Screen flickering while hot end PWM is on

So in order to avoid off-topic discussions in this thread I decided to start a new one dedicated to the issue.

The backlight of the LCD screen starts flickering the moment my hot end gets to the desired temperature and the PWM kicks in. So to be more specific, the process is as follows:

1. Turn PSU on, backlight is steady, the green LED on the RAMPS lights up
2. Turn on the hot end, target temperature for example 240 deg. C, hot end heater starts at 100%, backlight steady but gets a litte bit darker.
3. At some point the heater power gets regulated from the PWM control, the green LED on the RAMPS starts flashing and the backlight starts flickering in phase with the flashing of the green LED.



How could I stabilize the voltage when PWM is running? The bed heater which is connected to the RAMPS via a simple MOSFET circuit has absolutely no impact on the screen brightness even though it is quite powerful (350W).

- 24V 800W PSU
- Ramps/Arduino running at 12V via separate buck converter
- 350W heated bed running at 24V via a simple Mosfet switching circuit
- The voltage regulator of the Arduino (AMS1117) has been bypassed. The 5V rail is powered directly off the PSU through a LM2596S

Edited 2 time(s). Last edit at 11/12/2016 11:23AM by icefire.

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

Your PSU is probably in trouble. The only way to check this is by using a voltmeter on the 24V line to check for voltage fluctuation or drop.

There are voltmeters on the 24V, 12V and 5V lines. There is no voltage drop on either line when the heater starts . Furthermore, the heated bed which is like 10 times more powerful doesn't lead to any screen flickering. And let's not forget that the PSU is rated for 800W so I don't think it is a PSU issue.

I bet the flickering would disappear if I wired it also through an external mosfet instead.

Has anyone else experienced such an issue? Could it be a RAMPS problem?

Edited 1 time(s). Last edit at 11/12/2016 02:43PM by icefire.

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

In my experience the buck converters don't have great regulation, so it might be worth using the bucks to drop to a couple of volts above the required voltage and then use a linear regulator to finish the job. I feed 7.5v in via the hole where I removed d1 and that works well.

What I can try is to power the RAMPS with a computer PSU and see of there is a difference. But what does this have to do with the screen? The LCD Backlight gets power from the Arduino 5V line which is connected to the LM2596S.
The hot end heater and the RAMPS get 12V through the buck converter. Since D1 is removed this has no influence on the 5V rail. So how does this lead to the screen flickering?

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

I'm not sure what the graph shows- is the light blue line the bed temperature? If so you've got a problem with the wiring to the thermistor. It should not jump around the way it is. I recently had a problem where the hot-end temperature reading was jumping when I switched on the 24V supply to the printer (the uC was running via 5V from the USB connection). It turned out there was a high resistance short (2-10 MOhms) between the hot-end heater leads and the 24V supply.

How did you connect the wires to the power supply? If you daisy chain the grounds or even the + voltage lines, there can be a voltage drop that will cause the flickering. The power to the bed should go directly back to the supply and there should be no other connections to those wires.

Edit: the short was between the 24V supply via the heater leads and the thermistor.

Edited 1 time(s). Last edit at 11/12/2016 09:27PM by the_digital_dentist.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
icefire
What I can try is to power the RAMPS with a computer PSU and see of there is a difference. But what does this have to do with the screen? The LCD Backlight gets power from the Arduino 5V line which is connected to the LM2596S.
The hot end heater and the RAMPS get 12V through the buck converter. Since D1 is removed this has no influence on the 5V rail. So how does this lead to the screen flickering?

So the only thing I can imagine is that, somehow, PWM switching is actually causing interference on the 5V rail. I would apply extra filtering on the 5V rail for testing purposes. One capacitor of 1000uF and one of 100nF.

My guess is that you have a voltage drop on the RAMPS ground line due to the extruder heater current, and you have referenced the ground pin of the 5V regulator to your PSU instead of to RAMPS ground. So although your 5V is stable at the output of then 5V regulator, it isn't stable at the RAMPS, which I presume is where you are getting the backlight power from. Try measuring between 5V and ground on the LCD itself.

These problems can be avoided by using a decent controller board that provides an adequate 5V supply.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

@ the_digital_dentist: I apologize for not explaining the graph more clearly. It shows the hot end temp. and the PWM output of the hot end heater. The bed heater has no influence on the screen flickering. Here is a better version of the graph:


I guess I could have some high resistance short between the thermistor leads, I am going to investigate it asap (currently printing so I cannot touch it). But I guess this issue is unrelated, although I find it suspicious that the temp reading starts jumping the moment PWM kicks in.

@ dc42: How could this be? The PSU, the 12V buck converter, the 5V step down module, the RAMPS, the Arduino, the MOSFET circuit for the bed heater and the PC USB have all common ground.

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

Quote
icefire
@ dc42: How could this be? The PSU, the 12V buck converter, the 5V step down module, the RAMPS, the Arduino, the MOSFET circuit for the bed heater and the PC USB have all common ground.

Because ground wires and ground connections have resistance (and inductance, but that is unlikely to matter here). Have you checked that the screws on the RAMPS power input terminal block are still tight?

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Yes, this I have checked. Not the problem here.
As soon as my print is complete (just started another huge part aaaaargh) I am going to measure the voltage between 5V and LCD ground as well as the resistance between the LCD ground and the PSU ground and connect them directly with an additional wire if it need be.

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

What's the exact sequence of connections for your various voltage converters? Do both the 12V and 5V converters connect to the 24V supply, or have you daisy chained the 5V from the 12V?

They are not daisy-chained. All voltage converters connect directly to the 24V PSU. Here is a quick illustration of the setup:

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

Ah, OK. In that case I think David is right (as usual), in that the problem is with the ground connection. That seems like the only way that the 12V load of the heater can cross-talk to the 5V LCD while the 24V bed does not.

Probably. David tends to always be right
Like i said I am going to wait for the print to finish and write back after I have checked the ground connection.

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

So the ground connection was actually fine. Connecting the RAMPS ground directly to the PSU ground brought nothing. Powering the RAMPS through a separate PSU and the Arduino via USB also didn't quite solve it.
However, disconnecting the 5V step-down module from the 5V rail and powering the Arduino with 7,5V through the barrel converter led to no visible flickering.

So I guess the LM2596S doesn't do a flawless job maintaining 5V. I could leave it like this (7,5V through the barrel connector + small heatsink on the AMS1117) or find a suitable 5V regulator to use instead of the LM2596S.

Could someone recommend such a regulator? I wouldn't prefer to start the countdown until the voltage regulator of the Arduino fails.

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

Unless you have servos or fans powered from +5V, the current draw from +5V should only be about 200mA. In which case, 7805 linear regulator plus two capacitors and a small heatsink would be hard to beat.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42
Unless you have servos or fans powered from +5V, the current draw from +5V should only be about 200mA. In which case, 7805 linear regulator plus two capacitors and a small heatsink would be hard to beat.

I have the reprap discount full graphics controller. I thought the backlight would definitely result in higher current. The AMS1117 wich is fitted on the Arduino is rated for 800mA, if the current were so small it wouldn't have been such a problem on all Arduinos...

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

Quote
icefire
I have the reprap discount full graphics controller. I thought the backlight would definitely result in higher current. The AMS1117 wich is fitted on the Arduino is rated for 800mA, if the current were so small it wouldn't have been such a problem on all Arduinos...

It's not just the current, it's the combination of the current and the voltage being dropped across the regulator. The power being dissipated is (Vin-Vout)*I and becomes a problem when people are drawing heavy loads whilst feeding 12+ Volts via the ramps. Also, a lot of the problems with the regs on arduinos in printing come from the unfortunate endstop connector design that shorts the supply if a microswitch style endstop is connected in reverse.

By feeding an input voltage closer to the output voltage the regulator will be dissipating a lot less heat, and providing you don't short the output will likely run fine for many years.

... Which is why I am now feeding it 7.5V and have attached a small heatsink on top of it. But maybe it is still a good idea to play it safe and use a LM7805 instead, just to be sure...

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

ok, the screen flickering is now solved (either using the stock Arduino voltage regulator by feeding it 7,5V and using a small heatsink or with LM7805 and two capacitors (0,1 and 0,33 uF + small heatsink) - both solutions work fine).

But I just noticed that the LEDs I have attached next to the nozzle to illuminate the object being printed also flicker in phase with the PWM regulation. The LEDs are powered off the 12V buck converter through a resistor. Is there any way to stabilize the output of the buck converter? Off the top of my head I could just change the resistor and power the LEDs from the 24V PSU or use a LM7812 just for the LEDs.. but there are all workarounds.. If I could improve the regulation of the buck converter I wouldn't need anything else.

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

Quote
icefire
... Which is why I am now feeding it 7.5V and have attached a small heatsink on top of it. But maybe it is still a good idea to play it safe and use a LM7805 instead, just to be sure...

Feeding it 7.5V instead of 12V is a good solution. Putting a heatsink on it is likely to be ineffective at best because the plastic top of the regulator does not conduct heat well.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].