If you swap the entire electronics (Mega+Ramps) with the one of your own printer ... does the problem move with the electronics or stay with the printer?von enif - RAMPS Electronics
This seems really to be a difficult one! So just some "far-off" ideas... - check that the Y-belt is well centered and aligned (no rotational force when pulled by belt) - exchanging the X- and Y-motor to see if the problem is motor related - exchanging the X- and Y-stepper drivers to see if the problem is stepstick related - reducing the baud rate to ensure that it's not a (strange!) communicvon enif - RAMPS Electronics
If you go with the cheap 40A mechanical automotive relay (as I do), you can change the BED_CHECK_INTERVAL from the default 5000 to 15000. This reduces the clicking sound mentioned by cdru to max. once every 15 seconds, so it's no longer an issue (at least for me). But the temperature of the heated bed will still be kept to within a few degrees from the target value. In Marlin the BED_CHECK_Ivon enif - Controllers
This is how I replaced the 11A polyfuse:von enif - RAMPS Electronics
Have you tried to lower the acceleration values on th Y-axis?von enif - RAMPS Electronics
Is diode D1 installed on your RAMPS?von enif - RAMPS Electronics
USB 4.75V will be fine for the 5V rail. Now connect 12V for motors and hotend and try to see how far you get the printer actually working.von enif - Sanguino(lolu)
If you suspect that the 7805 is the problem, then why not just take it out and run the 5V rail only from USB. As long as you have both USB and 12V connected the Sanguinololu should still work fine. You can then connect the bare 7805 to the 12V and GND (leaving 5V output open), if it still gets hot, you'll know that the 7805 is the problem.von enif - Sanguino(lolu)
Pins 18/19 are the MOSFET gates, they are connected via 10k to GND. So if it's 9.9/10 kOhm (not Ohm!) then it's normal.von enif - Sanguino(lolu)
Ooops! Of course not. Mine is removable on a socket... With no power connected, you could try to measure the resistance from all possible output pins (1-8, 14-20, 21-29) to GND and 5V (any why not 12V, too) and see, if you find a short somewhere.von enif - Sanguino(lolu)
When I remove the AtMega644/1284 and the stepsticks, I measure 6mA on the 13.9V input. How much you get? I still wonder if the problem is not due to some short (to GND or 5V) of an output pin of the AtMega...von enif - Sanguino(lolu)
135mA through the 7805 will roughly give 1W (7V*.135A), so it should not really get that hot... I just measured my Sanguinololu with the stepsticks installed, but with nothing connected: Powered via USB I measure 31mA and via 13.9V on the 12V input it takes 50mA. Might it be that one of the digital outputs of the AtMega chip is shorted? If you take out the stepsticks, do you still get the 13von enif - Sanguino(lolu)
200mA for the 5V rail should in principle(!) be fine. I use BAT43 here (which have the same max. current) without ever having had any problems. BUT if your regulator gets hot, you might have some drain somewhere on the 5V rail, so that might cause a problem for the BAT42. Have you measured how much current the board draws from the 12V input when no load (motors, extruder, bed, fans...) is connvon enif - Sanguino(lolu)
@maso: As long as the termistors are read while the printer is running (which is the normal case), the diodes do not matter, since then the 5V are coming from the 7805 regulator. On the contrary, using the diode on the USB 5V lowers that voltage to below 5V, even if the USB port delivers a bit more than 5V. Thus, there is no current flowing into the Sanguinololu, so that the 5V are coming entirvon enif - Sanguino(lolu)
At the input of the 7805 the forward voltage is no problem at all, since anyway we have to drop ~7V at the linear regulator. With only the USB input, the 1V drop is probably still ok, if the BOD (brown-out detection) fuses of the AtMega644/1284 are set to 2.7V. If BOD is set to 4.3V your USB port would have to deliver 5.3V. In any case, as soon as the 12V are connected, it should certainly wvon enif - Sanguino(lolu)
Sure you can still try with the diodes. The Schottky diodes are preferred because they have a low forward voltage. But you can also try with standard diodes, it should still work, as long as the voltage does not drop below your brown-out threshold.von enif - Sanguino(lolu)
This is how I replaced the MF-R1100 PTC fuse by an 10A car fuse on my RAMPS 1.4.von enif - RAMPS Electronics
Instead of putting one of those RadioShack LEDs with integrated resistor onto you MK2B bed, I think it's better to put one resistor and two normal LEDs (I use a red and a green one) in opposite direction, as foreseen on the MK2B. This way, the active LED protects the other one from reverse breakdown, while the resistor limits the current through the active LED. As the LED doesn't need to shinevon enif - Controllers
It's good to see that we both approached the problem in (almost) the same way If the diode is on the 5V side of the regulator, it decreases the 5V by the diode's forward voltage, whereas if it's on the 12V input side, it doesn't affect the 5V regulated output. But, of course, the diode we both put in the USB 5V line also decreases the 5V in that same way. One advantage of having the diode onvon enif - Sanguino(lolu)
I'd just like to document here, how, for my RapRaps, I solved the known issues of the Sanguinololu 1.3a board with respect to the 5V conflict between the USB port and the 7805 regulator, and the possible backfeeding of the USB 5V into the 12V input line. I don't know if others have posted the same ideas before (I searched, but didn't find). So in case I missed something, please be gentle with mvon enif - Sanguino(lolu)