Jump to content

Some comments on the design of UM2


Recommended Posts

Posted (edited) · Some comments on the design of UM2

Hi again,

Every bird signs with its own voice. Maybe I'm not the most diplomatic guy, I focus on the possible solutions. Please read the following in the positive spirit it has been written.

I got triggered to make this post because my UMO does not regulate again, this time it is not the "UMO PID formula error" but the FET transistor driving the heater. So it will be the second time I have to change this FET.

I looked into the diagram of the UM2 to see if the cause of the failing FET has been corrected there. It was unfortunately made the same way.

The curcuit is an AVR uC pin driving the FET gate through a 100ohm resistor.

This is not a good way to do the job.

This is basically a RC circuit. When the pin goes from 0 to 5V in a short moment there is 0 volts over the "gate capacitor" thus giving 5V or something near over the resistor.

5v/100ohm = 50mA.

According to the AVR ATmega2560 datasheet, the absolute maximum rating for a pin is 40mA.



Stresses beyond those listed under "Absolute Maximum Ratings" may cause permanent damage to the device. This is a stress rating only, and functional operation of the device or other conditions beyond those indicated in the operational sections of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability

From my experience exceeding specs has never been rewarded with reliability.

I have made approx. 30 designs with AVR processors since 2001.

A good rule of thumb is to be at 50% or below of max ratings.

This means that the resistor should be 270ohm. The FET will not like that, because of increased thermal stress due to long switching time like on my UMO.

I looked at the pcb layout, and fortunately there is space to solve this properly at a low cost and little effort, as there is pcb space available for 3 drivers.

In order to make a FET swith properly with low switch losses we are not talking milliamps but Amps of driver current.

The MOSFET gate driver  "ZXGD3004E6" is able to deliver up to 8 amps of gate driver current. It is very compact in a SOT23-6 case. I have used this many times and it works well. Here is an example:


Here the L6387 is a bit too weak to drive the FETs properly so the ZXGD3004E6 assists.

In your case replace the L6387 pin 5 with the uC pin and suppy U5 with 5V instead of 12V.

R18 can then be 220ohm and R17 and R19 can be selected for your taste in driving the FETS on the UM2.

This is a tested and functional circuit.

I will recommend you do a change like this or better, if you find a better solution.

Like my UMO has failed twice, the UM2 will fail also eventually. After all we are here to print, not to repair.


One of the first things I learned after I graduated, was from an older engineer. He said that whenever you use a mechanical switch, be sure to draw 5 mA through the contact set to prevent unreliable operation due to oxidation. The 5mA keeps the contact set clean without wearing it down.

I have never seen this mentioned in a book, and nobody told us at the university.

Since I have repaired many "temporary" button problems by adding resistors.

I will recommend you add pull-up resistors to the 3 contacts in the rotary knob, and also the card detect and the other switches get a smaller resistor in order to draw 5mA, i.e 1Kohm instead of 4K7.


I have a suggestion for the feeder wheel. I would like to have a knob on the wheel so I can feed filament manually like on the UMO. The stepper motor is so weak that you can turn the knob easily.


The very nice stainless steel thermal isolator, has only a few threads at the bottom.

It would be nice if the nozzle also had just a few threads near the block with the heater and temperature sensor. Then there no longer will be the need of so much turning and turning and turning..... to disassemble and assemble the nozzle.


The heated bed is all too weak mechanically for my taste. I have seen resonance with approx. 1 mm amplitude on the UM2 caused by the UMO printing next on the table.

I have seen up to 0.5mm bed amplitude measured at the front leveling screw when printing overhangs. This was measured with a dial gauge with 0.01 mm resolution.

I am going to make the heated bed mechanically more stable on both printers.

It would be nice if you changed this. It will surely reduce banding effects due to vibrations generated by the printhead causing over/under extrusion.


I would prefer a design with shorter distance between the x and y axes and the nozzle tip to minimize the position error due to free play in the linear bearings.


Each toothbelt introduces a positon error. I would prefer a design with stepper motors driving the rotating axle directly to eliminate half the position error.


If you made it to here, thank you for caring :-)


Edited by Guest
  • Link to post
    Share on other sites

    Posted · Some comments on the design of UM2

    One more idea.

    If all the linear bearings were adjustable, then we can minimize free play and get a better print quality.

    It can be made as an upgrade. New bearings and plastic parts for the printhead, and replacement bearings for the buildplate.

  • Link to post
    Share on other sites

    Posted (edited) · Some comments on the design of UM2

    One more thing caught my eye.

    The 3 instrumentation amplifiers ( INA826 ) for the PT100 sensors are loaded too much capacitively.

    The datasheet says:


    Load capacitance stability: typ. 1000pF

    Any amplifier with feedback is a regulating system in itself, where the bandwith and loop gain must be designed to have a sufficient phase margin at gain = 1 to avoid noise, or worse that the amplifier acts as an oscillator.

    The ouput are loaded with 100ohm in series and 10nF in parallel.

    10nF = 10000 pF.

    Even with 100 ohms in series, the instrumental amplifier is loaded too much and challenged to keep stable.

    The ADC input on the ATmega2560 has 100Mohm input resistance, so using 100ohm is not necessary.

    I will suggest using 2k2 near the amplifier output and 1nF near the ADC input pin.

    That will reduce the stress on the output stage in the amplifier and probably also reduce the noise on the temperature readings.

    Edited by Guest
  • Link to post
    Share on other sites

    Create an account or sign in to comment

    You need to be a member in order to leave a comment

    Create an account

    Sign up for a new account in our community. It's easy!

    Register a new account

    Sign in

    Already have an account? Sign in here.

    Sign In Now
    • Our picks

      • UltiMaker Cura 5.4 stable released
        The full stable release of UltiMaker Cura 5.4 is here and it makes it easier than ever to remove brims and supports from your finished prints. UltiMaker S series users can also look forward to print profiles for our newest UltiMaker PET CF composite material!
          • Like
        • 58 replies
    • Create New...