Duet3D announces new tool board at FormNext

Surgikill

@fcwilt I watched the video that was linked. I don't think the entire bed was measured. I saw 5 dial indicator test points, one at bed center, and then one at each corner of the bed. Maximum measurement was 0.04mm, minimum was -0.03, so total flatness from those 5 points would be 0.07mm (Flatness is defined as two parallel planes on either side of the flat surface as the tolerance zone for the surface. All the points on the specified surface must lie between these two planes.)

I'm not sure if the entire bed was measured in a grid to determine any high spots, but you can't really determine the flatness of a material by measuring 5 points and then assuming the rest of the bed falls within that tolerance.

On the other hand, I believe the tooling plate comes in 6'x12' stock sizes, and I believe that flatness tolerance is for the entire 6'x12' stock, so you might get a tighter tolerance than specified seeing as we are only using a small portion of that large stock size, but I wouldn't bet money on it.

fcwilt

@Surgikill

I haven't seen the video.

I will have to call Midwest and see if they can clarify the issue with regards to the size of the piece.

And when they cut a piece do they perhaps deform it in the process?

The custom bed I had made was measured by the machinist and it was, as I mentioned, flat with 0.09. I did not ask him how he measured that. I should.

I have another bed on order being made from a stainless steel which the machinist says will produce the best results. Time will tell on that.

Thanks much.

Frederick

Surgikill

@fcwilt That guaranteed flatness is before and after sawing. They are most likely giving themselves some buffer in that number. I would imagine that your machinist probably used a surface plate, something like this setup. https://www.youtube.com/watch?v=WwUqPiQ9JAQ

fcwilt

@Surgikill

Thanks for the video. It was very interesting and educational.

Frederick

deckingman

It matters not a jot to me whether you believe me or not because I know for a fact that my bed is both flat and level within +- 0.03mm at any point over the entire surface, and that I can print edge to edge with no firmware compensation of any kind.

Except that maybe you are calling me a liar to which I object very strongly.

This video was shot many years ago using the exact same build plate and before mesh compensation had been added to the firmware. Also, I use a single motor and continuous belt driving 3 lead screws so automatic bed leveling is not possible.

https://youtu.be/U733PMTou7M?si=cgGvQHLgwxZdXike

How do you explain that if I'm the liar that you are implying that I am?

deckingman

@fcwilt said in Duet3D announces new tool board at FormNext:

@deckingman said in Duet3D announces new tool board at FormNext:

Either you haven't used cast machined tooling plate

The flatness of typically available plate is appx 0.40mm for plate less than 12mm thick. For 12mm or thicker it drops to appx 0.14.

So what did you use to obtain 0.03?

My machinist has nothing as flat as that.

Frederick

I used this https://www.aluminiumwarehouse.co.uk/collections/plate-cut-to-size/products/12-mm-aluminium-cast-tooling-plate

The tolerance is usually quoted as the cast plate tolerance prior to machining and covers the entire sheet size (usually 1.2M x 2.4M). So a 400m x 40mm machined section will always be more flat than the quoted tolerance.

Most "keyboard engineers" think that anything bought to a specified tolerance WILL vary by that tolerance whereas the tolerance is the MAXIMUM variation that the supplier uses to be able to guarantee the product.

According to my DTI, mine is actually flat better than +_ 0.01mm. The +_ 0.03 variation is due to imperfection in levelling (which I could get better if I spent more time tramming the gantry).

o_lampe

@Surgikill said in Duet3D announces new tool board at FormNext:

Toolboards are primarily made for tool changer setups, so multiply your cost difference by the number of toolheads.

I'm not entirely sure if you'd need a scanning probe on every tool head? If I understood Tony correctly, they are also working on a nozzle coil which would be stationary and after a toolchange you'd level your nozzle there (just like the contact probes on CNC machines)

Surgikill

@deckingman in your video that you linked your dial indicator shows +0.04 and - 0.03 as the maximum and minimum values. It's there pain as day.

I never said you were lying, but your flatness tolerance is wrong by definition. It's either 0.07mm, according to the video, or 0.06mm according to you. A +/- is not applied to a flatness tolerance. If you only measured those 5 points in the video then it's really a moot point.

I'm not trying to sound pompous, but I'm a degreed mechanical engineer, not a "keyboard engineer". Using the wrong terminology can be misleading at best and disingenuous at worst. Instead of getting emotional about being incorrect, it would be beneficial to everybody if you used it as a learning opportunity. Nobody is perfect, myself included.

deckingman

@Surgikill So what if I made a mistake and it should have been 0.07mm rather than 0.06mm? If you had watched the video to the end, you'd have seen the part where I explained that I thought this would be tram and level enough (as was later proven to be the case) to get perfect first layers over the entire area of the build plate, but that I could spend a bit more time getting even better if necessary.

The reason for only showing photos of those 5 points is because the video was about tramming the gantry with respect to the bed and not specifically about measuring flatness. The clue was in the title "...............installing the Z stage and getting it level and "tram"..........." I had already established that the plate is as near dammit perfectly flat.

You can split hairs all you like but the point is that I can print edge to edge with perfect first layers without using mesh compensation. I can't help it if people like you don't believe me.

fcwilt

@Surgikill

I went and checked the video.

With the DI attached to the XY gantry how would you tell the what was responsible for whatever reading was displayed? Could not imperfections in the gantry contribute to the reading?

Thanks.

Frederick

fcwilt

@deckingman

I forgot to ask you what sort of surface do you print on? Directly on the metal plate or something added on top of it?

Thanks.

Frederick

deckingman

@fcwilt said in Duet3D announces new tool board at FormNext:

@Surgikill

I went and checked the video.

With the DI attached to the XY gantry how would you tell the what was responsible for whatever reading was displayed? Could not imperfections in the gantry contribute to the reading?

Thanks.

Frederick

You could direct your questions to me rather than via a third party.

But for sure if the gantry sagged it could imply that the bed was bowed or "dish shaped". My calculations showed that any deflection of the gantry due to the weight of the hot end was insignificant and the rails were known to be straight. But the important thing to note is how flat and level is the build platform with respect to the full range of movement of the nozzle. By fitting the DTi to the gantry where the nozzle is mounted, one is checking both the gantry and the build plate. If there was an unacceptable deviation, then one would carry out further measurements on the individual components to determine where the problem was. But it's a lot quicker to check the entire mechanism in its normal working configuration and only break that down into individual components if an unacceptable deviation is observed.

deckingman

@fcwilt said in Duet3D announces new tool board at FormNext:

@deckingman

I forgot to ask you what sort of surface do you print on? Directly on the metal plate or something added on top of it?

Thanks.

Frederick

Glass. 6mm thick plane float glass. I learned long ago that the process of having the glass toughened leads to unacceptable distortion. And yes before you ask, I checked the flatness. I've tried most other build surfaces but haven't yet found one that beats plane untreated, "hot swappable" float glass (with a little help from 3DLac now and then) and I'm still using the same 3 sheets that I bought about 7 or 8 years ago.

Surgikill

@fcwilt the gantry can contribute or offset the readings. Aluminum extrusion tolerance is terrible, straightness of 0.0125" per foot. Hence why measuring at 5 points and then assuming that those points will simultaneously be the highest and lowest measured points isn't really the right way to go about it. Even measuring the flatness of the build surface while on the machine isn't really correct, because you don't know if it's your gantry or your plate causing the inconsistencies.

If you are printing at 0.2mm layer height, there is a pretty large variation of z height you can get away with. As you try to increase resolution the problems will begin to show themselves. This is why mesh compensation is used.

Surgikill

@Surgikill said in Duet3D announces new tool board at FormNext:

@dc42 can you give us the exterior dimensions of the board? I'm trying to see how it will fit in my printer.

I threw some pictures up in CAD and scaled them, looks like board width is around 45mm without connectors, and board height is around 50mm without connectors.

dc42

@Surgikill I expect we will publish drawings and models next week.

nikscha

@elmoret Nothing stopping you from using a 24V heater at 36V XD
You'd get 2.25 times more power out of it!
(and a fire hazard lol)

Surgikill

@dc42 any update on this?

dc42

@Surgikill not yet, Tony normally does the drawings and STEP models but he is very busy with Formnext follow-up and SMRRF preparation.

edsped

@Surgikill said in Duet3D announces new tool board at FormNext:

@fcwilt said in Duet3D announces new tool board at FormNext:

@Herve_Smith

You consider the prices to be an issue?

The money I spend on hardware is dwarfed by the money I spend on filament.

Frederick

Toolboards are primarily made for tool changer setups, so multiply your cost difference by the number of toolheads. I'm currently building a tool changer, and each toolhead is going to run me close to $400. At some point, cost becomes a favor. When my toolhead alone is over double the cost of an ender 3 you have to take a step back and reevaluate.

I run RRF-36's on my toolchanger and wish I set it up that way sooner.