Prusa Mini is a flawed design – urgent action from Prusa required.
 

Prusa Mini is a flawed design – urgent action from Prusa required.  

Page 1 / 4
  RSS
graham.l2
(@graham-l2)
Eminent Member

Having received my new Prusa Mini a few days ago I have experienced some of the issues others have reported  I have a Mk3i which has been a useful benchmark and have tried to analyse the common issues in detail which are…

  • Z offset stability
  • Nozzle blocking

It appears that the two are related.

With the textured sheet on the Mk3i best results are obtained when the first layer is well squished into the sheet picking up the texture.   However if the same Z offset to achieve this finish is applied to the Mini two design deficiencies show themselves.

The first is the variation of the actual Z offset and appears to vary with temperature as others have pointed out and almost certainly due to the thermal stability of the MINDA probe, which is just not as good as the compensated PINDA used in the Mk3. 

When the first layer is squished more than desired due to these Z fluctuations the second issue kicks in and that is the force available from the extruder which is a single driven traction wheel as opposed to the twin arrangement on the Mk3.   This will offer half the available thrust of a Mk3 on the filament and when the first layer Z is to small the single wheel readily strips the filament loosing grip and appearing as a jam.     The third party upgrade for the extruder will help but it is not the whole solution.  ( https://www.3djake.uk/bondtech/prusa-mini-extruder-upgrade-kit)

In writers opinion this reduced performance of the extruder over the Mk3 MAY be ok IF the Z level stability is good which it is not, and no amount of firmware tweaking will fully overcome this. 

Prusa no doubt are aware of the above and I ask them as a loyal customer to bite the bullet and upgrade the MINDA sensor asap as hand tuning the Z level before each print is not acceptable.

Coupled with the unrelated issue of providing networking but no way of transferring Gcode – What!

Come on Prusa you are better than this!

A frustrated customer

Graham Law

Posted : 19/09/2020 3:37 pm
charles.h13
(@charles-h13)
Famed Member

No so sure I would agree.  There are varying experiences.  I wish we could figure why some have problems and others do not.  I often wonder if it is tied to certain rolls of filament.  I had the issue with CF PLA.  I took a little time to overcome.  

Posted : 19/09/2020 5:54 pm
graham.l2
(@graham-l2)
Eminent Member

Well top marks to Prusa for allowing my post as I was a bit grumpy when I wrote it but do mean it as constructive feedback.

As for filament variants - if my conjecture is correct Id expect different filaments to complicate the picture.  I tried to eliminate this be taking half a spool of filament from my Mk3 that was printing flawlessly and used it on the Mini and it was just problematic.

Fo me all jamming experiences were on the first few layers where the Z offset is pertinent.   Ie causing a transient nozzle block when the extruders single traction wheel would tear the filament and stop feeding.   If others are experiencing similar but only on the first few layers of print then is consistent with what I believe is the main issue is, ie poor performance of the MINDA prob exasperated with reduced thrust available from the extruder.

This could be hard for Prusa to fix as the control board probably does not have an input spare compatible with a Thermistor to provide PINDA temperature compensation support.

Im happy to be wrong but it sure looks that way to me.

Graham

Posted : 19/09/2020 8:36 pm
bobcousins
(@bobcousins)
Reputable Member
Posted by: @graham-l2

For me all jamming experiences were on the first few layers where the Z offset is pertinent.   Ie causing a transient nozzle block when the extruders single traction wheel would tear the filament and stop feeding.   If others are experiencing similar but only on the first few layers of print then is consistent with what I believe is the main issue is, ie poor performance of the MINDA prob exasperated with reduced thrust available from the extruder.

This could be hard for Prusa to fix as the control board probably does not have an input spare compatible with a Thermistor to provide PINDA temperature compensation support.

I checked the PCB design, there is a thermistor pin on the MINDA connector, and it is routed to an ADC on the CPU. So no PCB change is needed, firmware and obviously upgrade MINDA.

AFAIK,most common jamming issues are heat creep in the hot end, not lack of traction in the extruder. However, the teeth in the extruder clog quickly if it does skip, and require regular checks to make sure they are clean. Also, you might want to check the tension setting.

I find with the textured bed, a low Z height is not needed, but I can imagine that a low Z would cause additional back pressure in the hot end which might lead to the extruder chewing filament.

After having jams for the third time, I have upgraded my Mini with an E3D v6 hotend, works like a charm. The retail cost of the genuine E3D is about the same as the Mini, but I guess Prusa save a lot of cost with their in house design rather than paying E3D.

I have not owned a Prusa before, only several cheap types, but I would only rate the Mini as "OK". I probably wouldn't recommend to a friend, unless they are prepared to do some mods. It's a pity, because it could be great with a few improvements. I have not seen any indication that Prusa staff monitor this forum or act on any feedback, apart from one guy dealing with firmware support.

Posted : 19/09/2020 10:16 pm
graham.l2
(@graham-l2)
Eminent Member

Thanks Bob

good news on the PCB front I can only assume the removing the thermal feedback option was a cost thing but they have hedged their bets in case it was needed which we believe it is &  asap.

You may well be correct on the hot end I have seen some evidence of that but I tend to pre heat so that would help.  As I was focusing first layer issues the symptom was a failure part way through ie the filament was flowing then stops as the Z gap fluctuated toward zero stalling the drive and stripping the filament at the drive wheel.

With potentially three weaknesses here:

1) Unstable Z offset due to the MINDA

2) Dubious (imo) single wheeled drive 

3) Hotend thermal bleed (Bob)

Its no surprise that the symptoms will be very variable and hard to nail down definitively if focussing on one them.

Im interested in the E3D V6 you mention,  can you provide a link please and comment on compatibility with fixing it to the Mini, thanks

Also on textured finish it just so much better when you can really drive into the plater with a stable low Z offset that the Mk3i is excellent at compared the Mini which is just hit & miss by comparison.

Graham

 

 

 

 

This post was modified 1 month ago by graham.l2
Posted : 20/09/2020 10:55 am
ChillyPenguin
(@chillygeo)
Eminent Member

It would be great if Prusa just enable temperature compensation in the official Mini firmware. Then anyone that wants to upgrade to an PINDA can do.

 

I understand why they have chosen to use the MINDA to keep to the Mini price point, as its save 33% on the part cost (based on replacement cost).

Posted : 20/09/2020 11:43 am
Texy
 texy
(@texy)
Reputable Member

Nothing new in this thread that hasn't already been mentioned, as you say Graham. Prusa don't necessarily monitor or moderate these forums - the best way to vent your frustrations is via chat or email to Prusa. I too am hoping that a PINDA upgrade is supplied. In the mean time I just live with it, although I am seriously considering buying another Mk3 to replace my MINI.

Texy

Posted : 20/09/2020 1:12 pm
charles.h13
(@charles-h13)
Famed Member

I understand the anger also.  I am not casting blame, but I find it interesting that only some users are having this problem.  I have had the problem with jams when printing certain filaments on my Mini and especially when switching filaments.  It is not alleviated with cold pulls most of the time.  Most of the time, I am able to clear it by removing the filament and running some cleaning filament through the hotend.    

I can see how the MINDA / PINDA issue may be the issue, btu why is this not more widespread?

Posted : 20/09/2020 2:29 pm
zoltan
(@zoltan)
Member Moderator

Hi @texy,

there are moderators who moderate the threads and are doing their best to help users. As @charles-h13 mentioned there are many users with different experience. Happy users are posting very rarely here. So we cannot asume, that all users are unhappy. The threads are monitored and especially the official github pages are always responded. On the other hand those pages require a thorough description of the problem. There are many experienced community users helping the newcommers.  I have tried to post here a simple guide https://forum.prusaprinters.org/forum/english-forum-general-discussion-announcements-and-releases/we-want-to-help-you-you-have-to-tell-us/ for users to describe their issues, so the helpers can spend their time on help not on additional questions generation. The sites help.prusa3d.com is growing from day to day. So I would not say Prusa is not listening to the community. Some users just do not search for solution here, there is easier to post the question for which there is a standard answer on many threads.

 

 

 

This post was modified 1 month ago 2 times by zoltan
even an old man can learn new things 🙂
Standard I3 mk3s, FW 3.9.1, no closed box, sketchup , fusion 360, PrusaSlicer, Windows 10
PRUSA MINI FW 4.2.1, technical background...
Posted : 20/09/2020 3:55 pm
flo.kohl
(@marlin)
New Member

@charles-h13

I would like to second that experience!

It must be said that although I have a lot of experience with FDM printing (in an open lab), the MINI is my first self-owned 3D printer so I am not yet totally familiar with troubleshooting.

However, I recently bought a new PLA filament from RYKE on Amazon only to find out that I cannot use it because either the nozzle jams randomly and sequentially the extruder grinds the filament, or the other way round, where the extruder grinds and the underextruded filament jams the nozzle. 

I am verry frustrated because I already tried everything I could (cleaned extruder, adjusted idler-wheel tension, did cold pulls, changed the nozzle, tried to move the hotend up a bit to avoid heat-creep). I was even able to print 3 pretty decent parts with this new filament (mostly before all of my fixing attempts) but now nothing seems to work anymore. I do not even get past the second layer without grinding.

This is irritating, because my prusament works splendidly. (after reversing the corrections for the RYKE filament of course).

I have also noticed a significant disparity between the slicer settings for prusament and for generic PLA. I tried the same model in prusament with the prusament profile and the generic PLA profile and they came out vastly different (generic one was underextruded)!

What the heck is going on here?

I am not good enough yet to adress all these issues! This is my first 3D-printer! Why does this happen?

 

I would appreciate some advice, but I am almost ready to accept my defeat ); 

 

Marlin

This post was modified 1 month ago by flo.kohl
Posted : 20/09/2020 7:57 pm
mcjamez
(@mcjamez)
Active Member

Can you describe the behavior you see from the MINDA? I have been running G29 quite frequently lately in support of the silicon tubing mod and have been getting very consistent results from the MINDA when the bed is pre-heated to 170c. 

Posted : 21/09/2020 4:52 am
bobcousins
(@bobcousins)
Reputable Member
Posted by: @graham-l2

Im interested in the E3D V6 you mention,  can you provide a link please and comment on compatibility with fixing it to the Mini, thanks

I ordered the E3D from the E3d website, 1.75mm 24V Bowden variant. To save a little time, I paid more for the assembled version. https://e3d-online.com/products/v6-all-metal-hotend I used the mount here https://www.prusaprinters.org/prints/31006 with a R1 version of the MINDA probe mount.

Installation was quite easy, the only adjustment required is to make sure the MINDA probe is at the right height. I am using the same profiles I was before.

 

Posted : 21/09/2020 12:22 pm
bobcousins
(@bobcousins)
Reputable Member
Posted by: @charles-h13

I can see how the MINDA / PINDA issue may be the issue, btu why is this not more widespread?

I'm not sure why people think that a product either always works in all circumstances, or never works. There are so many variables, it is never a binary thing.

I've spent many years in product development. We spend weeks testing a product in the lab and it passes all the tests. Then the first day it goes for EMC testing, I get a call to say our product is "not working". After some troubleshooting, we find it is a configuration we had never tested in the lab. We fix some things then continue. Then we send the products for field trial, and customers find things we never encountered before. Then we get to production, and making thousands per year, we find a 0.5% failure rate at production final test. Most of those are manufacturing issues, but occasionally there is a genuine design issue.

It is a continuous process of iteration and improvement. Even years after product introduction, a new issue may be found, and we wonder, how come we never noticed this issue before? The only way to improve is to keep analyzing the faults, and implementing corrective actions. 

Given the stage of the Mini in it's lifecycle, we should not be surprised that there are issues seen by a minority of users. How many versions of the Prusa i3 have there been?

Posted : 21/09/2020 12:43 pm
charles.h13
(@charles-h13)
Famed Member

@bobcousins

I understand.  I guess what I was saying is that I don’t understand why a faulty MINDA probe would not cause more wide spread problems.  I guess I must not fully understand the problem.  

Posted : 21/09/2020 1:34 pm
graham.l2
(@graham-l2)
Eminent Member

Bob I think thats very charitable you are clearly you a glass half full sort of chap.  🙂

Im also a development engineer (electronics/ software) and if this was Prusa first printer I could side with you on this, but it is not.

Its a boiled down 'variation' of the Mk3 in which many man-years of development have gone into and with the Mini decisions to take bits out have been taken too far.  Ie take the PINDA on the Mk3: it was the result of development where temperature compensation was deemed worthwhile.  The Mini's MINDA without such compensation could only be to save say $4 and the detriment in performance deemed acceptable and presumably fully understood.    Prusa could not have been 100% comfortable with the decision or they would not have designed the control board to accept a thermistor it was solely to save $4.     

Its that sort of thing that is... disappointing lets say. 

I shall leave it others to rant,

I suspect its mostly users knowing & expecting Mk3 performance that will be the most grieved - new users may just accept its 'flaky personality' .

Graham

 

 

 

 

This post was modified 1 month ago by graham.l2
Posted : 21/09/2020 2:25 pm
charles.h13
(@charles-h13)
Famed Member

Can someone explain what proof we have that Prusa cut corners to save money?  

As I see it, a small minority of users are having issues.  Another group has replace parts or used cold pulls and cleaner filament to get a use able product and quality prints.  What am I missing?    

Posted : 21/09/2020 4:29 pm
mcjamez
(@mcjamez)
Active Member

The claim here is that there is an inherent design flaw, not that there are edge cases were it fails. That's big difference Bob. 

I think design flaw is an awful big claim that has yet to be proven with evidence. 

I'd like to see the result of multiple G29's proving non-reliability of the MINDA to start. 

Posted : 21/09/2020 6:17 pm
jsw liked
bobcousins
(@bobcousins)
Reputable Member
Posted by: @mcjamez

The claim here is that there is an inherent design flaw, not that there are edge cases were it fails. That's big difference Bob. 

I think design flaw is an awful big claim that has yet to be proven with evidence. 

I'd like to see the result of multiple G29's proving non-reliability of the MINDA to start. 

Not really. Where I work, there can be a design "feature" that affects 1 in 1000 users.  Whether you call that a flaw or an edge case is just semantics. In our company, we call it a design flaw. I think people say "it's just an edge case" when they mean "don't want to fix it". I'm glad to say our company takes it's reputation seriously - especially for certain Very Important customers 🙂

I've been building and using 3d printers since 2011, I've designed and built controller boards, contributed code to Marlin and other firmwares. I'm satisfied that the modifications I have made to my Mini fix issues in the design. I am not one to add bling just for the sake of it.

I think there are some people who will always refuse to accept Prusa has made a mistake, honestly I see no point arguing with them. No amount of proof will be enough for them. I can help other people fix their printers, I'm not interested in changing the minds of people who don't see problems or think there are none.

Posted : 21/09/2020 9:26 pm
bobcousins
(@bobcousins)
Reputable Member
Posted by: @graham-l2

Its a boiled down 'variation' of the Mk3 in which many man-years of development have gone into and with the Mini decisions to take bits out have been taken too far.  Ie take the PINDA on the Mk3: it was the result of development where temperature compensation was deemed worthwhile.  The Mini's MINDA without such compensation could only be to save say $4 and the detriment in performance deemed acceptable and presumably fully understood.    Prusa could not have been 100% comfortable with the decision or they would not have designed the control board to accept a thermistor it was solely to save $4.     

Its that sort of thing that is... disappointing lets say. 

I agree with you there. There are some features of the Mini that are a useful leap forward, e.g. LAN connection, fast color graphics. But then the MINDA seems to be step backwards, so I find that a puzzling decision.

I guess the Mini hotend was an attempt to create a more compact version, and maybe reduce cost, but hotends are notoriously hard to get right, and even harder to cost engineer. I think the Mini was a good attempt, but not quite there.

Posted : 21/09/2020 9:38 pm
mcjamez
(@mcjamez)
Active Member
Posted by: @bobcousins
Posted by: @mcjamez

The claim here is that there is an inherent design flaw, not that there are edge cases were it fails. That's big difference Bob. 

I think design flaw is an awful big claim that has yet to be proven with evidence. 

I'd like to see the result of multiple G29's proving non-reliability of the MINDA to start. 

Not really. Where I work, there can be a design "feature" that affects 1 in 1000 users.  Whether you call that a flaw or an edge case is just semantics. In our company, we call it a design flaw. I think people say "it's just an edge case" when they mean "don't want to fix it". I'm glad to say our company takes it's reputation seriously - especially for certain Very Important customers 🙂

I've been building and using 3d printers since 2011, I've designed and built controller boards, contributed code to Marlin and other firmwares. I'm satisfied that the modifications I have made to my Mini fix issues in the design. I am not one to add bling just for the sake of it.

I think there are some people who will always refuse to accept Prusa has made a mistake, honestly I see no point arguing with them. No amount of proof will be enough for them. I can help other people fix their printers, I'm not interested in changing the minds of people who don't see problems or think there are none.

I've been on this site for a month, you don't know me, and I have no reason to defend Prusa in any way. You offer zero proof and just make general statements about the way your company does things. You are doing a good job at not changing peoples minds.

I asked for evidence that the MINDA is not working correctly. Even gave a method for showing clear proof. I have no idea what the result of that test will be with the OP's printer. However, I do know what the results are with my particular printer. It would literally take 10 minutes to do 20 runs of G29. Please perform the G29's at 170c like the Prusa slicer does. If the results are inconsistent then we'd have proof and not just rants. 

Maybe there is a flaw? I don't know. Prove it one way or the other.

-Jim

 

This post was modified 1 month ago by mcjamez
Posted : 21/09/2020 10:52 pm
Page 1 / 4
Share:

Please Login or Register