Biodiesel and SVO Forums Home
Burned out my SSR

This topic can be found at:
http://biodiesel.infopop.cc/eve/forums/a/tpc/f/4441089311/m/7167008573

August 26, 2013, 02:55 PM
Ryan P.
Burned out my SSR
So I built my temp control circuit on my methanol still using a Fotek 25A SSR and a Digital PID Temperature controller as recommended by a few folks on here. Pretty sure it was this exact kit off Ebay.

I did the math, the system should have drawn about 16amps, so a decent factor of safety (36%) under the 25amp rated SSR. But the SSR blew Saturday; shot a nice little 4" flame out of it and smoked up my shop. 220V and a measured resistance across the element of 14ohms; that comes out to 16amps and a 3500 watt element.

After it blew, I rewired the element directly to finish the methanol recovery batch I had started and just watched it like a hawk and maintained temp. manually; the element continued functioning perfectly.

Other than a part defect, I can't see what happened. I ordered 40amp SSR as a replacement.

I DO see that I can get a heat sink for the SSR's; at what point does that become necessary? My SSR was in an electrical box; so no air flow over it; I never realized that might be necessary. This isn't the first batch I have done this way; I did at least 2 batches last summer with no problems before I had a pump go out on the system that I just finally fixed this last weekend.
August 26, 2013, 03:05 PM
Ryan P.
OK, commenting on myself for info: in trying to do some research I find this;

quote:
Please note that in order to achieve the rating of your SSR e.g. 25amp you must use a suitable heatsink. If you do not then the rating is dramatically reduced. For example a 40amp without heatsink is only rated to around 10amps.


ALRIGHTY THEN!!
August 26, 2013, 03:44 PM
Paulus
As there is a (small) resistance across the SSR itself, as you pull X amps through it a voltage drop occurs.

If the resistance were 1 Ohm, as V = I x R, Vdrop = 16*1 = 16V. V x I =Watts = 256W dissipation by the SSR itself.

Even if the resistance were 0.1 Ohm, dissipation would still be 25W.

Heat sink essential in high power semiconductor devices.
August 26, 2013, 04:01 PM
Ryan P.
I think its just funny that in all the discussion I read on setting this up this way, a heat sink never came up or showed up in any of the pictures or wiring diagrams!
August 26, 2013, 04:22 PM
Paulus
Some good info here. It says an SSR will need to dissipate 1W per amp of current so yours was producing about 16 Watts of heat. With nowhere for that heat to go when it hit the boundaries of the SSR package, thus temperature rise. There will be a flat surface for mounting, and it is a heat transfer surface requiring thermal grease. Thats the first clue.

http://www.automation.com/libr...g-solid-state-relays
August 26, 2013, 10:54 PM
Jon Heron
Yep, you answered your question, heat sink needed. I have always mounted mine directly to the metal backplate in my enclosure, a chunk of plate aluminum will also work...
These things do just fail out of the blue too though, usually in the closed position but sometimes violently too...
Next time just get a contactor, they are more reliable and not as prone to failing in the closed position nor violently...
Cheers,
Jon


___________________________

Simple schematic for a pump and heater control with a high limit
Sensor for the biodiesel/glycerin layer
August 27, 2013, 04:01 AM
Johann Cape Town
I seem to recall that a PID / SSR could be set up to drive a normal heater element in a manner that would replicate a low watt density element. Can this be done?
August 27, 2013, 11:43 AM
Ryan P.
quote:
I seem to recall that a PID / SSR could be set up to drive a normal heater element in a manner that would replicate a low watt density element. Can this be done?


I wouldn't be surprised. That PID can do all kinds of things I am not asking it to do. The coolest part, that I don't even have to try to program in, is that is does some basic self-learning of the system. It figured out the delay between turning on the element and seeing the temp. change, on a basic level, and learns how fast the temp drops when the element is off. Then it starts turning the element on in anticipation of each event to maintain the set temp more tightly. Pretty slick.

What I am learning to love is that the thermocouple measures the head temp so precisely that I can watch and see when I think it is switching between methanol entering the condenser and the first of the methanol-water mix. On my still, and where I have the thermocouple probe mounted, when I am getting pure methanol, the head temp hangs at 33-33.3 C for an hour or more. As soon as I start to see it climb to 33.5 and then 34C and definitely by 35C, I can check the pot temp with my IR thermometer and see that it just broke over 116C or so and the reflux column entrance temp is creeping over 100C; I am getting water vapor in the column, so time to shut it down.

I could probably push a little further until the reflux column EXIT was getting closer to 100C, and get a bit more methanol out, but it risks putting water into the condenser and then the methanol drum, so I call it "good enough" and stop.
September 02, 2013, 07:40 PM
Ryan P.
It has a good heat sink now!



And a fan to pull air through the box.


November 10, 2013, 03:41 PM
dkenny
yes to the heat sink..always..
and a fan is a good idea too..

if you're looking heatsinks with fans..think computers..all the computers of today
have heatsinks and fans..'of today' think since about 1995..anything after than..the newer the
larger the heatsink/fan..

yes..mine are heat sinked and fan cooled..

-dkenny


'84 bluebird school bus, DD8.2L turbo( 4/2011, the bus tranny has died..Frown 8.23.11 bus driven to scrap yard Frown )
2006 Jeep Liberty CRD Smile - the wife's
99 dodge 2500 5.9l 24v..-mine Smile
everything run B100 when its warm enough Smile
November 10, 2013, 07:21 PM
Ryan P.
Yep, that heat sink and fan both came off of junked computers. That ~3"x5" heat sink was on a computer processor that was about 1.5"x1.5"!