Melzi vs. RAMPs?

Sorry for newbie question.
I have a Chineese i3 copy.
My melzi board, delivered with the printer, works OK from SD card, but often disconnects the serial connection - and by often I mean every ~25-60 minutes.
I want to use OctoPrint, but due to the disconnecting serial port, it's unusable.
And, as Melzi does not have serial port pins, I am stuck to SD card printing for now.
However, I bought RAMPS 1.4 from aliexpress form ~$16 and it's sitting on my desk right now.
I read that Melzi is superrior to RAMPS, but I have difficulties to find out in what exactly is the superiority.
So, the question - would I be loosing a lot if I move from Melzi to RAMPS 1.4?
What is the actual difference other than the parts are directly soldered on a single board with Melzi and replacable with RAMPS?

There is little difference

The main difference is because the A4982 stepper drivers are on the main board it has better thermal dissipation, and its a one board solution

Its also its draw back, since you can't replace anything if they die (unless you can do SMD re flow work)

Melzi does have serial a port... the usb cable connects to a FT232RL this connects to the serial port on the Micro-controller (pins 9 and 10)
You can bypass this converter and use a usb to ttl rs232 converter cable.. if your good at soldering etc...

You don’t lose anything going to a ramps... I don’t know what you read, but its sounds like it was just marketing bs.

NB you do know a ramps has to sit on a arduino mega and requires pololu modules to work?

Sure I know - the $16 was for a pack of RAMPS+Mega+5 pololu modules
My problem with the FT is that it is connected to the ATmega directly without resistors, so I kinda hesitate - if the FT for whatever reason reports logical 0 on TX pin, I'll not be able to send anything. But, perhaps, this is not a big deal as there is a how-to on reprap wiki to connect BT module without disconnecting the FT...

Apart from the better heat dissipation of the stepper driver chips, the other advantage of the Melzi is that its 5V regulator is capable of higher current. With Arduino/RAMPS, the regulator on the Arduino is adequate if you don't connect anything else, but once you connect either a servo or a graphic LCD, is it more or less guaranteed to overheat. There are many posts on these forums from RAMPS users suffering from this problem.

Your USB disconnections are unlikely to be caused by a hardware fault on the Melzi. They are more likely to be caused by some other USB device connecting/disconnecting on the Windows PC, or by noise on the ground loop that is formed when you connect a PC and a printer via USB. Try the following:

- Power the PC and the 3D printer and nothing else from the same mains power distribution block
- Use a USB cable with a ferrite bead on it
- Don't turn on or off any other USB-connected devices while printing.

However, printing from SD card is recommended anyway when using atmega2560-based boards, because it's often difficult to maintain sufficient data rate over the USB connection to keep up with printing.

---

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].

Thanks a lot for the power warning! I totally missed that as I for now only tried to find difference between Melzi and RAMPS
To be honest, I don't care about the display at all - I have it for Melzi, but I rarely use it - only in case of serial failure in the middle of the print.
I can overcome the power issue by supplying the 5V to Arduino directly using the pin and providing it from a voltage regulator outside of the board, couldn't I?

However, the other tips I already tried - I tried 3 different cables (2 with ferite cores).
I'm not using Windows PC, rather Banana Pi.
The only other device on the USB is a webcam.
I tried using powered and unpowered hubs (2.0 and 1.1, in total 6 different hubs).
I have an extremely stable 3A power source for the Banana with 1mF capacitor for filtering any bigger voltage spikes.

I can try to disconnect the webcam and move it to other device as a last resort - but the same happened when I connected the printer with OpenWRT-running router with OctoPrint installed and no webcam connected.

I am doing all of this to have the ability to create extensive timelapse plugin - to have PiP information about temperature, progress, g-code processing, etc. But I'm failing on the very first step here and it starts to be frustrating...
Thanks for all the help!

Oh, and one more thing - my idea was to connect the Melzi serial directly to Banana's serial port (with voltage converter, of course) - to avoid the USB bus at all. But as Melzi does not provide serial header and I am not keen to break the warranty, I am a bit hesitating to solder on it...

Do you have the Banana Pi powered from the same mains distribution block as the printer? If the power brick for the Banana Pi doesn't use the ground connection, then the distribution block must be surge protected (this is a good idea even if it does use the ground connection), otherwise transients on the mains will feed capacitively through the transformer in the power brick and then through the USB cable to ground.

Yes you can feed 5V power to Arduino/RAMPS from an external 5V regulator. Another option for using a GLCD is to power the backlight from 12V through a series resistor instead of from 5V.

---

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].

Hmm, this is a bit more than I can understand
Yes, the Banana and the printer share the same wall plug as I have it in the cellar with single connection only (and, honestly, the whole house seems only to use 1 phase for all wallplugs).
The power brick (meaning the AC to DC adapter) for Banana uses the 2-wire connection to mains (line and neutral), not the protective ground.
I do not understand the process of capacitively charging the transofrme in the power brick, to be honest - I understand the movement of the motors can induce some currents that can, possibly go through the main AC-DC converter on the printer back to mains as a "noise", but that's as far as I can get with my understanding - I thought that the power brick and the filtering capacitor for Banana will filter it out...

How can I get around this issue? What if I power an USB hub between Banana and Melzi from Melzi's 12V (using DC step down converter)? Will this help?

"With Arduino/RAMPS, the regulator on the Arduino is adequate"

This depends on your Mega, I fully agree if you get a cheap standard mega this is the case, personally I prefer a Taurino Power, can take up to 35v, and no need to remove D1 and generally has a much better regulator

Well, I did a bit of soldering. I used the Bluetooth-how-to, but instead of BT module, I used the wires and plugged them to Banana.
I disassembled an old laser printer and salvaged some ferite core and wired few rounds of the serial wire around it.
And, finally, I connected it to Banana's ttyS3 (UART7). Of course, I only used 3 wires (RX, TX, GND) and also used the 3v3 conversion as for the BT.
Now, after 3 days of permanent printing, the board works OK, banana did not disconnect once, I am able to do timelapses and everything is great again.
Thanks for everyone who shared the knowledge with me to help me overcome the issue.