just wondering how people are connecting heated beds to there electronics?

from my research i see that the bed takes just under 7A of power

now heres were i see problems

the header and pin connection are only rated at 3A

even splittling them like the Sanguinololu does over 4 pins instead of 2 only gives 6A

then the other problem i see is the wire needed does not support the crimp connectors correctly.

im thinking its safer and cleaner to build a small board and run the heated bed separate to the board.

It's worse than that. The bed takes 10A when it is cold.

I wired 12V direct from the PSU to the bed and connected the other bed lead to the tab of the MOSFET using a ring terminal.

---

[www.hydraraptor.blogspot.com]

in reality they should of designed the baords so that the mosfet was not on the board and on a smaller inline board with decent 10A terminal connectors to be safe.

might have to look into deeper into doing this as iv already had a connector melt from too much power.

just for notes when i got an ameter on my heated bed i was getting 6.8A from cold. But was using very thick wire so might of knocked the power draw down.

looking at the ramps board it uses the correct terminal pins for both extuders and also heated bed.

its starting to make sense now why the Sanguinololu comes in much cheaper.

Most electronics use screw terminals for the heated bed and the hot end. I think this was a problem with 1.2 or earlier RAMPS so you need to run +12V directly from the PS to the heated bed, but I thought the later versions handled this much better with separate power supply input for the bed. But the MOSFETs probably need heatsinks and they don't seem to come with it.

I've been using the Azteeg X1 and it works quite well with small heat sinks on the MOSFETS. I'm about to test the Azteeg X3 which uses lower resistance MOSFETS so it doesn't need as much heat sink.

There really isn't much of an issue on having a FET for this purpose *on a printed board, as long as the subcircuit
- uses a dedicated and isolated supply line (so you're not forced into sharing the power supply for the processor/etc with that of the heater)
- has sufficiently large traces that are properly placed on the board
- uses a sufficient FET for the purpose (imo over-engineered rather than marginally)
- provides sound thermal management constraints for the FET (sufficient room for a heat sink and/or edge mounting so it can be attached to an enclosure wall)
- uses proper termination at the user interface (ie large enough screw terminals or connectors, both to accept the wire and to *comfortably accommodate the current).

There are a couple of additional things I might do were I to design a board, but that covers the lion's share.

Based on the boards used by this community that I've seen so far, I tend to agree that an offboard temp controller might be a good idea, at least for the bed heater.

I use a 12V relay (with a diode across the coil to kill the back emf) to isolate the currrent to the heated bed. But I only use bang bang for this. it works for me.

Quote

in reality they should of designed the baords so that the mosfet was not on the board and on a smaller inline board with decent 10A terminal connectors to be safe.

This reality: [reprap.org] ?

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

I don't think all of this is really necessary with a Mk1 PCB or equivalent at 12V. My RAMPS 1.2 has been working fine for over a year after I added a heat sink to the bed MOSFET. I've printed over 60 lbs of ABS with it. For my second printer, I've been using the Azteeg X1 without any problems and it looks like the Azteeg X3 does even better with heat management.

yes i was thinking of using a relay but the clicking would do my head in

i dont think ramps has a problem as its has correct terminal headers

Sanguinololu electronics are okay and fine its just the connections which are the problem the connectors cannot take that many amps safely.

ideadly we could do with a 12v reading mosfet which then handles the output to save modding the board.

for techies its easy enough to mod the board and solder wires and put the Sanguinololu mosfet to terminals but would be nice to have a simple plug in solution.

The simple solution is earlier in this thread: [forums.reprap.org]

---

[www.hydraraptor.blogspot.com]

? the diode solution?

if so can we get an diagram of this on how to wire it up

EnlightX

I think nophead was referring to the use of an offboard switching solution, like the SevenSwitch PCB that Traumflaug noted above. (The IRFZ044 is rated at a far higher max current and voltage than is used for the heater, so it should be VERY solid.

Or he could be referring to the solution in the post he linked directly too.

Quote
nophead
I wired 12V direct from the PSU to the bed and connected the other bed lead to the tab of the MOSFET using a ring terminal.

hmm...good point. Odd, I could have sworn the link took me too traumflaug's post earlier when I used it. Oh well.

Yes the simple solution is not use the bed connector. One side is +12V so take that direct from the PSU. The other side is the drain of the MOSFET which is also the screw tab. I connect the other wire of the bed to that with a ring terminal .

The current rating of the MOSFET is largely irrelevant as you would struggle to find a low voltage TO220 MOSFET rated at less than10A these days. The important things is RDSon at Vgs a bit less than 5V. That determines how hot it gets.

IRFZ044 is not a logic drive FET so not guaranteed to have low RDSon when driven from 5V. The FET on my Sanguinololu gets quite hot but survives. The ring terminal and thick wire I use take some of the heat away. A heatsink would be a good idea. Better still replace it with a logic drive MOSFET.

.

---

[www.hydraraptor.blogspot.com]

Quote

The FET on my Sanguinololu gets quite hot but survives.

As does the IRFZ 44N.

I mean, I've put quite a bit of research into this issue and for me it's simply impossible to destroy this type with an 8A current sink. The data sheet says 4.0V (worst case) are sufficient for switching. Similar types with a lower gate voltage behave about the same. I'd appreciate to see measured facts if one disagrees.

That said, applying the Gen7 v1.4.1 upgrade to the SevenSwitch is the obvious next step.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

The worst case gate threshold is 4V for IRFZ 44N. That means it will guarantee to pass 100uA at 4V, not 10A required by a cold bed.

Logic drive fets specify the RDSon at 4V and have a max gate threshold around 2V to achieve a reasonable figure at 4V. The only reason your IRFZ 44Ns work is luck that you haven't encountered the worst case yet. You see people in these forums that are less lucky and report that their MOSFET melted.

Why do you think logic drive FETs exist if it is fine to use ordinary ones with logic?

---

[www.hydraraptor.blogspot.com]

I'm sick of this discussion. If there were all these people showing melted MOSFETs and if you had actually run such a MOSFET yourself I 'd trust you. But in fact many hundred people are running Gen7s successfully and complaints come mostly from those who didn't even try. Maybe there are some who were out of luck and fetched a sample so much worse than everybody else. Maybe they just forgot to mount the heatsink or ran high-frequency PWM on their bed. Or both. You can point them to the reports about melted Sanguinololus.

Anyways, it's history. Gen7 v1.4.1 is out. With a MOSFET so much oversized, making even the heatsink obsolete.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY