Thanks for the ideas. I have managed to get a compiled version of marlin 1.1.8 to run from less than 130,000 bytes of progmem, with bilinear bed compensation included. I found that commenting out the G26 mesh validation tool saved me just the little bit of memory I needed.by DjDemonD - General
How about a small mosfet switching current from the 3.3v rail on the expansion connector via one of those low current signal pins.by DjDemonD - General
Hi can anyone suggest or point me in the direction of any guide to creating a build of marlin 1.1.x which cuts down on the amount of progmem it requires? I am trying to build for a machine with 130048 bytes of progmem, and have everything necessary working in 126558 bytes until I try to enable bilinear probing, then I run out of memory.by DjDemonD - General
Looks good let us know how it works, or if you need any help tuning/tweaking it.by DjDemonD - General
So I was using a 3 point (2 corners and the opposite side centre) mounting system with adjustable height piezo mounts. I tried disconnecting my front centre mount piezo from the circuit but got very poor probing accuracy when I was near the front edge, there just wasn't enough movement in the system to get reliable triggers. I also changed probe type on RRF to P8, the unfiltered mode, and it impby DjDemonD - General
Certainly using bigger pulleys, and then a reduction to restore the resolution would make quite a nice solution.by DjDemonD - General
I just bonded the noise cancellers on solidly with superglue. I figure since they are there to get rid of vibrations the more solidly they are in contact with the carriage I am trying "noise cancel" the better. And it does work. On this particular machine its nowhere near as accurate or sensitive as the hotend mounted piezo plate that was there before, but then its a moving-z bed, so noise cancelby DjDemonD - General
Well, I had 16t pulleys on my kossel XL which originally had PU/steel belts when I built it Sept 2016, by Christmas the belts were fatiguing causing a very odd issue that I couldn't track down for days. I also installed them on my corexy with 20t pulleys, they lasted longer maybe 5-6 months, no drama when they failed as I noticed it pretty quick and knew what to look for. I'm sure you could reducby DjDemonD - General
I've tried a few belts and basically steel is good and very inelastic compared to neoprene/fibreglass, but yes 16t and even 20t pulleys will fatigue the wires especially with 180 belt path. 30+ tooth pulleys are better but that's resolution lost in favour of stiffness. I've got some genuine gates 2gt power grip 6mm belts on my kossel Xl and it really makes no difference. Buy ebay belts with confby DjDemonD - General
Ahh cool yes I did see that I just didn't necessarily link your name to it. That's good I suppose diagonally opposed 2 piezo mount wouldn't work. I might change my 3+3 setup to 2+2. After doing a bit of a baseline to see how accurate its is first.by DjDemonD - General
Where do you position your two?by DjDemonD - General
I have started doing a bit more probing when cold. I think it does lower noise and perhaps improve accuracy, but it's harder to clean the nozzle and of more benefit if you can switch off your fans, which on more basic controller boards isn't usually an option. If your bed deforms when heated and you're probing a grid then heat it. Consider a noise cancelling piezo. I just bonded 3 under my coby DjDemonD - General
Tighten the 4 corner (or your equivalent) screws down a bit.by DjDemonD - General
I think thats the first damaged piezo disc which wasn't the result of bad drilling. Thankfully with this tech the sensor itself is cheap. Let us know if you need any assitance with it.by DjDemonD - General
I think you can raise the probing speed and get more reliable triggering, but its very dependant on the setup/mounts and how the piezo is actuated. You don't always get more accuracy, but the multitouch probing is indeed a big improvement for accuracy and reliability. My microdelta underbed setup is so sensitive I think I could probe at 1 mm/s reliably (if I have time Ill try very slow probing)by DjDemonD - General
That's working really well. There is definitely something in this noise cancelling setup that has some merit, where achieving a good range of operation in a mechanically noisy machine is concerned. I must find the time to revise my corexy underbed setup and see if it improves it.by DjDemonD - General
If theres no mount on thingiverse then that sounds like an awesome idea. The prometheus is a groovemount so shouldn't be too hard to adapt or fit into the existing Piezo20 design.by DjDemonD - General
Just set your z min software endstop to - 4mm then hit G29. You can always set it to something more reasonable like -0.2mm afterwards. There was a change in marlin around 1.1.0 when the way software endstops or travel limits are handled. So they don't always turn on and off as you would expect Eg send a G92 command which would allow movement beyond the travel limits and the movement won't be allby DjDemonD - General
Yes sorry its z probe connector. X and y movements will trigger the probe but this doesn't matter when actually probing as the x and y movements stop before the probing dive. You have a pause configured in (M558 R) so this allows the frame/axes to settle before probing. RRF does not monitor the probe unless actually probing. We toyed with the idea of using it to detect rough layers and crashes bby DjDemonD - General
We leave you to chose how to wire the system to your board otherwise we would be liable for miswiring etc.. So take a look at the duet documents There is 3.3v,gnd and signal. The pin that can be left unconnected on the duet is mod.by DjDemonD - General
Sorry it took a while to arrive we have a choice with shipping of cheap and slow or very expensive and quick. What stickers? Yes we have changed the pinnout to make it match a standard endstop cable so it's red, black, green (vcc, GND, TRG). I think documentation might be lagging behind hardware development I apologise if documentation isn't quite there yet but I can assure you the changes toby DjDemonD - General
Thanks, sorry if I've misinterpreted how you are using it. So you have 3/4 corner screws/adjusters and a central piezo bed mount?by DjDemonD - General
I think we have two cases where underbed works well: 1) delta, the stationary bed results in the least issues with false triggering you can use 1-6 Piezos depending on how you want to set it up. 2) using one piezo under the bed the way Stef/Chowa uses it. This works more in "microphone" mode (at least I presume it does). EDIT seems it is not using microphone type sensing but instead a single pieby DjDemonD - General
I was thinking have the coils in the track and permanent magnets in the carriage.by DjDemonD - Delta Machines
If you can make a linear motor that would be awesome. But... Isn't the point to get rid of wires, if the forcer is still in the carriage then you'll have to run cables to it.by DjDemonD - Delta Machines
The h parameter in G30 denotes a height correction which can be used to compensate for effector tilt by measuring the difference between probed height and actual height if using an offset sensor. If they were all the same it would make no difference (only possibly to the computed height) but if one was missing the algorithm thinks one probing point was massively higher or lower than the others. Yby DjDemonD - Delta Machines
Yes prices for cylindrical transparent plastic are very high. I'd suggest that a transparent cylinder like a fishtank would make a very good delta enclosure placed over a more conventional frame, but using it as a frame makes for a tricky build. Although the next step if I was going to make more would be figuring out how to make assembly easier.by DjDemonD - Delta Machines
So I added some anti-backlash nuts and reprinted the carriages in carbon fibre PETG, makes no difference to the calibration deviation, it remains at 0.25mm. However in terms of real world results I just printed this Marvin: and the layer registration issues seen in the previous cubes seem to have vanished, in fact, he has visible facets (Nebbian's test of quality - where is he these days?)by DjDemonD - Delta Machines
Maybe there's a way to encode the position information onto the power lines like homeplug type systems. On the other hand maybe some wires to the motors isn't that much of a problem.by DjDemonD - Delta Machines
I like the idea of the motors moving as its novel and whilst at first it might seem unnecessary, some benefits may emerge so its worth trying. My main concern is that your leadscrews will never be straight enough to prevent horrible wobble as you move up and down the axes. So you'll need to add a linear rail or rods or something in addition to the screws. Which begs the question could you use theby DjDemonD - Delta Machines