The water cooling Part 3 and z-axis problem

Finally the last missing piece arrived.

When I screwed the fitting in the water pump I recognized that there is much play in the thread. One option is to use teflon tape, but this could split the plastic body from the pump if there is to much pressure on the threads.
So I decided to try an O-ring to seal the connector to the pump. I used a 10x1.5 mm o-ring that fits in the pump and seals towards the front of the pneumatic connector.


I had ordered an adapter cable with a plug that fits on the connector from the Water pump. As this is originally made for PC fans, this is pretty much standard. I was only interested in the part on the right in the picture, as this was hard to find.

As this is all standardized for PC fans, this is the pin layout for the 4 pin adapter. 

I want to run the pump always at full speed, so I don't need to connect the PWM signal(control). Anyway, I don't have any logic board to control it, I strongly recommend to not try to adjust the pump speed based on the water temperature. I explained this in detail in an earlier post.

The tacho signal(sense) will be used as input for the Einsy board, So the Einsy board will monitor if the pump runs, just like it usually does with the hot-end fan.

This is the logic diagram of the cable I made.

For the Temp/Fan control board I use a separate cable from the power supply.

This cable basically connects the ground from the power supply to the water pump and the fan 2 output on the Einsy board.

You will need a 3 pin Molex connector for the connection to the Einsy board. This cable need to be long enough to connect back to the Einsy board.

The 12 Volts DC from the power supply connects to the power (pin 2) on the water pump. The sense or tacho signal from pin 3 on the water pump connects to the fan 2 output (pin 3) on the Einsy board.

I also consider to use the 5V signal for the hot-end fan to switch the pump on/off. But I have not decided yet if I want to spent the effort. I guess then I would make sense to come up with a little perf board to connect everything together.
For now pin 2(middle) of the connector to the Einsy board fan 2 output is empty.

Make sure you don't mix up the pins, that might damage the Einsy board! You can see the 5V+, - and S(sense) for the fan  connections in the diagram below.

I have also made a small connector in the connection tot he power supply. So I can detach the cable from the power supply for my convenience.

This is the back of my printer now. I need to clean up the cabling a bit. The temperature sensor is just clamped between the radiator and the mount. I still need to find a better way to measure the water temp of the return flow. But the diameter of the sensor is 5, if it would be 4 or 6 I could simply use a T-shaped pneumatic connector.

Some words about the Molex connectors. I have got some from Aliexpress, but I cannot recommend them.

This is an original Molex connector. There is this little pin that locks the connector in the plastic housing.

On the ones I got from Aliexpress there is this pin missing. But you can bend over this part backwards to lock the pin in the housing. I am not sure if this intended to do so, but that worked for me.

But you should rather get the original Molex connectors. Usually you need an expensive crimp pliers, specially made for those Molex connectors. But with a little dexterity you can crimp them with some pointed pliers as well.

Not it was time to test if everything worked as expected. So I heated the hot-end up, but had the cable with the tacho signal not connected tot he Einsy board.

After some seconds the printer turned off the hot-end heater and displayed extruder fan error.
Just as expected and how it should be. This will be my fail-safe if the printer is on and the water pump is not running, the printer will automatically turn of the heater.

Then I connected the cable to the Einsy board and repeated this test. As you can everything worked just fine.

I was happy so far, but not with the "silent" water pump. This pump is anything but silent, at least when you run it on 100%. But see for yourself. I need to get a real silent pump or this thing will drive me crazy.

Now I wanted to see if I can run any of the calibrations of selftest from the original Prusa Software.

Here the problems began with the z-axis. While I can move tit around without any issues when I manually move the z-axis. Every attempt for full or just z-axis calibration the z-axis did not move, the z motors just hummed.
The problem is that Prusa uses no limit switches, even though the Einsy board does have connectors for it. The original Prusa software measures the motor current to determine that they reached a hard limit. They also seem to limit the current used for the upward movement during the z-axis calibration.
In my case this was not enough starting torque to get the motors moving. With a little help they started to move. So I tried white lithium grease on the z-axis shafts and the lead screws, then I moved the z-axis up/down over 20 times. But that didn't help much.
Then I loosened the upper mount for the lead screws and tried without them. Just to rule out that the width of the x-axis was ok and not bending the lead screws while moving.
Now it works every second try. But I have only problems with the calibration, if I move the z-axis from the menu everything runs smoothly. I guess the brurolls motors just need a higher current. I have read reports from other user that have problems with the Einsy board and the z-axis stepper motors. They reported to wire them in series and use just one output on the Einsy board improved the problem for them. At least with marlin they got it to work this way, but not with the Prusa firmware.

For the next build I am going to order the new stepper motors from LDO, same brand that Prusa is using. They announced an optimized version of motors that don't get as hot as the ones that Prusa uses on their original Mk3s.

So finally the printer started the z-axis calibration, but on position 3 of the bed leveling it failed again. But this time everything was ok, the P.I.N.D.A probe was already past the end of the heat bed.

As the x-axis calibration fails as well, I am not sure if I have the same motor current issue on the x-axis as well or if this is caused by the position of the P.I.N.D.A probe in the Titan Aqua mount.

The next time I am going to get the source code of the Prusa firmware and make the adjustments required for my Titan Aqua Klon hot-end. Then compile the hex file and test it on this printer.

Prusa i3 MK3 aluminium X axis carriage

Today arrived my new CNC machined x-axis carriage that I ordered from Aliexpress.

It is a 2 part item, just like the original 3D printed version from Prusa.

They use some plastic washers from Nylon or POM as spacers.

It came with 2 LMU bearings and had alle the required holes and cavities for screws and cable management like the original one.

As all the other CNC parts I ordered, this one came also well made and the anodization was flawless.

So why is this one better then the other ones, like this one I own one already.

Well as you can see the form is much simpler and optimized for a Titan Aero or Titan extruder.

So this one is not bad, but as always it depends what you want to attach to it. In my case I am already sourcing parts for my next project. An MK3 klon with the new Bondtech BMG Extruder Upgrade for the Prusa printers. And this Bondtech upgrade requires to have the same x-carriage as the original from Prusa. Of course you could make an adapter, but why spent the effort?

You may also ask why the 3D printed part is not ok and I want the alu version. Well, as with all the other parts, they are just more stable as the 3D printed versions. Not that there is a lot of force in 3D printing required, but you have vibrations. So a more stiffer setup means less vibrations.
This is the reason why many do the Bear or Haribo upgrade on their Prusa printers.

One disadvantage is that the alu parts are heavier than the 3D printed versions. But you cannot have everything, right?

I have also asked triangle lab if they will klon the Bondtech upgrade, but they replied that there is not enough demand that it would justify a mold for injection molding of the extruder parts.
Given the price of around 110€ for the original upgrade kit form Bondtech including a pancake stepper motor for the extruder, I guess there is not enough money to make for the Chinese knock-off manufactures.

Power on for the first time...

I needed to make my own power cables as the ones from the kit I ordered are too short for the PSU to mount it under the Ikea Lack enclosure.

Once I had solved this little problem it was time to power up the printer for the first time. I connected everything except for the hot-end heater cartridge and the hot-end temp sensor.
But before I connected the power cables I checked the correct assignment of Vcc(+) and Ground(-) and the Voltage of 24 Volts with a multimeter.

When I turned the PS on, the screen lightened up and the green LED on the Einsy board as well. But no text on the screen. But the reset button seemed to work as the print fan started up when I pushed it.
I turned off the PS again and disconnected most of the cables, to ensure that not a problem with the cabling was preventing the printer to boot up.
then I double checked the 2 cables from the LCD, everything seemed to be right.
So I gave it a 2nd try, but still no text on the screen.

Then I thought that the LCD display might be broken. I did have another one lying around here(for the next project) and connected it. Still no change.
As next I connected one of the chinese LCD displays to my original Prusa MK3 and tested this.
Nope, no text.

I started to wonder if there is anything different on the LCD displays that Prusa is using, from the ones that I got from the Chinese seller on Aliexpress.
Just for comparison. On Aliexpress you pay less than 8 € including free shipping for the 2004 LCD Control board including the adapter for the Ramps 1.4 boards.

Prusa charges you over 58€ for the same board with no adapter(which is not required for the Einsy board) plus shipping! There must be a reason for it, right? NO!

Everyone that is not an electronic expert right fall for this and order one, don't. But also no reason to be ashamed as I was considering this as well.
But don't panic, I am an electronic engineer!

The solution is very simple. On the backside of the display there is a little potentiometer to adjust the contrast of the display. No idea why, but on all of mine the contrast was turned down until there was no text visible on the display.

You need a very small screwdriver with a Phillips head to adjust it. Fortunately you can reach this with the LCD installed on the printer. Just lay the printer tp the side and adjust it carefully until you can read the text on the screen. Make sure you don't break it or shortcut any connectors on the LCD board while doing that.

Once I did this my Prusa wanted to take me thru the configuration wizard.  But I declined as I need to make more changes to the printer.

But I could test that the stepper motors for the 3 axis where moving correctly. I have seen some people complaining that they needed to change wires on the steppers. For me everything was working out of the box.
Then I checked the filament sensor, which was recognized. This one is tricky, be careful when you connect it to the board as the connector does not fit to the one on the eons board. It is the same problem as with the original Prusa.

Then I tested if the P.I.N.D.A probe was working. Compared to the original one you can barely see the LED from the sensor.

You can test with any metal, like a screwdriver if the sensor works. If you hold it under the sensor and it is detected the light goes off.

I need to finish the water cooling before I start testing the hot-end. But I got the last missing coupler for the tube today. So you can expect to see another update soon.

Once everything is tested with the original firmware I am going to show you which changes you have to make to get the Titan Aqua klon to work properly and how to compile the firmware.

Enclosure Part 2

After comparing prices and doing some research I finally ordered my panels for the enclosure made out of MAKROLON® or Polycarbonate. This does not break as easy as Plexiglas® or Acrylic.
It was also slightly cheaper. Even more cheaper would be PET pannels, but I have no experience with them yet.
For the panels I paid >50€, but I forgot one door. As I extended my enclosure for 150mm in height, I needed to change the height of the panels accordingly.

3pcs: 440 x 590 mm, thickness 3mm
2pcs: 220 x 590 mm, thickness 3mm

To my surprise the panels did not align with the sides of the table legs. I spent all the effort to extend the flute for the panels over the full length of the extensions, and then they barely touch them.
Yes, I cold have ordered panels 4-5 mm wider, but then spent another 50€? I was not happy with the construction anyway. I felt that those panels are a bit wobbly, especially as I extended to 59cm.

So I decided to print some mounts for the sides. As the Prusa cannot print that large object in once I decided to go with a 2 piece approach.

I printed them in Prusa Orange which gives me a nice contrast to the black Ikea Lack table.

The other thing I changed was the feed through for the cables. The original Prusa design has one feed through on the left back corner for the cables from the power supply to the printer.

I basically mirrored the parts to have another feed through for the water cooling tubes to the hot-end and the 12V cable for the LED lights.

Then it was time to mount the parts under the lower Ikea Lack table from the enclosure.
What you can see from left to right are the 24V printer power supply, the radiator with the fan on the back, the temperature control box for the radiator fan and on the right there is the 12V power supply for the water cooling and the LED lighting. You cannot see the water pump and reservoir, but they are behind the right table leg.

This is the backside. You can see the water pump which is directly screwed on the water reservoir on the left side. There is a 8mm PTFE tube from the outlet of the radiator to the inlet of the water reservoir. The 4mm nylon tubing from the hot-end will be directly attached to the radiator and the water pump. I was thing of using a bigger diameter tubing and then reduce the diameter closer to the hot-end, but I think that would not make a lot of difference in flow and pressure that this would have any effect on the cooling.

This is how the water cooling tubing is guided out of the enclosure.

And of course there must be a problem again. This time it is the cable kit for the power supply. As you can see the cables are too short for the PS to be outside of the enclosure as intended. They can barely reach the Einsy case, even if I don't use the feed through. So I have to make my own here as well.

But overall this looks already nice. Maybe I have to shorten the water cooling tubes. But I fear they might put some force on the hot-end if they get too short as they are so stiff.

Yes, you see right. The MMU 2 is already on top. It is not yet connected to the printer, neither electronically nor mechanically to the hot-end. But I is ready for testing already.

Update on my negative experience with "Dave 3D TECH Store" on Aliexpress.com.

So after all the back and forth over the dispute, every time he responded it requires my action and then he got another 3 days time to react again, he finally missed a deadline to respond.
Then the customer support of aliexpress jumped in, but only to inform me that they have extended the time to act for the seller again.
I was upset, but there is no way to escalate the dispute. Then they set a refund of € 47,02 or a full refund if I send back the part I already got(sure have fun with that, no address of the seller and the shipping cost might be enormous).
So I agreed to the proposal, but felt not really happy that I get not even 1/2 of my money back.
Then he got time to agree, which he obviously did not. But after some days the support closed the dispute and ruled in my favor. So at the end I got at least some of my money back.
One week after the dispute was closed I got a refund on my credit card.

Well nothing left for me todo, but leave a bad rating for him and hoping that others do not fall for his line.

New Parts and Fan Problem Part 2

In a previous post I did mention that you should not use the radial fan from my initial parts list, as it does not have a tacho signal(it has only a 2 pin plug instead of a 3 pin). As I could not find a 3 pin version in 5 Volts, I ordered mine from Prusa. That was the only original Prusa part so far.

Another User from the reprap forum send me a link that you can get 3 pin fans meanwhile from aliexpress. I ordered these, but you find more listings on Aliexpress now(seems like the sellers read my blog? :-) )
2 x fan5015-50x50x15mm 5V 11,50€

So that problem is solved, you can get a clone part for this fan as well. I did not get mine, so I cannot tell anything about the quality yet.

Another item I found is the Prusa X-carriage in alu. I know, there are plenty, but this one seems to be an exact copy of the 3D printed part from Prusa, not just a square block.

1 set Prusa i3 MK3 aluminium X axis Prusa MK3 25€

I ordered one, not sure if I will use it i this build or the next one(oh yes, I am gathering parts for the next build with the Bondtech BMG Prusa Upgrade hotend, so stay tuned!).

One thing I can say about those aluminium parts on this built is that the everything or more stiffer than the 3D printed parts. Ok, usually there is no force on the hot end, and this should not matter. But in some cases it might help a lot with vibrations. I tested some infill pattern on my original MK3 and the whole Ikea enclose was shaking.

The water cooling Part 2 and new cable holder

In a previous post I explained how I wanted to set up the water cooling. As I want to have a very quiet printer I decided to use a PWM controlled fan from Noctua with the RPM controller based on the temperature of the water flowing back from the hot end.

First I wanted to design and build my own controller, but then I found this one on ebay.
DC 12v 24V 48V 2-Way 4-Wire PWM Temperature Control 7,85€

I cannot beat the price and so I ordered it. It came with 2 temp sensors. You can attach 2 sensors, 2 fans and a buzzer for an alarm that you can configure.

But now I need a case for this, one that I could simply screw unter the Ikea Lack table, so that I can see the digital temperature display. Well, I have a 3D printer and Fusion 360!
So I designed a simple case that could be mounted in a 90° angle under the table.

Then I printed it. Does not look nice, but I was in a rush. Also I have problems with the black PETG Spool from Prusa. Not sure what the problem is, blue, white and Prusa orange prints fine.

When screwed the circuit board on it with M3x6mm screws, I need to be careful not to bend the circuit board and break it. Usually I would have added some stand offs, but then it wold be more difficult to print.

The cover fitted perfectly, even the holes for the fan and sensor connection had the perfect location and size. Measure twice, print once!

The other element is the radiator with the fan. I need to print a mount for the radiator as well. Again I designed one in Fusion 360.

Printed it looked like this with the radiator on. As you can see, the idea is to clamp the mount between the radiator and the fan. The radiator has 4 holes with M3 threads and the screws are supposed to go thru the fan and the mount in the radiator.

And this came directly from Austria, the Noctua 120mm NF-F12 PWM fan. I explained already in the earlier post that why I decided to use this fan. It has more static pressure and is designed for the usage with radiators.

The way I mounted the fan to the radiator the air will be sucked thru the radiator. I am not sure if it would be better to push the air thru the radiator. In that case I need to flip the fan around.

I also tested the water pump meanwhile. It is a PWM controlled pump and has a 4 pin connector like a PWM fan. So in theory, if there is not PWN signal the fan/pump should run on full speed. This is why you can connect a 3 pin computer fan to a 4 pin connector on the motherboard and it will work, just without speed control.
The pulse-width modulation (PWM) control signal is a square wave operating at 25 kHz, with the duty cycle determining the fan speed.
The pump is supposed to run always at full speed. If I would PWM control the pump, then the hot water might get stuck in the hot end while I measure the water temperature somewhere in the return flow. But without the pump running the temp would rise in the hot end, but not the return flow.
But the fan for the radiator can be RPM controlled, based on the temperature of the water in the return flow.
And as a side note. The sense/tacho signal from the pump will be connected to the signal input from the Einsy board that usually monitors the speed of the hot end fan. If the pump fails, the printer will shut down the heaters. Just like the original design with the air cooled hot end.

I was basically read to put everything together, but then the PFTE hose came with hard knicks that makes them unusable. This is not the regular 2/4mm PTFE tube that is used for the 1,75 filament in 3D printing. This one is 3/4mm. I was hoping to gain more flow by the bigger inner diameter.
But this made me think twice if the material and wall thickness of the hose is the right choice.

So I ordered a 2/4mm Nylon hose. Honestly this look like regular PVC hose that you normally use in fish tanks. I did find some PVC hose with a Nylon mesh for reinforcement, that is used for compressed air, but not in 4mm. The picture from the seller on ebay shows Nylon printed on the hose, but signs of a mesh.

Then I thought about how to attach the hose to the printer. My idea is that both hose go upwards from the hot end, over the frame and down in the back. As my power supply with be mounted under the Ikea Lack table anyway, I need the bracket to enforce the frame. So I designed one that could clamp the hose and bend them backwards without hard knicks.

And mounted to the printer like this, instead of the power supply.

With Meshmixer I found the best way to print my new cable holder design. My design leads the cables in a 45° upwards instead of straight back to the printer. Why? Because if you printer several objects in series, there is a lower risk you hit something with the cables.