MakerBot Print is our newest print-prepration software, which supports native CAD files and STL assemblies,
allows you to interact with all your printers via the Cloud, and many other exciting new features.

Download Now

Replace Zrib mosfet or replace Zrib?

Please Login to Comment

Hi all. I purchased a something called an EG4 XL which is an upgraded Tronxy X3S by a company out of Montreal, Canada.
Safer as I can get warranty and support, right? Lmao! Nope!
I know. Save it. I know better now! Lol
Zrib board with Ramps 1.4 and an ATmega2560 as the brains.
24volt 360w psu, pushing a 330 X 330 heated bed.

I really took my time to do it all right but when I started it up and tried to level the bed the hot bed turned on and all the white smoke that makes the mosfet work came wafting out within seconds. Shut off immediately.
I will go back over everything and make sure all is good. No shorts etc. Again!
But, what would the next step be if it was you?
Would you replace the 855nf mos with a better one? Replace the whole board? Throw the damn thing in the garbage and buy a different printer?
This is my first 3d printer but for the life of me I cant see the mosfet genie escaping for no reason. I just cant find the reason.
Any advice or answers to the above questions would be appreciated. Thanks.

Myself, I would make one attempt at repair, realize quickly that it was more trouble than it is worth, and either get an external mosfet like was suggested of pitch it and move on. If you go the RAMPS route, check out the 1.6.

These may be of interest (great channel also):

Is this your board?

It may also pay to check the resistance of the heated bed since it seems that is what blew the mosfet.

Thanks for the reply. That's pretty much my board with only slight differences in the fan pins arrangement.
I will check the resistance on the board and I'm thinking of upgrading the soldered connections.
It doesn't seem like a cheap board to me but I guess they all have the odd didd.

Cool, it does not look very easy to swap the mosfet on that one unfortunately. Depending on the wattage and resistance of your heated bed, it may be marginal to use the onboard mosfets (or you may have some short in the bed itself if resistance is too low). Worst would be to blow the next one out due to an undiagnosed problem with the bed :-(

Biggest problem I have with my RAMPS which I use with a RE-ARM is keeping the RE-ARM cool since the RAMPS covers it up. The integrated boards like MKS and yours seem much better in that regard and often have a better layout for wiring.

If you plan to go the RAMPS route, the PICA is another option that seems to be a high quality derivative (edit NM, looks like they are sold out for a long time):


I'm thinking of keeping the zrib and using external mosfet at the bed and the extruder as suggested by valand70 while also ordering a board that better supports expansion.
I really like the all in one boards but getting the best board for the buck is tricky and the 24v requirement lowers my choices.
The mks boards were suggested and they look good. Not sure what manufacturer to go with. The due and some of the other smoothieware boards also look tempting. $!
Is it common for people to buy a board from a brand name printer manufacturer and use their boards?
I would think quality of parts and manufacture might be better.

Not sure if it is common to buy a board from a name brand printer manufacturer, but I don't think that is a guarantee. Some boards, like the one in my Ender3 have a lower spec than a RAMPS+Arduino 2650 and are more difficult to upgrade since it has less memory and no bootloader. Stock boards may also not break out all the pins if they are not needed in the particular printer, and they may not easily support different stepstick drivers. Myself, I would look at boards that are both generic and popular since those will make it both cheaper and easier to get help with :D

Aside from support, the biggest selling point of the Smoothieboards and Duets in my view, is that they are 32bit and offer a web interface. But I don't think 32bit would be a big deal on a Cartesian printer (I have not studied it), and I use a Octopi for the web interface. I use a 32 bit Re-ARM with Marlin2.0 on my delta since I feel it could use the extra help, my ender3 runs a stock board.

I did not read through the whole thread, was there any problem found with the bed (or it's resistance) that could have blown the mosfet?

I cant thank you guys enough for taking time out of your day to help me out.
The resistance, measured from the unconnected wires at the board, was 1.4 ohms r after adjusting a tenth for my multimeter.
I have been searching for some kind of short for days now and I just cant find one.
1.4 ohms resistance sounded right to me but in all honesty I haven't done the math yet.
Finding the short that caused a brand new board to smoke a mos, on the very first use, is where most of my time has been spent as I figured it was a mistake of some kind in my build.
I included a pic of my board just because it is an oddball and you might be interested.
I hate the idea of pitching something that may still have some kind of use so I will save it. Maybe it will work in a project later that isnt so demanding.
I need a replacement that is 24 volts, lots of room for expansion and reliably good quality.
At the moment, no matter which board I get I will use external mosfet for the bed and head.
Now, if I could just figure out where I went wrong in my build I would feel a lot better about starting with a new board.

If the thermistor was not connected may have caused your fault. Having no feedback from the thermistor should have triggered thermal runaway but being a powerful bed it could be that the timer wasn't quick enough. "WATCH_TEMP_PERIOD" and "WATCH_TEMP_INCREASE" in "Configuration_adv.h" are the parameters involved...
These parameters are there to avoid damage in case of thermistor fault and being yours a powerful hotbed, it may be that standard settings are not quick enough to avoid it...

As you are getting a new board, I would first do another try with the original board swapping to the second extruder mosfet and use it for the bed. It is simple, just open up pins.h file in marlin, find these lines and change them:

define HEATER_0_PIN 2

define HEATER_1_PIN 3


then swap HEATER_1_PIN with the HEATER_BED_PIN and connect the bed wires to the second extruder connector. Leave the thermistor where it is.
Before doing these changes, just disconnect the hotbed, turn on the printer and find out if the thermistor is working...just heat up the central area of the hotbed and watch if the temperature on the display changes. You can heat it up with an hair dryer or even just placing your hand on the bed should be enough to see if temperature values change. Then if it works, go ahead and swap the configuration pins. Better to play around with this board before the new one arrives and get to know how to flush firmware and find the correct configuration of your hardware.
Make sure the soldered connections of the hotbed do not touch the aluminum plate, you can test with the meter between the aluminum plate and one of the hotbed wires if there is any short.
Then just upload the firmware and try...

That's perfect.. You answerer the questions I wasn't sure how to ask.
You guys have been great, thanks for all your help.

If your board is atmega2560 based, pin numbers to change are in 'Pins_GT2560_REV_A.h'

If you download new Marlin firmware, follow this

The best thing would be to get configuration files from the company you bought the printer from, so you just overwrite marlin ones, edit the lines HEATER_1_PIN and HEATER_BED_PIN and upload...
Otherwise, there are example configuration files in a folder of Marlin for the most common printers you can use to start from.
If yours match one of them you can use them instead, but you must check if the board selected in configuration.h is the same. If not, it is not good for you. The board you select in configuration.h is vital because if you select the wrong one you can screw it up badly.
Your board should have written somewhere what model it is, start from that and find config files that use it...

I realized that pins configuration may be in a different file depending on your board configuration...do you have any firmware that came with the printer ?

I recieved an email from the supplier of the kit. The eg4xl is a 24v kit but the heated bed is 12v.

Here is a quote from the email.

"All our EG-4 XL are sale with a 12V heatbed. The controler board adjust automatically the voltage between the power supply and the heatbed. So you can use 12v or 24v power supply."

Gotta say, that would have been good to know.
Maybe the disconnected thermister did have a part to play.

Yes it's the atmega2560. I have the config files that came on an sd with the printer.

Then the only thing you have to do is to change pins numbers and upload the firmware. Make sure the termistor is working like I said, and hopefully you will have a working printer again...

As I found one end of the thermister was never soldered and just placed together under the heat shrink, I doubt it would have been able to do its job but the fact that the moss blew so quickly, maybe 3 seconds, I assumed it would not have had any involvement in blowing the moss. Of course, I'm also learning I'm not as smart as I thought I was . : )
There seem to be a few reasons this moss may have blown. The sad fact is I cant find the issue and I'm super hesitant to replace the board without knowing exactly what happened. It's an expensive way to troubleshoot!
Thanks for the help though. I will go through the code to see what I can find.

What terminals are you measuring to get that low a resistance and what sort of meter are you using?

MOSFETs commonly fail short. What happens is a breakdown in the thin oxide layer of the gate. You might have popped it inadvertently with static discharge. I don't see the a reason to get external FETs if you just get a proper board with FETs that have a low Rds on value. The higher the voltage you're running the less current you're going to be using for the same amount of power because P=V*I and the power loss across that FET is going to be I2R, R being that Rds value when saturated. These are logic level enhancement mode FETs that saturate at 5v and should show a very low resistance. A proper MOSFET has an Rds value on an order of mΩ. Do that math for your, let's say 300 watt bed, and we get 300/24 = 12.5 amps. The math for the power loss across the FET is then, assuming the data sheet value of 10 mΩ, would be 12.5a*12.5a*0.01Ω = 1.5625 watts power loss. I use more than 2 watts during my morning piss and external FETs with enormous heat sinks are a waste of time and money. Then consider that they aren't always on either, they're pulsed on/off to achieve desired temp.

Like I said, get an MKS gen 1.4 and don't look back.

I am placing an order later tonight for an mks. There is a dual z version I may be able to get....just looking at drivers now.
The PSU is 360 watt. At 24v that's 15 amps. I'm not sure it really pushes that, though.
My meter is a 35$ hardware store ching chow! I use two of them and so far they have been good but I admit, they aren't exactly Fluke brand.
I measured at the bed and the ends of the cables. I did find the one thermister lead was never hooked up under its shrink tube covering but that shouldn't have contributed to the moss blowing.
I must have gotten a bad board or mosfet. I noticed today some tiny bubbling just after the blown moss where the traces lead away.
I assume it's good for little more than salvage.

Any of the 5 driver boards will work for a single extruder/dual z and there's a way to add a second extruder with a 6th driver. You have to mirror the z in firmware to use E1 as a second z. It's best to use a driver per motor. Otherwise you're splitting the current between the two motors. Grab 10 drivers while you're at it too. You might pop one setting the current limit, I've done it a few times.

I've measured my 200w bed at around 13 amps when it's at around 90 C. You might also want to make sure that the heated bed isn't a direct short and can handle that 24v. If it came with the kit it can probably handle it. What kind of heated bed is it? PCB or a proper silicone pad?

The bed is the typical aluminum mk2a. I have made sure to isolate the bed with plastic washers during the build as I was nervous of the thin coat of enamel that separates the two sides. Its attached to an acrylic table underneath so it isnt conductive in any way.
The bed came out at 1.5ohms and I can't see anything in the way of a short.
I really hoped to find a definitive reason for the moss blowing but shorts have pretty much been ruled out. Unless the detached thermister had something to do with it. That's the only incongruentcy I could find. All the code and settings came from the dealer on an sd card and should not have required any adjustment from me.
I think all I can really do is cross my fingers and try another board. I hate doing things that way but this issue with the blown moss doesn't seem to have been caused by anything obvious and didn't leave a trail of crumbs to follow.

That doesn't sound right at all. The heating pad shouldn't be attached directly to acrylic, it should be attached directly to the underside of the aluminum. Did you account for the resistance of the probes on your ching chong meter when you measured? Last time I bothered trying to use a ching chong HF meter, if you can even call them DMMs, the lead resistance was around half an ohm.

Dont worry. I didn't do that. The mk2a is installed properly. Something I said confused you. My bad.
I did account for the 0.1ohm r of my meter when I calculated the 1.5ohm r of the bed.
I've been wanting to buy another meter but a Fluke is out of my range. I will have to familiarize myself with other brands first.
Anyway, going over this printer, no shorts! Cant find a thing that could have caused the mos to smoke. Nothing!
I'm terrified to hook up another board but sooner or later I'm going to have to face installing the new one.

Take a look at the aneng an8009. It's a reasonably priced meter that's surprisingly accurate. Don't bother trusting it for high energy applications though, nor for anything nearing even 10 amps. EEVblog on YouTube has multimeter shootouts that compares varying meters in different price ranges side by side. Fluke is really only for industrial use so the fellas don't laugh at you.

Can't edit my previous response or the bots will get me. Some text to avoid them SOBs. Some more text to avoid them.

I think you've gotten the 12v version with your kit. Several sources, including this one, list the resistance for the 12v being around that 1.5 ohm, whereas the 24v is listed as a resistance of 3-4 ohms.

That's some good internet sleuthing Sir!

Didn't make much sense to me but I couldn't put the finger to it. At 24v the current through a 1.5 ohm heater is going to be 16 amps giving off 384 watts. Time to trash those cheapy PCB heaters and get a proper silicone impregnated heater with proper resistance wire in it.

Those slicone heaters look nice, but the only ones I have seen are for AC and would require a relay or SSR, so probably will be a while before they make it to the low end printers.

Hope that the OP gets his EG4 XL bed replaced with the correct 24V version. Really the seller should also swap his board since if they screwed up and sold him a 24v printer with a 12v bed, it was not his fault that the mosfet blew.

I have no idea what you're talking about. I have a 200w silicone heater on my <200$ kit and it works very well.

Will PM you about the heater (interested in more info on the DC version).

I think you misunderstand resistance heating. It's just a wire, such as kanthal or nichrome, through which current is passed. It doesn't care if it's a DC or AC. In the case of resistive wire for the purpose of heating all energy is 'lost' across the wire, which is the goal in the first place.

Yep thanks, that's how I make toast in the morning too.

I would recommend that you educate yourself on ohm's law if you still don't understand the relationship between voltage, current and resistance. That video you linked me is extremely ridiculous and seems to be the product of someone trying to prove themselves to the internet. I would advise against playing with mains voltage, leave that to trained professionals.

Hi, I never claimed to be an expert, but I am really puzzled how you draw that I don't understand Ohms law, can you explain specifically what my error is since I can only assume that either I don't, or there is some misunderstanding? The video was simply an explanation of the AC silicone pad setup in response to your "I have no idea what you're talking about" when I mentioned them above. Here is my PM in full with the link so folks can understand what you are referring to:

" registeredthing
registeredthing 8:26 am

The AC silicone heaters for 3d printers look a lot like oil pan heaters. The big advantage is that they heat up quickly.

This video explains the AC silicone beds with an SSR used to control it:

I had not seen the silicone DC beds, but after looking around on ebay some more, I found some in 12V which look similar (my printer is 24v however). I'm OK with the speed of the 24v (PCB) bed on my Ender 3 but may add some insulation on it when I rebuild it with linear rails."

You seemed to be convinced that there's a difference in heaters using AC vs. DC.

'Those slicone heaters look nice, but the only ones I have seen are for AC and would require a relay or SSR, so probably will be a while before they make it to the low end printers.'

That statement right here. A resistive wire doesn't give two shakes of a rat's ass if the current alternates or not. You want to see a proper silicone heater that doesn't require mains? Here is one. Saying that a resistive element requires an AC is like saying a wind chime only rings when a southerly wind hits it. It doesn't make any sense and goes against the most basic principles of EE.

If you double the voltage for the same resistance you quadruple the power. This has nothing to do with AC/DC. For AC you can find the RMS value, for mains 120/240v depending on country, then use apparent power to calculate the power loss. Resistance is resistance unless you're calculating and adding the inductive or capacitive reactance cause by the 50/60 hz; a negligible value by the way for a non-solenoid wiring scheme. Those 'AC heaters' you're talking about just have a higher resistance to them to limit the current and therefore power loss. You can use some 120v DC just the same way, it doesn't matter. Kanthal, for instance, would react just the same at 10v AC as it would with 10v DC. If you don't understand that, you don't understand ohm's law nor basic analog circuitry.

That man is a danger to not only himself but others as well.

Haha, I am starting to think you don't like me for some reason. But allow me to explain.

Most printers use a 12V or 24V (DC) power supply and therefore the heating pads for 3D printers are sold in reference to 12V (DC), 24V (DC) or sometimes 120V (AC - when using an external bed relay or SSR), along with the power rating of the pad. I know of no printers that run on 12VAC, 24VAC or 120VDC, but I am no expert as you well know. As well, most sellers don't usually list the silicone heater pads as 1 ohm, 4ohm or 29ohm, 40ohm etc, though that would be more accurate. The listings do not state that they are for AC or DC for the reason you well described above, which is largely obvious. I think you extrapolated a lot based on my using AC and DC "versions" as a shorthand for the heating pads that are sold for 3D printers. I mean, I guess I could use a 40ohm heating pad with my 24V Ender3, but it won't every get to temp, so I'd probably just get the "24V DC version" and go back to eating cheetos.

You can judge them by the power rating at the voltage given. My 200w 200 mm² bed runs on 12v and gets up to 90c with ease after PID tuning for reference. You'd have to get one that your PSU can run with some safety overhead. A good way to look at them is watts per cm².

Don't like you? I don't care one way or another, I was trying to explain something. I'm too crass sometimes, I get that, nothing personal.

If I upgrade mine to a silicone mat I'd probably need to stay around 200W also, not sure what (actual) mosfets they have on the stock board, and right now it is about 150W (6A), I think 8A may be about as much as I'd trust going with it. I've got too many damn projects and not enough time though.

Find the part number on the FET and see what the Ids max value is on the datasheet. Logic level FETs are usually pretty high current gain devices with low Rds on values.

Here's another

And another

And another

And from eBay

another from eBay

They're all over the place, they have many, many uses. There are manufacturers that will make these custom to order and those are dirt cheap. Diesels commonly use them to prevent the fuel oil from slushing up and not flowing correctly. Diesels aren't connected to mains either. I have to wonder how you missed them all.

Thanks, I will have a look at that meter.
I thought the only difference for the mk2a bed was just the difference of the number of soldered attachments. Single pad being soldered for twelve volt and to both pads for 24v. That's for the ones I've seen, anyway.
Although, it would explain the ohms issue.

If your board smokes is not good for its health ...:))

Do you have a Zrib board or a 2560 with Ramps ?

Anyway, if it is the mosfet that is gone you may get away with replacing it, maybe with something that can handle some more power than the original one. There are external boards that you wire to the power supply and to the hotbed output signal and handle the power side relieving that from main board.
A common problem on many boards is that it is not the components that blows up but the power track of the board. If this is the case, you may get around with one of this boards I mentioned, you can search for "hotbed mosfet module" online. Another possibility is to get a new ramps board that is pretty cheap, or there are also complete boards with everything on, like the mentioned Zrib board or any other atmega 2560 based board.
You have many choices as most of the printers runs on Marlin firmware and use almost all the same components. If you have this Zrib board replacing the mosfet on the board may be tricky but wiring up an external one is easy, just find out which pin of the mosfet is the signal and use that to control the external mosfet...
Good luck

Thank you for the advice. My original intent was just to get a board that has the options of the original board but better built.

The board I have is described this way on the site.
A mix of MEGA2560 + RAMPS1.4 + 5 Motor drivers.
Firmware can use the similar configuration as RAMPS1.4.
Improve 3-D machine upgrade into twin extruders and lightweight engraver.
Improve upgrade with filament run-out detect and resume serve as.
One of the crucial driving force can connect to 2 Z motors.
4 PWM capable energy mosfet outputs with 12 – 24V wide voltage vary for MainPower.(Mattress, Extruder0, Extruder1, Fans)
Recoverable fuse for brief-circuit coverage.
USB driving force chip is CP2012.

Voltage: 12V – 24V

Thanks again!

Well, as it looks like you have an extra mosfet for a second extruder, you can swap the pin number in Marlin config files to use it for the bed instead of the burned one, but you may still have a problem if the bed is shorted or use too much power for the mosfet and burn that one too...
As you have an all in one board, you cannot use a ramps board that normally plugs on top a mega2560, so your options are to replace only the mosfet or use an external one or get a new all in one board or get a 2560+ramps.
An external mosfet in my opinion is the best way to go, pretty cheap, easy to wire, and they are full size mosfets with a real heatsink that can handle way more power than the smd ones of most boards. If you go this way, consider using one for the hotend too...

Thanks, Those sound like a great ideas.
The more I look at the way the mk2a heated bed is designed the more I worry about it shorting out or, me getting a shock. They seem to work for most people so maybe I'm being paranoid.
I've been jolted before and immediately decided I didn't enjoy it at all so I want this "upgraded tronxy x3s / electronicgeek clone" to be "right".
While I wait for the parts I will focus on trying to make this thing safer.
If anything jumps out at you about using the 30 x 30, m2ka bed at 24v feel free to share!

These RAMPS boards use knockoff FETs rated for Chinese amps, to get another one would be absurd.

I don't have a ramps board so that may be true, but many printers around runs on these boards. Maybe in this case a 300x300 heatbed is a bit out of the mosfets capabilities...
I have a Geeetech printer and the board is similar to the MKS gen and it's working fine for over a year but I bought two external mosfet boards because I was afraid the power tracks would not hold. Still working fine without them so I didn't install them yet but I plan to do it.

The size of the bed is a bit problematic. Once I get it going and the enclosure is done it will likely present other problems that will need to be addressed. One thing at a time.

I would purchase a board like the MKS gen 1.4 that doesn't just have FETs flapping in the breeze straight from the Wun Hung Lo factory. I once owned a RAMPS 1.4 for all of two hours until it melted itself into oblivion. The power connections were piss weak. I've had my MKS gen for over two years now and the magic smoke is still in the FETs where it should be.

You can take a look at the Repetier online config tool when you get a new board. Don't throw the printer out or you'd waste those motors and frame for nothing.

Thanks for replying. I've looked into a couple mks versions.
Mostly gathering info and opinion at the moment. I've considered the smoothie platform too but it is a little pricey.
I'm not that sure of Zrib quality but I thought about trying to find a better quality mosfet for it and going with a new board, as well. I could always use it in another build if this is too demanding for the Zrib.
I know it's a big build plate and it takes a lot to power them. It was the thing that made me most cautious when deciding on this printer but there didn't seem to be many complaints, at first.
Thanks again.

Your PSU is a 24v one right? For the same power you're running less current which makes that board even more suspect given that Vds on should be pretty low for a MOSFET during saturation. Though, I was under the impression you were using a RAMPS board. Now I'm confused, which is it?

Descriptions like this can confuse a new guy, a bit.

"Parts & Accessories
Zrib board ..Zonestar
A combination of MEGA2560 + RAMPS 1.4 + 5
Motor drivers; Z driver can connect to 2 motors.
MEGA compatible for Atmega2560 processors
and for all RAMPS classics firmware.
Compatible for dual-extrude printer, support
upgrade into light engraver and filament run-out
detect and resume function."

I found this a bit confusing and wasn't sure whether to refernce it in my description or leave it out.