loading

A RAMPS board is one of the most common boards used to control RepRap and other hobbyist 3D printers. RAMPS is essentially an Arduino shield that is capable of controlling stepper motor drivers, taking temperature readings, powering heating elements, and other tasks necessary for 3D printing.

Unfortunately, some of the boards on the market are of sub-par quality - the components and connectors used are not rated for the amount of current they're required to pass. For instance - a 12V heated bed can draw of 8A by itself. On some of the cheaper Chinese boards, the power MOSFETs aren't rated for nearly this current, and will get dangerously hot, which can pose a fire hazard. Similarly, the power connector the brings power to the board might not be rated for the upwards of 10A the board draws as a whole.

This Instructable covers the replacement of the MOSFET that powers the heated bed (labeled "2" in the picture) with one capable of handling the high current drawn, as well as replacing the power connector with 16 gauge wires soldered directly to the board (labeled "1" in the picture).

Components you will need:

  • RAMPS Board
  • 16 gauge stranded wire
  • 4 pairs of (at least) 15A rated Anderson Power Poles (the ones shown are rated for up to 30A)
  • IRLB3034PBF Power MOSFET

  • (optional) Appropriately sized heat sink for MOSFET

Tool you will need:

  • Wire Cutters/Strippers

  • Crimping Tool

  • Solder Pump/Desoldering Braid
  • Soldering Iron/Solder

Step 1: Removing the Power Connector

First, unclip the male end of the power connector, and set it aside.

Next, slowly work the female end of the clip (the one that's soldered to the board) up, to the point where you can slip your wire cutters underneath. Carefully cut the four leads, allowing the plastic bit to detach.

Now, you should only have the 4 leads which are soldered into the board.

Step 2: Desoldering the Power Leads

Using your wirecutters, carefully trim the leads down to size.

Using your soldering iron, heat up the solder around the remaining nubs, allowing you to work them free from their holes. Once they have been removed, it's time to remove the excess solder and clean up the holes.

With your soldering iron on one side of the board and your solder pump on the other, heat up the remaining solder until it is completely liquid. Pressing the solder pump flush against the board, use it to suck up the solder. What should remain is a perfectly clean hole, allowing you to put in new wires.

Repeat this process until all four holes are cleared, and then move onto the next step.

Step 3: Prepping Your Power Wires

Cut four short lengths of wire, two red and two black. Strip the ends so there's a workable amount of wire.

Step 4: Soldering the New Wires In

Now, carefully insert the wires back into the board as shown, making sure to match the original polarity (which should be indicated on the board). As you do this, make sure all of the strands go cleanly through the hole, and that none split off which could later cause a short circuit.

Using the soldering iron, solder the wires into place, making sure the insulation on the other side it firmly butted up against the board. Once all four wires are securely soldered, use your wire cutters to trim the excess wire off.

Now is a good time to make sure the wires are not shorting to each other before proceeding. This step is critical, as a short circuit in a high-current circuit poses a significant fire hazard.

Step 5: Replacing the MOSFET

Locate the MOSFET that powers the heated bed, as seen in the pictures. Make note of the orientation - installing the new MOSFET backwards could potentially fry your board.

Using your wire cutters, carefully cut the 3 leads connecting the MOSFET to the board. Following the same procedure as before, desolder the remaining leads and clear out the holes.

Matching the orientation of the original, carefully insert the new MOSFET into place. Solder into place, and trim the excess length of the leads as before.

Step 6: Installing the Connectors

Now we install the power poles that allow us to connect the RAMPS board to the 12V power supply.

First, strip the free ends of the four power wires as before.

Crimp on the connectors that came with your power poles, making sure that you have a secure fit. Slide the contacts into their housings according to the instructions that came with your chosen connectors. Double check to ensure they're completely seated and secure.

Step 7: Optional: Adding a Heatsink to the MOSFET

If you chose to use a heatsink with your MOSFET, now is the time to add it.

Sandwiching an electrically insulating slip between the MOSFET's body and the heatsink, attach the heatsink and secure properly. Make sure the metal body of the heatsink isn't touching any other components that may potentially cause a short circuit.

Now, your upgraded RAMPS board is ready to hook up and run a test print.


Good luck, and happy printing!

I have just made this mod and the mosfet was cool during the operation. I also have added a heatsink to the mosfet just in case. I hope mosfet can handle it because I don't have any idea about the "saturation" thing that @acolomitchi mentioned below. Thank you @AndJoeG.
<p>I'd be careful with the choice of the MOSFET - the input capacitance and the rise/fall + turn on/off for the IRLB3034PBF are absolutely humongous, keep a eye on its temperature, the controller you just modified <em>may</em> not be able to drive it properly.</p><p>For example: if the controller <em>can</em> supply/sink current fast enough to charge/discharge the gate, the minimal duration of an on-off switch which will let the FET reach saturation (instead of just do the switching deep in the resistive region) is 1344ns=1.3 us. Suppose now a PWM with 256 levels, in which a duty cycle of 1/256 takes 1.35 us minimum. This mean the period of PWM-ing will be 0.34 ms, or a max PWM frequency of just about 3kHz - if the firmware with RAMPS tries to go with a higher frequency (why should it?), heat will develop inside your FET rather than the heatbed.</p><p>Now, if the controller <strong>cannot</strong> supply/sink the current fast enough, then the trouble is bigger. If the replaced FET was (say) IRFZ44N then through the same gate switching logic (usually involving a resistor) the time required for IRLB3034PBF to switch to the same level as the IRFZ44N will be 7 times longer (ratio of input capacitances - see Ciss in the datasheets). Are you sure the wimpy board you wanted beef up is able to actually command the IRLB3034PBF behemoth without causing more troubles?</p><p>Besides, what kind of heat bed are you using to be prepared for a 195A (package limited) max current supported by the IRLB3034PBF? <br>12V*195A=2.3kW - surely you don't intend to use a RAMPS controller to cook baked beans (just kidding, not being malicious here). If you really-really need that much, perhaps using a 12V supply isn't the best idea.</p><p>You say &quot;For instance - a 12V heated bed can draw of 8A by itself.&quot; - for 8A, an IRFZ44N (49A max rating) is more than enough (49A) and 2.5 times cheaper at the same source you quoted ( <a href="http://www.mouser.com/ProductDetail/Infineon-IR/IRFZ44NPBF/?qs=sGAEpiMZZMshyDBzk1%2fWi5%252bqVgN3%252bWS8JmdArXmg%2fHY%3d">http://www.mouser.com/ProductDetail/Infineon-IR/IR...</a> )</p>
<p>Hi acolomitchi, thanks for the phenomenal feedback!</p><p>That's something I hadn't considered, and you make some excellent points. In terms of my chip selection, I chose the 3034 not for the max current, but for the very high power dissipation value. As far as I can tell, for the bulk of the time the heatbed is on maximum when heating up, and is only occasionally on during operation (almost like a bang-bang system) to maintain heat - it's not a continuous current.</p><p>That said - you seem to know a lot more about MOSFET selection than me, so I'm curious to hear your thoughts and I'm more than happy to modify my 'ible and board if you have any better insights than me.</p><p>At the end of the day, I've been printing on an identical board for quite a while now, and when I first modified it I very carefully monitored the temperature of the MOSFET, and haven't even felt it get warm after several hours of operation.</p><p>Looking forward to hearing your thoughts!</p>
<blockquote> <p> At the end of the day, I've been printing on an identical board for quite a while now, and when I first modified it I very carefully monitored the temperature of the MOSFET, and haven't even felt it get warm after several hours of operation.</p> </blockquote> <br> Well then, if it works it works, this is all that matters. I'm not familiar with how the RAMPS controller drives the heatbed logic, thus I can't offer pertinent suggestions (all I written on the above deals on hypotheticals).<br> <blockquote> <p> That said - you seem to know a lot more about MOSFET selection than me</p> </blockquote> I wouldn't bet on it. My only knowledge on MOSFETs not fit for a job came when attempting to build this 'ible https://www.instructables.com/id/Arduino-based-Switching-Voltage-Regulators/ and experiencing a heating MOSFET <em>under particular circumstances</em> - turned out that a too short (repeated) pulse was not driving the FET into saturation.<br> You know... once bitten twice shy.
<p>Ah, fair enough. I guess if it works it works, and if I'm doing anything beefier than this I'd be switching to a solid state relay anyways.</p><p>But definitely something I'll keep in mind for future projects, it's not something I would have considered off the bat.</p>
<p>That's a thing to keep in mind indeed - when you deal with BJTs, it's seldom that you need to think of timing, current limits is your first worry. </p><p>With MOSFETs, once you select some with a max current/voltage that satisfies your power needs, the next thing you need to worry is how hard/fast you need to drive the MOSFET to get it out of the resistive region - typically, the input capacitance and the turn-on/off times are the limiting factors.</p><p>As a rule, at higher max fet saturation currents, you can expect higher gate capacitances - naturally, large saturation currents require large gates, large gates require more charge to fully open. <br>Large charges require either large currents to fillup the gate capacitors and/or longer times to do it.<br>Longer times to open the gate means lower switching frequencies are possible (and switching is what MOSFETs are good for)<br>All the above translates into: the larger the max saturation current, the lousier switching time you can expect. </p><p>If you really-really need to switch large currents, sometime it pays to put two or more MOSFETs in parallel - your only care would be to make sure your gate driving circuit can open now more gates in the same time. </p><p>Consider the gate(s) as capacitors see what times you need to switch on/off, compute what resistor values you need to use to have those capacitors charged/drained in the switching times you need. Then see if your driver circuit can deal with the current imposed by that resistor - if not, then you won't open that gate fast enough (and have your fet running hot) or you'll blow the gate driver circuit or both.</p>
<p>Ah, fair enough. I guess if it works it works, and if I'm doing anything beefier than this I'd be switching to a solid state relay anyways.</p><p>But definitely something I'll keep in mind for future projects, it's not something I would have considered off the bat.</p>
<p>Good project. You could easily remove the power MOSFET(s) to a remove heat sink assembly and run proper gauge wires back to the leg placements on the PC board for a more secured heat dissipation effort, even a fan, if necessary, over the heat sink. And one thing about soldering those wires through the PC board. If this is going to be a long term use project, after cutting the excess wire off, re-solder the ends to stop metal corrosion, in fact use a good liquid flux to make sure the solder joint is good. And make certain the PC board runs can support the extra current or it will fry them eventually. I was a micro-miniature NASA certified solderer before retiring. Thumbs up!</p>
<p>I'm glad you liked it! I definitely means a lot to hear that a pro enjoyed my work.</p><p>Thanks for the tip on the wire tips (hah!), I didn't even think to do that. Next time I'm in the lab I'll make sure they're nice and tinned.</p><p>In the future I'll be building a printer that uses a bed that draws even HIGHER current - which I'm planning on powering with a beefy solid state relay externally - no way a tiny MOSFET would be enough.</p><p>Trace wise, I think I should be good, although I'm putting my faith in the open source community. The board layout itself, which is open source, was made by some pretty smart people. The problem is the suppliers who make the board from the design, and oftentimes cut corners on components. I'll have to dig out my EE textbook and double check the widths myself.</p><p>-Joe</p>
Awesome 'ible! Those yellow polyguses need the same treatment - they'll likely trip under a standard mk2b heat bed, un-necessarily!
<p>Definitely, may be my next project.</p><p>I haven't had any problems with them so far, and my main concern with this was not having my printer catch fire haha. If I start having problems I'll explore replacing them as well.</p>
*polyfuses