So, what if there's a way to print TPU at the  same speed or faster than you already print PLA? 

In this video I will show  you how I managed to do this,  

and how I managed to actually  improve quality at the same time.

-- intro ---

Are there really TPU cheat codes? Well,  honestly, kind of - inasmuch as a set of  

rules that you can follow to dramatically  increase your quality and reliability  

when it comes to printing TPU on a bowden  style printer like the ender 3, 5, and so on. 

But they're not magic of course, this is about  understanding the intricacies of printing  

TPU and leveraging it to your advantage. If you haven't seen my first TPU video then  

add it to your viewing list and check it out after  this, a card should appear about...now for that. 

OK then, we've got a lot to cover  in this video so strap in, let's go.

pressure ========== 

flexible materials are flexible, the hint is in  the name honestly, but if you imagine trying to  

push an elastic band down a tube and expect to be  able to control how much comes out the other end  

then you get the idea of what  we're up against when printing TPU.

Of course a lot of the process is  engineered to make that work better,  

which is one of the slipperiest materials  known to man, and the width of the tube is  

practically the same as the diameter of  the filament giving it nowhere to hide.

If we have a look at this badly drawn  representation of a bowden setup you can  

see the general process, the point I want to make  here is that you have a setup that works almost  

like a hydraulic system because the filament  is liquid at the business end, and so if you  

think of it that way then you have a large cross  section at the extruder where you are applying the  

force - effectively 2.4 square millimetres, and  at the liquid end you have a cross sectional area  

on a 0.4mm nozzle of 0.13 square millimetres,  so from what I remember of science, that means  

around a 20 to 1 ratio between  the extruder and the nozzle.

Now, importantly, pressure is the enemy  of flexible filaments, or rather excess  

pressure is the enemy. Too little pressure and  it will just compress inside the tube, but too  

much pressure build up while printing - which is  the more common problem in a bowden setup - will  

eventually cause a failure, as in the diagram  the first weak link in the chain is the extruder,  

so you get filament popping out, looping  round the gears, or being ground up.  

It's actually a sign of a problem, not a  problem itself. If your filament is doing this,  

it's kind of an emergency pressure valve, a  sign that you should manage the pressure better.

So - Well, one thing we can do to lower the  pressure that we're going to look at today  

is that nozzle cross sectional area. So  yes, in a way this was a really long way  

of asking if a 0.6mm nozzle will enable us  to print TPU faster and with less errors.  

I just think it's important to explain  why, as well as pose the question.  

mini delta. ===========

I want to introduce you to the monoprice  mini delta. If you are subscribed to dr  

vax's channel...he's changed his name  but that's what I remember him as,  

then you will already have seen this printer.  In fact they've released a new version. 

The monoprice mini delta is to printers what  a hammer is to precision machine screws.  

With absolutely no bed levelling and a bowden  extruder that would work better on TPU if you  

had the tube sticking out the side where  it invariably spits it, these machines are  

ideal to test our theory out. Don't worry, we'll  get to the Ender printers in a bit, humour me.

I'm fitting one of my deltas with a  0.6mm nozzle and the other with a 0.8mm.  

I'm already well aware that neither of these  printers can print satisfactorily on 0.4mm so  

there's little point having a control. The control  is literally spitting filament out the side.

Let's see how these printers  cope with the larger nozzles.

Filament types. --------------

Before we get the results I want to discuss TPU  filament. If you've bought multiple brands then  

you probably are aware that there is a massive  range of what's sold and described as 95A. 95A is  

a hardness rating that's ascertained by poking  a sort of thing into a block of the material  

and figuring out the displacement, so either it's  not a good measurement of how filament is, or  

more likely most of these filaments have never  been hardness tested or rated at all. I couldn't  

tell you what true 95A feels like but it might be  one of these, I guess. I don't know if it comes  

across on screen how much stiffer some of these  are than others. And I want to point out that  

Noulei is stiffer than the rest by quite a margin.  Noulei filament is easily the most bowden friendly  

TPU I have come across, which isn't necessarily an  endorsement of it as a material, but more that it  

is the least likely to fail under pressure, and  it can typically print faster than the others,  

basically due to it being stiffer. Of course,  it's stiffer printed too, that's the downside.

Results --------

On the noulei blue TPU and the 0.6 nozzle  the results were night and day better on the  

mini delta. I was printing at PLA speeds quite  happily, I even took the time to make some feet  

for the printer since I've been balancing  it on random objects ever since its original  

sticky blob things fell off. This might make  it less noisy too, silent drivers this aint! 

This is why I recommend the noulei filament, being  able to print at speeds like 30mm/s on a 0.6mm  

nozzle with decent results - for this printer -  makes printing TPU a breeze, this is undoubtedly a  

game changer. And while my fast draft settings may  look a bit meh in places on pikachu here, this is  

due to the settings, if you used more conservative  settings you would get better results.

So this looks really promising for 0.6mm nozzles.

Unfortunately not so much for 0.8mm nozzles, and  the experiment did take a dark turn from here. 

First I tried the blue filament on the 0.8  nozzle machine, and to my surprise the results  

did not improve, in fact on the contrary. To be  fair, the pressure situation was not a problem,  

the extruder never failed even at high print  speed, but the output was...unsatisfactory.

I tried different filaments and the  results just weren't getting any better.  

This finish, if you wanted it intentionally  it would be fun, but I don't. What happened?

Well. This is where the experiment diverges.  I couldn't understand what was happening so  

went back to the 0.6 printer and tried the  other filaments. I was sure to print slower,  

around 10mm/s or under, to be on the safe side. Confusingly none of the other filaments would  

perform nearly as well as  the one reel of blue noulei. 

And a temperature tower of the  green noulei was...well, horrible.

It took me a long time to work out what's going  on here because it's not what you'd expect.  

Believe it or not, this is not a problem with  the machine or the nozzle size, it's a problem  

with established facts. We'll come back to this,  because the next thing I did, was assume that if  

my green noulei was printing like this, and  making popping sounds when it printed too,  

then it must be wet. Established internet  knowledge says that if your filament pops  

and looks like this then it's wet. Never mind that  I'd only just opened it, noulei must be at fault,  

supplying wet filament, how dare they?  Into the dryer with you for 12 hours.

[12 hours later] And it didn't work at all.  

Because it wasn't wet. It's never moisture.  It's kind of a running joke I have in my head  

with filament. It's never moisture.

Next I put the same filament  onto the ender 3 v2 and chose  

my best TPU setting and...well...this came out.

I'll cut a long story short by just explaining  

without the series of boat  bits and head scratching. 

It turns out that TPU has a speed limit. No, not an upper speed limit. Of course  

it has one of those, too much speed is too  much pressure is chewed up in the extruder. 

No, it has a LOWER speed limit. If you print TPU  at very low speeds, well, let me demonstrate.

Do you see what it's doing under  the macro lens? Exactly. It's kind  

of....dripping. The pressure,  believe it or not, is too low! 

Of course this is going to be  worse with larger nozzle sizes.

I told you it was not what you expected.

I never did manage to find  a balance between pressure,  

speed and print quality for a 0.8mm nozzle  so I can't recommend it, but I did for 0.6mm.  

So let's talk time. But first lets talk math..s

geometry and why 0.6 is bigger than you think ============================================

0.4 to 0.6mm is a difference of what, 50%? Right? Wrong. 

Did you ever see 10 inch pizzas sold  for £6.99, and 12 inch for £8.99? 

Well, thanks to...basic maths, you should  always go for the 12 inch and here's why.

The cross sectional area of a cylinder,  which a pizza is..a really short cylinder,  

is pi r squared obviously. We all  know that. Well, we do now anyway.

So a 10 inch pizza has 5 squared times pi  (remember radius is half of diameter...they dont  

sell pizzas by the radius annoyingly). 78.5 square  inches of pizza. Sounds like a lot doesn't it.

And a 12 inch pizza has 6 squared times  pi which is 113 square inches which is  

44% more! Nearly half again  for £2 more? Yes please.

Anyway there is a point to this,  

the same applies to nozzles. The conversion  from 0.4 to 0.6 increases the cross sectional  

area from 0.126mm^2 to 0.283mm^2 which  is...yeah that's more than double isn't it.

And this is why the pressure is lowered  so much when you switch up to 0.6mm, and  

this is why the real benefit here, is the  time you save. So yes - let's talk time.

let's talk time ===============

Now yes, I know some of these results are  scruffy, the prep for this video took a lot  

of twists and turns and I didn't want to repeat  timing experiments and waste material, and in  

some cases just whatever, draft mode. Focus on  the time, the rest will fall into place. Honest.

So, A 0.4mm 3 wall benchy  on my standard settings with  

0.2 layer height takes around  three and a half hours. 

Whereas the very same settings on a 0.6 nozzle but  with a 0.4 layer height takes only 1 hour 8 mins. 

And on a 0.8mm nozzle, with  0.6 layer height, 43 minutes.

If you go more conservative  with the layer heights,  

you could say, 0.4noz / 0.2h = 3h22m 

0.6noz / 0.3h = 1h31m 0.8noz / 0.4h = 1h03m

So you can see that when you look at time savings,  you're printing the same model in less than half  

the time even with more conservative settings.  There are massive time savings to be made.

The BMG extruder ===============

I had this thing kicking around and in a moment of  anger I decided to fit is to one of the mini delta  

machines. Surprisingly it actually fit, though  configuration was more involved than I'd have  

liked, for one it ran backwards and also it's  geared around 3:1 compared to the old one,  

both of which can be fixed in config, but I also  had weird buzzing of the stepper which I solved  

by just retracting slower, I suspect this is due  to the 3 times larger step size but this video  

is long enough without looking into that. I will  probably do a vid just on this extruder maybe, but  

it is way more resilient when it comes to soft  TPU, and I was able to print even the softest  

on this printer which I hadn't before. I even swapped back to a 0.4 nozzle and  

the extruder still handled it really well, so it  definitely is better than stock for TPU at least,  

if you can overcome the technical aspects. Oh and also when you're putting it together,  

the gear goes on like this - not the  other way round. I made that mistake  

and it chewed up the plastic gear a bit, no  massive harm done but a heads up I guess.

The Ender 3 ===========

Having learned a lot on the mini delta end of  the experiment, I then took this and applied  

it to my two ender 3 printers. Essentially  there's no difference between the 3 and the  

3v2 from the perspective of TPU printing,  but I did also fit a 0.6mm nozzle on my v2.

Results were exactly as expected. The  ender printers take a lot of flak for  

being low end but they really do perform.  I will let the output speak for itself.

And by the end of the prep for the video, this  TPU had arrived, so I had to try to print it  

out with a 0.6 nozzle. It looks fantastic  considering it printed in literally one hour.

It's controversial but in my profiles for  these prints, I have the outer wall speed  

FASTER than the infill and in  some cases faster than inner wall.  

My theory is that I want to be  able to get shiny outer walls,  

while allowing pressure to level out a bit  during the infill, which can print a bit slower.  

It seems to work, so for the time being this is  what I recommend doing. This seems like a cheat  

code to me, it's the exact opposite of defaults.  As you can see it certainly seems to work.

settings used for TPU ====================

I'm going to show you my settings now,  pause and copy them, but also I'm sharing  

them on the website, link below in the  description. These settings work on the  

ender 3 and v2 on any bed texture because TPU  is very tolerant to just about any surface.  

The settings also give me good results with  ALL my filaments and I've tested it out on  

both ender 3 models with glass and textured bed. The settings work well out of the box for all  

three colours of the noulei filaments that I tried  in the video, and also for the blue ziro filament.  

If you find yourself looking for these filaments  after watching this video then links are in the  

description for each, and you can apparently  support the channel if you buy through them,  

although I just set this up  so who knows if it will work.

And of course if you want to  perfect your prints even more  

by fine tuning the retract settings and  temperature, go check out my other TPU video,  

link in the description or the card I  already showed you at the beginning.

Tidy up and conclusions ======================= 

Don't switch off yet, this is the most  important bit, I want to tie everything  

up into some fairly neat advice based  on everything I found during the video.

first layer slow irregardless ============================ 

you're "squishing" which adds extra  pressure, newton's second law,  

the bed is pushing back at you, so  don't push your luck on layer one.  

It typically doesn't matter if layer one looks  awful so it's absolutely a good idea to put your  

first layer down at 10mm/s or less. I tend to do  a 50% thicker first layer at half normal speed,  

this doesn't make it thicker because cura  will compensate by lifting the head further,  

but what it does is gives you more filament  to work with so less chance of poor adhesion.

don't obsess over strings ======================== 

I'm going to show you a trick that most  of us know already but I think few people  

actually do in reality because it requires you  to have a heat source. This is a butane torch,  

be careful with these, they are pretty gnarly,  but I'm using this as opposed to a lighter because  

quicker, directional and no soot. Just turn it on  and wave it over the piece and strings disappear,  

and ok, bigger ones turn into warts. But yeah  that's a cool trick for dealing with those  

messy hairy bits that you always seem to find in  your food an hour later, never understood that.

Summary ================= 

OK so this is important, to kind of find some  semblance of advice to give you out of all this.

Firstly a 0.6mm nozzle is actually a really good  way to print TPU on a bowden, you get faster and  

more reliable prints, it's a game changer. You  of course sacrifice some resolution but remember  

you can still print lower line heights so  really the loss of resolution is largely  

restricted to the tops and bottoms of things. So would I recommend a 0.6mm nozzle for TPU?  

Yes, especially if you either print mostly  TPU or you have a dedicated machine for it.

Secondly, don't waste your time with  0.8 or bigger. At no point did I get  

satisfactory results. You might, but I  guarantee it's not worth the sacrifice.

Thirdly we learned that for good results when  printing TPU you need the outermost wall to  

be printed at between around 15 and 30mm/s,  this is what makes it smooth and shiny. Lower  

than this will give poor or inconsistent  results, depending on the brand I guess.

Fourthly and finally you can increase the  range of filaments you are able to print with  

if you upgrade to a BMG style extruder, although  I, perhaps irrationally, don't think I'd want to  

do this on my main machine. I am happy to have one  machine with this extruder on, but I just prefer  

the stock extruder on the ender, and calibrating  e-steps was a hassle. While it is a game changer  

of sorts, I would only recommend this on a  dedicated TPU machine, but it's up to you.

So this is the end of another video that proves  that you absolutely don't need a direct drive  

printer to print TPU. And with that, I'm off  to check out ender 3 direct drive upgrades.

For science.

As usual let me know what  you think in the comments.  

I'll tell you what I think - I think this  has given me a headache. But, what a result.

Thank you for watching!