Three PID - SSR powered elements on one 240V - 30A circuit..

I knew that would get some folks attention....



OK, what if we use logic based 'priority' to only send power to one element at a time... in my system, with a 50A breaker and cord, I did not see many times when 2 4500W elements were on at one time.. only when heating strike water or initial boil is any one element on 100%, my RIMS is maxed at 25%... I even think I could do consecutive/stacked brews and not have an issue other than the MLT taking a little longer.......

I believe that if you had HLT, MLT(RIMS), and BK, the priority would be MLT, then BK, then HLT...

MLT:
If we hooked the MLT PID signal direct to the MLT SSR.

BK:
Run MLT signal into a logic inverter and then that inverted signal and the BK PID signal int a logical AND, and send that to the BK SSR

HLT:
Invert the BK PID output and AND it to the inverted MLT PID signal, AND that signal with the HLT PID output to drive the MLT SSR..


PID outputs:
MLT 0 1 0 1 0 1 0 1
BK. 0 0 1 1 0 0 1 1
HLT 0 0 0 0 1 1 1 1

SSR inputs:
MLT 0 1 0 1 0 1 0 1
BK. 0 0 1 0 0 0 1 0
HLT 0 0 0 0 1 0 0 0

only one element powered at a time, and the PID functions will take care of *small* missed power pulses that the BK or HLT miss...


Obviously, circuit breaker to protect things from a failure that tried to power two elements and overload circuit....


I thought this would be really helpful for folks not wanting a huge investment in 50A power cords/plugs/receptacles/load centers... thoughts?


-mike
www.clearwaterbrewery.com

A

interesting concept... so would you use relay logic or is there some PID programing that i havent heard of?

Someone on here had a system designed to run all at once using the elements at 120V to reduce power consumption when they were all on... I'll see if I can find that thread.

OK, I see what you did there. The "PID outputs table" is just a counting pattern from binary 0 to 7 (to cover all the possible outputs from the PIDs).

Then theres some logic between the outputs and the SSRs to ensure that only one turns on, and the result is shown in the "SSR inputs" table.

Seems like a perfect application for a 3-to-8 decoder. See this:
http://www.fairchildsemi.com/ds/MM/MM74HC138.pdf

It'd be easy to do in code with a microcontroller, but the part above would be the minimal solution.

Interesting post Mike.

Yes,.. and the output table was an afterthought... I filled it out by counting binary... I figured it would help people see what I was doing...

I was actually going to use 7400 series TTL inverter and AND chips powered from my BCS-462 5V out, for real-time system, not relying on the BCS ladder logic if that was even possible...

I was hoping someone would point out a more elegant solution, I should have known it would be Andrew...

thanks for the 74138 idea, I will search through my 1980's chip stash ;-)

edit - darn... outputs are inverted and I would have to OR the 8 outputs down to 3... an hex inverter and quad AND chip it is..

clearwaterbrewer said:

edit - darn... outputs are inverted and I would have to OR the 8 outputs down to 3... an hex inverter and quad AND chip it is..

Click to expand...

The inputs to the SSRs are optocouplers. Just tie the '138 logic output to the (-) side of the SSR input, and tie the SSR (+) input to the same voltage as the '138 supply. That will get the invert for you.

Pick the logic family that will get you the current you need. 'ACT or something.

[edit] or look for an open-drain outputs to tie to the (-)
[eidt] to combite outputs, isolate each inverted output with shottky diode - high output does nothing, low sinks current through diode
[edit] or maybe there is a non-inverting logic part ('238?).

I found the other thread that was similar... but after reading these other comments it dosent sound like you are going the direction as this guy did. Both sound sweet and I like the direction you are going... I would love to stack brew days and this would allow it. Keep us posted on how it turns out!

PS: for reference this is how someone else did this https://www.homebrewtalk.com/f170/3-vessel-2-tier-1-pump-e-rims-30a-203926/

One thing to look out for here is that the AC SSRs switch on zero-crossing, which means there is up to half-cycle delay ( 1/ 120 second) between the input changing and the output turning off. Probably not a problem, but something to keep in mind.

Looks like it would work you just wouldnt get any pulses to the HLT SSR while you BK was at 100% getting up to a boil. Only matters if you are doing double batches.

I distilled the idea into having the RIMS be actively controlled and the remaining duty applied to the BK and HLT. The question I have is how would you measure for HLT temp? You wouldn't want it running fully manual like the BK.

cyberwollf said:

Looks like it would work you just wouldnt get any pulses to the HLT SSR while you BK was at 100% getting up to a boil. Only matters if you are doing double batches.

Click to expand...

You could just fill the HLT (lowering temp and having PID turn on) while Lautering to the BK, HLT should be able to be up to temp by the time you start boil

Bobby_M said:

I distilled the idea into having the RIMS be actively controlled and the remaining duty applied to the BK and HLT. The question I have is how would you measure for HLT temp? You wouldn't want it running fully manual like the BK.

Click to expand...

All have their own PID, temp sensor, and SSR. All we are doing is not allowing the signal to the SSR that powers the BK while the RIMS is being powered ( routinely .25 sec every few seconds for my system) and not allowing the MLT PID signal to get to the SSR if either the BK or the RIMS is powered..

I got the diagram down to using 2 inverters and 3 AND gates, or 5 NOR gates, or what I will most likely use, 8 of the more popular NAND gates. 2 cheap, common 4011 chips, a little project board, and a 8-position terminal block..should be able to test this weekend... I bet there is a way to use 555 timers or clocked registers or something similar to do some sort of thing that is closer to true TDM...

I think I have forgotten more than I ever new about this stuff but very cool idea. Keep us posted on how it works out.

clearwaterbrewer said:

All have their own PID, temp sensor, and SSR. All we are doing is not allowing the signal to the SSR that powers the BK while the RIMS is being powered ( routinely .25 sec every few seconds for my system) and not allowing the MLT PID signal to get to the SSR if either the BK or the RIMS is powered..

Click to expand...

Nevermind, nothing to see here. Duh.

I guess the purpose would be to squeeze out every bit of available power from the supply. I get that. Even with 50a to spare, I'm considering dual 4500w in the HLT and BK and using a manual selector for full power to either or half power to both. However, it's a cool idea to be able to let either device take duty cycle as it needs to as long as that time period is available.

It sounds like something BCS could easily be programmed to do also.

I know that this is an old thread but I am curious if you ever got this to work correctly. I don't think you could ever achieve this with 3 separate PID controllers and actually get what you want. I am also not familiar with the BCS so I don't know if you could implement it their either. You need to synchronize the outputs based on the next higher priorities current output. Let me try to explain what I mean.

Say your PID outputs are calling for MLT at 25% duty cycle, BK at 50% and HLT at 100%. If priority goes MLT, BK, HLT and everything is on a 1 second period then what you would want would be MLT to be on the first 1/4 sec, BK on the second and third 1/4 sec and finally HLT on the last. If you simply take the PID outputs assuming they are synchronized and tie them with discrete logic then you would get MLT on first 1/4 sec, BK on second 1/4, and HLT on the last 1/2 second. this is because each PID output would turn on at the same time. You need to shift the start of the period of the BK to the time when the MLT turns off and similar for the HLT. This would have to be done dynamically at each cycle based on current duty cycles.

Clearwater, does the BCS have enough flexibility and control to set something like this up? The only easy way I can think of doing this would be to program it in a microcontroller.