Quicker nozzle swaps  

RE: Quicker nozzle swaps

As you still have to take out the whole hotend assembly for the swap, it would not save a lot of time to have multiple heater blocks. The handling of the heater block for the swap is a lot easier and straightforward compared to the MK3xx. You can hold the parts in your hands and see clearly what you are doing without contorting yourself. All the required tools are included with the printer, and the risk of damaging any wires with them is minimal compared to the procedure on the MK3xx.

RE: Quicker nozzle swaps

I'd say it's easier undoing 2 thumbs screws and 2 wire clip 

RE: Quicker nozzle swaps

Not to mention that you have to take really special effort to burn your fingers ... 🙂

RE: Quicker nozzle swaps

I don't mean to swap it hot. Based on the images and manual, you are supposed to release thumb screws and remove the heat block and the nozzle. At this point you need to take a tool and remove the nozzle assembly from the heat block and then screw in the new one. On my idea I would have all the option nozzles permanently connected to their own heat block, so to swap I would just need to release the cables and thumb screws, remove and insert a new. Although after checking the shop, just the nozzle is 29,90 €. If I want to have a heat block for each nozzle it would be around 75 € per nozzle.

RE: Quicker nozzle swaps

I’ve swapped on my MK4 a number of times and find it quick and easy. The only fussy part to me is getting the wires back in the cable channel when re-inserting the assembly. Probably I have clumsy fingers. I’ve found if I work them into the cable channel as I’m re-inserting the assembly it goes easier. The instructions say to do that after re-insertion. 

RE: Quicker nozzle swaps

A thought came to me that when buying new nozzles also by a heat block, heater and thermistor for it. So, when going to change the nozzle it is just an operation to release the cables and swap the whole thing. Do you see any problems in this idea

I maybe see a potential problem in changing the complete thermistor+heater+heaterblock unit without doing a new thermal calibration.

Because i have read about so many thermal anomaly errors on MK3S machines i have not used this firmware on my MK3S yet.

No idea how sensitive this thermal monitoring is on the MK4. I am still waiting for my MK4 kit to be delivered. 

RE: Quicker nozzle swaps

Posted by: @roland85

Hi, I'm currently waiting for my MK4 part and I was looking in to the nozzle swap process. I have MK3S+, but I have never changed a nozzle in it.The process for MK4 look every easy, but still requires tools. A thought came to me that when buying new nozzles also by a heat block, heater and thermistor for it. So, when going to change the nozzle it is just an operation to release the cables and swap the whole thing. Do you see any problems in this idea or I'm just overthinking this too much 😀

Of course this way every optional nozzle will cost more, but it should make the swaps very fast right?

This has been my plan since I first learned of the MK4. Just have different nozzle size assemblies each with their own heater, heat block, and thermistor.

I would have ordered these parts along with different nozzles when I got my kit but they were out of stock.

RE: Quicker nozzle swaps

From my experience with the XL and Mk4, I don't see much difference in the time it takes for swapping a complete assembly vs just the nozzle. While the nozzle swap is arguably easier and safer than on the Mk3S, it not what I'd call "quick". And not without its own perils as I've managed to break a thermistor cable when pushing it back into the XL. I admit, I miss the Dragon hotends on my Mk3S and Minis. Yes, you had to hot release and tighten the nozzle but the actual nozzle swap was a cinch. My hope for the Mk4 was something similar, with a heater block fixed in place, but no luck, I guess due to patent issues. 

RE: Quicker nozzle swaps

the nozzles (0.25, 0.6 and 0.8)  will come in the same order as the printer

Posted by: @roland85

I don't mean to swap it hot. Based on the images and manual, you are supposed to release thumb screws and remove the heat block and the nozzle. At this point you need to take a tool and remove the nozzle assembly from the heat block and then screw in the new one. On my idea I would have all the option nozzles permanently connected to their own heat block, so to swap I would just need to release the cables and thumb screws, remove and insert a new. Although after checking the shop, just the nozzle is 29,90 €. If I want to have a heat block for each nozzle it would be around 75 € per nozzle.

RE: Quicker nozzle swaps

Is there anything special needed to assemble the 3 components into a working unit? Meaning the heater block, thermistor and heater? I changed nozzles a lot at first but started feeling a bit cautious until I could get a replacement for the hot end.

RE: Quicker nozzle swaps

I wonder if you need boron paste or something for he heater carstridge ?

RE: Quicker nozzle swaps

Posted by: @kevman

I wonder if you need boron paste or something for he heater carstridge ?

I just got my spare heater block/core/thermisters in today, Boron Nitride paste coming in tomorrow. I'll be testing exactly this when I get a chance this weekend.

RE: Quicker nozzle swaps

Is that the same stuff Prusa sells for thermal paste?

RE: Quicker nozzle swaps

Thank you! keep us posted!

Posted by: @carlmmii

 

Posted by: @kevman

I wonder if you need boron paste or something for he heater carstridge ?

I just got my spare heater block/core/thermisters in today, Boron Nitride paste coming in tomorrow. I'll be testing exactly this when I get a chance this weekend.

RE:

Posted by: @ronguest

Is that the same stuff Prusa sells for thermal paste?

Nope. Prusa includes/stocks ThermoPasty AG Silver, which is more similar to CPU cooling thermal compound. The normal issue with these compounds is that they're not rated to high temperatures, which is clearly indicated in the official data sheet for this compound: -50c to 250c. For comparison, the Boron Nitride paste is rated for much higher temperatures, specifically up to 850c for the Slice Engineering stuff.

For assembly, the only place the use of thermal paste is mentioned in prusa's guide is to mate the heat brake to the heat sink in V6 (and possibly other) style hotends (i.e. here, step 14). For this purpose, temperatures do not exceed 250c even for the upper range of possible nozzle temperatures, so there's no risk in breakdown of the material for the intended use.

I cannot find any mention of using any type of thermal paste for any official Prusa assembly documentation for the actual heater core/thermister installation though, even for the XL (same design/components). Everything right now seems to only refer to a full hotend assembly instead of individual core/thermister for replacement. Going back to the previous design with the mk3 linked above, that design is fundamentally different as it keeps the heater core in place with near full contact sleeve-type retention, as opposed to this style that is held in place with 2 grub screws -- that means if there's no compound used, there's only a slim contact patch for conduction which will vary by machining tolerance.

So, that's what testing's for.

RE: Quicker nozzle swaps

any updates ?

Posted by: @carlmmii

 

Posted by: @kevman

I wonder if you need boron paste or something for he heater carstridge ?

I just got my spare heater block/core/thermisters in today, Boron Nitride paste coming in tomorrow. I'll be testing exactly this when I get a chance this weekend.

RE: Quicker nozzle swaps

FWIW I asked Prusa CS "when putting the Heater 40W into the heater block is something like Boron Nitride paste required". They responded that "For the Nextruder you won't need any other items to assemble it." I had mentioned this thread and he asked for a link to it which I sent. 

RE: Quicker nozzle swaps

If you’re using the nextruder nozzles, then it is fairly easy to remove the nozzle/block/heater/thermistor assembly and swap the nozzles.  I took the approach of fully assembled replacement sets though, using the longer wired XL heater on all, and they work equally well on both XL and MK4.

RE: Quicker nozzle swaps

Testing done. Definitely took a while, and it's been an interesting journey. Let's get into it...

Purpose: Determine the necessity of thermal paste/grease (specifically Boron Nitride) usage for spare Nextruder hotend assemblies (specifically MK4)

Control assembly: Pre-assembled hotend from factory (included with fully assembled printer)

Setup:

• Nozzle is torqued to 1.5Nm for each heater block [**see notes afterward]
• Same 0.4mm original Prusa nextruder nozzle is used for each block
• Filament used is eSUN PLA-LW black
• Firmware 5.0.0 alpha 3 All .gcode is run directly off the USB drive (octoprint only used for data capture)
• Original Prusa Enclosure

Variables:

• 4 separate complete hotend assemblies (heater block, heater core, and thermistor)
• Grub screw tightness (loose vs reasonable tightness)
• Thermal paste application for heater core
• Thermal paste application for thermistor

Tests:

Static test

Purpose: determine steady-state ability to target temperature without any heat flux other than ambient (also acts as safety check after assembly for testing)

• Filament is unloaded
• Temperature is stepped from 170c -> 270c in 20c intervals with 2 minutes to stabilize
• Temperature is lowered to 210c with 2 minutes to stabilize
• Temperature is raised to 250c with 5 minute final stabilization

Benchy @ 69% scale

Purpose: determine temperature stability for high-flow variability through nozzle

• Printed using 0.4mm IS profile (before new profiles added to slicer)

Tempreature tower

Purpose: determine temperature stability for differing temperature targets with constant flow rate

• Printed in vase mode (note: even though it's vase mode, there is a pause at each layer causing a noticeable bump -- I don't know what is causing this)
• 200c -> 270c in 10c increments every 5mm

Note: this also provides empirical temperature results at-the-nozzle by use of the PLA-LW filament for visual comparisons

MK4 nozzle block testing results

Alright. That's a lot of data. Let's talk...

All of the graphs were taken as screenshots from octoprint after each test run. As mentioned before, every .gcode was run off the USB stick so that there were no anomalies introduced from octoprint issues. Nonetheless, there were plenty of complete system crashes (happened randomly starting a new print, apparently this is possibly due to a Z-axis homing bug) -- thank you alpha firmware.

The images to the very right are actual photographs of the temperature tower test. Photos were taken in a locked off setup with a Sony a6500 using a manual lens at 1000 iso. For anyone who hasn't seen PLA-LW, this stuff foams up to different densities based on the temperature it's printed at. For our purpose, the surface finish lets us see a comparison of the actual temperature at the nozzle.

The first thing to note is the Stock results in the Other section with "Initial test without checking nozzle torque". There is a very specific aspect of the heat block design that needs to be addressed:

The natural wire tension will act to untorque the nozzle when installed.

Securing the heater block in the extruder assembly is only done by securing the heat brake section of the nozzle. The heater block itself will be tensioned by the heater core and thermistor wires (until they're fully trained, they will act as wire springs). This creates a torque in the direction to unscrew the block from the nozzle, which in this line of testing occurred a few times:

• Initial stock test results (listed under other)
• Heater block 2 (noticed after thermistor-only test run)

Because of this discovery, I decided to check the impact of nozzle tightness on the flow towers. Going off the banding, it appears that there's easily a difference of 10-20c between torqued and not torqued... not to mention, the horrible surface quality of the original stock test.

For this reason, heater block 1 and heater block 2 results may be inconclusive. You can see a huge anomaly in heater core 2 thermistor-only results with the flow tower -- the first 2 bands are perfectly smooth, indicating a temperature drop of at least 10c at the nozzle.

Moving forward from that discovery, I made sure to re-torque the nozzle after every setup run (i.e. when taking it out to tighten or add paste). The nozzle was unscrewed first to make sure it was still tight, then re-torqued.

Now that the validity of the results have been addressed, there's one other point to mention. Because we're testing thermal paste application, there's no true way to have a test using the same heater block where thermal paste is applied to only one component, and then only the other component. Thermal paste gets literally everywhere, and I can't say that removing thermal paste after application would be the same as returning it to the base condition.

So, that's why each heater block has one specific "only on ..." section. Half of the blocks had paste applied to the heater core first, the other half had paste applied to the thermistor first. Every block did still get thermal paste applied to both at the end.

The detailed analysis of each test is included in each separate sheet underneath the comparisons. Here's the overall takeaway from each test:

Grub screws "loose"

This test specifically was done to test the worst-case condition to see what things would look like with the worst thermal contact between components. As expected, temperature response suffers, and noise is noticeably higher for both readings and overall temperature seeking. Response time was also delayed for temperature shifts. With that said, it still results in nearly flawless print results. This was completely unexpected, and this test specifically is the reason the "static" test was included, as that would let me know if there's a catastrophic failure before risking an actual print situation. I DO NOT recommend this for obvious reasons, but it is a definite indicator of the machining tolerances involved if we have these results already. Output temperature does look ~5c hotter than stock (more on this later).

Grub screws tight, no thermal paste applied

The "recommended" configuration apparently (going off the response that @ronguest received from Prusa support). Overall noise is reduced, but is still noticeable. Benchy test now looks basically identical to stock. Still looks ~5c hotter than stock for the temperature comparisons.

Thermal paste applied to ONLY the heater core (thermistor still unpasted)

I wanted to see if there would be any significant improvement for pasting one component over the other with these tests. For the heater core only, it seems to slightly improve the fluctuations when changing temperatures. Benchy tests still look identical to stock (now completely indistinguishable). Still ~5c hotter than stock.

Thermal paste applied to ONLY the thermistor (heater core still unpasted)

This is where we see the major anomaly show up with heater block 2 -- the nozzle had become untorqued, resulting in temperatures ~10-20c below stock according to the temperature tower results. With that said, the actual temperature response graph appears unaffected. Both of these blocks appear indistinguishable from stock -- all temperature fluctuations over stock have been eliminated. Block 3 actual temperature does still look to be slightly hotter than stock, but not quite ~5c.

Thermal paste applied to BOTH the heater core and thermistor

Full pasting. Everything looks as good, if not tighter than stock results (less overshoot, more stable stabilization). Still ~5c hotter than stock for temperature tower tests.

Conclusion

Is Boron Nitride paste required? No.

Non-pasted results show near identical real-world response compared to a full-paste application. There is slightly more temperature variation when stabilizing to different temperature targets, but nothing that would realistically affect anything, especially compared to the stock heater block performance.

Would I recommend applying Boron Nitride paste anyway? Yes, if you want to be absolutely sure of the conduction.

I cannot speak to the non-pasted results here being representative of every configuration of machining tolerance between all 3 components, but especially considering the "loose" results still show very respectable results, it seems that the machining tolerances are good enough that it's a non-issue. That said, for my own peace of mind, I am a lot happier having fully-pasted blocks.

But what about that +5c temperature difference over stock?

I suspect that this is something to do with the stock heater block itself, and possibly some effect of actually being used for 500+ hours already. Any reduction in thermal contact at the nozzle is already shown to have a significant effect on actual print temperature (just going off of the torqued/non-torqued results). All of the tested blocks are very consistent between each other, so the comparison between themselves for paste vs non-paste shows there is no shift in output temperature.

Extra thoughts

The convenience factor of actually doing any heater block swap is obviously a major question. For myself, after going through all of these swaps for every test, it's now a trivial thing. I also very much enjoy the fact that I don't have to worry about doing anything hot -- yes, it takes time to heat the nozzle, unload filament, cool back down, swap the whole block, then reload filament... but I can just go about my day and not feel pressured to do a hot-tighten ever again if I don't want to. It's the same reason I love the Revo system.

For those that don't mind hot-tightening and are using a combination of nextruder nozzles and V6/adapters, then you really only need 1 extra block assembly for full convenience and time efficiency. One you can dedicate to the nextruder nozzles (swap nozzles while cold), and the other can be dedicated to the adapter setup (swap hot).

If anyone is looking to do this and wondering how much paste they actually need... I picked up a 5cc tube and haven't even used 1cc for the 4 blocks fully pasted.