Constant drive on a heating element?

[BrunDog]

You cannot control an AC powered heater without on/off pulsing. The AC voltage and current go to zero 120 times a second (in the USA), which means at those points the element is off. Phase angle power control, is still on/off switching, but at a cycle time of 8.33 msec. So, everything discussed so far in this thread is still on/off pulsing. The only thing changing is the cycle time, which for the options discussed varies from 8.33 msec up to 2 sec.

The only way to get zero on/off pulsing is to use a controlled output voltage DC power supply. If you want to go this way, a switching power supply is much more efficient than a linear supply (I don't even know if you can find a linear supply at ~5500W.)

Looking for a solution implies that there is a problem that needs to be solved. Many of the replies to this thread are asking what is the actual problem that needs to be solved, and noting that there is no real unsolved problem related to powering heating elements in homebrewing applications. One of the basic tenants of engineering is that if you don't understand and state the problem correctly, you are very unlikely to find a good solution.

Brew on

Doug beat me to it - this would have been my exact response. AC, “Alternating Current”, by its very definition, works via “on” and “off” periods, so saying you don’t want that is conflicting.

Also as Doug mentioned, stating your goal can help drive the solution. It sounds to me you are down for the exercise of testing a unique pathway. But as a guy with a bit of Arduino/Software experience, I’ll tell you this isn’t how I’d skin that cat if practicality, cost, reliability, etc. were all being considered.

 

[maisch]

So what is your conclusion?

1.) It is possible to drive a heating element without on/off switching of the AC and the most viable solution is the one AJDelange offers using a saturable core reactor but those are custom built components and aren't cheap.

2.) There's a difference between a controller driving an SSR and the SSR driving the AC. Eliminating the controller switching the SSR on and off (or at least hiding it from the user) is done by using a proportional SSR which is driven by a variable potentiometer or rheostat - or a variable low voltage input (SSVR?). Eliminating the SSR switching the AC on/off isn't possible as that's how a triac circuit works, however you can get a "zero crossing" SSR which will switch at 0 degrees and be a bit more efficient than chopping off the sine wave.

3.) Either implement one of the #2's with an Arduino or purchase a PID unit, and I'm not sure if they can drive proportional SSR's or not. Then there's question of PID algorithms, etc... which was touched on in the other thread.

4.) From my point of view, the fact that AC works by alternating periods is too low level and really beside the point of this conversation.

 

[kydan47]

Not sure if we’re going for the same thing, but my Auberin Pids have 2 control settings (outL & outH) which allows me to raise the crossover point so as to get a more level boil. I usually use 60 & 70 percent.

 

[ajdelange]

1.) It is possible to drive a heating element without on/off switching of the AC and the most viable solution is the one AJDelange offers using a saturable core reactor but those are custom built components and aren't cheap.

Even a saturable core reactor uses AC and so is subject to the microseconds of off time when current goes through 0 at the end of each half cycle. Now a three phase saturable core reactor followed by a bridge rectifier would give you reasonably smooth (low ripple) DC with no zero crossings. So would a single phase rectifier with a capacitor to "fill in" near the 0 crossings. So would a linear power supply with a regulator (we want control here too) but we eschew those because of the power loss in the series pass transistor. How about a switching power supply? Much more efficient and also regulates?

The reactor solution was offered sort of tongue in cheek but the switching power supply can be offered seriously as it meets the requirements of giving you an adjustable continuous DC voltage to the elements without the losses of a pass transistor. Of course something is switched in a switching power supply but it is the input.

 

[maisch]

Even a saturable core reactor uses AC and so is subject to the microseconds of off time when current goes through 0 at the end of each half cycle. Now a three phase saturable core reactor followed by a bridge rectifier would give you reasonably smooth (low ripple) DC with no zero crossings. So would a single phase rectifier with a capacitor to "fill in" near the 0 crossings. So would a linear power supply with a regulator (we want control here too) but we eschew those because of the power loss in the series pass transistor. How about a switching power supply? Much more efficient and also regulates?

The reactor solution was offered sort of tongue in cheek but the switching power supply can be offered seriously as it meets the requirements of giving you an adjustable continuous DC voltage to the elements without the losses of a pass transistor. Of course something is switched in a switching power supply but it is the input.

*Not* talking about the sine wave of the AC, as noted numerous times that can't be changed. It's about the device providing constant power from 0W to 5500W to the element and being able to adjust it dynamically without turning or switching the (AC) power on and off for certain periods of time. A smooth transition from 0W to 50W from 50W to 1000W from 1000W to 500W from 500W to 4326W, etc... Dynamically adjustable power without a "period" or "cycle" time in which the AC is literally disconnected or turned off from the device and then turned back on. A 1s cycle time in which the AC power is turned on for 0.75s and turned off for 0.25s. <- is *not* what we're after.

As I understand it a saturable core reactor provides variable impedance to AC power, which seems to be inline with what is needed.

As an after thought I haven't researched DC powered heating elements nor what happens if you apply dc power to an ac element. It probably just has reduced capacity (i.e. doesn't provide the rated wattage).

 

[ajdelange]

When we talk about 240 VAC and AC current we are ignoring the fact that the AC voltage is a sine wave that varies in amplitude from +339.4 V to - 339.4 V and that 10 amperes of current means that the current is varying from 14.1 amps in one direction to 14.1 amps in the other. Instead we figure out how much DC voltage and current would be required to produce the same average heat and publish those numbers. Thus a heater supplied 240 V AC will produce the same amount of heat if it is supplied with 240 V DC except for tiny effects related to the inductance of the wires connecting it and "skin" effect (to which DC isn't subject).

 

[maisch]

So, build a 240VAC to 240DC power supply (bridge rectifier), smooth the signal to perfectly flat +240VDC w/ some additional components (maybe a capacitor).

Now what?

Are there digitally controlled variable rheostats, potentiometers that can handle such high voltage? How would one smoothly control the 0VDC - 240VDC input into the heating element?

I've noticed there are DC specific heating elements but not sure what advantage they have. They're marketed for solar power.

 

[doug293cz]

So, build a 240VAC to 240DC power supply (bridge rectifier), smooth the signal to perfectly flat +240VDC w/ some additional components.

Now what?

Are there digitally controlled variable rheostats, potentiometers that can handle such high voltage? How would one smoothly control the 0VDC - 240VDC input into the heating element?

I've noticed there are DC specific heating elements but not sure what advantage they have. They're marketed for solar power.

In it's simplest form, a switching power supply is just a bridge rectifier on the AC input, followed by some filtering, followed by a pulse width modulator, and finally more filtering. There is also a feedback circuit that monitors the output voltage and feeds an adjustment signal to the PWM to maintain the desired output voltage. The interesting thing about the switching power supply is that instead of PWM working at 8.33 msec cycle time, or slower, the PWM works with cycle times from about 100 micro sec (10 kHz) to 1 micro sec (1 MHz.) The higher frequency operation allows the use of smaller capacitors and inductors in the output filter circuit to obtain a given amount of output ripple.

You could use use voltage control of the feedback loop to control the output voltage, much like a PID uses the difference between the set temperature and measured temperature to control an SSR.

Notice that you still have pulsing voltage and current in the system, just usually at a higher frequency. The thing about pulsed power is that if the cycle time of the pulsed power is much less than the thermal time constant of the element, then the difference between DC and pulsed is not noticeable at the element output (which is what most of the posters in this thread have been hammering on.)

Brew on

 

[ajdelange]

In the simplest manifestation you would use IGFETs or IGBT's in the bridge instead of plain diodes and gate them from 0 to 1/120th second synchronous with the power line. You wind up with exactly what you would have with a triac gated this way except all the current pulses are in the same direction. Thus you can put a capacitor across the output of the bridge and smooth some of the ripple. It's going to have to be a pretty big cap, though, and in this application there is really no reason for doing this as it confers no advantage.

The other alternative is to use diodes in the bridge, put a more modest sized capacitor across the output and connect that to another capacitor through an IGFET or IGBPT gated by a controller which turns on the transistor with a PWM signal whose duty cycle is determined by the difference between a desired output voltage and the voltage measured at the second capacitor (i.e. across the load) and whose frequency is in the 10's or 100's of kHz. This is a feedback regulated bucking DC/DC converter. With high enough gating rate it can take out much of the ripple from the diode bridge rectifier thus giving much lower ripple at the output. But to what end? Again, there is in this application no advantage to doing this except as a learning exercise.

 

[maisch]

The guy who said Variac was on to something:

https://www.alibaba.com/product-det...m=a2700.7724857.normalList.137.3f0f1510Rk2Jz4

Now if that was only able to be controlled by an Arduino.

 

[ajdelange]

But of course it is. One buys a Variac with a motor instead of a knob and uses a servo circuit to position the motor shaft to keep the output voltage at a desired set point. I actually had one of these in the Middle East. It wasn't controlled by an Arduino though.

 

[BrunDog]

I keep checking the date to see if this thread's posts fall on 4/1.

 

[Bobby_M]

What electronics are used to constantly drive a heating element eliminating the rapid on/off switching of a PID?

I was thinking a frequency drive might do this but then how would the FD be controlled from say an Arduino?

Is this even possible without burning out the element faster?

Any electronics whizzes out there who can help?

Can you please state for the record why you want to avoid rapid switching? We all want to know If you have an interesting end goal or just didnt know that rapid switching is the way this is done and that there are no function downside.

 

[Temp81]

I’m not going back to reread this whole thread to make sure this was said at one point so don’t quote me on it, but I thought at one point it was mentioned that the rapid cycling of on off power from the duty cycle setting of the PID may shorten the life of the element. All that aside I think this thread got a little too techy for what the OP was asking for. I’m an electronics technician so I understand electricity, duty cycles, and all that stuff. If I understood the OP correctly he just wants smooth power transition with out the pulsing from the PID. We effectively covered multiple ways to vary and control the power, but I think all he was trying to do was turn the voltage down. A variac does this well, but as already mentioned, it would be kind of impractical. As far as the ardunio control, that is really separate issue in my mind.

 

[maisch]

All that aside I think this thread got a little too techy for what the OP was asking for.

Absolutely. Yes.

If I understood the OP correctly he just wants smooth power transition with out the pulsing from the PID.

Absolutely. Yes.

I can do without the insults and overly technical explanations.

 

[maisch]

Can you please state for the record why you want to avoid rapid switching? We all want to know If you have an interesting end goal or just didnt know that rapid switching is the way this is done and that there are no function downside.

When I learned rapid on/off switching of the SSR by the controller was the method used, I thought there must be a better way. The interesting end goal *is* constant drive or power to the element. Learning that wasn't possible, I thought maybe there was an abstraction that could be used to hide the rapid on/off switching of the SSR by the controller. Thus the suggestion of the proportional SSR.

In terms of algorithms (PID, AI, Fuzzy) used to control the SSR, I don't know if there would be any difference between one designed for rapid on/off switching and one that "thinks" in terms of a smooth linear percentage of power 0% - 100%. Perhaps that algorithm would also be an interesting end goal.

 

[Temp81]

@ maisch
I’m glad I interpreted your post correctly for what you were asking. Sometimes we all have ideas in our head that don’t seem to come out of our mouths or go to print the right way to ask the right questions. Everyone will always have there own interpretations to every post. Also, there are a lot of really smart and technically sound people in this form that come up with ideas, solutions, and contraptions that sometimes make my head spin. Take replies to posts with a grain of salt sometimes. I don’t feel anyone was being malicious in this thread, they were just trying to understand and provide help. I think they just didn’t feel you were directly answering their questions and got a little frustrated. Anyway, I’m not going to sit here and preach, and I hope you can find a viable way to do the idea in your head. Brew on!

 

[ajdelange]

The interesting end goal *is* constant drive or power to the element. Learning that wasn't possible...

This may be where you are hung up. The goal isn't really constant drive power to the element at all. It is constant heat delivery to the load (wort). Constant heat is indeed impossible but we can approximate it sufficiently closely by supplying pulses of heat as long as they are sufficiently closely spaced such that a plot of temperature vs time is not discernibly different from a sloped straight line until boiling is reached after which we want boiling to be continuously maintained at a uniform level of vigour.

The thing that allows us to do this with pulses is a thermal time constant that is greater than the pulse spacing. Think of a can with a hole in the bottom. Water flows out of the can at a rate proportional to how full the can is. If the can is big enough and the hole small enough we can maintain a virtually constant flow of water through the hole even though we dump a cup of water into the can, walk across the room to the sink to refill the cup, come back and dump it in, and repeat. If we hustle the water level in the can will rise as ON AVERAGE we are filling the can faster than the water is running out. Of course the fuller the can becomes the faster the water flows out and we have to keep hustling. If we are leisurely the opposite is true. Thus the flow from the can can be pretty uniform and at any rate we want which we control by how frequently we dump in a cup of water and/or by how big a cup we use.

The water flow is not completely uniform but if the can is big enough it can be made to be pretty smooth.