Project Logs

Collapse

• Discoloration

  Dominik Meffert • 08/13/2025 at 06:07 • 0 comments

  When mixed with pure ethanol, the sandarac ink has a slightly yellow color depending on the mix ratio, but over the last few months, the ink inside the printer turned green, which shouldn't be happening.

  Since yellow mixed with blue gives us a green color, I suspect that the blue color from the housing of the ink tank got dissolved by the ethanol and mixed with the ink. This already happened before when I used blue colored PU tubes for testing out the falling ball viscosimeter. After a while the transparent ink turned blue from the color of the tubes.

  Since the green color is an unwanted ink contamination, I started cleaning the system by flushing out the green ink.

  Here you can see the green ink compared to highly concentrated sandarac ink.

  To prevent discoloration in the future, I got a new tank made out of white Polypropylene.

  Unfortunately, the new one is no longer transparent. To still see the fluid level of the tank, I added a fluid level indicator to the side of it (just a 4mm PE tube).
  

  When I disassembled the old tank, I found fittings with a suspicious blue color on top of the PTFE tape close to the tank.

  After I connected the new tank I flushed the system with pure ethanol and loaded it with fresh ink.

  After loading the new ink I let it sit inside the machine for a week of vacation.

  When I came back, I saw that a certain part of the tubes had turned green again.

  The chances that the blue color from the anodized aluminum had also turned the ink green were lower than that of the tank housing, but to eliminate all risk, I ordered new fittings to replace the blue anodized parts.

  When I disassembled the fittings, I saw that there was also a lot of blue residue inside them. Since the ink is not corrosive I could rule out that it was a sign of brass corrosion. Instead, I think it could either be color from the anodized fittings or just a buildup of the blue pigments from the tank housing that got stuck on the fitting walls.

  When the new fittings arrived I once again cleaned the printer and loaded new ink into it.

  I mixed 500ml of new sandarac ink and loaded half of it into the printer.

  But when I came back to the machine a day later, the same tube turned green once again.

  I'm currently suspecting that there is still some residue on the walls of some fittings, which turns the ink green, and that after more flushing, it will disappear at some point, since there are no more colored parts in the system.

• Motorized Pressure Regulator

  Dominik Meffert • 07/23/2025 at 05:15 • 0 comments

  For the last weeks, I was searching for a way to adjust the ink pressure via software, so that the printer can react to changes in room temperature and viscosity to keep the jet velocity constant.

  Until now, I used a pressure regulator to adjust the pressure by hand while looking at the jet velocity indicator. For controlling pressure via software, it was necessary to either get a new pressure regulator that could accept a control signal or add a motor to the existing pressure regulator. Since I had problems with other pressure regulators getting stuck from dried ink in the past - the ethanol sandarac ink is very sticky and could easily be used as a medium strength glue - I preferred to stay with the pressure regulator that worked and so I started looking for a way to add a motor to it.

  While commercial CIJ printers usually use a PWM-controlled ink pump in combination with a fixed size restriction in a feedback line, this is currently not an option for this project, since these specialized pumps are very expensive and the cost for building this printer should be as low as possible with no specialized parts. Because of that, I'm currently using a cheap ULKA EX5 24V vibratory pump in combination with a relief valve for setting the pump pressure (usually to 50 psi), a pressure regulator for setting the ink pressure (varying between 30 and 40 psi), and a damper that helps keep the pressure stable.

  I started by testing out some geared DC motors I had at home and quickly realized that the pressure regulator model I'm using needs more torque to turn than the motors I had could provide.

  Because of that, I looked for something that could provide more torque and got an old windshield wiper motor.

  The motor had enough torque to turn the pressure regulator, and I could control direction and speed with an H Bridge. After trying the motor out for a while, it turned out that the speed of the motor could only be reduced to some extent until the motor failed to start up, and that the motor coasted for a bit when not actively braked, so that the precision was not that high with this setup. Another problem with it was that there was no way for the software to know the motor's position, so that the motor could overtighten the pressure regulation screw and damage the pressure regulator if the pressure sensor reading or TOF reading were incorrect.

  Because of these problems, I decided to use a Nema 23 stepper motor in combination with an NMRV030 50:1 gearbox and a 3590S multiturn potentiometer to get more precision and position feedback.

  When the gearbox arrived, I 3D printed a shaft adapter and mounted the motor onto the gearbox.

  With the adapter in place, the gearbox was ready for testing and designing a mounting bracket for the multiturn potentiometer.

  I 3D printed a plate with a hole in the middle for holding the potentiometer, and four holes on the side for mounting it on the gearbox with the help of some spacers to have enough space to fit in a 3D printed adapter that connects the potentiometer shaft to the gearbox's output. The potentiometer was now able to track the exact position of the gearbox over multiple rotations, and it was possible to use an Arduino Nano for controlling the stepper motor driver and reading the potentiometer to limit the gearbox's movement to a certain range.

  I uploaded some code to the Arduino that continuously moved the shaft two turns back and forth to see if everything kept working over a long time, and it turned out that the PLA shaft adapter failed after a while due to wear, so I had to buy one that was made out of steel. Because the steel adapter...

  Read more »

• Vertical Layout

  Dominik Meffert • 06/03/2025 at 05:27 • 0 comments

  New Layout
  

  Over the last few days, I completely disassembled and reassembled the printer to tidy up the wiring, plumbing, and component arrangement to use previously unused space. To reduce the overall footprint of the machine, I changed the layout from a horizontal to a vertical design, placed the tank and bottles in front of the machine, and also removed all parts from the backside of the grid. 

  With the new layout, the machine now has a similar width and depth as the typical 3D printer - it's just a little taller.

  All of the components are now mounted on the front of the grid, and with that, also all of the weight, so that the machine only needs support feet at the front side since it wouldn't tip towards the backside.

  This way, it's now possible to place the machine pretty close to a wall to reduce the space it takes in the room.

  Components

  With no longer any components hidden at the backside, it's now possible to see all used components at once, which I hope will make understanding the function of the machine easier.

  Here is a small list of the machine's components, starting from the bottom of the machine:

  Flush bottle, solvent bottle, ink tank with vent and return line inlet at the top and in and outlet at the bottom of it, silicone tube with check valves, damper, pressure pump with fan, gutter filter, fluid distribution assembly made of low-pressure side with inlet from relief valve, TDS sensor, temperature sensor and outlet/inlet for draining, maintenance valve, and high-pressure side with inlet from pressure regulator, pressure sensor, connection to damper and connection to ink valve.

  Starting on the left of the picture:

  Relief valve, pump pressure gauge, and pressure regulator. Valve assembly on the right with the solvent pump and return pump above it.

  Valves from left to right are the ink valve, vacuum valve, gutter valve, flush valve, and tank valve.

  Almost all of the electronics are located in the middle of the grid to keep it organized. The only exemptions are the lines leading to the sensors, pumps, valves, and printhead.

  Starting on the left of the picture:

  TDS sensor amplifier, BTS7960 for pressure pump, LM2596 for solvent pump, LM2596 for return pump, WAGO distribution for I2C, WAGO distribution for 24V, USB power supply for LCD, AMS1117 3.3V for 3.3V to 5V level shifter, LM2596 for 5V logic, WAGO distribution for 5V, ADS1115 for the sensors, DS3231 for time and date, MCP23017 for digital IO, 8ch relay module for switching the valves, WAGO distribution for 24V and GND of the valves, main power supply 24V 10A.

  I think this is the first time I listed all of the electronics used for fluid management together since they were spaced all over the machine before.

  At some point in the future, the printhead driver electronics and high-voltage power supplies will also be added to the machine.

  I'm currently working on these but it will still take more time to get them ready. TOF is working; Phasing will be next.

  The LCD is now mounted on a tablet holder at the top of the frame.

  I think with the new layout the CIJ printer prototype now looks a bit less experimental than before. My plan for the future is to place the machine on the floor next to a plotter or other machine on which the CIJ printhead can be mounted for applying ethanol-based ink, varnish, or glue to a surface.

  With that, optimization of the fluid management part is done for now and I will continue working on the printhead driver electronics, so that I can hopefully soon get the phasing to work.

• New Valves for a compact Design

  Dominik Meffert • 05/19/2025 at 01:28 • 2 comments

  For the goal of a more compact design, I spent the week modifying some 3V1-06 valves.

  For that, I drilled 5mm holes through the two mounting holes which are usually meant to be used with special screws, to be able to stack five valves and connect them with two M5 threaded rods.

  Far accessing either the inlet or outlet side I cut M5 threads either close to the edge for accessing the outlet side or in the middle of the valve's front side for accessing the inlet side.

  I then connected M5 fittings to them for the tube connections.

  The ink, tank, and flush valve are closed with 1/8-inch plugs, which are sealed with PTFE tape.

  To not block the new connection on these valves I placed a cut-in-two washer at the bottom of the thread so that the plug stops far enough from the bottom to keep the new connection open while sealing the 1/8 inch side connection.

  The vacuum and gutter valve are connected via the 1/8 inch side connection which is sealed by a 12*3mm O ring.

  Two tubes are used for connecting the ink valve's outlet to the vacuum valve's inlet and for connecting the flush valve's inlet to the tank valve's inlet.

  On the inlet side, I used 1/8-inch plugs with M5 threads cut into them as adapters.

  To suppress the voltage spike that occurs when the valves are turned off, I added an IN4007 diode to each valve's terminal connection. 
  

  After assembly, I connected everything and checked for leaks.

  For that I let the machine run while tidying up the room.

  After sealing one leak, new valves seem to work just as well as the old ones while looking much more organized and compact.

  While connecting the new valves I also cleaned up the wiring a bit.

  I think this update is another step towards a compacter design.

• Reducing Complexity

  Dominik Meffert • 05/12/2025 at 20:02 • 0 comments

  While there is still much work to do on the charging electronics, the current fluid system design can be considered complete. It's been working reliably for about a year now, with only minor changes from time to time to make it even more reliable.

  So, I thought now would be a good time to remove parts that are not essential to reduce cost and complexity. Currently, the frame has a large size of about 100*50cm, which I plan to reduce to at least half its size over time.

  Falling Ball Viscosimeter
  

  One thing that sometimes caused problems was the falling ball viscosimeter. When not used for a while, it happened that the steel ball got stuck at the bottom of it, and it was needed to disassemble it, clean all parts with ethanol, and assemble it again. Since the falling ball viscosimeter became redundant by using a ratio of pressure divided by jet velocity for the viscosity reading, I thought it would be better to remove it to have one less point of failure on the machine. So, I removed all parts of the viscosimeter, including the 2 inductive sensors, the polycarbonate tube with steel ball, and the fittings. 

  Peltier Heater / Cooler

  With these parts removed, there was still the brushless pump for circulation and the Peltier cooler/heater left. The circulation line is used for mixing the ink with solvent when the jet isn't active. It also has a higher flow rate than the inkjet and is used for pushing the ink through the main filter to keep it clean. So, the circulation line had to stay.

  The heater/cooler was used for keeping the ink at a reference temperature of 25⁰C when using the viscosimeter because viscosity changes with temperature. With the viscosimeter removed, there is no need to keep it.

  To still be able to take temperature changes into account, I finally added a temperature sensor to the nozzle assembly which reads the temperature as close as possible to the ink jet before it performs the TOF reading.

  Maybe this can later be used for temperature compensation of the viscosity reading.

  Brushless Pump

  Now the big question was what to do with the brushless circulation pump.

  At the time, I decided to remove it and replace it with another peristaltic pump to only have two different sorts of pumps on the printer instead of three. Later I found another solution to it.

  External Pressure Pumps

  Another problem that I wanted to solve for a long time was that the pressure pumps were located outside the printer frame. I used four small pumps instead of one large pump because when used as intended, they are only allowed to run for two minutes until they need a one-minute break to prevent overheating.

  To solve this, I decided to use one large 24V vibratory pump and run it on a lower-duty cycle while cooling it with a small fan.

  The pump I used for that is the ULKA EX5 24V, which is normally powered by AC. However, since these pumps already come with an integrated diode, they use only one half-wave of the AC cycle and can, therefore, be driven by pulsed DC without any problem. For driving it, I used the same BTS7960 H-Bridge that was used for driving the Peltier module before, because it was suitable for driving the pump, and I already had it wired to the printer controller and power.

  I used a signal with 50Hz and a 30% duty cycle to drive the pump. To keep it cool, I used a 40*40*10mm 24V fan, and to not transfer the vibration to the printer frame, I used two short pieces of gt2 belt and some zip ties to let the pump hang from an angle bracket. In addition to that, I used a silicone tube for the inlet and an NBR tube for the outlet because the vibration of the pump was transferred through the PE tube I used before, which caused a lot of noise.

  To protect the pump from overheating in case the pulse signal would freeze and expose the pump to DC for some time, I installed a bimetal switch that would...

  Read more »

• Jet Velocity - Time of Flight Reading

  Dominik Meffert • 03/22/2025 at 12:40 • 0 comments

  Relation between Jet Velocity and Pressure, Viscosity

  In CIJ printing, there are two important key values that can change during operation, so it's important to track and control them.

  These are viscosity and pressure which both affect the jet velocity:

  - An increase in viscosity leads to a decrease in jet velocity while a decrease in viscosity leads to an increase in jet velocity when the pressure stays constant.

  - An increase in pressure leads to an increase in jet velocity while a decrease in pressure leads to a decrease in jet velocity when the viscosity stays constant.

  Viscosity is affected by temperature changes and solvent evaporation. A higher temperature leads to a lower viscosity and a lower temperature leads to a higher viscosity. A higher solvent content leads to a lower viscosity and a lower solvent content leads to a higher viscosity.

  Relative Viscosity is measured by a falling ball viscosimeter and the ink temperature can be controlled by a Peltier module. The solvent content can be controlled by adding solvent via a peristaltic pump or by adding pre-mixed ink to the ink tank. Adjusting the temperature and solvent content usually takes some time until the solvent has mixed with the ink and the set temperature is reached.

  In contrast to that, pressure is not affected by external factors and can be measured with a pressure sensor and controlled with a pressure regulator or a PWM-controlled pump, which makes fast and precise pressure adjustments possible.

  Since a constant jet velocity is important for reliable operation and good print quality I added a way for measuring the jet velocity to the printer.

  Great thanks to Robert H. and @Paulo Campos for helping me with that.

  The jet velocity is measured by counting the Time of Flight (TOF) that an ink droplet needs to travel between two sensors that have a fixed distance between each other. The distance is then divided by the TOF to get a reading in distance per time.

  This new TOF sensor should make it easy to keep the jet velocity constant by looking at the reading while adjusting the pressure. Currently, I'm using a bar with a "proven to be good" value in the middle and a small good range around it while adjusting the pressure regulator by hand. In the future, this could be automated if needed.

  S0 and S1 are timeout indicators that can help with troubleshooting by indicating which sensor has a problem.

  I also added another reading which shows the ratio between pressure and jet velocity to get information about how much changes in viscosity affect the jet viscosity.
  

  Recent tests have shown that with the viscosity of the currently used resin ethanol mix, small changes in viscosity caused by temperature variation have very little effect on the jet velocity. Since the printer is used inside the house the ink temperature should also not drop much below 20⁰C and rise not much above 30⁰C.

  Based on this, I think changes in room temperature across different seasons will have less effect on the viscosity and jet velocity than solvent evaporation over time.

  In contrast to that, each psi difference has a noticeable effect on the jet velocity, so that a more precise pressure control would have a positive effect on reliability and print quality.
  

  TOF Sensor as Functionality Test

  In addition to measuring the jet velocity, the TOF sensor can also verify the function of breakup and charging. It relies on charged groups of ink droplets for time measurement, which is only possible if both processes are working correctly.

  When no signal is detected on the TOF sensor this can indicate the following problems:

  - A clogged nozzle

  - A misaligned jet

  - Problems with the piezo

  - Problems with the charge electrode

  - Problems with the TOF sensor

  This can help with troubleshooting and could be used for pausing the print in the future if a problem is detected.

  Building the TOF Sensor...

  Read more »

• New Ink for the Printer

  Dominik Meffert • 02/25/2025 at 06:17 • 0 comments

  Until now, I used a mix of polyvinyl butyral and ethanol as ink for the printer, and even though PVB worked well for regulating the viscosity of the mix, the dried-up ink layer had bad adhesion properties so that it was constantly peeling off into flakes which gave it a very dirty look. 

  Because of that, I started searching for other ethanol-based inks.

  Shellac
  

  The first thing I found was shellac, which was often used as an alcohol-based furniture varnish in the past. 

  So, I went to the hardware store to get a small bottle of it for testing out the surface finish.

  To test for the surface quality, I poured a small amount of it into the drip tray and let it dry.

  After waiting some time, I checked the surface quality, and it looked very promising.

  The shellac made a smooth, clear, shiny, and nonsticky surface with good adhesion to the stainless steel of the drip tray.

  Since the varnish is alcohol-based, it was also possible to fill up the bottom of the tray with ethanol and let it dissolve the shellac, which makes cleaning it up very easy.

  So, I ordered a bag of shellac for loading it into the printer.

  It turned out that the shellac flakes were easy to mix with ethanol, and depending on the concentration, the color reached from orange to dark brown. It also worked well to change the viscosity based on the concentration.

  When testing the charge + feedback signal it also worked as well as before.

  All in all, it can be said that shellac-based ink is superior to PVB ink with a nice surface finish and an easy preparation process.

  The only drawback of it could be the brown color and the fact that shellac is made by shellac lice, which could maybe raise ethical concerns.

  For this reason, I continued testing out other inks.

  Plant-based Varnish

  The next thing I tried was colophony/pine rosin.

  The good thing with colophony was that it dissolved very well when mixed with ethanol and the color was much more clear compared to the shellac.

  But when I then tested out the surface finish in the drip tray as with the shellac before, I quickly realized that this mixture would not be suitable for using it in the printer.

  While the surface finish looked ok at first glance, it turned out that it was rather glue than a varnish since it remained very sticky even when fully dried.

  For fast printing of text or binder jetting and also to prevent a sticky mess when working with it, it would be better to select an ink that fully dries as fast as possible while forming a smooth nonsticky surface. So, the search continued...

  Just to mention it, I also tried out gum dammar, which turned out not to dissolve in ethanol, and elemi, which dissolved, but the surface was more like wax than varnish or ink.

  Trying out solvent-based Varnish
  

  To leave nothing untested, I also tried out solvent-based varnish from the hardware store, which I think was important for gaining the experience, but I wouldn't recommend it or try it again, especially not in winter when opening a window brings the temperature down to frosty levels.

  I started with mixing 1l nitrocellulose-based varnish and checking how the mixture ratio influenced its viscosity.

  It turned out as expected - the varnish-to-solvent ratio influenced its viscosity.

  After working for some time with it, the smell got so horrible that I decided not to continue testing with it, since I'm planning to use the printer inside the same space where I live, like any other 3D or desktop printer.

  The next thing I tried was using alkyd resin-based varnish, which I mixed with white spirit "Reinigungsbenzin".

  It turned out that the...

  Read more »

• Test Stand and DIY Amplifier

  Dominik Meffert • 01/31/2025 at 04:32 • 0 comments

  My next goal is to measure the speed of the ink stream by using two sensors and counting the time the droplets need to travel from sensor to sensor.

  To make work on the electronics easier, I built a new printhead test stand with a drip tray for catching ink drops that miss the gutter.

  After everything was in place, I tried building an amplifier for measuring the small charge that would be applied to the droplets by a Time-of-flight test signal.

  To verify that the amplifier works, I did a quick test of the new amplifier by applying a 0.1V 20kHz signal and looking at the amplifier's output.

  The test showed that the amplifier is turning the sent 0.1V square wave signal into a 0.25V triangle wave.

  While I didn't expect the output signal to be a triangle wave and I also thought that the amplification would be a bit higher, the test showed that the built circuit was working.

  So, I built a sensor out of a UHF connector that I connected to the amplifier for detecting the charge on the droplets and applied a test signal to the charge electrode.

  Conclusion after the first Test:
  

  For now, it looks like it's possible to send a test signal to the charge electrode and receive a feedback signal through the amplifier.

  This is an improvement compared to the AD620 amplifier that I have used so far since the old amplifier only worked randomly, and most of the time, it didn't work at all. 

  The new amplifier, however, is working pretty reliably so far.

  Here is a short video of the test:
  

  At the moment of writing this, I already built another one and added a TDA2030 amplifier and a high-pass filter to each one of the two to further amplify the signal and filter out the 50Hz mains frequency.

  I also built a new sensor assembly made out of a PCB. The new one provides two sensors with a fixed distance between them so that the jet velocity can be calculated by using the known distance and the measured time delay between the signal appearing on the first sensor and the signal appearing on the second sensor.

  The next step would be building a peak detector to get the signals ready for feeding them into a microcontroller to measure the time delay.
  

  When that's done, it should be possible to get a stream velocity reading, which would be the first real progress on the printhead driver electronics.

  After that, the same circuit could be used to test for the phase shift that works best for charging, but the current amplifier design is too slow to detect groups of charged droplets at the needed frequency. 

  So, a redesign of the amplifier circuit would be needed before it's ready for this test.

• Making PCBs

  Dominik Meffert • 01/31/2025 at 01:39 • 0 comments

  Some time ago, I bought a pulsed 1064nm IR laser (Sculpfun IR2) for metal engraving and I recently figured out that this laser is the perfect tool for removing spray paint from copper PCBs without any residue.

  A year ago I tried the same with a common 450nm laser, but this laser rather burned the paint off while leaving a lot of residue on the PCB. 

  My guess would be that the pulsed operation mode of the new laser is the decisive factor here, which causes the paint to chip off or evaporate when hit with the laser instead of just getting burned away.

  PCB Manufacturing Footage:

  I think having figured out a way of PCB manufacturing that works well for me will help move the project forward and will also be very useful for future projects.

  Until now, I always had to build circuits on perfboards and connect the components with solder wire. This was a lot of work and also looked ugly compared to the new design.

  It's also an advantage that the circuit and PCB layout can be designed on PC for lasering it 1:1 onto the copper PCB. 

  This makes the process similar to 3D printing, which will make it easy to share the design with others and will ensure that each copy of it will be identical. 

• Quick Update

  Dominik Meffert • 01/17/2025 at 09:11 • 0 comments

  Recent Updates:

  Here is the progress since the last build log:

  - I figured out a way of designing and manufacturing PCBs that works well for me so that it will be possible to build the needed circuits for the project as proper PCBs.

  - I built a TL072-based amplifier by myself to replace the AD620 amplifier that I used so far, which never worked reliably. The new amplifier seems to work much better, while the amplification is currently very low.

  - I built a test stand for the printhead, similar to the drop breakup test stand, but out of 2020 profiles and only 500mm high.

  - I tried out many different resins for replacing the PVB and ended up using gum sandarac, which is plant-based and has better solubility, finish, and adhesion than the PVB.

  - I added filters to the printer to keep the ink clean and prevent the lines and nozzle from clogging.

  - I got a new oscilloscope that has 4 channels so that a trigger signal + the signals of the 2 new amplifiers can be displayed in the same timescale to verify that both work correctly.

  Will try to write more buildlogs with text and footage of the work I did over the last months when I can find the time for it.

  Future Plans:

  I'm currently working on a time-of-flight counter for measuring the ink stream's velocity. This will ultimately be used for manual and later automatic pressure adjustments to compensate for changes in viscosity which appear when the machine is used after being off for some time.

  Currently, the printer can measure a relative viscosity reading while heating/cooling the ink inside the printer to 25⁰C since the viscosity changes with temperature.

  Hardware-wise, there is also a peristaltic pump installed for adding ethanol to the tank to compensate for evaporated ethanol and keep the viscosity steady.

  While the feature is currently not activated, it would already be possible to use the viscosimeter reading for automatically adding ethanol.

  The stream velocity reading + (auto) pressure adjustment will be used for a finer adjustment in addition to the viscosimeter which can bring the ink stream velocity to the right level immediately.

  This, in combination with the viscosity regulation, the right piezo frequency, and the nozzle drive, will ensure that the stream breakup will always occur at the same position (inside the charge electrode). 

  Once a favorable piezo frequency, nozzle drive setting, and stream velocity are found, these settings can stay as they are since the TOF-based pressure compensation will keep the stream velocity constant at all times during operation and also when the machine is started again after being off for some time.

  Even the older CIJ printer models use a TOF counter, so I think this will be a very important step in getting the machine running.
  

  Once this is working, the next thing will be trying to get the phase test working again.

  When I tried that last time, I had no way of knowing the ink stream velocity, so I worked with random conditions and only got random results.

View all 61 project logs