1

I have a Kossel mini set up with a heated bed

I had it printing well, but recently it will often get part way through a print before one of the belt motors starts making rapid “knocking” noises and the extruder head buries itself in the print

My first thought was that this was an error with the G-code, but I have ruled that out

My second thought was that the heated bed was causing the motor drivers to get too hot, but again I have ruled this out by monitoring their temperature - and also by ensuring there is plenty of thermal insulation between the bed and the circuitry

This problem seems to come and go - for instance it may happen on layer 1 or layer 100, and may disappear just as quickly. After the last occurance homing initially didn’t work (problem persisted)…then all of a sudden it worked perfectly.

I have checked the connections on all motors, and all seem fine. Unfortunately since the problem is intermittent, I haven’t been able to nail down for sure which of the three motors is at fault…

I wondered if the problem could be down to the PSU…I am running the heated bed (12v, 120W) and the motors from a single Dell DA2 PSU rated to 18A. This doesn’t seem to be overheating either…however I can still get the motors to skip with the heated bed disabled.

Measuring the 12v rail with the motors off gives me a steady 11.66v

This drops to a steady 11.45v when the motors are on - not sure what the tolerance here is…could be an issue?

Does anyone else have any suggestions?

2 Likes
2

I would replace the 18A PSU to a 30A one.

The dell PSU won’t overheat or get hot as designed not show that to the user as this will cause this

(Customer: “my laptop’s PSU is hot to touch, will it catch fire? I better phone Dell”)

It could hitting it cut-out value for current aka Amps as I would say 18 is bit low.

Each 17nmea stepper can pull 2A, your heated bed is 120w so using ohm’s law that 10A, your hot end could be about the same, so you see we are up to 20A just with hot end and heated bed.

Yes, it doesn’t pull 20Amps all the time but at times during the print if might, which could cause the MCU to shut down and reboot due to the voltage drop, thus causing the steppers lose where they are (go to zero)

3

Agreed The fact the voltage drops 0.2v when motors are energized would seem to hint at current limits… Will see if I can find a beefier supply

4

I have seen a Rep Rap Huxley run OK with a 19v (some low amp current, can’t remember exact value) laptop type brick PSU.

It was designed to provide a 40A peek for a few milliseconds, it got hot but that’s what made it work, as it didn’t skip a beat. I always stated that it was underpowered but other disagreed and it was fine.

So read the data sheet of the PSU to see if it mentions a peek current value as finding 12v 30A supply can be tricky. A good quality one that does 12v 20A might work for your set-up.

5

No luck finding a beefier supply in my box of bits I had expected a dell da2 18A supply to be enough - others have used them… Possibly the supply is not to blame, but the volt drop suggests it may well be. Running the high and low current tasks of the RAMPS from two seperate PSUs seems problematic… So ill try a beefier supply and cross my fingers EBay here I come…

6

Maybe you should also check;

- Have you used the right steppers?

- Check or replace - Stepper Drivers (A4988 or DRV8825)

- Check jumper configuration of the stepper drivers 1/8 1/16 etc…

- Firmware?, did you compile it yourself?, is the torque and speed set properly?

7

I will check all that

The steppers were fine until I added the heated bed and changed Marlin versions (to one that supported calibration) - so I don’t think the hardware is at fault.

I’m using this branch of Marlin:

And this hardware (slightly modified - Z-probe replaced with an inductive sensor, heated bed added)

https://www.think3dprint3d.com/Kossel-Mini-3dPrinter-Kit

I will try the new PSU later this week, if necessary start rolling back until I find the cause - firmware, heated bed etc…

It was printing perfectly onto heated glass right up until the motors went odd - I’ll get there again :slight_smile:

8

Might be a bad wire connection on the stepper if its always the same stepper doing it. If one of the phases becomes disconnected, the driver only powers the other phase, and the motor hops back and forth instead of rotating.

9

Hi there,

I had a very painful experience setting up my Micromake Mini Kossel. Though you may not have the Micromake version, there are a lot of other similarities in all chinese Kossels. Your PSU sounds fine to me. Here are some things I learned along the way, perhaps one or more will fix your issue:

1. Do not use the Micromake (in case this is what you bought) firmware. Grab yourself RichC Marlin (make sure its from his Testing Branch on github)

2. Replace the crap diagonal rods with some proper TRaxxas + CF tube, or magnetic rods.

3. Twist all of your stepper motor connections, and keep the end stop wires away from your stepper motor wires.

3. REDUCE the x,y,z acceleration in the firmware from the default 9000 to something far more reasonable like 500.

4. REDUCE delta segments part of firmware to 100. You won’t see a difference between 100 and 200.

5. Keep printing speeds low while troubleshooting. Increment them slowly. Mega2560 based controllers can’t handle the processing required for fast moves.

6. Check the current on all your steppers. This is usually a voltage signal that you get from stepper drivers. You can adjust it by turning the potentiometer. For instance on my DRV8825 1.0 Volts = 0.5 A. I set all tower steppers to 0.5 A and extruder stepper to 0.8 A. I might increase both by 0.2 A for when I turn up the speed.

7. Ensure that your belt tension is tight, but not too tight!!! It is very easy to over tighten them. If they are too tight you will notice a lot of resistance when trying to move the carriage up and down by hand. Don’t make them so loose that you skip steps. There is a fine middle-ground you are trying to achieve here. Also, don’t try to make them so tight that the belt gives you a low “guitar note” when you strum it.

8. Use RichC’s firmare to run an auto calibration by sending gcode G30A, and save with M500. You can type M666 L to list all of your constants, so you can post them into the firmware. Before every print, run G29 with the hot end at printing temperature.

Let me know how it goes!