burned contacts on bed heater connection

[ivymike1031]

I completed my PrintrBot build on 21-Nov-2012, and I've been printing things with reasonable success up until now (approaching 2kg total filament used). My heated bed is installed on top of 8mm of Neoprene foam, and the foam is on top of the birch table. I have some metallic tape on the top surface of the neoprene to distribute heat and reduce oxidation of the foam. On the top side of the bed I have a 1/8" sheet of glass. I usually warm the bed up to an indicated temperature of 90C before I start printing, but I have it set to 81C (indicated) while printing I have not ever attempted to measure the temperature of the top of the glass.

I had noticed instances where the printed bed temperature would fall off while printing... sometimes down to 69C or lower... other times it would hold 81C without trouble. I had attributed this to a software problem and hadn't pursued it. Yesterday while printing I noticed that my bed temperature dropped off all the way to 29C. I didn't have anything else to print, so I shut down when I was done and didn't mess with it. Today I went to print something else, and the bed wouldn't warm up significantly. I got it to rise from 18C to 18.67C, then eventually 19.4C. I have the printrbot neatly mounted in a wooden box, so I loathe pulling it out and unsnipping all the wire ties, etc., to service it...but I went ahead and did it. What I found was that the heated bed connector at the Printrbot main board was charred and blackened around each of the 4 pins . I disconnected the bed and gave a warmup command, and did not find any voltage at the pins . I don't usually trust a single measurement, though, and after four or five power cycles and re-measurements I started reading 12V at the pins (so either I did a crappy job measuring at first or the pins came back to life).

I'm considering installing an external relay in the heated bed circuit so that the PB mainboard provides only a switched voltage and the board is directly powered by the PSU.

I'm also wondering whether a higher-power PSU would help (after reading the above) but that seems like a fix to a problem other than the immediate one.

Just thought I'd mention it in case someone else is having a similar set of symptoms.

[scantrontb]

My heated bed is installed on top of 8mm of Neoprene foam, and the foam is on top of the birch table. I have some metallic tape on the top surface of the neoprene to distribute heat and reduce oxidation of the foam. On the top side of the bed I have a 1/8" sheet of glass.
>>much snippage<<
What I found was that the heated bed connector at the Printrbot main board was charred and blackened around each of the 4 pins . I disconnected the bed and gave a warmup command, and did not find any voltage at the pins . I don't usually trust a single measurement, though, and after four or five power cycles and re-measurements I started reading 12V at the pins (so either I did a crappy job measuring at first or the pins came back to life).

I'm considering installing an external relay in the heated bed circuit so that the PB mainboard provides only a switched voltage and the board is directly powered by the PSU.

have to ask the obvious... is the ckt board face up or face down? if the side that has the electric connections and the traces on it is closest to the METAL tape, or face DOWN, you probably shorted out some of the traces with that tape. the various motions of the bot versus inertia caused the glass and ckt board to move around a tiny bit rubbing the metal tape and the ckt board traces together wearing out the varnish? coating on the ckt board, causing the actual current-carrying traces to touch the metal tape directly in multiple places causing short circuits, then the resistance would be lower, and Ohms law says if you keep the voltage the same, but lower the resistance, then the current goes up proportionally... thus you get burned contacts on the connectors where it can't handle the increased current...
However... if you have the traces on the board closest to the GLASS, meaning face UP, then at least I'm kinda stumped as to why you have burned connector pins, as there is no logical reason why you'd have an increased current thru the connector pins...

taking multiple measurements is always good, you probably just had a bad connection the first time due to charring getting in the way and the later attempts had gotten rid of it for where you put the meter probe.

as for the power relay idea, that has been mentioned here before, but i don't think anyone has actually done it yet and reported their findings here...

[ivymike1031]

I did/do have the board upside down in that sense - but there is ample clearance around the connections themselves and no evidence of rubbing when I peeked underneath. As the neoprene, metal tape, circuit board, and glass are all very firmly clamped together I doubt there is much relative motion... but it also wouldn't likely hurt to flip the board over so I'll go ahead and do that in a bit.

I measured 1.4 Ohms across the terminals of the heated bed. With 12V applied directly, it pulled around 8.5A, which lines up with the resistance figure and also should give around 103W heat output (I don't know what the bed spec is, but that power sounds about right to me just eyeballing it).

I installed an automotive-style 40A fog light relay ($5.99 @ Autozone) in the circuit, such that 12V from the power supply connects directly to the heated bed when the PB main board sends current out the (formerly) bed connections. It makes an obvious "click" when I hit the "set" button in Pronterface, which will be a nice audible confirmation that things are moving in the right direction. I'm also going to solder in a little LED strip @ the power connections on the bed, so that I get a light that turns on when the bed is powered. I'll do that after I flip the bed (because the contacts will be more accessible).

[ivymike1031]

update: LED strip installed & bed flipped. hot bed works, and it cycles off/on every few minutes (set at 80C) with a click and a blink of the lights (on time ~3min, off time ~20 sec, so I'm probably near max temp).

one possibility for why the terminals may have gotten hot - not that it necessarily happened this way
- if the contact at one of the terminals was intermittent, the other terminal at the same potential would have to carry all the current, which would cause it to heat up more than if it was carrying only half. Any local contaminants which might impede the connection would also tend to drive an imbalance between the pins. If I had a weak/intermittent connection at one pin, it might have forced another pin to carry the full 8A (and they look a bit small for that load), and after a bit of charring/melting it would have become a regular occurrence at each pin up until they stopped working for me. An intermittent contact at the connector would be consistent with the observation (noted earlier) that the bed temp would sometimes drop off when printing started (and vibration increased).

My wiring was no longer neat & tidy after the recent batch of repairs and as I moved it around a bit just a second ago, I noticed that the relay for the bed would click off and on rapidly, so I do seem to have an intermittent connection at the printrbot main board (the other connections are all soldered). I'll have to see whether this is an ongoing thing while printing, or if it's limited to larger translations of the wire. It at least lends a little support to the idea above (although my connector was not black & charred from day 1).

[ivymike1031]

here's a wiring diagram showing how I added the relay ... it was a fog light relay from the auto parts store (12V 40A SPST/NO).

[lwalkera]

You'll also want to add a reverse protection diode across the coil too. Otherwise the FET on the board will fail after sometime if the coil is too large.

[holmes4]

Is that needed with a solid-state relay? I just put one of those on.

[lwalkera]

It's probably not needed for an SSR given that it doesn't have a coil. Depends on the SSR though.

[ivymike1031]

good thought... I've ordered one (50 actually, but one to use) and will install it when it comes. That was $2.25 including shipping (total for all 50).
I'll update the drawing accordingly when I get a chance, for others who come along later.

[chad]

I ordered a better power supply and am planning on wiring it up with a relay as mentioned on several threads. I plan to use a mechanical relay since I can get one quickly from Radio Shack or an auto parts store. I wanted to find out a little bit more about adding a diode. It sounds like it should be wired into the circuit with the relay coil. My question is what is the purpose of doing so and since I know very little about diodes is there sizing or other considerations when choosing one? Thanks to all who have helped me get this far. I appreciate the help.

[ivymike1031]

See page to this document for information about what happens when the coil switches off. Ignore the fact that the document talks about SSR's as the effect is the same on the Fet.

http://www.crydom.com/en/Tech/Newslette ... 20SSRs.pdf

[jr57k]

wish I had seen this.... before today.... Been printing for 10 months.

viewtopic.php?t=2432&f=94

had this happen. I've got a handful of SS relays, I'm going to slap one of them in. Assuming the FET is still good.

[holmes4]

I ended up pulling the SSR I was using as it overheated and dropped power. I am using a coil relay now with a diode across it for protection.

[Nonomojo]

Can you share the model diode you guys are using? When using a relay, it just goes across the pos/neg side of the heated bed, right?

[ivymike1031]

I still havent receiver my diodes yet.

The diode goes on the coil/pb circuit of the relay, not the bed circuit. Put it across the coil terminals of the relay, so that the voltage always has the right sign, short-circuiting any induced current back to the coil.

[holmes4]

1N4001, Radio Shack part 276-1101. It goes across the relay coil, NOT the hotbed! Polarity mark band on the positive side so that it normally does not conduct.

[RetireeJay]

Excellent discussion about the relay and protection diode.

But I'll add one more data point concerning loss of power to the heat bed and concentration of current into one of the connector pins.

When I was doing some servicing on my PB Plus, I discovered that one of the two black wires had broken away from the board.

I happen to have accumulated some knowledge about this kind of thing in my job. When you have a stranded wire that is soldered to a junction point, the solder wicks out into the wire for a short distance and stiffens the wire without really reinforcing it. Copper wire by itself has poor fatigue resistance, but stranded copper wire with solder in between the strands is much worse. If the stranded wire is not supported in such a way that this wire-to-solder zone is held absolutely stationary in a fixed configuration, then repeated bending will definitely break the wire. This can happen with as few as a dozen bending actions, where the bend is only a few degrees. If only one black wire and/or one red wire breaks, you have a circuit remaining that can still heat the board, but with less than full efficiency. And instead of sharing the current between two "black" pins or two "red" pins at the Pronterboard, the current is forced to flow through only one of these pins.

Ideally, the design should have a mechanical support for the wires leaving the board which is independent of the solder joint. In other words, separate the functions: (1) mechanical anchoring with some kind of clamp, and (2) electrical connection with solder -- in separate location (at least a few mm apart). In a completed Printrbot, this is accomplished by using a zip-tie on the wires to connect them mechanically to the wooden bed. BUT, before and during assembly or during any servicing (any time the heater PC board is not anchored to the wooden bed), it's extremely difficult to avoid bending the wires where they are soldered to the board. Breaks are likely.

[Nonomojo]

Great, thanks for the quick response. The diode placement makes sense. I already bought the relay, but I'm going to run up and get the diode today before I install it. Trying to prevent a connector burn-out and possibly killing a FET on the Printrboard. Again, great diagrams and explanation of the components.

[ivymike1031]

updated diagram to show diode

[izaklast]

Thanks for posting the diagram, I'm going to build one this weekend. Also thanks to the member who recommended the diode. Where do I add the diode? Near the relay terminals or further down the line, exposing some wire and soldering there?

[RetireeJay]

The best place for the diode is right at the relay terminals.

[izaklast]

Thank you Jay. The diode is directional correct? Is it similar to an LED where the long leg is the cathode? If that's the case, would the long leg go to red?

Also, since I have your attention, you mentioned (on a different thread that I can no longer locate) that a multimeter set to ohms will show a connection between the ground pin on the main power jack and the positive on the heated bed output showing that the MOSFET is still good. I did that last night and I do still have a connection. Last weekend I soldered new wires and connected the to the bed only to get no power coming through. I thought the circuit was blown when the molex plug nearly melted. Last night I plugged the board in and experimented with turning the heated bed on and off while the multimeter was in place. I get a steady .4 to .48 volts at all times. Does the circuit require a loop to operate correctly? Shouldn't I be getting either 0 volts or something like 12?

Thank you for the post about mechanically anchoring the new wires to the chassis somehow, this is valuable advice and I'll be sure to heed it.

[RetireeJay]

The diode is directional correct? Is it similar to an LED where the long leg is the cathode?

Yes, absolutely it's directional, and hooking it up wrong will burn out the diode (or something!)
I think that a 1N4001 diode has equal-length leads. But if you look at the schematic symbol for a diode, you see a triangle (or arrow) pointing at a line perpendicular to the direction of the arrow. On a physical diode, that line corresponds to a ring printed around one end of the cylinder. Following the convention that electricity flows from positive to negative (forget about those pesky negatively charged electrons), that means that the diode will freely conduct electricity if the end with the ring is more negative than the end without the ring. It will block electricity if the end with the ring is more positive than the other end. So when you connect it to the coil, you want the ring end of the diode connected to the +12V supply and the relay; the other end gets connected to the other side of the relay coil, and it gets switched to ground when the board is calling for power. The diode only springs into action when the current gets shut off: it provides a local path for current to keep flowing temporarily while preventing a voltage spike. (The current very quickly decays to zero.)

Also, since I have your attention, you mentioned (on a different thread that I can no longer locate) that a multimeter set to ohms will show a connection between the ground pin on the main power jack and the positive on the heated bed output showing that the MOSFET is still good.

No, that's not quite right. When the board is powered up and calling for heat, then the MOSFET should be acting as a switch connecting two of the pins on the heater connector to ground. If the board is not powered up, or the board is not calling for heat, the MOSFET should be acting like an open circuit.

Last night I plugged the board in and experimented with turning the heated bed on and off while the multimeter was in place. I get a steady .4 to .48 volts at all times. Does the circuit require a loop to operate correctly? Shouldn't I be getting either 0 volts or something like 12?

Assuming that you have the heater connected, then the voltage at the switched (MOSFET) side of the heater circuit should go to 12V when the board is NOT calling for heat. The other side of the heater is already at 12V, so that means zero net voltage across the heater. When the board IS calling for heat, then the voltage at the switched terminal should go very low (while the voltage at the other terminal going to the heater remains at 12V).

[izaklast]

I completely understand, thanks. My board was only showing a steady .4 volt reading unless the board was off. I'll try again tonight and see what happens. Do any other components (steppers, hot end thermister, etc) have to be attached to make the heater function?

[RetireeJay]

Yes, you need thermistors connected for both hot end and heatbed if you want to heat the heatbed. The hot end doesn't have to be heated up, but the temperature reading has to be within a sane range. Open circuit makes it look like 0C (or below). If you don't have the thermistors handy (working on the board apart from the printer) then substitute 100k ohm resistors, which simulate room temperature.

[izaklast]

Ok, I didn't have the hot end thermistor plugged in so that may explain my results. I'll try again in the morning and report back.

Jay if nobody every told you you're the man then let me be the first; thanks for the help, you're the man!

[izaklast]

Board is putting out 12V just fine. Plugged the hot end thermistor and the heater circuit works as Jay described.

New, related question; can I solder the diode in at any point in the board feed such as where the harness wires will meet the new wires I'm going to solder into the board?
Here's the harness and relay with diode.

[RetireeJay]

The closer the diode is to the relay coil, the better job it will do "snubbing" the voltage spike and preventing glitches from getting into other circuitry.

[izaklast]

What size fuse would be appropriate? My gut is telling me 10A but I want to be sure.

[RetireeJay]

That should be fine, unless your heater will pull over 100 watts. Power = Volt X Amps. Presumably you are using 12V, so you can calculate your amps if you know the power. Or, if you can measure the resistance of the bed accurately, you can use Amps = Volts / Resistance.
Generally, you use fuses that are 20% higher than the expected load, so you don't get "nuisance" blowing, but they protect you if there's a serious fault.

[izaklast]

I measure 1.8 ohms on the bed leads. That would give me 6.66 Amps (12/1.8) so 10 should do nicely. Thanks for the help Jay, as usual

[izaklast]

This is my relay harness setup with the diode exiting the board control lines. I'm doing something wrong: the board shuts down when I turn the heated bed off, relay is not tripping. Is my diode not preventing the back currently correctly? .

20150130_174722.jpg

[RetireeJay]

the board shuts down when I turn the heated bed off, relay is not tripping. Is my diode not preventing the back currently correctly? .

Do you mean, "the board shuts down when I turn the heated bed ON"? That would make sense, and yes, it would mean that the diode is backwards. But you can correct that easily by just reversing how the control wires going to the relay are plugged in to the board. That will reverse the polarity of the signal across the diode.

[izaklast]

Thanks for the swift answer Jay. When I power the bed ON, nothing happens. When I cut power to the bed, the board shuts itself down. I hardwired new wires into the board so I may have to remove the diode and install a new one. The ring end is closest to the positive wire, the configuration I see on the diagram and also what I understand from your thorough description of its use.

Does anything about the way I wired the diode in strike you as being incorrect?

[RetireeJay]

Do you have a voltmeter? The diode leads are exposed, and you can measure there. What we want is for the end with the ring to be more positive, like 12V positive relative to the other end (when calling for heat). If the voltage on the end with the ring is negative with respect to the end without the ring, it should never go above 0.6 volts or so.

I don't understand how the board would shut down when you are NOT calling for heat. In fact, how could you ever get the board to fire up in the first place? When the Printrboard is first powered on, it is NOT calling for any heat (or motion). You have to issue a command to make it start heating. (I hope you are NOT plugging the relay in to the board after powering it up. Never do that; have it connected before you turn on the power.)

Perhaps there's a deeper issue with the wiring. Perhaps you have mistaken the wiring diagram for the relay, and it's not really hooked up the way you think it is. Maybe the relay contacts are shorting out the supply, instead of sending power to the heated bed.
Note that if you are using the diagram from 03 Feb 2013 in this thread, do not trust the wire colors! Normally we use red for positive and black for ground, but that's not how the folks at Printrbot did it. You need to look at how the wires connect to the board.

Looking at the board itself, the two pins (or the screw terminal on a Rev F board) closest to the edge of the board are switched to ground through the MOSFET. The two pins (or the screw terminal) closer to the power connector on the board are CONSTANTLY connected to +12V. That's the wire that should be connected to the banded end of the diode.

[izaklast]

Voltmeter: -check-. I have a little digital pocket-sized one and a 9V hobbyist level box style one as well.

I am not plugging the relay in while the board is on, I power it up with relay harness plugged in. At this point the power end (Power supply feed and bed feed) of the replay is not wired up, I'm just trying to get the relay to fire so I know I have it set up right.

Based on what you described, I have the diode backwards. I'm using a "REV B" board so I have the two pins furthest from the main power jack wired to the left side of the coil (using the diagram supplied above). I have the two nearest the edge of the board Wired to the other side of the coil.

Thank goodness I only repeated the shut down trigger twice, hopefully I didn't fry the board.

[RetireeJay]

Here, let me give you a picture that may be a little more clear.

PrintrboardDiodeWiring.png

[izaklast]

Very helpful Jay, thank you. I'll remove the diode and wire a new one in there.

[spider2k]

question. this may have been answered but i didn't see it.

the temp of the heat bed is controlled by varying the output on the board in relation to the temp read by the thermistor correct?

if that is the case, how does the board control the temp when the variable output of the board is just used to trigger a relay feeding straight voltage to the bed heater.

or is it not a variable output from the board and it simply feeds full voltage constantly and shuts off at a point then goes back on when it drops x degrees.

thanks in advance and i hope i made my question clear.

ryan

[RetireeJay]

Yes, it controls the temperature by on/off cycling. Basically like the thermostat in your house: your furnace or air-conditioner runs a "full power" whenever it's on, and the thermostat turns it on and off in cyclic fashion to control the temperature within a reasonably narrow band.

In the field of temperature control, there's also a method called "PID." When implemented on a Printrboard, it also uses on/off cycling... but it uses a much faster pulse rate, so the control is based on the on/off proportion of time. That's an option for the heated bed, generally not suitable for relays (at least not mechanical relays). But PID is always used on the Hot End. (PID = Proportional, Integral, Derivative - referring to the type of math used to control the heat output in order to achieve the best possible control over the achieved temperature.)

[spider2k]

ok so basically the added relay will be cycling constantly and is just to take some of the load off of the board?

just asking as i'm building my simple metal and am going to add one in now instead of later.

[RetireeJay]

ok so basically the added relay will be cycling constantly and is just to take some of the load off of the board?

just asking as i'm building my simple metal and am going to add one in now instead of later.

That's exactly right. The small current required to control the relay passes through the board, and the big current going to the heater passes through the relay.

Just a note, however: the new Rev F boards have screw terminals for the heated bed. These terminals have much higher current ratings than the previous header used on Rev D boards. In effect, the heated bed circuit on the Printrboard becomes a good quality Solid State Relay all on its own; no external relay is required.

[spider2k]

it is a new rev.F. i might add the relay anyways just to keep some heat off of the board.

[DukeEuphoria]

Hi all, recently had this happen to my (very aged) PrintrBot LC+ (rev1)
I've installed a solid State relay but am now finding that the bed is even slower to heat than before, and cannot makie it much above 55 degrees C.

The SSR on the other hand becomes toasty warm...

Anyone using an SSR with any degree of sucess ?
If so, any clues ?

[RetireeJay]

What's the current rating on your SSR? It seems that you are losing some of the voltage that should go to the heatbed inside the SSR because it's under-rated. Hopefully you use an SSR with a rating of at least 20 amps.

[Mooselake]

This looks like the right size, for a little under 7 quid.

Kirk

[plexus]

After replacing the connector on my PBrd rev D with a fatter connector, I have had no issues or signs of heat damage. I am thinking the FET on the board is spec'ed ok for this application, its just the connector that isn't. so adding the whole relay thing seems like over-kill. it seems that just replacing the connector is sufficient. Just saying. I am running a 200W heater and its been working for over a year. Just recently actually the heater died. seems 1/2 of the resistive element opened. however I had a loose main's plug issue that flickered the power on and off a bit. next thing my bed header was 1/2 dead so I think the power fluctuation may have caused a surge that damaged the heater.

[craigj]

Looking for help with this type of issue with a Printrbot Metal using a Rev F board and the wide bed upgrade.
I have been running this bot for about 2 years now. several times a year I have to rewire the bed heater because of either wire fatigue failure or a scorched wire.

During my last repair I rewired with 16 gauge stranded wire and installed a 4 pin molex connector (heat pos, heat neg, thermistor pos, thermistor neg) in the middle of the harness and firmly fixed the bed side section of the harness to the bed to prevent any movement in the section thus hopefully limiting the possibility of future fatigue failures to the section of harness that could be replaced without removing the bed.

This worked well for several weeks of constant use until the heatbed pos pin on the molex connector scorched and failed. This occurred at some point during a 6 hour print where I had the bed set to 70 C.

So far these failures have only been a nuisance - maybe the print fails when the bed stops heating and, of course, I need to take the time to rewire the bot which affects my print schedule. But I am worried that at some point this will cause a catastrophic failure that destroys the printrboard or starts a fire.

I welcome any suggestions on how to permanently resolve this problem.

[RetireeJay]

I recommend using silicone-insulated, 14 gauge wire with very high strand count for the power leads. A cheap and available substitute (but not silicone insulated) is sold in the big box stores as lamp cord - but be sure to get 14 gauge. If you want to put a connector in-line, use XT-60 connectors like in this post: viewtopic.php?f=15&t=2789&p=31652#p47448

For non-power wiring like thermistor, you don't need heavy gauge wire; 24, 22, or 20 gauge will be fine. Again, a very high strand count will help avoid fatigue failure.

And of course, allow plenty of extra length of wire so that you can "dress" the wires to have no sharp bends when the 'bot is in operation. All flexing should be gentle, over a very large radius.

[craigj]

Thanks Jay. Does this look like the right stuff?
https://www.amazon.com/BNTECHGO-Silicon ... icon&psc=1

[RetireeJay]

Thanks Jay. Does this look like the right stuff?
https://www.amazon.com/BNTECHGO-Silicon ... icon&psc=1

Yes, that looks great! A bit pricey, but it will definitely solve your problems with fatigue failure.

Remember that when you solder stranded wires, solder wicks into the strands for a short distance, so make sure that you are not imposing ANY bending stress on the immediate vicinity of solder joints. The wire there should be immobilized.

Good luck!
-RJ

[novice]

Dang I'll say that looks a lot pricey. Our local Ace Hdwe sells bulk Iron/Appliance 3 wire cord that can be stripped for the individual silicone insulated wires.