Boil/HERMS Element Switching Frequency

Homebrew Talk - Beer, Wine, Mead, & Cider Brewing Discussion Forum

Help Support Homebrew Talk - Beer, Wine, Mead, & Cider Brewing Discussion Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

KuntzBrewing

Well-Known Member
Joined
Aug 22, 2011
Messages
630
Reaction score
13
Location
Kokomo
For those of you who code your own herms/rims systems what frequency do you switch the element at? I have my system at .95 Hz using the 16 bit pwm with a 16MHz crystal and prescale of 256. My concern is how the instantaneous power surge of a random turn on SSR could affect the system. Worst case scenario you could switch on at 340 volts (peak value of 240 RMS) so switching slower would extend the life of the hardware, but how slow is too slow? Switching too slow and you can have the case of fluctuating boil intensity.

Thoughts? Suggestions?

Boiler Up!!
 
Most SSRs only switch on/off at the zero crossing so there will be no instantaneous power draw by turning it on. Secondly most SSRs have a response time of 10ms or more. Switching the SSR on/off does not affect the life of the hardware like it does for a mechanical relay. That being said I use a period of 1 second.
 
Craine is right regular SSRs have a zero cross detection , thats why they have 10ms response time for 50Hz or 8.3ms for 60Hz. Because in worst case .scenario zero cross detector may wait almost full half-period.
But even switching in a peak of amplitude doesn't create any problem for resistive load. SSVR works exactly this way, they "cut" part of every half-period(ie 120 times per second) and nothing bad is happening. The only problem - some extra RF interference.
 
First off, I like the 1 HZ PWM frequency. It keeps heater variation within a 1 second interval - no way you need faster than that. It also gives you 120 half cycles of control for varying power. 120 steps is 0.8 % steps - why would you need finer than that???

IMO, 16 bits is WAY overkill for resolution. In theory that's 0.0015% steps! More importantly, that resolution only works if you use a random trigger triac/relay (i.e. trigger at any phase). All those steps won't work for any zero crossing triac/relay you drive. Those 65 thousand steps will reduce to the 120 half cycles that exist in 1 second - 120 is all you get (only 100 if you're french ;-) ).

If you use a zero-crossing triac/relay, then you can't get anything more than 120 steps at 1 HZ PWM (since they must be ON in an integer number of half cycles - so no such thing as 1.5 half-cycles).

The positive? Zero crossing is lower EMI, AND lower stress on the triac (and surge to the heater). Did I mention, half-cycle control is more resolution that you really need?

Go with a zero-crossing triac/relay. Stick with 120 steps at a PWM rate of 1 Hz. If you really must have finer control, use 240 steps at 2 Hz PWM (0.4% steps).

-G
 
FWIW, someday, someone is going to force me to hook up a thermocouple to my heater, drive it with 1 to 4 Hz PWM and plot the temperature response to prove my point. And if by some slim chance I'm wrong, I will then have to put it in 40 pounds of water and repeat the test. And if I'm wrong after that, I'll have to move the thermocouple a couple of inches away in the water (now it will resemble a real homebrew setup).
I'm betting a keg of beer I'm not wrong. :yes:
🤣🤣🤣
 
Elements have mass and cannot heat and cool instantly... there is a point of diminishing returns by switching faster and faster. I think 1 second cycle time should be perfectly adequate for the elements we use in our applications. In my testing, no noticeable pulsation is seen at 30% duty with a 2 second cycle time.
 
Back
Top