A cooker ring (resistive element) and driving it

[negativ3]

Hi all,

I am building a PLA burnout furnace as a first step for aluminium casting and am looking for some advice regarding the control of a re-purposed cooker ring.

I currently hook it up to 240V via a relay, and pulse (PWM effectively) it via a micro-controller with a thermocouple and a PID routine as if I want to maintain a certain temperature, however I was wondering if it's possible to control the current via a transistor circuit?

My thoughts are that pulsing it like I am doing (relay) is going to shorten the rings (resistive element) life.

Any advice would be appreciated.
Andy

[helius]

More likely that your relay will give out first.

[3db]

Perhaps a  triac and phase control or zero voltage switching.

[negativ3]

So this is like a very fast AC switch to the resistive load... interesting and has started further investigation. Thank you.

[Ian.M]

Phase control really isn't needed for large thermal mass heating elements and causes excessive RFI.   While you could use an optocoupler with zero crossing detection driving the gate of a TRIAC to replace the existing relay, you can get all of that in an affordable easy to use module.  Simply use a solid state relay - that will give you the zero crossing switching, TRIAC and snubber in one easy to hook up unit, then control the heat output by cycle skipping:   To minimise DC imbalance you should update the controller output at the same frequency as the mains supply.

[negativ3]

Phase control really isn't needed for large thermal mass heating elements and causes excessive RFI.   While you could use an optocoupler with zero crossing detection driving the gate of a TRIAC to replace the existing relay, you can get all of that in an affordable easy to use module.  Simply use a solid state relay - that will give you the zero crossing switching, TRIAC and snubber in one easy to hook up unit, then control the heat output by cycle skipping:   To minimise DC imbalance you should update the controller output at the same frequency as the mains supply.

This is excellent Ian, thanks. So an SSR does not suffer the same issues as a traditional relay, i.e. moving/switching mechanical components. Guess the answer to that is in it's name. Are any issues such as inrush current still an issue?

Added DC imbalance to the list to read up and learn.

[Ian.M]

DC imbalance is when the load draws a net DC current from the AC mains supply.  Its generally frowned upon by the utility company, as an excessive number of high power loads only drawing current on the same halfcycle can cause enough DC current in the secondary of the utility's distribution transformer to nudge it into saturation and cause it to overheat.   A single cooker element is unlikely to *EVER* cause a problem even if you connect it via a series diode, but its still good practice to avoid DC imbalance.

An AC SSR is an optocoupler, TRIAC and snubber all in one pre-packaged module.   Inrush current is still potentially a problem, and SSRs tend to fail shorted if their peak or average current ratings are exceeded, so in a frequent switching application, you should calculate the first cycle current from the cold resistance of the element and the supply voltage then choose a SSR with a generous margin on its continuous current rating.

[negativ3]

DC imbalance is when the load draws a net DC current from the AC mains supply.  Its generally frowned upon by the utility company, as an excessive number of high power loads only drawing current on the same halfcycle can cause enough DC current in the secondary of the utility's distribution transformer to nudge it into saturation and cause it to overheat.   A single cooker element is unlikely to *EVER* cause a problem even if you connect it via a series diode, but its still good practice to avoid DC imbalance.

An AC SSR is an optocoupler, TRIAC and snubber all in one pre-packaged module.   Inrush current is still potentially a problem, and SSRs tend to fail shorted if their peak or average current ratings are exceeded, so in a frequent switching application, you should calculate the first cycle current from the cold resistance of the element and the supply voltage then choose a SSR with a generous margin on its continuous current rating.

Fantastic, thank you Ian. I will absorb all the information then go with some real world measurements and testing.

[IanMacdonald]

Agree with IanR. Use a triac arrangement. Add a thermal cutout in case of triac s/c failure. Ideally should be zero-voltage switched but probably not too important if it just switches at a random point in the cycle. No worse than what mechanical thermostats do all the time.

No need for phase control (and could cause a lot of RFI on this size of load) and elements are designed to be pulsed on/off by the usual mechanical simmerstats, so no worries there.

[amyk]

The thermal mass is likely large enough that a regular mechanical relay will be OK.

[negativ3]

The thermal mass is likely large enough that a regular mechanical relay will be OK.

amyk, kindly explain?

[Ian.M]

Sure, a mechanical relay will be OK for occasional use if you slow down the control system response enough and don't need to hold the temperature closely, but its going to wear out fairly quickly due to contact burn if its breaking a large proportion of its rated current several times a minute.   If you can afford an appropriately rated SSR, its an easy drop-in replacement for the existing relay.   As you are driving it from your own code in the MCU, it would be fairly easy to add isolated load current sensing to detect any failures and add an audible alarm for it.    Whether or not you need to add a thermal trip depends on the consequences of  severe over-temperature.   However as the most problematic time to have uncontrolled full power would be while still melting the PLA out of the mould due to the risk of igniting the PLA in the drip tray, you might want to simply keep the existing relay in series with the SSR so the controller can cut the power if it detects a SSR failure or other fault. The mechanical relay woud be constanly on in normal operation. Redundant temperature sensors would be a good idea - if there is too much of a difference between them, treat it as a fault and shut down.

If you aren't going to use the burnout kiln regularly,  you may wish to build up the controller as a detachable unit with a socket for controlled mains out, and plug in connectors for the temperature sensors so you can use it for other heat control applications.   You may even decide to add the capability to load settings and temperature profiles from SD card or via USB to make it more versatile.

[negativ3]

Sure, a mechanical relay will be OK for occasional use if you slow down the control system response enough and don't need to hold the temperature closely, but its going to wear out fairly quickly due to contact burn if its breaking a large proportion of its rated current several times a minute.   If you can afford an appropriately rated SSR, its an easy drop-in replacement for the existing relay.   As you are driving it from your own code in the MCU, it would be fairly easy to add isolated load current sensing to detect any failures and add an audible alarm for it.    Whether or not you need to add a thermal trip depends on the consequences of  severe over-temperature.   However as the most problematic time to have uncontrolled full power would be while still melting the PLA out of the mould due to the risk of igniting the PLA in the drip tray, you might want to simply keep the existing relay in series with the SSR so the controller can cut the power if it detects a SSR failure or other fault. The mechanical relay woud be constanly on in normal operation. Redundant temperature sensors would be a good idea - if there is too much of a difference between them, treat it as a fault and shut down.

If you aren't going to use the burnout kiln regularly,  you may wish to build up the controller as a detachable unit with a socket for controlled mains out, and plug in connectors for the temperature sensors so you can use it for other heat control applications.   You may even decide to add the capability to load settings and temperature profiles from SD card or via USB to make it more versatile.

Thanks Ian, I plan to install an RCD, traditional fuse and a body thermal trip inline with the live line but re-using the relay is also a good idea.
I'm already fairly far on with the uC part including a learn cycle to profile the heating element plus loading profiles from SD...

[negativ3]

And my first circuit of this device... no PSU for the uC yet.

[IanB]

Hi all,

I am building a PLA burnout furnace as a first step for aluminium casting and am looking for some advice regarding the control of a re-purposed cooker ring.

I currently hook it up to 240V via a relay, and pulse (PWM effectively) it via a micro-controller with a thermocouple and a PID routine as if I want to maintain a certain temperature, however I was wondering if it's possible to control the current via a transistor circuit?

My thoughts are that pulsing it like I am doing (relay) is going to shorten the rings (resistive element) life.

Any advice would be appreciated.
Andy

The traditional power control of a cooker ring was a simmerstat (Google it), a mechanical device that periodically turned the power on and off to the ring, regulating the on time according to the setting of the control knob. So turning the power on and off like this would be no problem at all, and also should not be a problem for a suitably rated electrical contact such as the one in the simmerstat.

[negativ3]

a suitably rated electrical contact such as the one in the simmerstat.

IanB,

When you say the above, do you mean an SSR rated the same or better than the original simmerstat in the oven? The ring is being recycled.

Thank you.

[Ian.M]

Calculate the RMS current in the first cycle from cold using the nominal supply voltage +10% (for line voltage variation) and the measured cold resistance.   To allow enough margin for reliability, you need a SSR of at least 25% more rated current, (50% is better) *AFTER* any derating factors for ambient temperature specified in the SSR datasheet have been applied.   

N.B. if the element burns out and partially shorts or shorts to ground it is highly likely to blow the SSR before it blows the fuse.  If the element has a grounded shell, put the SSR on the Neutral side of the element so that the ground fault current in case of element failure does not pass through the SSR

For a mechanical contact, you'd want a much larger margin - maybe a factor of two or three to get enough contact surface to stand up to the frequent cycling.   The simmerstat is designed for frequent cycling as that's what it does.  If you do a teardown, you'll almost certainly find the contacts are oversized compared to a relay of a similar current ration.

Power for the MCU is not a significant problem - just hack a USB charger or other small wallwart to extract the board and hard wire it, keeping enough of the wallwart case to make a safe mount for the board.   Check the PSU board has adequate creepage and clearance distances - if in doubt post photos here for comment.

I'd add a neon indicator on the load side of the SSR.  Seeing it blink on and off is good confirmation that everything is working as expected.

[Zero999]

Use an optically isolated zero-crossing TRIAC, such as the MOC3043, and a couple of SCRs or a TRIAC.
http://exa.unne.edu.ar/ingenieria/sysistemas/public_html/Archi_pdf/HojaDatos/Optoelectronica/moc3041.pdf

[Ian.M]

That is basically what is inside an AC SSR.    A universal SSR (AC/DC) however is more likely to use a pair of photo-MOSFETs.

Yes the TRIAC circuit will do the job, but if you are breadboarding, a chassis or heatsink mounting Zero Crossing SSR, with screw terminals for the load circuit is vastly preferable, if only for the safety it provides.   

[negativ3]

That is basically what is inside an AC SSR.    A universal SSR (AC/DC) however is more likely to use a pair of photo-MOSFETs.

Yes the TRIAC circuit will do the job, but if you are breadboarding, a chassis or heatsink mounting Zero Crossing SSR, with screw terminals for the load circuit is vastly preferable, if only for the safety it provides.

Thanks Ian.M

Yea I had a surprise in 1994 when a piece of equipment I was working on needed flipping over on the bench. I had not isolated the power so needless to say, where hands needed to go had all the open-cage power supplies. Since then the greatest respect for mains voltages and live equipment has been observed.

[negativ3]

Use an optically isolated zero-crossing TRIAC, such as the MOC3043, and a couple of SCRs or a TRIAC.
http://exa.unne.edu.ar/ingenieria/sysistemas/public_html/Archi_pdf/HojaDatos/Optoelectronica/moc3041.pdf

Interesting however the website link does not respond for me.

[Zero999]

Use an optically isolated zero-crossing TRIAC, such as the MOC3043, and a couple of SCRs or a TRIAC.
http://exa.unne.edu.ar/ingenieria/sysistemas/public_html/Archi_pdf/HojaDatos/Optoelectronica/moc3041.pdf

Interesting however the website link does not respond for me.

Try this:
http://www.mouser.com/ds/2/308/MOC3041M-1121232.pdf

I think the dual SCR circuit might be better suited to this, as it's easier to get them in high power rating and the power will be distributed across two devices.