Where to get a replacement Mightyboard Rev H for Rep 2x?

[Gian Pablo]

So I've been having persistent problems with my R2X, and today I noticed a strange behavior: the right extruder fan turns on immediately when switched on, but not at full speed.

Checking the voltage at the motherboard shows it is delivering 18V, not 24V.

I suspect something is wrong with the Mightyboard.

Where is a good place to buy a replacement Mightyboard for a Replicator 2X?

[Jetguy]

Well, the circuit doesn't work like that so your measurement is probably not accurate.
The + lead or pin of the fan connector is constant on 24V. It is not through any transistor or FET. You should to comparing that to input ground.
The - lead of the fan connector is the side that is switched via the MOSFET. During operation- compared to actual source ground, that might be a fraction of a volt above ground.

The only good place to get replacement is from MakerBot or a distributor and the only reason for distributors was other countries- hence why https://www.bilbycnc.com.au/ is the only other place I know of mentioned.

"the right extruder fan turns on immediately when switched on, but not at full speed."

This indicates a failed and blown into the shorted state MOSFET.

We've shown threads of replacements or I'll do it for you as a favor since I owe you one.

[Jetguy]

We are talking a part that cost pennies to replace VS several hundred if they charge you for the board.

[Jetguy]

Here is the picture marked up from another thread

I've circled the MOSFETs in question in red. Specifically, Q11 and Q9 control the extruder fans, Q7 controls the main blower duct on a Rep-2

I found the specifics here:

Dan Newman posted in the Tips and Tricks group in reference to a user with a 2X obviously a Rev G or H board https://groups.google.com/d/msg/3dprintertipstricksreviews/Vgq2z2_x1zc/9PfZqflsbBIJ

"The MOSFET they use for the cooling fan is a 2N7002K which is good for 60V but 
only 0.3 A. So make sure that the two fans combined use do not exceed that 
current rating @ 24V.

[Joseph Chiu (Toybuilder)]

So, uhm, I ended up with 1,398 2N7002's...  Well, I have maybe 1,395 left now...  If anyone wants one, drop me a line...

Line No: 2	Stock No:	Manufacturer Part No:	UOM:	Quantity:	Price:	Extended Price:
	65W8792	2N7002-7-F	Tape and Reel Cut 1	1398	$0.017	$23.77
	Customer Part Number:
	Description: MOSFET, N CHANNEL, 60V, 1.2OHM, 115mA, SOT-23; Transistor Polarity:N Channel; Continuous Drain Current Id:115mA; Drain Source Voltage Vds:60V; On Resistance Rds(on):1.2ohm; Rds(on) Test Voltage Vgs:10V; Threshold Voltage Vgs:2.1V

[Dan Newman]

On 29/07/2015 3:04 PM, Gian Pablo wrote:
> So I've been having persistent problems with my R2X, and today I noticed a
> strange behavior: the right extruder fan turns on immediately when switched
> on, but not at full speed.

You have a failing MOSFET for that fan. They cost only a few dollars and are
easy to replace if you have soldering skills. There's been past posts about
replacing them. The part is a 2N7002K,215 (or equivalent).

> Where is a good place to buy a replacement Mightyboard for a Replicator 2X?

MBI is the only supplier. They are not readily interchanged with a rev E
mightyboard such as you can buy from FlashForge or WanHao. If you replace
the board, then you need to replace the LCD / Keypad / SD unit also. (The
wiring is ever so slightly different and will short 3.3V to ground if
you plug in a mightyboard rev E.) But here's the fun part, the LCD display
is a slightly different size between the Rep 1 and Rep 2X and so slight
case modification may also be needed to put in a different LCD display.
Or not. All depends upon what exact size the replacement is.

Dan

[Ryan Carlyle]

Wow Joseph, that's pretty awesome :-) I'd take a few but I don't think they're worth the postage!

[Joseph Chiu (Toybuilder)]

Yeah, it was the same kind of thinking that made me go from ordering 100 to 1,000.  Then I realized that I might as well be nice and just take the rest of the reel they had left...  I figured I was already buying some...

[Gian Pablo]

Thank you for your help! I think I will be trying the replacement route.

In the meantime, is it feasible to use the motherboard port for the R2 fan, and just turn it on with GCode?

[Jetguy]

That's a work around but better would be hard wiring the fan constant on. The concern is forgetting to turn it on- hence the automatic feature of anything above 50C.

[Jetguy]

While on the topic of repair and the fact we seem to have plenty of spare MOSFETS and they are cheap, I'm thinking of paralleling at least 2 by stacking them vertically.
Given the weak nature and the failures we hear of all the time- that simple fix might be enough to stave off future failures.
It's trivial to stack these tiny 3 pin SMT devices. Maybe not pretty to some folks- but functional and cheap beats "pretty" IMO.

[Jake Clark]

Gian,

We, Fargo 3D Printing, can get a 2x board for you. Just shoot me a message and I can get a quote made up.

-Jake

[Ryan Carlyle]

You know, I was just wondering last night if there's a good drop-in replacement with higher rating. I suppose stacking two could work.

I'm due for some R2x board work anyway, mainly swapping the 1280 for a 2560, so I suppose I could try the stacked FET mod. Unless somebody has a better idea. 

[Jetguy]

I "think" there is a 4A variant in that package size with the obvious limitation is really the overall die size, size the solder legs and lack of any heatsink other than the leads to pads connections.

 

The RAMPS-FD from GEETECH uses the same size FETS (I think - no guarantees) and they brag they had higher ratings. (the 1.2 version I have, not the 2.0 version you got)

As with all these smaller parts- the absolute peak instantaneous max is often quoted without stating all the limitations around that max, so de-rate by 50% or more as good engineering practice.

 

As always, at best we have 5V gate drive and that normally is another de-rating value that needs thrown in. Rarely do you find the triplet combo (High voltage rating, high current rating, and ultra low gate threshold or full RDS ON at 4.5V or less). By design, those characteristics are a function of how the FET is made in the silicone. Without even going into deep theory, the fact you don't want the device to "self turn on" means that high voltage (relative here- 30-60V) VS a full on state at 5V or less is opposite ends of the spectrum. High current means  larger actual transistor zone AKA physical junction size and copper substrate. Given the package size and the polar opposites of rated voltage VS turn on voltage- that's the limits.

[Joseph Chiu (Toybuilder)]

Perhaps the right 2N7002 needs to be spec'd?  

http://www.digikey.com/product-search/en?FV=fff40015%2Cfff8007d&k=2n7002&mnonly=0&newproducts=0&ColumnSort=-612&page=1&stock=0&pbfree=0&rohs=0&quantity=&ptm=0&fid=0&pageSize=25

There's a pretty wide variation -- the Id(cont@25C) varies from 115mA to 475mA (and Rds from 7.5 ohms to 1.6 ohms) -- it may simply be that the ones on the MBI Mightyboards have low current limits...

So the 2N7002K that I bought in bulk* are rated only for the 115mA which is perfectly fine for where I bought them for, but may be marginal with the 24V fans that seem to vary between 50mA to 200mA.

[*] If I have repaired your Mightyboard, don't worry: the 2N7002K I have purchased for Mightyboard Repairs are good for 300mA.

[Jetguy]

If we play around and work backwards, we say we know we have SOT323 sized and spaced mounting solder pads, we know we want a minimum of 30V rating given a 24V source, and we know we have 5V gate drive and the target is what is the maximum current rating @ those limitations.

http://www.digikey.com/product-search/en?pv7=2&FV=fff40015%2Cfff8007d%2C97c0005%2C97c001b%2C97c001d%2C990071c%2C990110c%2Cefc0005%2C142c01bb&mnonly=0&newproducts=0&ColumnSort=0&page=1&stock=1&quantity=0&ptm=0&fid=0&pageSize=25

 

I knew I wasn't crazy and had seen a 4A "rating" in that SOT323 form factor. Now if that is a real rating or if somebody somewhere in some lab under ideal conditions said- that's the max rating is another discussion.

 

Other fun facts dug up after reading the fine print in a data sheets. Digikey sells both part numbers and it's cut tape so you can buy 1. However, do you personally care how big the reel it was it came off of if you don't buy a full reel??

Part Number                     Case            Packaging

DMN3065LW-7                SOT323         3000/Tape & Reel

DMN3065LW-13              SOT323       10000/Tape & Reel

 

I know this is probably a lot above folks heads but buying "the best FET" is not easy, not by a long shot. There are thousands of package sizes that are within reason to each other. As Joseph pointed out, there are dozens of ratings even within the same basic part number. Every manufacturer puts a spin on the information in the data sheets. It's really kind of crazy.

 

Technically, the 2N7002 is an SOT23 (TO-236AB) sized device. However, I knew that SOT323 is within all reasonable size specs, the same thing.

And knowing the circuit is 24V and 5V gate, and we want the same pinout, same soldering layout, opened up a much larger and different list of part numbers in the Digikey search.

And that's not to say their aren't other parts that might fit the pads and have even better ratings. Being honest, it seems there is always a compromise. We are replacing a 60V part with a 30V part- except we in theory only need 24V rating with some "headroom" but by compromising the overrated voltage down to 30V, our current capability jumps by a huge factor.

http://www.digikey.com/product-detail/en/DMN3065LW-7/DMN3065LW-7DICT-ND/4810993

 

Now what is also crazy is the price break. 1-9 they are $0.51 each. 10 is the magic number they drop to $0.362 each and it gets lower quickly. So just saying, I call this a 10 unit minimum.

 

I make no guarantees this is the best part. Based on what I know, reading a data sheet, doing a bunch of comparisons, this appears to be the highest current rated device at a gate voltage of 5V and meets the bare minimum safety voltage spec of 30V which is 6V above our expected operating voltage of 24V. Based on that- and the relative cost, this is worlds of current handling above the stock 2N7002s.

 

 

 

 

[Joseph Chiu (Toybuilder)]

Well, I just recently spec'd out it's smaller brother SOT-23 P-FET that goes up to 4.3 A, but only 20V... (http://www.digikey.com/product-detail/en/DMP2100U-7/DMP2100U-7CT-ND/3829371) -- It's got Rds of only 38mOhm!

But SOT-323 is not an ideal fit for SOT23-3 pads.  I would instead go with the DMN3033 in the SOT23-3.  It drives 6 Amps.  It's 51 cents in onesies, and 43 cents in ten-sies.  http://www.digikey.com/product-detail/en/DMN3033LSN-7/DMN3033LSNDICT-ND/1774327

[Joseph Chiu (Toybuilder)]

Oh, and you'd ideally want Vgs down in the low 2V range (2.1V in the DMN3033)...  You mentioned 5V gate voltage -- which I think you really meant to say Vgs compatible with 5V logic...

[Ryan Carlyle]

You guys sure 30v is adequate? Don't BLDC fans cause voltage spikes if you stall them? Or am I mis-remembering that?

"Insert hex key to destroy FET"

[Jetguy]

Well, I don't know that that's the best way to search.

 

I agree, yes, we probably are going to end up with a low voltage threshold. No argument there.

However, the way manufacturers are doing this and more so- how Digikey puts them into the catalog DB and makes the fields searchable don't line up.

For example, searching on gate threshold alone- well that's not the whole curve. In theory yes, the sooner we turn on- the more likely we are fully turned on at 5V.

However, the database doesn't give a range unless you select multiple values- then leaving you some huge list. The part I identified listed it's current at 10V gate (just like thousands of others).

There was a search option for 1V gate threshold- but that part did not have the highest current rating. Choose TTL or logic level gate may omit perfectly good parts that might even be "the one" because that is ONLY if the manufacturer listed those words and the poor soul entering the parts into the DB happened to add that "flag".

 

Also, not trying to make a big deal but according to the data sheet, the pin spacing, part size and solder pads exactly 100% lined up between the SOT23-3 and the SOT323 part.

And, looking at this data sheet, the Z spacing (Front 3rd leg pin to the back 2 pins across the chip length) is actually longer than the spec I saw in the 2N7003 sheet for SOT23-3.

 

I'll admit, I'm not an expert, I don't do this for a day job- nor do I play one on TV.

 

I do think this on paper the http://www.google.com/url?q=http%3A%2F%2Fwww.digikey.com%2Fproduct-detail%2Fen%2FDMN3033LSN-7%2FDMN3033LSNDICT-ND%2F1774327&sa=D&sntz=1&usg=AFQjCNFNDAsEiSTVaS9aq3ay4AuhOHQ0AA is a better FET than the one I linked. I think it will probably solder to the pads and the fraction of a mm I saw in the data sheet means nothing.

But that's the problem, there is a lot of "room" for opinion. Both end up being 51 cents in single and a significant discount in 10 count or more.

 

What I'm trying to show folks is that building and designing electronics takes a LOT of background and experience. It's not just clicking a few option boxes in a Digikey search. 2 folks might come up with different answers and we both could be wrong (not saying you are or that either of us is).

And this is literally why electronics have failures. Some bean counter gets in the middle and swaps in part substitutions without knowing the whole picture or why the engineer who designed the circuit chose any one part. A substitution happens and one of the specs (voltage, current, thermal, threshold, resistance, material type) gets' changed and it puts the entire circuit out of whack.

[Jetguy]

Ryan, you absolutely are 100% correct in asking and that's why I caveated this whole thing in that from a testing perspective- we are completely in an unkown- unknown.

We don't know what we don't know.

 

A brushless fan must have a brushless controller internally. That means that it has some sort of transistor internally to commutate the 2 or more phases of the brushless motor. That likely means it has an FET that has a Voltage rating higher than the labeled input and diode to arrest the "kick" from the inductor the instant that phase is turned off. Most all brushless fans have some capacitance across the input to buffer both incoming power and reduce any surges/spikes/EMI back out of the input. Or so that's the theory. But we don't know that fan does have all that. It could be a really shoddy fan with no capacitance and is EMI and power surge prone.

 

In general, I don't expect significant back spikes from any fan. But to say it doesn't exist- that I cannot prove.

[Joseph Chiu (Toybuilder)]

It'll probably work, but that ~0.5 mm difference in pitch (Dim G and C) is enough to cause grief.    Uhm, not that I'd make that kinda mistake, or soethin'... <whistle into the air>...  :)

Per Wikipedia:

• SOT-23 (SC-59, TO-236-3): 2.9 mm × 1.3/1.75 mm × 1.3 mm body: three terminals for a transistor ^[30]
• SOT-323 (SC-70): 2 mm × 1.25 mm × 0.95 mm body: three terminals ^[31]

 

 

SOT-23 DMN3033SN

 

SOT-323 DMN3065LW

[Joseph Chiu (Toybuilder)]

You're right - you can have a much higher Vds from an inductor as it's being shut off.  

But in this case, I would expect that to be handled by the controller on the fan, as it is already doing the work of chopping the motor coils on the fan...

[Jetguy]

What you might be thinking of is the statement regarding the stock FET in a 2 or 2X printer.

All we know is it's a 2N7002 "family". We say that because due to the way these are marked, it might be nearly impossible to determine the actual specs and real world current rating.

And, that current rating is temperature based on airflow and ambient temp of the board- maybe even how big the trace or pads are for one FET or the other.

 

What we do know is the very best rating on paper is probably a realistic 300mA which is could be exceeded by many fans when stalled.

And because there are lower rated parts- as low as 115mA and those are under the normal current draw and actually being pushed to the breaking point in theory- maybe the stock one is somewhere between the lowest and highest rating for the general part number. Given the reported failures- all this seems valid.

 

So the theory is- trading a voltage rating for a higher current rating might prevent future burnouts and improve the circuit.

We think we are exceeding current- not voltage and therefor need to fix the current rating in the best way possible.

 

Of course we could always dead bug some bigger FET into the circuit- part of this is just as much about form factor and installation along with long term reliability.

Hence- why the search for a better part isn't "easy".

[Jetguy]

And yes, I see where I may have looked at a dimension incorrectly and the one you linked is obviously a drop in fit- it's exactly a SOT23-3 just like the 2N7002.

I have several data sheets open at the same time and probably compared the wrong line or something. SOT323 is narrower by 0.5mm pin spacing as you pointed out.

[Gian Pablo]

All right, so what's the consensus? Should I get some lower voltage / higher current FETs and do the surgery myself?

In the interim, I'm just thinking of getting a bench supply and running both fans directly from that. I'm doing experiments with higher flow cooling anyway so this might be a good thing to do in any case.

I'll just stick a post-it note on the supply to remind myself to turn on the fans every time :)

[Ryan Carlyle]

If you're going to bypass the FETs, I would hard-wire the fans directly to the R2x PSU somehow, it's pretty dangerous to run the heaters without the fans on. Not the sort of thing you want to rely on memory / post-its. 

[Gian Pablo]

OK, how do you feel about the DMN3033 in the SOT23-3 package? Should I get those and do the experiment?

I have a backup plan, so I am comfortable taking the risk.

[Jetguy]

I say install DMN3033
30V is fine.
Worst in theory is that let's say it did flash over. Well, then it's stuck on like now.
Certainly no worse than the stock situation and not expensive.
Pinout and layout is identical. Simply desolder the old one and solder in the new.

Personally, I would use the blob method. I just get a blob of solder going around the tip and the pins to keep them all hot at the same time. The flux keeps the solder in a clean ball and then the bad chip I walk sideways off the pads and remove with tweezers. Wipe the pads one time with the cleaned bare soldering iron tip to "level" the pads but still have a tiny bit of tinning.
Use tweezers to orient the new one and tack one pin. Then tack the others and she's done. I don't even use desoldering tools. I use the fact solder tends to ball up like mercury in the presence of good flux. You control it with heat and it's like magic.

[Gary Crowell]

BTW & IIRC there is a 27V ESD zener on all outputs,  so there is some protection there for a 30V FET.

[Joseph Chiu (Toybuilder)]

I really think 30V should be fine -- the fan's motor control shop keep the inductance kick under control.