put drop of oil in the wrist pin oiler, into the cylinder, a few other places, and then oil the points.
I am NOT a master mechanic. I worked on cars with points from maybe 1970 until 1982. I never saw or heard of oiling points.
There is grease on the rubbing block, though.
The idea of the microswitch was an idea someone else suggested and made me curious after I said wouldn't work. The engine might run around 1000 RPM, which means the switch has to move in a fraction of that, so we're talking well under 100 microseconds. The only mechanical parts I know of that move that fast are reed relays. These are a small pair of contacts in a glass bulb that are controlled by driving a current into a coil to snap them together. It would be even a crazier redesign than this.
I should never do math in my head before my first cup of coffee has set in.
DeleteFigure the engine is running at 1000 RPM. The time for the points to close is really a small fraction of one revolution. I don't know how big a fraction, but I know we want to fire about 15 degrees before TDC, and the points have to start moving before that to be closed at that time. Give it maybe 3 degrees of arc to close? 3/360 = 1/120 of a revolution. 1000 RPM is 16.7 revolutions per second. One revolution takes 60 milliseconds, and 1/120 (3 degrees) of that takes 500 microseconds or half a millisecond, if you prefer.
So it's not well under a microsecond, it's well under a millisecond.
I could set up an experiment to measure the switch, use something like a drill motor to turn a cam and measure the speeds with an oscilloscope, but it's basically the same thing as putting them on the engine. So why not put them on the engine and see if they work?
I must have missed/forgotten the sticking points. It is possible that the socket head screw is a touch too large for the pivot of the contact set. Possibly causing the inner liner to swell against the pivot.
ReplyDeleteTypically, oil and grease are problematic to the operation of points. I've managed to over-grease the rubbing block once, and it caused all sorts of misfires. The grease got onto the contacts, causing erratic operation. Made for a very long heat race and poor starting spot in the feature.
Fortunately, your ignition box only uses the points as a trigger, and not a saturation device for the coil. I fully agree with your trial. You already have the parts and the implementation won't take any longer than experiments to see if they would work.
So; why not, indeed.
Leigh
Whitehall, NY
It is possible that the socket head screw is a touch too large for the pivot of the contact set.
DeleteThat was my suspicion, too. These points are really held by the adjustment screw. I actually had to enlarge the hole around the spring on the other side - the #10 screw wouldn't pass through the hole. I put a 3/16" drill through it. The #10 isn't torqued down, just held with LocTite 680 (green).
Yeah. 3/16" is a fairly tight fit on a #10 screw. The problem with drilling plastic is that you don't always get the size you are shooting for, and the material can extrude outward if the bit isn't extremely sharp. A #12 bit might provide an adequate amount of clearance, without having too much slop.
ReplyDeleteLeigh
Whitehall, NY
Hmmm...
ReplyDeleteThe first one seems to be a serious case of underextrusion. Did your filament bind or get hung up on something?
The second one looks a lot better.
The first one is why you watch your first layer.
But since the part didn't work out the way you wanted it to, the nice thing is that it only cost a few cents of plastic.
I wonder how long it would have taken you make it out aluminum?
Nothing that I could see. It seemed to just not stick well to the glass. The printer hadn't been on long and it was cooler than normal in my shop, but still 66F or so when I'm heating the bed to 60C. Is it best to let the printer sit idle for a while once it says it's at temperature?
DeleteThe other thing is that I haven't been able to find some PVA glue for the bed so it's bare glass.
Cura said the prints should take 14 minutes, but I didn't time them. My rough guess is that it would have taken about twice that to drill four holes in aluminum, including finding a piece of 1/8" aluminum and cutting the piece to size.
If it's not sticking, try cleaning the surface with alcohol.
DeleteMaybe you could reduce the gap between the nozzle and the bed.
As for the glue stick, you should be able to find some almost everywhere.
amazon.com/Elmers-Disappearing-Purple-School-E562/dp/B00143SNPG/
But with that first print, I would have cancelled it when it didn't stick.
Also, on something that small you might want to go with a brim vs the skirt.
If that doesn't work then you have a problem.
Oh, and the print time by the slicer is just a rough guess that can be off by ten minutes or more +/-.
Deletemaybe this
ReplyDeletehttps://www.amazon.com/ELECTRONIC-TRANSISTORIZED-IGNITION-MODULE-Engine/dp/B01JUBZSCE
Never seen that. How does it work? Someone said it used points and a condenser, so it’s not a CDI. That’s all I can say.
DeleteFrom what I gather from the page, it looks like it replaces the points in a magneto style ignition, by mimicking how the points operate.
DeleteProbably calculates when it should fire by RPM and the crank angle of the magnet from TDC. I know that newer small engines: ie -mowers, chain saw, trimmers, etc; have the ignition module built into the coil. They no longer have points at all. Other than that, my knowledge of how they operate is limited.
Leigh
Whitehall, NY