I only use CW to orient my model and to add support. It is then saved as an STL. That STL is then loaded into nanoDLP which actually slices the model drives the printer through a Raspberry Pi 3. The slicing and resin profiles are in nanoDLP on the Pi.

Ok, so CW is not the program used to print, therefore it's results are to be disregarded, but the NanoDLP results don't look right either.

Based on the settings shown, the first 4 layers will take 42.5s each, for a total of 170s.  Normal layers take 8s to print, then you have 1.5s wait before print, 0.01s wait after print and 1.9s wait after lift.  That adds up to 11.41s for each layer.

Now to add things up: 

Assuming you have the default raft that CW creates and nothing else, that's 1mm, and the model is 124.57mm.  At 100micron, that should be either 1255 or 1256 layers.

1251 layers x 11.41s = 14,273.91s + 170s (first 4 layers) = 14,443.91s = 240.73m = 4:00:45 hours

Then I noticed that for some reason NanoDLP thinks the model is only 418 layers.  That would be the number of layers if the model was being printed at 300microns, or if the model was shrunk to 1/3of its original size.

I could be wrong but the numbers don't add up, so I would say that Nano DLP print using your settings would be about 4 hours.

Actually based on the previous prints listed below Guan Yu, done with the same profile, it should take even longer than 4 hours, more like 6+ hours. 

Take a look at the three highlighted prints at the bottom, compared to the one at the top.  Based on the prints that were done before, the normal profile seems to average about 195 layers per hour.  So the full 1250+ layers should be 6:30 hours.  Definitely not 1:45h.

Here are profile shots from NanoDLP.

This last image is the time to print and as I said the slicing is done in the background so there is no slicing time. As you can see from the top plate in the list which is what this software calls a print that this model will take 1 hour and 46 minutes using my actual profile that I use with my current load of black Wanhao resin.

I'm having a hard time understanding why it does some of the things it does, and I fond a lot of things that don't make any sense.

Since this came preconfigured for the D7, I used the default settings for generating support.  The default "Bed of Nails", is pretty useless, as it only generates supports for the areas that face the bed.  If anything needs supports but is not overhanging the bed surface, it doesn't get supported.  Take a look at the shot below.  If the model is oriented right side up, the bottom gets supports and only the parts overhanging the bed. 

I can see why you rotated the model upside down, just so you can get more of it facing the bed.  Even so, there are still areas that are not supported.  Here are a couple that are immediately obvious.

I understand wanting to use the "Bed of Nails" profile since it very fast (10-15 seconds).  But why not use the "Adaptive" support profile?  It's not as fast (60-90 seconds), but produces much better support and it supports areas that the "Bed of Nails" doesn't.  You can also orient the model normally, which results in much more reasonable support structure.

Then there's the estimated build time.  Using the 100 micron slice profile, it always comes up in the 8 hour range, so I was wondering how you got the 4 hour build time, until I noticed you used a default slice profile and a default machine, rather than the Wanhao D7 profiles. 

When I sliced the model with the default profile, it came out to 2 hours build time, not only that but it was at 50 micron.  That simply didn't make any sense, so I took a look at the difference. 

There is a huge difference in exposure times, especially on the bottom layers, then there's another huge difference in the Lift & Sequence time.  That is a starting profile, and is probably to be used as a starting point for other printers.  I can't see it being usable at all on the Wanhao D7, given the huge discrepancies.

I mean exposure time for default is 1000ms, wheres the D7 profile is 7000ms.  The bottom layer exposure is 5000ms vs 60,000ms !!.  Then the actual lift sequence is 2000, vs 15,100.  That's why the default profile is 7-8 times faster than the actual D7 profile.

So while I can understand that you can tweak the profiles to some extent, and get better performance, I doubt that you can deviate too much without compromising the print.

carl_m1968 wrote:

Using Creation Workshop which is what comes with the D7 the slice time for .100 layers is 3 seconds and the build time is 4 hours and 45 minutes. I don't have enough resin to actually print it as I am doing a project for work. But the numbers reported have been very accurate so far.

Can you post a screenshot from Creation Workshop of what the sliced/supported model would look like?

Here ya go.

Using Creation Workshop which is what comes with the D7 the slice time for .100 layers is 3 seconds and the build time is 4 hours and 45 minutes. I don't have enough resin to actually print it as I am doing a project for work. But the numbers reported have been very accurate so far.

Can you post a screenshot from Creation Workshop of what the sliced/supported model would look like?

Out of curiosity, why do you think it's not a good model for comparisons?

too much support needed - but that is just my personal opinion. I don't typically do things like this, altho I know there are many that do.
most of the things I print need very little to no support, so very little to no post processing involved, lol

I can try slicing again with the updated model, just to see how much of a difference there is between yours and mine.

I have never used Cura myself, altho I understand they have made some pretty serious changes in the last couple of years.

Since no one else is stepping up to the plate with an FDM printer I will

I am going to preface this with the statement that I think this is a terrible model for comparisons between FDM & SLA type printers.

Out of curiosity, why do you think it's not a good model for comparisons?

Since no one else is stepping up to the plate with an FDM printer I will

I am going to preface this with the statement that I think this is a terrible model for comparisons between FDM & SLA type printers.

Onward...
I use RH and Slic3r v 1.2.9 for my SD4s. I have tweaked the time estimate setting to the point that my estimated print times are typically within a couple of minutes - no more than 5 mins off.

first thing of note: when I loaded the model, it came up with the "non-manifold" error and it is laid on its side (this is the "solid" version of the model, btw) after standing it upright...

the staff/sword seems to be the problem - i can slice it, but the staff/sword disappears. (and yes, there was an attempt at repairing with netfabb basic)

slicing time was 4m:26s - most likely due to the manifold issues and adding support structures.
Layer height was set to 0.254 (keep in mind this is for the SDs stock 5/16" threaded rod)
infill was set to 15% with supports turned on.

No, I will not actually be printing this.

Well, I didn't think the model had any errors, but I loaded it back into NetFabb, and it does indeed have a small hole in it.  It's not the staff, rather it's where his left boot intersects the robe.

But it's the type of error that doesn't affect the printability of a model.  So I opened it back into Repetier, and you can see the slice results below.

Using my SD3 profile I first sliced it with Cura @ 0.2mm.  Slicing time was 23 seconds, print time would be 3:57 hours.  I then sliced it again at 0.1mm.  Slicing time was 47seconds, print time 7:50 hours.

Next I sliced it using Slic3r, which I had stopped using a while ago, because of it's slow print times, and issues with some models.  The first slice produced similar results as yours, the staff missing altogether.  It turns out that Slic3r has a problem with models that are made up of multiple solids.  In this particular case, the staff is a separate object that simply passes through the hand and touches the base.  A lot of my models are composites, and I don't always Boolean them because the process does not always result in a nice clean intersection.

I've never had any issues with this before when using Cura, nor do I have any issues with PreForm.  But apparently Slic3r doesn't like it.  So for a final test, I did Boolean the 2 parts, so Slic3r can now slice the model correctly.

I also uploaded the repaired file and linked it in place of the original.

I am going to preface this with the statement that I think this is a terrible model for comparisons between FDM & SLA type printers.

Onward...
I use RH and Slic3r v 1.2.9 for my SD4s. I have tweaked the time estimate setting to the point that my estimated print times are typically within a couple of minutes - no more than 5 mins off.

first thing of note: when I loaded the model, it came up with the "non-manifold" error and it is laid on its side (this is the "solid" version of the model, btw) after standing it upright...

the staff/sword seems to be the problem - i can slice it, but the staff/sword disappears. (and yes, there was an attempt at repairing with netfabb basic)

slicing time was 4m:26s - most likely due to the manifold issues and adding support structures.
Layer height was set to 0.254 (keep in mind this is for the SDs stock 5/16" threaded rod)
infill was set to 15% with supports turned on.

No, I will not actually be printing this.

I have a D7 but out of resin. Once I get some resin I can run the test on it unless someone beats me to it.

I have the same.