Delta printers do home at the top.
This is a fair point, but Delta printers also have those long arms that connect to the effector so it’s not nearly as much travel.
1 Like
So your printer homing off by a mm or two from a “cold” to a “hot” home?! Wow. LOL
Plus you are assuming the steppers are disabled only because Klipper timed them out. Bad assumption. Could be intentional in the pause macro.
Mechanical is accurate to 1 um, optical to 2 um.
If you want that extra micron then you’ll have to write your modified kipper because I think no one else will for such a negligible difference.
If it’s intentional in the pause macro then remove it, it’s bad practice.
The easiest switch to use is a proximity switch. Those vary in accuracy
Thomas Sanladerer did extensive testing of endstop sensor accuracy - mechanical contact, capacitive, inductive of different types, optical.
His testing showed standard deviation ranged from several um to into the ten’s of um, cold vs hot.
So much for the “extra micron” nonsense.
True, there is temperature drift.
-
From Thomas’ video I’m not sure whether the 10 microns he mentions/shows for optical refer to deviation among the measurements hot, or to the deviation between hot and cold.
Since he mentions the variance increases with increasing detection distance it might be it’s the latter and it could be compensated with an offset with a macro… -
I’m not sure how much a 10 microns even matter (5% of a typical layer height? and for just one layer), but if you say it does, then it does (at least to you) and I see why you look for a better solution.
Likely you’ll have to fork klipper and code your double endstop solution, because I doubt anyone able to code (I can’t, at least to the level required) will think that this is more important than so many other issues/improvements in klipper (or you hire a cheap python programmer, likely 1k will be enough) -
I don’t see it happening, but if you want you MIGHT get the same feature by
a) connect the extra endstop to an extra pin in the mainboard
b) force klippern to move the tool head or bed BLINDLY, so basically force a motion of the kinematics without any safety, while manually monitoring the extra endstop (in a loop: move, check pin, move, check pin, …), then once you detect the endstop triggered call for homing g28,see Configuration reference - Klipper documentation
But if you use my suggestion you have to promise you’ll let us know (new topic, this one will autoclose at a certain point) how it goes, especially if you crash your printer 
#3 - A hard cpu loop… awesome, Klipper will just love that. I think I see why you are struggling with this. Did you miss my earlier post to you where I said I had decades of sw development experience? HW too. Don’t worry about me.
I’m not struggling, since I’m not interested in it.
I missed the information about the experience, so it won’t be difficult to implement
Now that I read the first post again, I notice there is actually no request at all.
You mentioned it as an idea, that’s it.
Later you mentioned that if you wanted, you would implement it yourself.
But since you haven’t yet (as far as I know), does it mean that even you don’t care about your idea being implemented?
I’m puzzled. What’s the goal of the discussion actually?
That’s a lot of posting and questions for something you claim to not be interested in.
Gee, I thought I was supposed to choose the timeframe to implement this in, not one you deem as timely. Is this like the expiration date on a milk carton, “Must implement idea by xxx or else expired.”? This is quite amusing.
1 Like
I made my mind about a hardware solution and came up with this:
After reset it is in this state:
When activating the Pre-Endstop, it is in this state:
After releasing the Pre-Endstop it is this:
After this, the Pre-Endstop is disabled:
Now the Endstop is free to go.
Nice! I considered the same but with a $3-$5 mcu board. 2 digital inputs, one digital output, and do the in/out logic with a few lines of circuit python. Your discrete chip will work but doing it in code gives better flexibility. Plus could use a sensor that was analog.
as a DIY’er\Tweaker\Problematic Printer owner - i feel you.
I actually thought of the same idea.
The problem usually is the MCU doesnt have enough ‘ports’ for things like multiple thermistors or ultrasonic sensors.
Ahh, but it is klipper/btt octopus related. I cobbled up a voltage limiter that satisfy’s the not more than 5 volts the octopus pro allows w/o blowing the probe port, but still allows the sw to use 24 volts. So that’s what I was talking about. A dependable Z_home, something the octopus’s aren’t famous for. A 4700 ohm R to +24 supplies the pullup, but the probe pin is only pulled a few mv above the 5 volt line by running into a schotkey diode to the 5 volt bus. Works great, won’t blow the probe pins opto.
Now since I’ve offended, and my next Q is what does prusa make the mk3s hot end duct out of, where do I ask that?
No one replied. I’ve since found on thingivese, a duct shoe which looks not unlike the early mk3S duct. plenty of clearance for the chc 300C hot end heater. So I made a batch of them out of pc. From an X-Max3 originally but that needed more bed heat and had a limit to the nozzle due to heat creep running up the heat break and freezing up in the extruder. After making one and putting it on Ender5Plus, I could transfer the whole thing to my rebuilt Ender5Plus which can do 280C at the nozzle now, and around 160C bed. So pc with its horrible reputation actually Just Works normally. I made about a dozen copies cuz the original 1 bolt attachment is weak. On a team one extruder and a chc heater, It just Works. While running at 120mm/sec+. It, OOTB gave Y layer shifts at 20mm/sec OOTB w/o any IS in either case. I posted, about 10 days ago, detailed pix of the Ender5Plus rebuild in the “I made a thing” channel but they were deleted the next morning. Why? That has turned into a does everything printer. I’m proud of it. Now going much the same thing to a tronxy-400.
But typically nobody is running some stock neutered board when doing things like this. Klipper running on a Pi can use the usb ports or IC2 or CAN to talk with additional mcu boards.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.