Noise on temperatur sensors - MKS Gen 1.4

Hi.

I have a problem with my controller. I just upgraded my printer to 24 volts. I am using a 24 volt LED power supply and an MKS Gen 1.4 controller. After I've upgraded to 24 volts, I discovered a lot of noise on my temperature sensors. When I use 12 volts, there's hardly any noise (there may be a bit of noise, but I've only noticed it after I've started looking for it). When I just drive it on 5 volts via the USB port, there is no noise (beyond what you can expect from a normal thermistor).

I first thought it was the 24 volt power supply, but I have now tried two different (of different brands) and the noise does not disappear. I have also tried a MKS gen L controller, but here there is also noise.

I would really like to avoid buying a new controller as I am waiting for the Duet Maestro to come to the EU. Is there anyone who has a good proposal for a solution. Could I possibly change the controller so that it only gets 5 volts via USB?
I have attached pictures of the hotend sensor when running on 12 volts, 24 volts and 5 volts via USB.

5V:


12V:


24V:


Edited 1 time(s). Last edit at 09/18/2018 11:26AM by flydeprutten.

I noted when I purchased my second 3D Printer .
The first being a 3DP11 12v 30amp supply
The second being a 3DP12 24v 20 amps supply

I too starting getting this ripple noise in not just the temperature sensors but the filament run-out sensor and the RepRap LCD.
I used some aluminum foil DUCT Tape to cover the inside of my controllers case and grounded it.
I also wrapped the LCD Cables with Blue Painters Tape covered with aluminum foil DUCT Tape just in case I ever wanted to remove the aluminum foil DUCT Tape from the LCD cables.

Problem was solved on my machine.

Hope this helps

---

Computer Programmer / Electronics Technician

This is a small test setup. Here I use my MKS Gen L controller. The problem is not so pronounced on this card, but it can still be seen clearly. As you can see, there should be no possibility of any particular interference.





Edited 1 time(s). Last edit at 09/18/2018 02:46PM by flydeprutten.

You may need to re tune the PID for the hotend see link
PID

With the basic setup you have in IMG_7733.JPG you have almost nothing to cause the noise.

Do you have line noise leaking into the LED Power Supply
Can you measure the voltage on the Supplies output leads.

Edited 1 time(s). Last edit at 09/18/2018 09:56PM by Roberts_Clif.

---

Computer Programmer / Electronics Technician

Quote
Roberts_Clif
With the basic setup you have in IMG_7733.JPG you have almost nothing to cause the noise.

It's the same picture with my other 24v power supply.

Quote
Roberts_Clif
Do you have line noise leaking into the LED Power Supply
Can you measure the voltage on the Supplies output leads.

It's 23.9 volts

ok, just as an update if there are others with the same problem.

Inspired by this thread, I resolved the problem by removing the D1 diode on my MKS Gen 1.4, thereby eliminating the power to the card's onboard buck converter. Then I installed my own buck converter between the 24 volt and 5 volt pins (right next to reset button).

In this way, I assume I have gained a significantly more stable 5 volt on the card. (Unfortunately, I don't have an oscilloscope). At least I have obtained a more stable temperature measurement.

Hi,

I am having the same setup:
24V PSU and MKS Gen L. I have 4 MKS Gen L controllers and still the same issue on all of them. After giving a print command the print won't start as the temperature does not stabalize.
Are you sure its the buck converter?

We put the scope on the +5V supply (after F3) on the MKS GEN V1.4 and found a fair bit of switch mode PSU output ripple (100mV p-p), so then soldered a 1000uF 10V cap and a 100n cap on the back side of the board on the track directly under neath F3. The +5V ripple was reduced by quite a bit (down to approx 10mV p-p).

It's no surprise at all because the chinese do not fit decoupling caps if they can get away with it. They tend to remove as many as possible until the board fails, then add just enough. So they tend not have any ability to reject external EMI/RFI etc, and also can radiate an increased level of EMI.

Also fitted a 100n cap from the CPU AVCC pin (pin 100) to ground as it has no decoupling at all near the pin.

Attached are the before and after scope traces.

All DC-DC PSU/regulators really need an LC filter on it's output, especially when you have analogue measurements going on (the ADC), but unfortunately it's rarely done. At least they fitted an LC filter on the input to 5V SM regulator.

The entire board is lacking decoupling caps and I/O filtering and protection, not just here and there. But that's typical of cheap units.

Edited 3 time(s). Last edit at 03/03/2019 05:50AM by Pippy.

This is the noise level I'm getting with the board using a 12V cheap PSU feeding a 150W boost DC-DC to 24V feeding the board/stepper motor. The peak-peak noise is within 1 ADC bit - the Marlin code averages ADC samples to obtain sub-bit temperature sensor sampling. It's still a little up and down with the noise so could be better, but it's more than good enough for the intended purpose really.

Rather than bypass the board DC-DC 5V regulator with a linear or other regulator, extract a toroidal inductor off an unwanted PC ATX supply (get the cheapest you can find off ebay) and fit it in the +V line to the MKS board, and also fit a couple of nice big and small caps after the choke/inductor to create an LC filter to the board. That will filter some of the noise from whatever switching power supply you use to power the board/steppers.

Also, twist all stepper motor coil wire pairs, the temperature sensor wire pairs and any power supply wire pairs together. Twisted pairs can reduce emitted and induced EMI/RFI by quite a lot.