Duet 3 Scanning Z probe

dc42

@Notepad the auto calibration is only possible if you have another way of establishing Z=0 first. This could be another type of Z probe such as a BLTouch, or manual probing.

We haven't yet tried using the scanning probe as the only Z probe. To do this you would need to do the following:

Set an accurate Z=0 reference, run the calibration command, and save the results in config.g
At several different temperatures, measure the Z height that produces the reference output value from the probe, and use these measurements to set up suitable temperature compensation in the M558 command.

Notepad

@dc42 said in Duet 3 Scanning Z probe:

the auto calibration is only possible if you have another way of establishing Z=0 first

Ahh, so reference point first, then it can scan the entire bed.

We haven't yet tried using the scanning probe as the only Z probe. To do this you would need to do the following:

Ill definitely be testing this out once I get mine, as a bit of pain calibrating it initially will mean I don't have to constantly carry a inductive/BLTouch on the toolhead.

fcwilt

@dc42 said in Duet 3 Scanning Z probe:

We haven't yet tried using the scanning probe as the only Z probe.

That is a bit of bad news.

It makes the product less appealing.

I need to think on this a bit.

Frederick

zapta

If I got it correctly, this is the third Duet3D Z sensor, each with a very different technology. The quest for the ultimate Z sensor continues.

This post is deleted!

Notepad

@fcwilt said in reference to @dc42 Duet 3 Scanning Z probe:

We haven't yet tried using the scanning probe as the only Z probe.

It definitely makes the switch over not as simple of a choice. Hopefully if I can figure out a test cycle or homing regime to use it as the only Z probe, Ill share that to everyone.

If it cant be used stand alone. The main selling point of fast and compact bed scanning looses the compact aspect as there would need to be a redundant probe just for initial offset measurements.

zapta

@Herve_Smith, what the scanner provides is similar to what the inductive probe provides with Voron 2.4 but at an order of magnitude faster. That is, height measurement relative to a known zero, which is provided by the 'tap' end switch or similar.

In other words, it takes a z=0 point at a given (x,y) and quickly provide zero points on the entire bed surface. The alternative is to use an absolute Z probe (e.g. BL Touch or Cliky) and slowly probe the entire bed, or wait until a fast absolute Z probe will be invented.

Wether this product has value to your application it's up to you.

fcwilt

@Notepad said in Duet 3 Scanning Z probe:

If it cant be used stand alone. The main selling point of fast and compact bed scanning looses the compact aspect as there would need to be a redundant probe just for initial offset measurements.

And the speed would allow creating a new height map for every print which would be nice for me since I use a variety of printing surfaces.

For me the speed is the real benefit. The 8mm inductive sensors I use take up very little space.

Frederick

Notepad

@fcwilt said in Duet 3 Scanning Z probe:

The 8mm inductive sensors I use take up very little space.

I predominantly use 12mm inductive probes, and while the size is bigger its not a massive issue. But the time to probe the bed is just soo slow. On a regular bed it takes ~3 minutes and on a larger bed at least 6 minutes.

Even If I have to mix both, The extra speed this probe will have is going to be well worth it.

fcwilt

@Notepad said in Duet 3 Scanning Z probe:

I predominantly use 12mm inductive probes, and while the size is bigger its not a massive issue. But the time to probe the bed is just soo slow. On a regular bed it takes ~3 minutes and on a larger bed at least 6 minutes.

Yes it takes a while but how often do you currently do it?

With the hours and hours spent printing things what is the "cost" of a few minutes creating a height map every once in a while?

Frederick

Notepad

@fcwilt said in Duet 3 Scanning Z probe:

With the hours and hours spent printing things what is the "cost" of a few minutes creating a height map every once in a while?

Before every print, but the over all time cost vs printing time is so negligible so im happy to do it.

I'm known to not treat my printers very well (as they are built to be absolute tanks) so probing for every print just gives me peace of mind that its perfect every time.

droftarts

@fcwilt @Herve_Smith When we say we haven’t used it, or don’t recommend it (currently) for use as a Z probe, we mean that we haven’t tested it to do that yet. We haven’t calibrated it against temperature, we haven’t tested repeatability, we haven’t written macros or implemented firmware changes, written documentation or whatever else is required, to enable its use as a conventional Z probe. So we don’t recommend it, at the moment.

Temperature affects inductive probes. But inductive probes have been used successfully as Z probes, eg Prusa Pinda, and the SZP board had its own thermistor on board. As far as I’m aware there is no technical reason why it couldn’t be used as a Z probe; it’s obviously very accurate and repeatable, at least under the conditions of the show, where I watched it scanning the bed for 4 days straight. We just haven’t had them for long enough to work out the best method to make it reliable and repeatable in that role, in a wide range of situations, when the focus was on bed scanning.

Ian

fcwilt

@Notepad said in Duet 3 Scanning Z probe:

Before every print, but the over all time cost vs printing time is so negligible so Im happy to do it.

OK - I'm sort of lost here.

Earlier you said the time to probe the bed was just too slow.

Also earlier you said the extra speed of this new scanning probe would be well worth it.

Here you say with your current probe, the time is negligible and you are happy.

I find the idea of spending time/money to solve a "negligible" issue confusing.

Thanks.

Frederick

Notepad

@fcwilt
Nope, thats on me, My wording skills are pretty poor at times.

I probe before every print because I want to reduce any risk of an imperfect first level. A failed print because I tried to save 4 minutes in skipping a probe sequence just isn't worth it IMO.

And when you take a step back and look at the average time to complete a print (say 4 hours) the probing time is literally only 1% of the total time. So I'm happy to do the probing before every print.

But on the flip side of the coin, from a user experience aspect, the initial wait for 4 minutes just feels sluggishly slow, especially when doing really small prints (say less than 1 hour) the probing time just eats at the mind as I wait for it to complete.

my feelings are juxtaposed between happy with what I currently have, but also disappointed by the completely acceptable 4 minutes it takes.

What I am definitely looking forward to is the ability to do increase resolution height maps without any time impact. And after using a beacon probe, it does become a tipping point where once you have experienced those speeds, you never really can look at the slower methods the same.

Hope that makes more sense.

fcwilt

@Notepad

Sounds good to me.

I'm going to get one just to work with it BUT I don't know that I will undertake the work involved to re-design the tool mounts for my five printers (all different) IF a "traditional" Z probe is still required.

I don't currently use any 1LC boards but if they were updated to include the scanning probe then I would certainly explore a re-design of my tool mounts,

If I went that way I would, sadly, be forced to upgrade the two of my printers that are using Duet 2 boards. I mean who wants the latest and greatest hardware?

I'm sure the wife would understand.

Still thinking on it.

Frederick

zapta

@fcwilt said in Duet 3 Scanning Z probe:

I'm sure the wife would understand.

Start the conversation with her printer.

This post is deleted!

dc42

@Herve_Smith I think it's a case of damned if we release early, and equally damned if we wait until we've spent a lot more time characterising the behaviour when the probe used as the sole Z reference.

We've spent a lot of time testing the scanning function, to the point where we know it works well. So we've released a product that is definitely useful to many owners of existing machines that already have a good way of establishing Z=0, and probably capable of being the sole Z probe in some (but perhaps not all) machines with metallic bed plates.

We know that inductive sensors are temperature-sensitive, and we know that in some cases at least it is practical to compensate for the change in behaviour with temperature. So we've included a thermistor right on the coil so that we can compensate for changes in coil temperature. This might not even be necessary - it may be that the change in bed resistance with temperature is the primary cause of temperature sensitivity, in which case compensating for bed temperature will be more appropriate. Or it could be that in some machines we will need to compensate for both bed and coil temperature. Currently RRF can only compensate a Z probe for one temperature, but we can extend that if necessary.

This is a case where community input is so important, because the community has a much wider range of machines than we do. If we were to state that on a particular machine with a particular make of bed plate it is possible to use the scanning probe as the only Z reference, how useful would that really be, given that we can't guarantee that users with other machines and different bed plates will have equally good results?

Machines vary enormously, and on some of them (e.g. machines with a heated chamber), it may be that sufficiently accurate temperature compensation will never be possible. In these cases, if mesh bed compensation is required then it it still useful to have a fast scanning probe even if a separate contact-based Z=0 reference must be provided.

fcwilt

@dc42

Thanks for the comprehensive reply.

Thoughts:

could you design a 1LC board with the scanning Z probe? I would be far more inclined to re-design my tool mounts to use a 1LC board like that.
you guys are smart - really smart - design a very small "contact Z probe" and sell the whole package 1LC with scanning probe and contact probe. I would definitely go for that.

Frederick

droftarts

@fcwilt The new Roto toolboard already has a scanning Z probe on it, and is a bit smaller than a 1LC (more details to come). You may also prefer the connectors on it!

Ian