you can set the allowed position error larger, or you can slow down your feed
rate. I’d suggest increasing the allowed position error.
The “can’t keep up” error was an idea that just hasn’t worked out as expected.
Without acceleration planning, the default ‘2.’ value is just too sensitive.
David Lang
I need position error on 2 because I’m cutting
https://www.facebook.com/marcin.przyborowski.10/posts/2418694425035282
At setting bigger than 2 GC is approximating straight lines into bows, and starts cutting into things 
Acceleration planning, how? Fusion 360?
Cheap dust collector test 
I’m not sure if that gonna work. I wont use any plexiglass because it shatters badly in low temperatures. The hole in the middle will allow airflow generated by my makita fan to help flush dust from groves.
I hope it wont get clogged with dirt that much 
Awesome build!
I am really excited to see how this dust collection system works. Keep us updated when you test it
Meh. It don’t get clogged. I’ve set my dust collector on low speed and its now removing only light particles or super heavy ones. I can feel fresh air but makita fan is blowing too hard, so some of paricles are escaping under dish. Groves are clean. It’s working as expected 3rd day w/o cleaning and no glue collection yet.
Bit is working in better temperature. At the moment i’ve got -10*C in my workshop so it helps taking away the heat. Good enough.
Low power setting is 600W out of 2.300W dust colector.
I measured the amperage on motors. It picks 0.5A on no load, 0.9A on heavy load and 1.3A on blocked. So L298N 2A bridges are enough to run.
Fun fact… i need to pre heat up the power supply to run maslow becouse its too cold otherwise it wont start(hair dryier for 15sec and its fine for rest of day)
Yup. Today i forgot to attach hose from my new dust collector system to sled 
It turns out that this cheap version really working becouse after 5h of cutting w/o collection, dust really piled up everywhere
So It’s good upgrade for everyone.
After that fail i see difference clearly. No clogging.
That looks amazing!! Nice work, what a well thought out build. Your English is fantastic, especially with the difficult technical words…but if you want to make videos in Polish that could be wonderful too, maybe we can grow the Maslow community in Poland.
I have not used the automatic z-axis adjustment myself, but I think the information you want is available here:
Thank you for the update, watching this was absolutely the highlight of my day.
I’ve found that 2 channel encoders. They got good reviews.
Yes, this family of encoders is known to work very well.
they are just not chep
David Lang
I just watched your video. I was super impressed with your ingenuity in sourcing all your own parts and building a completely custom, yet what appears to be very sturdy, system. Generating 100 euro a month off of it is great. My favorite parts were the track piece used for the chain tension and your bent rod/ring. It proves you can truly make your own setup. Well done. Your english is easily understandable. Were you able to run the system off of a pizero using webcontrol? I found my position error of 2 only had a problem when the bottom of the sled was sticking to the cutting piece. this was because my bit was dull and was fuzzing the upcut on the surface and the dust collection was poor so it all caked on the bottom of the sled and the top of the work piece. In the cold, do your motor chain drive bearings have issues sticking? Could that be a source of the movement error?
Bearing don’t stick and work well already. I had problem with weight. Chain tensioners doesn’t compensate the load motors had.
2nd problem is that Z axis started loosing steps. I had to resolder connector 2nd time. I hope it will work
Next problem. The Z axis encoder started to count less in one direction. I need to change encoder board because it started lowering mill more and more in progress of cutting.
Maybe i will prepare my own magnet encoder out of scrap xD
0.5$ for hall sensor. I will need ~6 or 9 of them. They must be reliable so i have two ways. Buy 3x20$ or DIY… answer is simple. I will both solutions.
Today i added another 6 kilograms to sled and 18 kg to chain tensioners. Hmm now each one weight 27 kilograms
I use 5liters plastic water cans filled with sand. Sand is wet so there is risk of craking in low temperatures.
Ok, i can make clutch between 6mm and 8mm shafts from long M6 screw by drilling it with 8mm bit. I want to mount it directly to output shaft.
Lets start with new encoders for top engines:
LPD3806 with 600 P/R
or
GTA3806 (random P/R)
Now Z axis. I’ve got no place to connect it to TR8 screw.
I will give it highest P/R because output shafts are slow.
And maybe i will be eable to change motos to somthing stronger thanks to that. I will try to use motors from windshield wipers if new encoders do their job. I bought some from local scrapyard and some of them looks really beefy.
My magnet encoders got 20500 PPR and that 600 optical encoders will probably have 2400 PPR.
Can some1 tell me is that enough to run maslow normally?
Hey. Where is part of code responsible for hall sensors bouncing?
That is handled by the encoder library in the firmware. If I remember correctly there is not explicit de-bouncing code in there, but there is code to validate that a valid transition has taken place and not a random one
I don’t believe that there is any.
the way the encoders are setup, only one channel changes at a time, so if it
bounces, you just go back and forth one encoder step.
David Lang
Switches bounce, because there are two pieces of metal being smashed together.
Hall sensors don’t bounce. They may be affected by noise at their changeover point, but usually there is enough hysteresis present so that the output doesn’t flutter. Hence no need for Hall sensor debouncing.
