Greater range for z-offset

Feature Requests & Ideas
1

Currently the z-offset feature allows you to offset the starting z-position from 1mm lower than default to 1mm higher than default. This request is to expand that range to something more like 1mm lower than default to 3mm higher than default. The reason for this is that even at 1mm higher than default with some tanks there is still noticeable compression of the early layers of the print. For users who are printing directly to the platform this broader range would allow them to in many cases minimize or remove the remaining compression, and the artifacts that come with it.

I have previously suggested this to Formlabs and they said they wouldn’t do it, So here are my responses to some of the arguments I have already hear against it. I am aware that Formlabs does not support printing to the platform, and suggests against ever doing it. I am also aware that printing to the platform presents even more complications than standard prints. However I also know some prints are best printed straight to the platform, Formlabs even admits and discusses this in their article about the new rook sample part. I also know that this feature will improve the quality of parts printed to the platform making those parts that are best printed to the platform print even better. Also in the past Formlabs suggested this feature might lead to users damaging their printers, I have specifically made this request to only expand the z-offset range higher not lower so it adds no more likelihood that a user set it too low and damage their tank.

This post is partially in response to the request “Ability to turn off the initial layer compression for flat printing”, as it seems this is a more feasible option. Please see that thread here for further details. I would like to point out that in this post multiple users discuss, for lack of a better solution, modding their tanks with thinner PDMS layers to accomplish this exact goal. I hope Formlabs will recognize they can, with PRACTICALLY NO WORK, give these users and others like them a simple solution that doesn’t require modding their machines, or adding a new and possibly confusing and or dangerous feature.

Also for any users that catch this request and would also like to see this feature please chime in and let Formlabs know that they have users that want this.

1 Like
2

My thought on the compression layers is to be able to have them off by default when supports are not used. I think this could be a simple software addition.

Playing around with the z-offset and extending it’s range might be good for those who re-coat their tanks and enable them to re-zero Z but might not be all that useful in the case of flat printing as they noted. I had problems with layers holding up after .5mm offset. The new version of the software seems to have a % rather than a fixed amount so fine tuning might be impossible or hard to figure out.

3

@RocusHalbasch, I want to let you know that we’re not ignoring you, it’s simply a more challenging question than it may appear. We’ve talked about these challenges at length, and I just want to summarize some of the discussions I’ve had to learn more about this request.

Some initial compression is necessary to ensure that a print adheres fully to the build platform. It is a necessary part of the printing process. When you print directly on the build platform, this compression phase is absorbed into the model itself, reducing part thickness. When printing with a base and support construction, this is not an issue—but as you note, there are many occasions when it’s convenient, fast, or useful to print directly on the platform.

Simply turning off compression would mean that your print will not adhere. That’s not what anyone wants. Providing more user control in the fine-tuning feature (which is designed to accommodate for some variability in tanks), could result in a number of manual non-adherence errors. We’re working to design a system that is reliable, repeatable, and easy to use, so we try to avoid settings like that.

To best balance the needs of accurate parts printed flat on the build platform, avoiding non-adherence, and eliminating support structures, moving the build platform up (as is done within fine tuning) is not the answer. The creation of a raft on the model, however, is an approach that could address these three needs. This approach can best be thought of like the support structures with a base and without the supports. This technique can be done quite easily in any CAD program, and leaves the thickness decision completely up to the user, who knows best about the specific printer and tanks in use.

Of course, we’re always listening to you guys, and working to make improvements both to the hardware and to the software, but at the moment, the fix doesn’t appear to be as quick or as easy as we would like, especially considering the many different needs of folks out there. We’ll keep an on this as development proceeds. I’m sorry I can’t give you more detail than that.

4

@KenCitron

You can’t simply turn compression layers off. They are the product of the combination of printer calibration, individual tank thickness, and z-offset. Printer calibration results in tanks that sit at different heights and angles. The tank position and angle is different for every printer and varies quite a bit. The tank thickness varies for every tank and isn’t as variable as calibration.

The printer has no idea where the surface of the PDMS is located. It has an optical endstop sensor at the top of the column the platform moves up and down on. When a print starts it homes to this position at the top of that column and then lowers 165mm +/- z-offset, no matter where the surface of the PDMS layer is located. The z-offset feature allows you to raise or lower the starting position the platform is in when the print starts. From this position it prints the first layer and raises the height of one layer and prints the next. It does this once again with no idea of where the surface of the PDMS is. Compression layers happen when the platform is pressed down into the PDMS when a layer is printed. The lower the platform the more it presses into the PDMS during the early layers of the print and the more compression you get. The higher the platform the less it presses into the PDMS during the early layers of the print and the less compression you get.

Just how much compression you get without z-offset depends on your printers calibration and your tray thickness as I described above. A z-offset of 0 on some printer and tray pairs results in so much compression that prints fail because the PDMS warps too much from the platform sticking in to far, on others it results in so little that prints won’t stick to the platform in some of the print area because the platform and PDMS are too far apart from each other for the first layers to attach to the platform.

One of the major reasons that the Form1 has compression layers is because every printer varies a lot in the height and angle of the tank position and every tank varies some in it’s thickness, and the printer has no knowledge of either of these variables so it just blindly lowers the platform down far enough so no matter what those two are it should be far enough for your base to attach and then they have you print with supports and a base thick enough so it can absorb the compression, this allows Formlabs to be less exacting about the height and angle of the tank position during calibration.

So the reason you can’t just turn of compression layers is because they are caused by the platform sticking too far into the PDMS layer, and the printer doesn’t know at all where the surface of the PDMS layer is, and the position of the surface varies a lot from printer to printer. The result is that in order to prevent compression layer you have to adjust the z-offset a little at a time until you find the right amount for your printer so that the platform starts the print just touching the surface of the PDMS layer instead of sticking down into it. This will result in prints that show no signs of compression and are close to the right thickness if printed directly to the platform on a printer with a fairly straight tray.

The reason for this request in response to yours is since Formlabs can’t make a feature to just turn off compression layers, I requested this which is something that would allow users enough z-offset to manually turn off compression layers. However to meet your specific request a little more closely the Formlabs team could consider having two z-offset values, one for prints with supports and one for prints without supports. You would still have to find the right amount for your printer but once it was set it would give you the results you desire.

5

@Sam_Jacoby

So I’m still a little lost as to what the problem is. I recognize that for dependable adherence some compression is necessary. In my past experience with a previous Form1 I owned the amount of compression required for something with a fairly broad base was quite small. On that previous printer I was able to get far less compression in some areas of my platform than I can on my current printer. I know from this experience that with a greater z-offset range I could print with less compression than the software will currently allow and still have dependable adherence for certain prints, so it would seem for these prints moving the build platform up is the answer, or at least it is an answer that will work to get me better results than I am currently able to.

I understand that if the z-offset range were extended some users could increase the z-offset too much and some of their prints might fail to adhere. However that does not seem like a good reason to handicap the printers from meeting their full potential.

6

I thought the initial compression layers were used to make sure the first few layers are solid against the build platform.
From what I have seen on my prints the very first few layers are squashed but then everything corrects itself there after. The new software no longer seems to have the tweak by mm to adjust the z but rather uses a %. So I’m a little lost on correcting. I will need to pad my models and some won’t be all that easy to part on a lathe and I may have to mill backs to size. Scaling the models sort of helped but the model was still squashed.

So what you are saying is true then I agree your way would be best.

7

My understanding was that these users have maxed out the z-offset and still have good adherence and excessive compression.
But anyway… Here’s the fix-all… The right way… The real fix… suspenseful music here… Wait, where’s my paycheck?!.. haha…
Instead of eating into the model to create adhesion, repeat the first layer to do it. So the first 0.6mm of each print is duplicates of layer number 1. Now they can still play with the existing z-offset, but now they can find dimensional accuracy without loss.

I say 0.6mm (theoretical) because you will need to do the following for consistency between resolutions…
If 0.1 resolution, then use layer one (6x).
If 0.05 resolution, then use layer one (12x).
If 0.025 resolution, then use layer one (24x).

8

@JoshK

You are correct this is the issue at hand. However your solution has several problems. First I’ll tell you a little about my Form1 printer history as it is all I have to draw real numeric data from. The first Form1 printer I received I had before z-offset was available, so all prints made on it can be said to have a z-offset of 0. The position and angle of the tray where such that prints on the left side of my printer printed with thick bases around the full 2mm in thickness and prints on the right side would not adhere to the platform. I was using clear resin and could even see when the print started that the platform didn’t contact the PDMS on the right side due to the tray tilting to the right. I tried multiple trays but got virtually the same results with each. I have seen evidence in the forums that other printers have gone out with the tray housing that low and that these days z-offset is often used to fix the problem. The second printer I got I noticed that my bases where much thinner especially the ones in the front. The bases in the front where so thin that they curled when curing. Also as a result of this high compression level some of my supports would not adhere to the base. Just after this I was told how to access the z-offset feature when it was still hidden in diagnostic mode, and told I was one of the first users to use it. Setting my z-offset to -1mm resulted in bases that where thicker. When printing the five butterfly test print from Formlabs my bases ranged in thickness from ~1.6mm in the front to about ~2.45mm in the back. Notice I got bases in the back over 2mm. The bases where still set to 2mm, my best guess is that the first layer just printed ~0.55mm thick and it still adhered thanks to over-cure and the broadness of the base, however this kind of extra thick base was not reliable. Also note that across the build area the difference in height was 0.85mm. Before I returned this second printer I measured the angle of the tray housing and it tilted about 0.4degrees toward the back of the printer which is what caused the difference in thickness. The last Form1 I got had even thinner bases and about the same 0.4degree angle of tilt toward the back causing the bases printed toward the back of the machine to be thicker. Even at -1mm z-offset the thickest bases were 1.9mm in thickness which was 0.55mm thinner than on my previous printer. Once again there is a lot of evidence in the forum that other printers with the tray housing that high have gone out and that z-offset has been used to fix the problem. And in closing I would like to state that with each of these printers I used multiple trays and found only small differences from tray to tray.

So the point of all of that history is that it documents the level of variability from printer to printer with some actual real world numbers. So with a little quick analysis since my prints still adhered on my second printer with a z-offset of -1mm and they didn’t on my first printer at a z-offset of 0mm it is safe to say that the lowest point of the tray on my first printer was over 1mm lower than the lowest point on my second printer. Also since the difference between the thickest bases on my second and third machines is 0.55mm when both are set to z-offset of 1mm we can assume the lowest point on the third printer is 0.55mm lower than the lowest point on the second printer. This gives a total of somewhere over 1.55mm difference in height of the lowest point of the trays between these printers. Then add the 0.85mm difference in height between the lowest and highest points on the trays at 0.4degree angles and you get a total of some distance greater than 2.4mm between the highest and lowest spots I have seen and measured on any Form1. Granted I have never owned a Form1 printer that printed reliably with the z-offset set to 0mm, they have all required adjusting the z-offset up or down just to be able to print reliably. I can not say how representative of the range of variability in printer configuration this selection is but I figure it touches on the extremes as I think most Form1 printers print reliably with the z-offset set to 0mm.

So now on to the problems with your solution. First 0.6mm is a lot of compression. It is way more than is needed for adherence. I have had bases on some prints print 0.45mm thicker than they were supposed to be because my z-offset was too high and they adhered fine. I know that -0.45mm of compression isn’t dependable but I assure you you don’t need anywhere near 0.6mm to get there. That much compression will result in bulging on the first layers and fringe from squished layers coming out from under the print. With this in mind you would probably want to make your raft a much more conservative thickness like 0.1mm to prevent artifacts. Unfortunately on some printers even at max z-offset more than 0.1mm is eaten up by compression so you would still have to expand the z-offset range. If you expand the z-offset to accommodate all printers in the wild given how widely they vary I don’t think the 0.1mm padding is going to be enough to prevent some users from setting their printers z-offset to a value that results in print failures. This is more than supported by the fact I have owned three printers that could not print reliably when the z-offset was left at 0mm. The last problem I see is that this raft feature would only be relevant for prints made directly to the platform so would need to be conditional on some setting or detail, which from my experience with Formlabs makes it far less likely they will like the idea and not say it’s too confusing.

All that aside I really like it and think having a raft feature in the software that just took the first layer and repeated it for some height if it was turned on would be an awesome feature. I would use it to move my prints up some, and if I had enough range in the z-offset setting I would use it to get more reliable adhesion.

9

@Sam_Jacoby

Please read my last response to JoshK as it addresses some of the problems I have with your proposed raft solution. It seems strange to me that in order to keep your software “easy to use” you suggest that I use another program to account for the lack of features in yours, and on top of that you specifically suggest I modify my model to compensate for where your software fails instead of making your software capable of printing my model correctly. Also despite the fact that adding a raft in a cad program isn’t necessarily all that difficult I’m sure if it where a feature of your software it would be much simpler. Lastly JoshK’s suggestion of just repeating the first layer to make a raft seems like a far more efficient and sane method for adding a raft for this purpose than modifying the model in some other piece of software.

To giva an analogy of why using another peice of software feels ridiculous. If I bought a 2d printer that only printed text and found out that because of variability in manufacture some of the printers would not print the first several lines of a document. Also that each printer that wouldn’t print them would always skip the same number of lines, and I asked the manufacturer to change their software to find a solution. Does it seem like it would make sense for them to say well we don’t want to over complicate our software, so it would be better if you used some other software to modify any document you want to print and add blank lines to the top to pad out the ones the printer will skip and then print that.

1 Like
10

@RocusHalbasch, that at is a monster of a reply :slight_smile: If your platform and tray is not parallel software can’t solve the root of your problem.
When the printers starts the platform goes below the “floor” of the tank by, let’s say -0.6mm. Any printing that happens during that time is just negative layers that don’t add thickness (unless there is excessive gap in that area of the tank). The problem is the software should not proceed to layer #2 of the model until it gets back to the tank floor. So currently the negative layers eat the model. I am saying repeat layer #1 during the negative layers. Trust me, using layer one repeatedly until you are back to zero will do the trick. @Sam_Jacoby, jump in anytime.

11

@KenCitron you are looking at the X/Y Scale settings. To change z-offset click on Help > Fine Tuning as you did before. When the dialog pops up look for the words Platform Height and click on them. The two boxes with percent signs should turn into a slider with millimeters. Slide the slider to the left to decrease the amount of compression and to the right to increase it. Please reply to let me know if that fixes your problem.

12

Ahh ok, didn’t notice you can change the settings from % to an actual value.

I just re-ran one of my models and padded the base an extra 1/16" so final measurement should have been 1/8" and it came out just shy of that at .108 but the top face was pretty accurate. I did get nice results, back had a little peeling but I will mount that on a mini mill and fly cut it to spec.

At least I’m getting the results I need even if it’s just a little more work. wonder if the compression layers are based on total number of layers of if it’s a fixed amount all the time?

13

@JoshK when you say “any printing that happens during that time is just negative layers that don’t add thickness”, you are not wholly correct. For each of these layers that prints a small amount of material is still cured. An interesting demonstration of this fact is that on a printer with the z-offset set so there is significant compression print an object with supports with both 100um layers and 25um layers. You will find the base on the 25um print is quite noticeably thicker. This is because each layer printed in those negative layers does leave behind a very thin layer, and the 25um print prints 4 times as many negative layers so accumulates more thickness during the negative portion of the print. If you where right that negative layers don’t add to the thickness both the 100um and 25um prints would have bases of the same thickness. The unfortunate thing about these negative layers having thickness is they can cause the artifacts I mentioned earlier.

As for software not fixing my platform and tray not being parallel I am well aware. However Formlabs has repeatedly told me they make no effort to have them be parallel and that it is not a problem for them to not be parallel, and the degree to which they aren’t varies from machine to machine, and they have no standards or tolerances governing how far from parallel is acceptable, and that this is corrected for in both the calibration and the software. When I asked for more details they told me they couldn’t detail their calibration process. If you look through the forums you’ll find others have had the same problem.

14

@KenCitron the amount of compression is fixed per printer different trays can have slightly different thicknesses which will also effect it some.

15

I’m not sure about that I printed the same model at the same location that was padded to .125" thickness.
50micron layer model came out to .108"
25 micron layer model came out to .139"

The detail was better on the 50 micron model that was actually undersized.
Very hard to control a finish product if I don’t know what the padding is.

17

@KenCitron yeah read my last response to JoshK for an explanation of why they are different thicknesses. And now I think I see what you meant. The amount of compression as in how far down into the PDMS the platform goes is a result of the calibration of the particular printer and slightly effected by tray thickness. The thickness of the print and the number of compressed layers is the result of both the amount of compression and the layer height.

18

Wondering how much of the slope is from the tray or the track they sit in? Simply squishing something that could be more precise seems like a lazy way out. I do like the idea of controlling the compressed layers and it would solve quite a bit and only be on the software side.

I have a couple of nice dial depth gauges and might take a reading across the surface to see if it’s the pdms is actually parallel to the tank. Still on my first tank so I have nothing to really compare it to.

I’ll polish the end of the gauge so it doesn’t damage the surface

Your idea will really be the easiest way out.

19

@KenCitron from my experience, which has included three separate Form1 printers, and five different tanks, the problem of the slope is in the printer itself. I have measured this using a Flureon DXL360S digital inclinomer. I specifically measured the angles between the track the tray sits in and the column that the platform moves up and down on. I have consistently found that the track is off by the same degree as my prints are, so at least from my experience it has always been the printer and not the tray.

20

@KenCitron also have you tried using the platform height setting instead of padding your model yet? Did it decrease the compression artifacts?

21

Prints just finished with automatic supports with the following settings with the lowest support height. The part that was at parallel to the build platform failed, the other that was at a slight angle did work and had the proper thicknesses. Basically I haven’t able to get a successful print on anything printed flat with supports.

At 10° angle I got a descent print but unfortunately the auto supports are all tied together making it impossible to remove them without breaking the part because the resin is just too brittle. Any further spacing between them makes this type of part have too many artifacts to be any good. Almost ready to give up.

I still don’t understand why if I pad a file the model is still squashed proportionally, I thought padding it brought it past the compression layers.

More I try to use the auto support system the more I hate the whole process. Getting old fast!