RE: SSR’s - I have been contemplating switch a large AC load with an SSR. An industrial electonics tech warned me that they will get very warm and need a proper sized heat sink - of which particle board is not. If you’re going to try the SSR route again - I suggest mounting it to an aluminum plate or heat sink with fan.
(Note that the back of these types of SSR’s are aluminum so that they can disipate heat)
@Alex_Wiebe If you look in the picture its mounted to a large aluminum heat sink. The one that the company sells for this relay. I also had a fan blowing across it as well with thermal paste between the heatsink and relay.
he did mount a heatsink, but i think it is a poorly made CHINA copy, it seems melted overheated from inside. that would suggest, it doesnt matter if its operated in -40degrees, it will still fail 1 day
I don’t think speaker wire is going to be any trouble in this application. Wire current ratings are extremely conservative the way we use them, because they are rated with multiple conductors in an essentially insulated conduit. Free air ratings are much higher.
I can’t find the exact datasheet for that SSR, only up to the 10A version. (http://fotek.com.hk/page1e.htm). If the voltage drop is the same for the version you were running, and your bed was pulling 15A continuous, you’d have 24W loss. That heatsink looks like more than enough to handle that.
How is this driven from the electronics? The datasheet looks like it’s rated for 7.5mA current to trigger. If the control circuits inside the SSR are not smart enough, and the drive strength to the control “coil” of the SSR was below 7.5mA, the optically coupled transistor may be underdriven and overheat, as described in 4. Arduino pins can apparently source up to 40mA of current though, so this probably shouldn’t be a problem.
@Eric_Moy is describing what happens when you drive a FET’s gate that’s rated at 10V directly with the 5V output of the Arduino. The FET channel will not act as a low resistance, but rather saturate at some current lower than your load demands, will get a larger voltage drop across it than it should, and therefore dissipate more power. There are FETs that are capable of being driven with 5V, but some FETs commonly used on RAMPS boards are not that type.
The junction transistor used in SSRs don’t have a lot of overvoltage spike (caused by the transistor turning off) handling capability. According to the datasheet, the manufacturer added an RC snubber to try to damp those spikes and protect the device. FETs are more reliable in this regard, having an avalanche energy rating, where the device will clamp the excess switching energy sort of like a zener diode or varistor would.
If you thought this was a lot of text, I can go into more detail at request (although I suspect I’ve overdone it).
oh btw, you can almost never tell what is a genuine SSR from a CHINA pc unless you pry it open. many parts retailers are fooled themselves (or maybe they fake it haha). my take is try look for something on farnell, mouser, etc. i think farnell has never failed me on genuine parts.
@Joe_Spanier My bad - I missed the last sentence, and my brain was seeing the SSR laying on the wood with some funky strips of tinfoil lying in front of it. Definitely a faulty unit (my brain that is…)
im a fet noob myself … will this work? http://www.avrfreaks.net/modules/PNphpBB2/files/bootstrapped_mosfet_108.png
this looks good but gbp$ 42 ouch … http://uk.farnell.com/crydom/h12wd4850pg/ssr-50a/dp/1200256
@Tony_Olivo Thanks for all the info. Ill answer some of the questions in there.
There was about a 3v drop when I checked the output of the relay to neutral of the bed. The power supply was putting out about 12.4v and the neutral was reading ~9.4. That was a much greater voltage drop then I expected.
I was driving the input side from the D8 output of the RAMPS board. I am still not 100% that this was the correct way but without remapping a bunch of pins in marlin I wasnt sure of a better way.
You need to use an SSR heat sink. The one in the picture is much. Too small. The SSR needs to be mounted vertically with most heat sinks. Read the directions! You will be amazed how big a heat sink needs to be to get the full rating off an SSR. With the heat sink shown I would not go past 10amps. Can you derate your heaters? 30 amps with an SSR will cost. If you can not derate your heaters how fast does it heat up? If it takes several minutes, you could look at a mechanical relay and replace it every couple of years. Make sure to tune the PID so the shortest on time is 1 or 2 seconds.
I’m using that exact same SSR with the same 12"x12" heat bed and 30A 12v supply. I’m driving it with PWM. I’m using much larger conductors and that added resistance might be problem. I started with the duty cycle at 128 and pushed it up from there to 255. No issues so far. Also no factory heat sink just bolted to one of the aluminum extrusions in my printers frame.
Double check the output voltage of your supply, if it’s low it will push the amperage up.
I suppose its possible I just got a bad one. The quality control on these are awful and the reviews of all these relays range from my experience to yours. So who knows. I ordered a 40 amp and got an RMA for this one. Ill mess with those until I can figure out FETs
@Joe_Spanier , bad unit seems possible. Is it on the same supply as everything else? Also from my reading up, SSR’s for DC are based on one or multiple MOSFET’s while an AC SSR uses a TRIAC(usually). So they should be able to switch super fast without issue.
SSR, or SCR?
Ssr
@Tim_Rastall is this the diagram you followed?
https://plus.google.com/photos/+ThomasSanladerer/albums/5953448723016351905/5953448728778867138?authkey=CI2ci8iI_YGWOQ&pid=5953448728778867138&oid=101189457349966331973
@Joe_Spanier nope, the diagram was from @Stefano_Pavanello
Hopefully he can repost it. Stefano, I’m talking about the heated bed FET circuit diagram you posted a link to a while back
SSR-55DD
@Tim_Rastall yes, I remember. I will search for the diagram and I will reshare it 
thanks a bunch!
