cloud printing, not LAN printing?
@cwbullet
From last week's Prusa Live, Prusa Connect will be all online. It looks like the only Pi needed would be a Zero W to convert from WiFi to the UART pins on a MK3 since it doesn't have a network port or WiFi itself. The user interface is all web-based, so hitting it from a Pi or tablet with a standard browser should not be a problem.
Wait, so it will be a cloud printing feature, not just a http-based communication hosted on a printer? Being able to print over LAN, no matter if it is using CLI or just over HTTP would be very helpful, but I'm not planning on opening up the printer to the internet.
Cloud
@cwbullet
From last week's Prusa Live, Prusa Connect will be all online. It looks like the only Pi needed would be a Zero W to convert from WiFi to the UART pins on a MK3 since it doesn't have a network port or WiFi itself. The user interface is all web-based, so hitting it from a Pi or tablet with a standard browser should not be a problem.
Wait, so it will be a cloud printing feature, not just a http-based communication hosted on a printer? Being able to print over LAN, no matter if it is using CLI or just over HTTP would be very helpful, but I'm not planning on opening up the printer to the internet.
Oh hell no. If this is true, I'm definitely out. My models are worth more than my farm of printers.
Latest updates - Prusa Live 30
Ok, from this week's Prusa Live #30 (the one with Printer Bunny as guest) at just after an hour into the video there was some talk about it:
Prusa Link (formerly referred to at least once as Prusa Connect Local), is "on your network talking to one printer, more like Octoprint". Prusa Connect will be online, supporting multiple printers.
Prusa Connect is in closed beta. They will expand the pool of beta testers at some point. They're looking more at large farms right now.
There is currently one Mini connected to their network via WiFi. No more features than currently on ethernet, and it isn't 100% yet.
Prusa Connect have monthly fees? "Free with the printer, the basic stuff." Looking at if they want to support other brands of printers and charge for that. There will be a version for schools and companies that want to self-host that can't use the cloud platform.
Is Mini suitable for farm?
I'm following "farm" related stuff just out of curiosity (3D print is only my hobby and it will stay that way).
I think, more important than "is farm possible with Mini?" is the question "does it make sense with Mini?". Of course, it depends on type of product that needs to be printed and the size of the farm -for some, having three printers is already a farm.. so be it.
Anyway, I don't see Mini being suited for serious business. It just wasn't designed for that. Even it can be usable for prototyping, it's a hobby (home) printer in first place. It has too many limitations, where print size and limited choice of filament types, being the most important. I'm aware that many use Mini's (among other printers) in their "farm". But I just don't see Mini (farm) as rational long term decision -as mentioned here (at the end of the video).
Just sharing my thoughts.
[Mini+] [MK3S+BEAR]
Is that fair?
Please correct me print farmers,
I thought the mini was partially targeted at farms because it’s big enough for printing many parts, and it’s half the price of an i3 so you can double your print volume. Granted I’ve not tried to print any ‘proper’ engineering materials with mine yet (also just a hobbyist), but I can’t see it being totally impossible. I imagine the greatest challenge to building a farm based on this platform is getting the printers in the first place.
RE: Farm
Please correct me print farmers,
I thought the mini was partially targeted at farms because it’s big enough for printing many parts, and it’s half the price of an i3 so you can double your print volume. Granted I’ve not tried to print any ‘proper’ engineering materials with mine yet (also just a hobbyist), but I can’t see it being totally impossible. I imagine the greatest challenge to building a farm based on this platform is getting the printers in the first place.
Print farm to me is not equated to engineering-type materials. A print farm has multiple printers printing at one time to produce products for sale or use. I have been using 10+ printers for hobby / commercial use for over 2 years. They have not paid for themselves but they have paid for expansion and maintenance. I expect to break even in 1-2 more years.
Of course, the mini can be used for a print farm, but you need to understand what you can and can not print on it. It can print 95% of what I need to print. I am sure most of us will find the same results.
--------------------
Chuck H
3D Printer Review Blog
Prusa Connect
Ok, from this week's Prusa Live #30 (the one with Printer Bunny as guest) at just after an hour into the video there was some talk about it:
Prusa Link (formerly referred to at least once as Prusa Connect Local), is "on your network talking to one printer, more like Octoprint". Prusa Connect will be online, supporting multiple printers.
Prusa Connect is in closed beta. They will expand the pool of beta testers at some point. They're looking more at large farms right now.
There is currently one Mini connected to their network via WiFi. No more features than currently on ethernet, and it isn't 100% yet.
Prusa Connect have monthly fees? "Free with the printer, the basic stuff." Looking at if they want to support other brands of printers and charge for that. There will be a version for schools and companies that want to self-host that can't use the cloud platform.
It looks like they may have changed direction to go online. I am excited to see how it ends up.
--------------------
Chuck H
3D Printer Review Blog
I'm perpetually waiting...
Please correct me print farmers,
I thought the mini was partially targeted at farms because it’s big enough for printing many parts, and it’s half the price of an i3 so you can double your print volume. Granted I’ve not tried to print any ‘proper’ engineering materials with mine yet (also just a hobbyist), but I can’t see it being totally impossible. I imagine the greatest challenge to building a farm based on this platform is getting the printers in the first place.Print farm to me is not equated to engineering-type materials. A print farm has multiple printers printing at one time to produce products for sale or use. I have been using 10+ printers for hobby / commercial use for over 2 years. They have not paid for themselves but they have paid for expansion and maintenance. I expect to break even in 1-2 more years.
Of course, the mini can be used for a print farm, but you need to understand what you can and can not print on it. It can print 95% of what I need to print. I am sure most of us will find the same results.
This was the whole purpose for me converting over to Prusa - I needed a standard printer to work from. My most significant issue right now is not the lack of farm software, as I can juggle it for now with USB-C adapters and cable extensions (Seriously, it's by-far the most convenient solution). The issue is the print area, I have two options from Prusa, a 32-bit 7"x7"x6.9" print volume (Prusa Mini) or a 8-bit gonna-freak-out-if-you-look-at-it-wrong 8.3"x8.3"x9" (i3 Mk3s) that won't work with the farm software without an upgrade. I outright refuse to purchase a Mk3s because it is obsolete (in comparison to the Mini), I have worked with a farm of Mk3s's building face shield frames and the problems of the printer are significant - overly static-sensitive 8-bit board, old PINDA probe required bed pre-heat, 8-bit responsiveness caused issues if you looked at it wrong during a print, and other odd errors (I've printed ~14,000 hrs on the Mk3s).
What I'm trying to say is that I desperately need an XL printer (32-bit) as I'm running off Ender 3 Pro(s) (with the upgraded mainboard) as they are more functional to me than the Mk3s currently.
I'd prefer an XL printer over the farm software at the moment.
RE: Mini still needs work
Please correct me print farmers,
I thought the mini was partially targeted at farms because it’s big enough for printing many parts, and it’s half the price of an i3 so you can double your print volume. Granted I’ve not tried to print any ‘proper’ engineering materials with mine yet (also just a hobbyist), but I can’t see it being totally impossible. I imagine the greatest challenge to building a farm based on this platform is getting the printers in the first place.
In my opinion, based on my 5 years of experience in adding 36 minis to my print farm to complete a 105 machine army driven by OctoPrint/OctoFarm.....
The mini as shipped is unsuitable for farm operation due to it's reliability for the following reasons:
1.) The new hotend is very fickle and needs constant adjustment. I believe this is due to the 3 bite grub screws not applying enough force on the lubed (heat sink) surface. With repetitive heating, cooling, and vibration I would have to adjust the PTFE tension (push the hotend up and tighten) every 10 or so prints. This compounded issue #3 below.
2.) The old probe made the machine very susceptible to ambient temperature variations that would bring the live Z adjustment required outside of the software limit. This was apparent in my non-temp controlled farm in which original cal was performed at 70F; if the temp dropped to 50F (winter) a manual adjust of the bed level probe would be required as the probe couldn't be lowered to get a good first layer. This has been fixed with the SuperPINDA probe from what I have experienced, but we did have to dump another $1K into the farm and 40 hours to do the install.
3.) The extruder has no exit for filament debris, which fouls the idler bearing over time. This was exacerbated by the MINDA as a good first layer adjust was required any time ambient temps varied by more then 4-5F in the shop. This was also very apparent with materials such as HTPLA as it tends to slough a fine powder. We had a rash of bearing failures that we resolved by redesigning the extruder to make use of the R6 design from the MK3S+.
4.) There is a race condition on the serial port that causes a random serial disconnect that kills 1-2% of prints. I'm hoping RC4.3.2 resolves this as it is still a bug that kills at least 1 print a day on my farm.
5.) The bed thermistor opens up every 200-300 hours and must be replaced frequently as compared to the MK3S+. We've mitigated this with bed chains on the X and Y axis, but it appears as if the lower heatbed mount channel and the proximity to the nylock nut on the bottom is creating a pinch point on the thermistor wire.
We have made the Mini reliable enough to support our 100KG/mo farm, but it took a redesign of the chain of custody from the spool to the nozzle. We replaced the Mini hotend with a original E3D V6, and made use of the R6 extruder from the MK3S+. We upgraded all of the probes to the Super PINDA. The farm is still plagued by random serial disconnects but it is infrequent enough it's just a write-off at this point. We have plans to redesign the heatbed mount in the near future once we get more trend data to justify the part/labor cost.
Bottom line: As shipped, these machines have been a headache since they landed back in November 2020. We've had to add $100 to each machine and about 3 hours of labor (108 man hours total, disregarding extruder and hotend design time) to get them printing reliably, but they are still plagued by random serial disconnects. As a veteran Engineer of 24 years, I was able to design around the shortcomings, but your average "Joe/Jane" probably won't be able to.
Mini
I do not entirely disagree. Your bring up good points.
I will add that there is a difference unsuitable and a headache. I have 5 in my smaller farm and they are reliable if you recognize the limitations and have the right environment. Four with the stock hotend and extruder and one with the V6. Mine are in an environmentally controlled environment. If used properly, they can be a force multiplier. I have two Bondtech extuder, but removed them when I saw no improvement. I will be expanding my farm with 3 Mini+ In the coming months.
--------------------
Chuck H
3D Printer Review Blog
Filament Sensor
I almost forgot....
6.) The filament sensor is disabled in remote. This was an unwelcome surprise, and we get by rolling partials out to the MK3S+ rack. Occasionally we air print if we misjudge the filament remaining. We're hoping 4.3.2RC1 fixes this, and the initial reports from the other users looks promising. We're curious as to why it has taken over a year to enable this. Looking at their code, the re-wrapped marlin firmware and elaborate event handler system they implemented to support the new screen are likely the cause.
Labor must be accounted for
I do not entirely disagree. Your bring up good points.
I will add that there is a difference unsuitable and a headache. I have 5 in my smaller farm and they are reliable if you recognize the limitations and have the right environment. Four with the stock hotend and extruder and one with the V6. Mine are in an environmentally controlled environment. If used properly, they can be a force multiplier. I have two Bondtech extuder, but removed them when I saw no improvement. I will be expanding my farm with 3 Mini+ In the coming months.
I have to upgrade to "Unsuitable" as I have employees that I pay for the maintenance. Knowing my margins to the penny, the increased labor cost in maintenance drives the cost per print to the point that it reduced my margins by about 10% as compared to the MK3S+. If it were an one-man operation, then it would only be sleep & BBQ time at stake.
To add to it, the errors are not reported correctly. With the bed thermistor open, often the machine will give a BSOD or just freeze entirely. This was confusing my employees as they thought it was the random serial disconnect. I would troubleshoot it and find the -15C displayed on a bad spot, effectively doubling my time in troubleshooting.
Maintenance.
I do not entirely disagree. Your bring up good points.
I will add that there is a difference unsuitable and a headache. I have 5 in my smaller farm and they are reliable if you recognize the limitations and have the right environment. Four with the stock hotend and extruder and one with the V6. Mine are in an environmentally controlled environment. If used properly, they can be a force multiplier. I have two Bondtech extuder, but removed them when I saw no improvement. I will be expanding my farm with 3 Mini+ In the coming months.
I have to upgrade to "Unsuitable" as I have employees that I pay for the maintenance. Knowing my margins to the penny, the increased labor cost in maintenance drives the cost per print to the point that it reduced my margins by about 10% as compared to the MK3S+. If it were an one-man operation, then it would only be sleep & BBQ time at stake.
To add to it, the errors are not reported correctly. With the bed thermistor open, often the machine will give a BSOD or just freeze entirely. This was confusing my employees as they thought it was the random serial disconnect. I would troubleshoot it and find the -15C displayed on a bad spot, effectively doubling my time in troubleshooting.
You are definitely having more problems than I am. I have 99% uptime and only 1% or less downtime. Most of the down time is routine monthly maintenance. Maybe I will get more problems as I buy more.
--------------------
Chuck H
3D Printer Review Blog
Scaling is the issue
I think that is the key to the experience. With only a few minis it was not bad at all. I originally tested with 4 minis on a shelf before I committed to the 36. On fielding the 36 these issues started to surface and I originally blacked out the racks for 3 months with intent to sell them off until the Super PINDA was made public.
In scaling up the operation, the required maintenance becomes a burden as compared to the Ender 3 or MK3S+ that I have in the Army. Mind you, I had to heavily mod the Ender 3 to get the same performance of the MK3S+ with $200 and 8 man hours per machine (a rant for another day). Luckily those 24 Ender 3's have paid for themselves 10x and are my overflow machines.
Mini long term issues
1.) The new hotend is very fickle and needs constant adjustment. I believe this is due to the 3 bite grub screws not applying enough force on the lubed (heat sink) surface. With repetitive heating, cooling, and vibration I would have to adjust the PTFE tension (push the hotend up and tighten) every 10 or so prints. This compounded issue #3 below.
3.) The extruder has no exit for filament debris, which fouls the idler bearing over time. This was exacerbated by the MINDA as a good first layer adjust was required any time ambient temps varied by more then 4-5F in the shop. This was also very apparent with materials such as HTPLA as it tends to slough a fine powder. We had a rash of bearing failures that we resolved by redesigning the extruder to make use of the R6 design from the MK3S+.
5.) The bed thermistor opens up every 200-300 hours and must be replaced frequently as compared to the MK3S+. We've mitigated this with bed chains on the X and Y axis, but it appears as if the lower heatbed mount channel and the proximity to the nylock nut on the bottom is creating a pinch point on the thermistor wire.
Is there a place I can read up on these issues? A print farm with 40 minis is an excellent way of finding out ahead of time, what might happen to my one mini. The hotend adjustment and PTFE tube tension seems like they'd be hard to diagnose, until they get really severe. Are there early warnings? As for a mitigation, maybe a use of high-temp thread-lock would help?
Also, I wonder if the termistor issues could be caused by the silly way the heatbed is connected to power, with over 10A current going through a wire and through a thread, the resistance of this connection would vary significantly, constantly heating the whole area up and down...
Nylock
1.) The new hotend is very fickle and needs constant adjustment. I believe this is due to the 3 bite grub screws not applying enough force on the lubed (heat sink) surface. With repetitive heating, cooling, and vibration I would have to adjust the PTFE tension (push the hotend up and tighten) every 10 or so prints. This compounded issue #3 below.
3.) The extruder has no exit for filament debris, which fouls the idler bearing over time. This was exacerbated by the MINDA as a good first layer adjust was required any time ambient temps varied by more then 4-5F in the shop. This was also very apparent with materials such as HTPLA as it tends to slough a fine powder. We had a rash of bearing failures that we resolved by redesigning the extruder to make use of the R6 design from the MK3S+.
5.) The bed thermistor opens up every 200-300 hours and must be replaced frequently as compared to the MK3S+. We've mitigated this with bed chains on the X and Y axis, but it appears as if the lower heatbed mount channel and the proximity to the nylock nut on the bottom is creating a pinch point on the thermistor wire.
Is there a place I can read up on these issues? A print farm with 40 minis is an excellent way of finding out ahead of time, what might happen to my one mini. The hotend adjustment and PTFE tube tension seems like they'd be hard to diagnose, until they get really severe. Are there early warnings? As for a mitigation, maybe a use of high-temp thread-lock would help?
Also, I wonder if the termistor issues could be caused by the silly way the heatbed is connected to power, with over 10A current going through a wire and through a thread, the resistance of this connection would vary significantly, constantly heating the whole area up and down...
The PTFE issue manifests as an uptick in jams and excessive resistance on the filament path. It begins with a click here and there out of the extruder and fails as an air print. By that point, the extruder hobbed bolt and idler gear get covered in filament debris (HTPLA). The PTFE adjustment as recommended by prusa works until the idler gear becomes fouled and offers too much resistance in the filament path.
You might be on to something with the heatbed wire. It doesn't get "too hot", but it could be the thermal expansion in the cavity is enough to wedge the thermistor wire against the nylock nut. I'll keep an eye on it.
I haven't been blogging/VLogging like I used to on these machines as the user base for this machine is fairly green and the online forums are flooded with routine issues veteran users resolve quickly. I will not save Prusa on support costs be providing free technical support. This is his design, let him suffer for it.
Grub Screw Problem info
1.) The new hotend is very fickle and needs constant adjustment. I believe this is due to the 3 bite grub screws not applying enough force on the lubed (heat sink) surface. With repetitive heating, cooling, and vibration I would have to adjust the PTFE tension (push the hotend up and tighten) every 10 or so prints. This compounded issue #3 below.
3.) The extruder has no exit for filament debris, which fouls the idler bearing over time. This was exacerbated by the MINDA as a good first layer adjust was required any time ambient temps varied by more then 4-5F in the shop. This was also very apparent with materials such as HTPLA as it tends to slough a fine powder. We had a rash of bearing failures that we resolved by redesigning the extruder to make use of the R6 design from the MK3S+.
5.) The bed thermistor opens up every 200-300 hours and must be replaced frequently as compared to the MK3S+. We've mitigated this with bed chains on the X and Y axis, but it appears as if the lower heatbed mount channel and the proximity to the nylock nut on the bottom is creating a pinch point on the thermistor wire.
Is there a place I can read up on these issues? A print farm with 40 minis is an excellent way of finding out ahead of time, what might happen to my one mini. The hotend adjustment and PTFE tube tension seems like they'd be hard to diagnose, until they get really severe. Are there early warnings? As for a mitigation, maybe a use of high-temp thread-lock would help?
Also, I wonder if the termistor issues could be caused by the silly way the heatbed is connected to power, with over 10A current going through a wire and through a thread, the resistance of this connection would vary significantly, constantly heating the whole area up and down...
Early-warnings for the grub screw issue is Z offset shift, if you're noticing that you are having to adjust your Z-offset in one direction over time, then you should perform the quick-fix.
Step 1: Unload Filament
Step 2: Cool Down
Step 3: Remove Bowden Tube (unscrew nut above extruder)
Step 4: Remove bowden tube adapter above extruder
Step 5: Loosen Grub-screws
Step 6: Carefully push & hold the hotend/heat brake tube upward into the extruder housing
Step 7: While still holding the extruder, tighten the grub screws
Step 8: Re-install / tighten the bowden tube adapter
Step 9: Re-install the bowden tube / tighten
Step 10: Test print
There are a few causes for this issue, the primary one being the design of the extruder. The extruder was designed to have a PTFE tube in compression at all times, this force comes from the top bowden adapter which presses downward and the grub screws hold the heatbrake tube in place. This is not ideal, they really should have notches cut into the tube so you don't have to rely on the grub-screws pushing against a flat surface. Additionally you're dealing with fine threads in aluminum, so they tend to strip if you overtighten them.
Best of luck!
Filament sensor issue.
I almost forgot....
6.) The filament sensor is disabled in remote. This was an unwelcome surprise, and we get by rolling partials out to the MK3S+ rack. Occasionally we air print if we misjudge the filament remaining. We're hoping 4.3.2RC1 fixes this, and the initial reports from the other users looks promising. We're curious as to why it has taken over a year to enable this. Looking at their code, the re-wrapped marlin firmware and elaborate event handler system they implemented to support the new screen are likely the cause.
Are you still having this problem? I was pretty certain it was fixed on my last update, unless I have installed a beta. I will check tonight.
--------------------
Chuck H
3D Printer Review Blog
Ender 3.
I think that is the key to the experience. With only a few minis it was not bad at all. I originally tested with 4 minis on a shelf before I committed to the 36. On fielding the 36 these issues started to surface and I originally blacked out the racks for 3 months with intent to sell them off until the Super PINDA was made public.
In scaling up the operation, the required maintenance becomes a burden as compared to the Ender 3 or MK3S+ that I have in the Army. Mind you, I had to heavily mod the Ender 3 to get the same performance of the MK3S+ with $200 and 8 man hours per machine (a rant for another day). Luckily those 24 Ender 3's have paid for themselves 10x and are my overflow machines.
My ender 3 would take 2 years of consistent printing to pay off the work hours my team has put into it. It is 2 years old and has been down nearly 50% of the time. It is amazing how different our experiences are.
--------------------
Chuck H
3D Printer Review Blog
Mini maintenance and "break even point" vs "breaking point" in a farm?
@djkirkendall:
I agree with @John-14 that your chances of seeing specific issues pop up rises exponentially with the number of Mini's you're running. It also rises the more you print on each one. But, you also benefit from the experience gained from repairing and issue on the first one or two machines to make the repair times shorter or prevent the issue entirely (through adjustments, preventative maintenance, etc.) on the remaining units.
There is also the concept that a cheaper machine is easier financially to replace at the end of its "production life" than a more expensive machine. Not to say that the Mini should be considered to be disposable, but all machines wear out, get less precise, etc. over time even with regular maintenance and / or overhauls. My first printer (a Wanhao I3) was finicky at first, got way better after I did some inexpensive upgrades (mostly self-printed), and then went downhill fast after about 3 years. Not great, but not horrible for a $300 printer (one of the cheapest reasonable-quality ones at the time) with maybe a $100 worth of upgrades and parts (nozzles included) over that time. As a hobbyist, I think that I gained a lot of value out of it during that time as far as personal prints, education, and overall fun. I still keep it around, and may give it an overhaul for the occasional (noncritical) part that is bigger than the Mini's bed.
I think the Mini is too new to have MBTF (mean time between failures) numbers or overhaul schedules worked out yet. And, since most are in use by hobbyists, we probably won't soon have hard data on maintenance schedules and what happens how fast if they aren't followed. That is where as a community we need to work that out for all of our benefit, particularly from the larger farms that see the issues more often, and therefore can, and probably do, have more insight on maintenance and life-span. So without giving away any trade secrets here, maybe we can all help each-other out with some basic data points:
Maintenance:
What do you for maintenance how often?
Do you lube the screws and guide rods every month, or is that too often?
How often do you tend to change the nozzles with each type of filament that you run? Or, do you wait to you start to see a problem?
Upgrades:
Have you changed over from brass to hardened steel nozzles, and has it been worth the increased cost?
Have you upgraded the extruder, hot end, heat break, or PTFE tubing? Did it clear up issues? Was it worth doing financially?
Utilization:
About how much are you printing on each Mini per week or month, in terms of hours or spools of filament? A machine running a solid 8x5 every week is going to see issues a lot sooner than a home user that might average just a few hours a week.
Are you running your Mini's with fairly stock setting, or are you maximizing movement speeds, acceleration, etc., that may be beating on rails, screws, belts faster?
Overhauls and end-of service:
Have you had to do a tear-down and overhaul of any of your Minis yet? Where do you see the most wear?
Have you had to rotate any of your Minis out of daily production yet?
Are you still using them for backups, prototyping, personal use?
Were they still good enough to donate to a local school / library / 4H / Boy's and Girl's Club / Robotics Club, or were they basically shot at that point?
Financial:
Did you get your money's worth out of them by the time you took them out of daily production? What do you consider that point to be? 40 filament rolls worth of prints? 60 rolls? more?
Ability to dedicate multiple Minis to just churning out the same parts every day versus changing between parts on fewer expensive printers more often.
Financial ability to buy Mini's more frequently as you need them vs having to save up for more expensive printers.
Life-time value vs cost of the Mini vs other printers you've run.
Less tangible benefits vs other printers in the farm:
More but smaller batches of parts per day.
Less space and power required.