randomia: a trick of solid state relays

 Solid state relays are often used when an application requires very frequent on-off cycling of a high power load, such as when a heating element is controlled using phase angle control* to achieve a high degree of control over the actual physical temperature of the heater, when the voltage cannot be varied - it can only be turned on or off.

 Solid state relays generally require very large heat sinks, this is often a disappointment when a new person sees how much power a relatively small solid state device can handle... which seems quite amazing until one sees the enormous heat sink that little device must be very carefully attached to, in order to achieve its listed performance.

 Now I have a project I've been working on half-heartedly for a long time, which doesn't quite require a solid state switch, but it would be smaller than a "contactor" (heavy duty mechanical relay) with the same ratings, except for that pesky heat sink required, and my artsy-fartsy little box hasn't got room for much of a heat sink.

 I figured you don't have to get rid of heat that you don't create in the first place, so the first thing I did was look for which relay lines from which manufacturers had the lowest forward drop (on resistance), so as to create the least amount of heat for a given current being switched.

 And that's when I learned something useful.  For while there is some variation between manufacturers, it's not enough to bother with.  What is worth bothering about are the only differences in forward drop between units with different current ratings.  This may have changed recently, but when I looked all of this up a few years ago, most of the big manufacturers seperated their SSRs into just a few - three to five - groups of current ratings; 25A, 60A, 100A, for example.  The only technical difference in the detailed specifications between two models with different current ratings was the forward drop; higher rated models had lower drop and dissipated less power.

 SOO, the outcome of this is that you can use a 100 amp SSR when you only need a 25A unit, because you are really only switching (say) 15A... and you will be able to use a much smaller heat sink.  The catch is price: higher current units cost a lot more because they can't make as many of them.  The current ratings are not arrived at by careful design, they are achieved by binning: they make a bunch of them, then test every single one to see how good it is.  Because it's hard to control the doping of junctions with the kind of precision necessary, and because the bell curve of what's produced roughly matched the demands of the industry... they don't try.  This is true for a LOT of semiconductors.

 This was a bigger deal when I started the project... before LED bulbs took off.  Yeah, it's been that long.  Now, the device doesn't need to switch as much current, because the load could all be LEDs which draw next to nothing.

_____________________________________________________________
* this is just a fancy term for the way wall switch dimmers in your home
work - these big SSR circuits just involve bigger parts to manage higher
currents and powers. Essentially, they work by turning the power off and
then back on again, very quickly, with each swing of the AC power cycle,
so, a hundred and twenty times a second.