Slicers and Print Speeds

I am looking a bit at print rates (don't we all!) and I have the simple object in the attachment in hand here. Using Gcode from Fusion 360 it takes 4 hours with medium quality--I am not very experienced with 3D printers and don't have high expectations for speed but I thought this was a bit much. So--I took the STL and sliced it with Cura and it now sits at about 9 hours!I have the walls and infills set up for what I consider reasonable values for this item (it is a sliding dovetail and must be reasonably smooth running and strong.) Changing infill did not have much effect. My printer head speeds are in the 50s in keeping with dogma for my printer (Creality).I was just curious if this is all expectable (the 4 hour print from Fusion) or do others find this a bit long?ThanksFritz

Martin--
In Fusion----If I match your parameters exactly I get 3:24.
In Cura - I get 3:35
(sorry for the confusion on Cura but I had infill at 100%---I am not that familiar with it)

So--you have verified that I (or the software) is not doing something crazy. We now have all 3 in agreement. My edging towards 50% infill was perhaps a bit overkill.

I know these things are slow but for this part I thought it would be a little faster--and then I am not certain how the narrow edges will hold up--nor how smooth it will operate.3D parts are pretty crude.
Probably this will go to my milling machine and some delrin ---only takes a half hour including finding the stock and grabbing a beer!.

I appreciate your time--it was pretty helpful as I am new to the 3D stuff.
Fritz

seems about right.

You may well have unrealistic expectations of print speeds :-)

What you could do is to increase the layer height to 0.3mm
cuts the print time by 1/3
And ten minutes with a file gets it as smooth as a 0.2mm print.

Depends on whether you value less post processing over more print time - or not :-)

IF I printed this at my normal settings for non-critical parts. And kept the 0.2mm layer height and upped the infill to 25%.
so:
3 top and bottom layers and perimeters. 0.4mm nozzle and I can do it in 2 hours.
But that's on the sapphire pro 2 at 150mm/s.
And most machines just can't hang with saffy for pure speed :-)

What printer do you have ?
It makes a LOT of difference.

It is a decent sized part, anything under 4 hours would not seem unreasonable.
I'd have been happy with anything under 4 hours for that, before I got saffy.

It would take a lot longer on the delta as I'd need to stand the part on it's end.

Actually it doesn't.
2 hours 38 minutes.
That's standing on it's end, 125mm/s print speed - but the 0.5mm nozzle does make a LOT of difference.

And to be fair the part will be a lot cleaner printed on end anyway.
Although if you have an i3 setup - probably not so much.
With delta, cartesian or corexy tall vertical prints can be printed on faster speeds than with an i3.

So, yep without using larger nozzle or - probably - a different design machine, 3-4 hours is pretty good.

I will add, that I am impatient in this respect and I value functionality over shinyness.
So as long as it's structurally sound, I don't mind a little sanding if it saves me several hours print time.

Although 150mm/s at 0.2mm on saffy is pretty bloody good.

Okay so for comparison I sliced it on the knp - a replicator pro clone.
60mms speed (she will do 75, but not happily) 35% infill, 3 perimeters and layers and 0.4mm nozzle at 0.2mm layer height.
Around 3 hours 50 minutes.

The other thing that really effects printing time - particularly with smaller models.
Is the speed you print the first layer at.

If you've got a really sticky bed - you can up the speed of the first layer. On smaller prints the first layer can be 50% of the total print time.

I think the fastest I've put down a - succesful - first later is oribabkly around 40mm/s.
Ordinarily i print first laters at anywhere from 5mm/s to around 20.
Just depends on the material, printer and print surface.

But optimising your adhesion and z-gap to enable a faster first layer - can really put a dent in average print times.

Thanks Curious.
I have a Creality Ender 5. Looks at a glance to be the same style as the Sapphire but we don't now about the print head.
Correct me if I am wrong but I would think that has at least something to do with it-
And of course the nozzle. I am using a simple 0.4 in brass. So-I am not really sure why some XY printers are faster than others--I will certainly play with mine at speeds over 50 some time if I get a chance.


But you are confirming with Martin that my times are reasonable--since I am an beginner at owning my own machine I had no idea what to expect.
We had some heavy duty ones at one of my work places and they would spend half a work-day printing 12x12 pallets of small parts..
In my case once I prototype (mostly optical stuff) I either machine it, send it to a machine shop, or get them professionally 3D printed if there are a lot and finish is not premium.
Something I believe is still prohibitive for small runs (100 items) is extrusion. But I haven't checked recently. When I worked for GE plactics we had extruders in our test labs and it was a major operation to set up for small test runs.


Your comments regarding "on end" orientation interests me. I played a bit with this early on but decided to let Fusion pick the "best" orientation. I will experiment with this.
What is "i3 setup"? New to me. I am aware of polar printers but don't want to go there.


Usually I opt for workable crude parts as you mention (I have some nozzles up to 1mm but the part here for example has to "slide" so finish matters
Once I have to go to sanding I would usually just machine the part on a mill. I have some parabolic reflectors that I need to make soon and the only easy way to get the correct equations in is via a plastics printer--I will be hand sanding them and them and sending them to an aluminum sputtering service for the mirror finish. 3D print shines here-aspheric optical structures can't be easily hand machines and our present lab has no CNCs.


Thanks again for the times--they comply in the ballpark with what I am seeing so I feel like I am not doing anything wrong.
It does amaze me that there are no intermediate books out there describing the more advanced issues of 3d printing-most of us just guess and go.


Thanks for the tip on 1st layers--another thing to investigate.
My Creality wanted to go with 200 deg/50 deg--and it did not stick. So I used the fusion defaults for my machine--220deg/60deg and everything sticks just fine.

Cheers
Frit

Hi FritzThe person who called additive manufacture rapid prototyping was obviously not talking about print speeds. Having said that in the last year 3d printing in house even with our crude machine has saved us an enormous amount of design time and thousands of pounds.My company develops and manufactures Aerosol Silver Machines for the AstraChrome process and a couple of years ago there was a period when there was a craze for replica sports trophies (Fifa Word Cup tha sort of thing) made by 3D printing (FDM and SLS). even after sanding Silver plating them did not give a really very good result because of the build lines which took a great deal of work to absolutely sand out. As a result we had to develop a special ultra high build primer for the job.Consisting of a 2 part 100% solids polyester resin and an exotic mixture of solvents it can be sprayed on up to more than 1mm DFT but is filled with micro balloons for easy sanding. If you would like to try some give me PM and I will see whether I can get a sample to you - depends where you are because of transport.PS I like the Base clef logo are you also a musicians !

Quote:

---

So-I am not really sure why some XY printers are faster than others-

---

why is a ferrari faster than a cheap station wagon ?
Design and build quality.

okay lets compare and contrast:

Sapphire pro2:
Quality linear rails - can't really overestimate how much difference this makes.
Corexy setup - x and y stepper motors are stationary - so lighter gantry = faster speeds - as both motors are required to be under tension at all times - you get very little vibration and none of the stop/start movement that characterises most 3d printers (deltas work differently, but benefit from having all three motors always engaged in head movement).
Frame build: Solid 7mm aluminium top plate machines from a single piece of metal. 12mm support rods for the print bed, 2.25mm steel sheet frame, for the base.
Direct drive extruder. Okay i did that - but all it took was a 4 gm self printed adaptor.

Ender 5 (NOT corexy):
Wheels on aluminium extrusion, cheapest and least precise movement available.
frame: made from bolted together aluminium extrusion
Standard cartesian movement - the x axis motor is attached to one side of the gantry for extra and off-set weight.
10mm printbed support rods and significantly longer than the sapphire's,
Bowden extruder, adds extra play into the filament and should not be necessary.

You get what you buy - and if you buy creality you get the cheapest designed machine they can build. But boy do they spend a lot on marketing to make people think they are buying something much better than the reality !

Don't get me wrong - the ender 5 is a functional 3d printer. Just made on the cheap and designed by people for whom print speed is not a consideration.

There are no doubt a number of other - smaller - but no less significant differences.
I've seen a LOT of very different 3d printers over the last 10 years and, honestly, the sapphire pro 2 is one of the best designed and made machines I've come across.

My all time favourite is still the leapfrog bolt - but that costs approx 21x more than the sapphire pro 2.
At it's price point the sapphire pro 2 is probably the best designed and built 3d printer on the market.
I have no idea how they are making money on it.

I own a ferrari you own a station wagon :-)

(drove past a ferrari dealership today - hence the repeated references, I'm not a gearhead)

So what can you do to speed things up ?
Well first - convert it to direct drive.
As the x gantry doesn't move vertically - it can take the extruxders weight with no problems.
Maybe add cross members to the sides of the frame, to make it more rigid.

The things that AREN'T slowing the ender 5 down.
The electronics, the extruder & the hotend.
You know, all the things most people seem to think need changing.

Nope, the key to a really good 3d printer is the build and stability of the frame, and the way the extruder is moved around the frame.
The rest is 'almost' window dressing.

Most current printers come with good extruders and hot ends and 32 bit motherboards with silent stepper drivers. Those are cheap (why most machines come with them) and changing them will have minimnal effect on overall print speed.

Likewise, brass nozzles are fine - all I currently use. I change the deltas 0.5mm nozzle about once every couple of years.
Saffy's is still original.

I'm anti creality simply because all their machines could be massively improved for less than $50. But they shave every last cent of and in the process make the worst mechanically designed machines currently available.
It really doesn't help that lots of other manufacturers see how much money creality are making and then decide to use the same cost saving designs.

But honestly, it's pretty much a one man crusade and I really can't be arsed to keep at it. Let people watch a sponsored youtuber lie to them for money and go buy the 'recommended' machine because 'bob the basement builder' said it was the best machine on the planet and only three dollars and 25 cents !

lol okay so it's not entirely out of my system yet :-)

Sorry for the late note Gambo- we had a problem yesterday and I spent the day working on a circuit board--turned out to be "user error" of course!

Yes-I guess the word "rapid" is in the servo motor or the beholder. At any rate I agree that I can now get a lot of work done quickly for prototyping which is the most of what I do (I work in research where almost every instrument that we build is a "1-off"). I also have a Sainsmart desktop milling/routing machine and I can make prototype -fluidic cells in polycarbonate that would cost $500 and take a month delivery. So I am impressed with what can be done (as I learn) and it saves me a lot of work.

Regarding our mirroring operations-we us a local firm here in the US: https://muellercorp.com/
I definitely see lines in the results that we get so I am guessing better mirroring could be done. (Also better sanding). I have not been doing the sanding so I want to look more into this. But if it is problematic I would look at other mirror options with a thicker build as you mention.

One big problem is that we only sell a dozen of our units/ year and with volumes like that it is not reasonable for even small companies to make things for us.

In this particular lab I am working mainly on the optics, electronics, and spectroscopy for a "Plankton Microscope". The system has a "darkfield" imaging ring at one end--this consists of 2 mirrored plastic rings surrounding a ring of several LEDs--these present a parabolic profile and focus the light into the required ring at the microscope. I cannot afford ray-trace software for this so I am re-designing the work of another using a slide-rule and some photometers in our lab to do actual measurements. I am also considering forgoing the mirrors and using some newer "direct LED" illumination but this will require a lot of tests on the optic-bench. Making small parts for the latter is where the 3D printers shine. I can make dovetailed glides, holders, lens mounts and a whole bunch of things that are not easily available.
So-I will see how things progress but if we stay with mirrors I would certainly be interested in what you do. Sounds like you do interesting work.

My logo is the Viola Clef (Alto Clef). You are the 2nd person in 10 years to realize this! I hardly qualify as a musician (I am a scientist) but I always had a great deal of love for classical music. I play violin, viola and harpsichord. (I built a few harpsichords). When I am not in the lab I like to drink beer and play string quartets--the more I drink the better I sound! My wife is a violinist and we founded a small orchestra here in Falmouth MA, USA. About once a year we get and inquiry about playing with us--from Falmouth UK!

Cheers and many thanks!
Fritz

Thanks Curious.
Mechanically your machine is superior-but regarding speed--I would think there is an upper limit for head travel set by the plastic, its temperature and the nozzle. E.g. if you have any printer draw a straight line at a certain speed it will, for a certain depth, begin "skipping". After that patterning is key and also you mentioned a lot of points in your prior post. I guess the CoreXY adds a lot.

I bought the Creality based upon a friends suggestion-he runs a lot of CNC in his shop but uses the printer for ancillary things. Neither of us needs much speed (my question was more in the way of finding if I was doing things right). I went with the Ender-5 because it has a sturdier frame. I do like that all the parts are quite generic and I can make my own repairs once the company fades away. I must say the customer support is abysmal. I think there are only 2 vendors--and they seem to have language issues-I deal with the whole world in my field and always deal with language-no problem-but Creality is really a low performer for service. Trying to buy a few parts I finally ended up looking at other vendors web pages.
At any rate I am fine with the machine for my needs--but looking up the Sapphire II pro--looks to actually cost a little less that then Creality. So I will keep that in mind -one of the labs I work for will be buying a machine soon. One place I worked had 2 large "refrigerator" size machines-I think in the $100K range. Not sure what they got for this price but they ran 7x24 on production parts. Did not seem exceptionally fast to me--I think they were printing ABS.

Ferrari? Ha- Don't laugh at old Ford cars. Back in the days of Big Daddy Garlitz, I had a friend in North Carolina that drove a cheap 1963 Falcon some years ago (same as me). When he needed a few beer bucks he would sneak off to the usual "secret" drag strip and bet a race. He beat more than a few hopped up Plymouths and I think a Vette one time. After the race he would proudly open up his hood and reveal a chevy 454 engine with a Holley 4 on top---all sandwiched between two cut out wheel wells. Its all skin deep they say!

Fritz

Pretty much all the car stuff, totally lost on me :-)
I literally only used a car analogy because this morning we drove past a ferrari on the road (going 'garumberrhum') and a ferrari dealership a few minutes later. And it's rare so it was in my mind.

The big machines sound like stratasys setups.
They are expensive - I mean really expensive. But while they don't live up to their price point - they are good.

They also use locked in propriatary filament cartridges.

I've never seen one in bits - but they do have a good reputation for reliability.

As for speed - I've seen machines printing at 300mm/s with a 600mm/s travel speed.
A few years back it was 'the thing' you did' if you were a manufacturer.

Then creality appeared and started making silly money and everybody went: speed ? what's your hurry buddy ? buy this cheap printer instead and slow down and chill out.
And they splashed so much money around youtube, that everybody just forgot how printers should be made and what they were really capable of.

So far I have found on all my machines that the speed limits are purely mechanical. That's with decent pla.
many filament types really won't print that fast. I tend to max out flexible tpu at 30mm/s. That gives me clean retractions, tricky on a flexible filament.

But with pla - it'll print as fast as the machine can move.

I have a very cheap delta with absolutely no frills. At 205c it'll happily pump pla out at 130mm/s and 0.3mm layer for pretty sharp and clean prints,
I mean if you want really dirty - at 215c it'll got up to 150mm/s at 0.4 layer height and uses a 0.5mm nozzle.
The quality is utter shite - but it's the mechanics of the printer that are the main issue - not how fast I can pump out pla.
There is a real limit on how fast you can poush filament down a bowden tube.
It's one reason direct drive is always preferable to bowden.

I could probably beef the cooling up and print at a higher temp, but I can't be bothered.
And there are a bunch of other reasons she won;t go any faster.

With the sapphire - my current speed limits are simplify3d. I simply can;t get it to go past 200mm/s.
I'm pretty sure the printer will hit the advertised 300mm/s actual print speed.
But simplify3d just refuses to go that high.

But yeah recidivist print speeds are definitely something I CAN lay squarely at crealitys feet.

Thanks--all good stuff. I bought the Creality on my own nickel as a trial basically-and at $350 bucks it has already paid for itself in making optical bench parts that can cost $1000 of more, But I am sure it is over-advertised and maybe a bit obsolete. As I mentioned that parts are all "generic" so it can be maintained forever--sort of like a model A ford. (;->
. But if we decide that 3D has use in the lab (and it does) we will probably opt for a better machine when the time comes. Your Sapphire shines here. And now that I have an inkling about what I am doing, I can pay more attention to the items that you mention.

Software is another issue--I work 2 contracts right now and am split between Solidworks, Inventor and Fusion. I only know option 3 and have ME"s do the designs on the other 2. Software breeding is a real problem in the engineering field-I have the same problem with electrical design. Someone landed an Altium circuit on my desk the other day and at $7000/seat I told them to redraft it. Besides I haven't used it in 10 years.
I never can be sure what software methods are really best--I think a lot of skill goes into that and I admire that vendors like Fusion even write it. I cobbled a crude python script to break out the Gcode, and comparing fusion to Cura for example reveals a lot of differences. Both make sense of course --so which is best! I certainly am not the guy to know. Simplfy3d is priced right and I would guess it is tuned pretty well since that is their main focus. All fun stuff of course-this took a whole new life since 50 years ago!

Thanks
Fritz

hi Fritz I thought there might be a music connection there some where - wasn.t sure about the name of the cleft you use as a logo I am a rather poor piano player as an English comedian remarked " all the right notes but not necessarily in the right order") and our music tends to have only the treble and Base cleft.It sounds like we have had very similar careers but on opposite sides of the Atlantic, but I did spend some time doing research in Boston and living in Concord MA.On the question of Aspherical mirrors a few years ago we did quite a bit of work on metalizing mirrors for automotive prototypes which had been produced by 3d printing. The main problem we encountered was build lines including slight layer shifts and surface texture. To cut a long story short we found that sanding and painting then sanding some more was very laborious. In the end with the aid of a polyester resin manufacturer we developed a special spray coating with a very small "shrink back" and ease of sanding. Though the quality of 3D printing has greatly improved we still use this process.Taking a automobile head lamp as an example the process consists of using conventional spray painting techniques to "bury" the surface texture the applying, again spray painting, an extremely thin film of an epoxy resin on the surface of the mirror.Once the epoxy is fully cured using a water based process similar to the traditional way of silvering say bathroom mirrors we deposit a reflective silver film which we then passivate with Tin to reduce tarnishing.

Recommended speeds range from 1800-6000mm per min or 30-100mm per sec and the most efficient speed will depend on what type of filament you're using. If you retraction is too low, you may find filament slowly coming out of the nozzle before your printer head finished moving.

Yes--I am finally getting that. I had to print some optical "easy" parts last weekend so I went all out at speeds up to about 80mm/s. Everything worked out fine so I will gradually adjust my speeds up in due time.
I was seeing "drip" from the nozzle as it ran between areas so I see what you mean there.

Other interesting thing is that I see a little "sag" -- holes for example out of round, flatter in the vertical direction----at higher speeds. I presume the extruder returns to the next pass while the last layer is still soft--but then I am bringing down my temperatures a bit so it may cure that.

All told I am realizing that there is a lot of adjustments that can be made-I expect a lot of this is worked out already but for lack of decent books on the topic I have to "re-invent" the wheel so to speak.

Thanks for your comment
Fritz

Yes holes in flat horizontal surfaces are usually OK but holes in vertical surfaces often distort if they are more than a few mm in dia. when designing a part for printing it is a good idea to work out the orientation you will be using for printing and using that a guide for placement of holes. 3D printing doesn't do overhangs well and so sometimes it is necessary to print a piece in more than one part with different orientations -just remember at the design stage to build in registers for joins between each part.
Almost all the polymers can be securely stuck together with one or the other specialist "superglues".

Usually I am not doing parts that are overly precise. I find that a 1cm test cube gives a size of about 10.03 to 10.08 cm on a side, the "shorter" dimension being in the vertical. I do think I am seeing some type of relationship between speed, temperature and "sag" however--so I expect if I can get that under control it will make things closers. I guess the one thing that is tricky about speed--is that it can be "too slow". That apparently gets into all types of dogma regarding cooling fan speed etc-. If I was running 100 parts maybe worth the time, but for one offs I am trying to get the best compromise.Thanksfritz

Yep - In find top and bottom shells to be problematic to begin with. Unlike the side shells--which are smooth within the realm of the layer height--the cross-hatched top and bottom always look rough. After 3 slicers I guess this is just a fact of life although I am not sure why--I would think a set of parallel lines of plastic spaced by the nozzle width would look nicer. I don't know of anything in the GCODE generators that controls this (or a lot of other things).Another issue I am having with speed is "sag". I find my bolt holes to be obligated --Maybe piling on on layer before the previous one has cooled---or fan speed---etcFritz