Nozzle Probing comes to BEACON

[Penatr8tor]

1 hour ago, kingstar said:

Both the Beacon and Cartographer have accelerometers in them, so probably using those to detect when the nozzle physically touches the bed.

Except that the touch works with both the Rev D and Rev H the Rev D having no accelerometer.

My guess is that because the sensor can measure distance extremely fast with very very high accuracy, it just looks for when the Z measurement drops from a steady state to stopped or slowed. For example: Say the nozzle moving towards the bed, 30mm to 20mm to 10mm to when the nozzle touches 0mm to -0.00001mm to -0.00002mm and STOP... You have touched something.

It's only a guess. My background is mechanical so that's my disclaimer. 

[Dirk]

2 hours ago, Penatr8tor said:

My guess

 

  

16 hours ago, chrisalbertson said:

I can't see

 

On 5/5/2024 at 5:51 PM, Penatr8tor said:

Probably not.

 

Man... you guys are like a bunch of old wives telling tales and gossiping around 

Stop speculating about the so called magic of how Beacon code works... They say it is MAGIC.. I say MAGIC is very middle-ages...

Some say they Believe it is... I say Believing you can do in the church, mosque, synagoge, temple or whatever building you go to for believing...

I placed a post yesterday, which gives you the source code of how it is done. Go take a look at THE SOURCE how it works. So you will know there is no magic. No guessing. No praying. No slipping. Just logics and electronics.

For those who can't read the code, read the comment from the klipper developer. And what @mvdveer has to say about it.

And of course @Penatr8tor does an educated guess and is right.

 

 

 

[Penatr8tor]

Yay!!!! Woot Woot.

My guess was right!

Finally 

[chrisalbertson]

6 hours ago, Dirk said:

  

Man... you guys are like a bunch of old wives telling tales and gossiping around 

Stop speculating about the so called magic of how Beacon code works... They say it is MAGIC.. I say MAGIC is very middle-ages...

 

Wow, it works exactly the way I suggested it could.   It looks for the impedance to stop changing.

Next, how to eliminate false positives.    I think using redundancy might help.  If there is an accelerometer on the toolhead then a weak signal for both at the same time would be a strong indication.   Think of this as a sensor fusion problem and do NOT use some kind of complex "if that then this unless that" logic.  Just multiply the probabilities and threshold the product.  This requires only some "very deterministic" math and only one "if" statement to apply the threshold, even if you are fusing 5 kinds of sensors.   Using the "multiply and threshold" method you can even use "dead-reconing" as one of the "sensors".   Dead reckoning is just saying "I command a collision from about X distance and about Y rate and I'd expect a collision about now."      

[claudermilk]

11 hours ago, Penatr8tor said:

Yay!!!! Woot Woot.

My guess was right!

Finally 

Quick! Print this page and show your wife! "See? A bunch of strangers on the internet said I'm right!"

[Penatr8tor]

27 minutes ago, claudermilk said:

Quick! Print this page and show your wife! "See? A bunch of strangers on the internet said I'm right!"

Exactly!!! 

[PRGeno]

Hmmm....  Wondering what is the definition of "Early" means, as in: "Beacon Contact is set to officially launch in early May"

[mvdveer]

10 hours ago, PRGeno said:

Hmmm....  Wondering what is the definition of "Early" means, as in: "Beacon Contact is set to officially launch in early May"

In 3d Printing terms, early may just as well be translated to "whenever it is ready"  Still waiting for my LDO Milo kit I ordered in March, supposedly to be delivered Early April - nearly end of May now and still now news on when it will be delivered. Sad thing, it is paid for already, so I will just have to wait for "Delivery Early April" to become reality. Maybe Early April 2025? 

[TitusADuxass]

Not wishing to hijack this thread but...

If this cheeses anyone off I can start a new thread - let me know.

Now we have a multitude of devices/methods to scan/mesh the bed. Which one is the most accurate?

I've done bit of googling but haven't found anything definite, maybe it's time for someone to do some comparisons. I'd be up for it once I have my Vz330 up and printing well.

At present I have CNC TAP on my v2.4 and it prints well and is nearly fire and forget. It does, however, have its inaccuracies. Yesterday I printed a first layer test of 5 squares placed around the bed (corners and centre). My TAP/KAMP setup does a print specific bed mesh before each print (the time taken is not a driving factor for me).

When I inspected the prints there was a noticeable diffference between ones at the front and the ones at the rear (looking at the underside of the print - on rear ones it was easier to see the individual filament lines i.e. the nozzle was higher at the rear than the front. There has to be an explanation for this, is it at all possible that the actual extruded material is lifting the toolhead - I doubt it but I have my suspicions.

The TAP is good but adds a fair amount of weight to the toolhead, and even the CNC TAP has a certain amount of uncertainty due to the fact that the toolhead is not fixed.

I also have the original Omron inductive probe, a BTT MicroProbe (and gantry mount), a Cartographer, a BLTouch clone and an I have an itch that I need to scratch.

The Microprobe is extremely light weight and can be mounted in place of the Omron probe. I question it's surviveability in the hot area where it's mounted.

BTT have also just brought out the Eddy - this also mounts in place of the original Omron probe.

The Beacon and Cartographer both require a unique mount on the tool head.

The Beacon, Cartographer and Eddy all use the spring steel sheet (or other metal surface) their measurements/calculations. How does this take the thickness of the PEI coating into consideration or does just make an educated (but accurate) guess? I've read the linked page about Tap detection with Idc1612 - it's very interesting but also has a lot of if, buts and maybes in it.

At the end of the day is this just chasing zeros and whichever one you use is more than accurate enough for our requirements?

[mvdveer]

1 hour ago, TitusADuxass said:

At the end of the day is this just chasing zeros and whichever one you use is more than accurate enough for our requirements?

We are striving for perfection - and that will never be achieved. Have been trying to achieve it all my adult life - I am 65 in a week and have NOT even come close to achieving it. Mesh probing with the Beacon and Cartographer has been more accurate over a large build plate than TAP, BL Touch or Microprobe . (Tried them all) . How many full plates of prints will you print? How important is "Near perfection". I think in answering those questions - your decision may be a bit clearer

But then you  are in Germany - precision  is paramount. (I can say it as my first wife (deceased, married for 27 years) was German)

 

P.S - I saw paramount comes up as I link - NOTE - I DID NOT ENTER THIS LINK, must have been populated by AI thought the forum

[Dirk]

27 minutes ago, TitusADuxass said:

can start a new thread - let me know.

Not to do with cheese, but I am sure this is in itself worth a topic cause it is much more important than and especially more interesting than beacon

[TitusADuxass]

1 hour ago, mvdveer said:

But then you  are in Germany - precision  is paramount. 

It may be paramount,  but it's rarely achieved.  And don't get me started on the efficiency myth.

Edited May 21, 2024 by TitusADuxass

[Penatr8tor]

A wise person once said to me...

Better is the enemy of good enough.

Eventually, you just have to stick a fork in it and say done.

[Dirk]

7 minutes ago, Penatr8tor said:

Better is the enemy of good enough.

Oh yeah... my professor told me the same.. so it must be only what wise people say "Dirk, jongen... Beter is de vijand van goed"

Hard to grasp for perfectionists.. But truth as a  cow.

[claudermilk]

3 hours ago, TitusADuxass said:

At the end of the day is this just chasing zeros and whichever one you use is more than accurate enough for our requirements?

Yes.

OK, I'll add a bit more.  We need to remember we are running glorified, computerized hot glue guns. All of the probing methods are plenty accurate enough, so it's pick your poison.

I am quite happy with Tap. I'm not going for #speedboatraces, so the weight doesn't bother me. The lost 3mm front travel is a little annoying, but not enough for me to do anything about it yet. I've been printing a number of fairly big boxes lately that cover the majority of my bed. With my Tap and things dialed in, I'm getting beautiful, perfect first layers all the way across the surface every time.

[Pappy3831]

I went back and looked at the release notice. Nowhere do they mention anywhere what year in early may.  Guess I should have waited to order another beacon. Oh well. Guess I could go ahead and use my cnc taps I have instead. 

[Dirk]

Well... Since @TitusADuxass won't make a new thread and @mvdveeris already in the night mode and not going to carry this to a new thread, I will want to share my experience.

I understand we are no more talking about the voron stock Omron probe, klicky-magnet-in-the-binny, or about the able probes.

The best data I still have that compares the cnc tap to the IDM/cartographer is the one from @Paul Trautner over here.

It shows clearly the speed difference while giving a same bed mesh.

with the newer insights we would have to say: if the pei cover is flat.

@claudermilk I loved windows XP too. Hated that I had to use windows vista. And then 7. Same goes for my cupboard sized TV. But if you don't adjust to the environment, you will become obsolete (or OLD).

So if you would install CAN on your printer and see how much it speeds up your toolhead, you will know how paramount ( @mvdveer put in that link! Not me! ) a light and compact toolhead is without a dragchain dragging it down. Not only speed, but also the vibration caused by tap disappears when you get rid of it.

I remember reading a comparison of measurements between tap and klicky from one of the voron forums regulars. It showed clearly that adding voron tap was slowing the toolhead down enormously.

But also the extra resonations tap creates requires you to adjust your acceleration and speed as advised by the input shaper.

So it is not just accuracy. But also speed and the effect it all has on your prints. 

Another point I think is very important, is the lifetime of a device. German machinery (bmw, Mercedes, Audi) are good and preferred cause they are more reliable and break down less easy, cause less problems compared to many other brands.

If you check out forums or voron discord, you will see how many people have problems with failure of their tap and how often they have to replace it. I bought an R8 that came with printed parts containing abs/cf and I had to reprint it twice in 3 months.

The Cartographer I bought last december is still in its package because the IDM I bought a few weeks earlier still works without a flaw.

 

Finally, like always I will also bring in the price in the comparison. My IDM cost me 45€. Cartographer 40€. Tap kit with printed parts 35€. I have seen the cnc you guys use costs 65-80€s.

If I wanted to spend more I could also buy a cnc mount for cartographer for 20€, which will make cartographer still a cheaper product that has a longer life and gives me not only a good first layer but also a nicer print.

 

 

[claudermilk]

22 hours ago, Dirk said:

@claudermilk I loved windows XP too. Hated that I had to use windows vista. And then 7. Same goes for my cupboard sized TV. But if you don't adjust to the environment, you will become obsolete (or OLD).

So if you would install CAN on your printer and see how much it speeds up your toolhead, you will know how paramount ( @mvdveer put in that link! Not me! ) a light and compact toolhead is without a dragchain dragging it down. Not only speed, but also the vibration caused by tap disappears when you get rid of it.

I feel attacked.  Guilty as charged (GOM). Et tu, @Dirk? Now @mvdveer is sending his minions to pester me.  

In all honesty I have nothing against those, it's just my printer is working really well now and I have no driving reason to updated it. I can already get it moving faster than my Revo can squeeze out plastic, to additional speed does nothing for me. Perhaps one day I'll grab one of those CHM high flow nozzles and find a need to speed up the tool head, but today is not that day. Since I updated my Tap to R8 it's been working flawlessly.

I'm sure eventually I'll swap over to these newfangled probes and communication, but with things working I am happy with the printer as is right now and am focusing on other projects.

[Dirk]

1 minute ago, claudermilk said:

my printer is working really well

Don't fix it when it aint broke you americans say, right?

And your printer going faster than your hotend can push out plastic, is a good argument too. But stay tuned... 

[D__J]

Has there been any update on this anywhere? I can't find anything. Early May and all...

[Dirk]

54 minutes ago, D__J said:

I can't find anything.

Did you ask in the Annex/ Beacon Discord? or forum? I hear they are very supportive.

If you can share with us, I am sure you will make many happy