Heated Bed wiring meltdown

[adamfilip]

I was getting a max temp error on my printer
When I investigated the cause I found out that the relay terminal block was slightly burnt and the wires coming from the heated bed that go into the Terminal block were melted together with the therm sensor.

The Therm sensor is dead and I need a new one. I tried splicing in a new wires but it will no longer produce a reading.
The relay switch also needs replacement as I can no longer open and close it

update : i ended up changing the terminal block with something similar from my local electronics store and replacing the thermal sensor with one from my 2nd hotend.
so im back down to one hotend.. for now

[adamfilip]

BTW these are the wires that come presoldered to the heated bed element. and not the wires i added myself from the power supply

[beerdart]

Loose the little ​Relay and get one from the auto parts store.

[sniffle]

i had the saem thing happen adam, i just soldered wires directly to the PCB on one side and resoldered the terminal on the other to make sure it was soldered well...

[N5QM]

So, what do we expect the root cause of this is, the little relay or?

[sniffle]

Slightly loose wiring?... Dunno really

[Stigern]

I had the same happen to me, short story long.

Changed wires to 12AWG and got me a beefy 30A car relay!

So, what do we expect the root cause of this is, the little relay or?

I believe the contact surface is too small in those green screw terminals.

[printbus]

So, what do we expect the root cause of this is, the little relay or?

If the relay was the problem, the relay itself would be undergoing the meltdown.

For everyone's reference - from the markings on the terminals at least on my relay board, we know the terminals are FCI series 12-762 parts. A two terminal version is likely FCI 20020705-M021B01LF. The part is rated for 30 amps. The datasheet is a bit confusing on wire capacity - first saying it is good for up to 10 gauge wire, but then specifying the allowable circular mils (basically the area of the copper surface viewed on end) for stranded wire at 4 mm^2. #12 AWG stranded wire (3.31 mm^2) is the largest common stranded wire that matches that spec.

As stigern suggests, contacts in some of these screw wire-to-board terminals leave a lot to be desired. Some, for example, seem better suited for solid wire than stranded wire. The contacts in these actually don't seem too bad - at least the wire strands can't "squish" out from under the screw - an issue I often have on smaller terminals.

Especially with the 12-inch printers, we're borderline on a number of things here, based on a worst case load of around 30 amps. The relay is rated for 30amps, the terminals are rated for 30 amps, #12 wire is already arguably a bit undersized for 30amps. Everything needs to be perfect. Wire needs to be stripped so that the bare wire will fully insert into the contact with absolutely no insulation protruding into the contact. If ANY wire strands are cut or nicked during the stripping process, the wire needs to be cut off and re-stripped. It has been argued previously, but industry standard typically says NOT to tin the stripped end of stranded wire when the wire will be put under pressure such as here. After wire insertion, verify that ALL strands are in the contact - that none have folded over or otherwise didn't get inserted. Screw terminals need to be tight. I didn't decode the torque spec on the FCI data sheet, but I torque things like these as hard as I can with my small-handled screwdrivers. Stranded wire can tend to shift with wire flex or with time, so screw terminals should always be retightened after a while. For these where we're borderline, I'd check them periodically, also inspecting the terminal block and the wire insulation for any sign of distortion or melting.

For those that are comfortable soldering heavy gauge wire, just soldering the wires to the relay board can bypass all these issues with the terminals. You can't do it with #12 wire, but a standard practice I usually apply is folding over the stripped end of the wire so that twice as much wire is being clamped by the screw terminal. That can be helpful when you're trying to use say a #24 wire on a terminal good for much larger wire.

There's also nothing wrong with bypassing the relay board entirely and going with freestanding relay like an automotive type.

A final comment on wire flexing - flexing wire is generally not good for it. Over time, it can lead to increased resistance and broken strands. Some flex is necessary - like in our wiring to the extruder and the y-bed. But don't allow that flex to occur at the wire connection points like these screw terminals. Use wire ties some where, some how so that flex is limited to an open length of the wire. I think there are Thingiverse designs for cable support brackets that extend rearward off the frame. For the relay board, consider raising the board on standoffs so you can loop the wires under the board as a strain relief. Anything to keep wire flex from reaching the screw terminals.

Once a terminal gets a little loose or we start damaging the wire end from flex, things will degrade rapidly.

FOLLOWUP COMMENT: When later removing the relay board from my printer as part of the migration to Smoothieboard, I noticed a new issue with the FCI terminal block that could explain the meltdowns. What I observed is that the design of the terminal block pins does not allow the connector block to sit flush on the board. The connector block is off the board by a portion of a mm. That may not be a lot and likely doesn't matter for most applications. In our case, wire pushing and pulling on the terminal block could rock the connector back and forth, especially for those who have used large gauge, typically low strand count wire. Over thousands of bed movements, this rocking will stress the solder connections of the pins on the board, and eventually cause the resistance of the connection to go up. That resistance will lead to heat. Reflowing the connector didn't help in how my connector was seated on the board - there's apparently something about the pins that prevents them from inserting farther into the board. The available options in improving your luck with the relay board work include...

• strain relief the wiring as mentioned earlier so that wire movement can't push and pull on the terminal block
• use high strand count wire like that intended for radio control hobby use to minimize strain on the connector (#12 R/C wire has something like 400 to 700 strands, and the silicone insulation helps keep it very flexible)
• add shims under the connector block body so that it can't be rocked back and forth
• glue or otherwise bond the connector body to the board
• solder the wires to the board instead of using the connector

[ex-egll]

Has anyone tried using a solid state relay for the heated bed?

[adamfilip]

What about using 2 lines of 12 guage for positive and negative. to help carry the load
I do notice that since i changed up the terminals.. the heated bed warms up slower.