SSR relay problem solution

[mdkmdk]

Hello everyone,

I have a goldsmith oven to which I had to change an SSR relay because when it reached 475 degrees centigrade the temperature increase stopped. Since the oven must reach 800 degrees in about two hours, I have to solve the problem.

I changed all the main components to the oven: thermoregulator, thermocouple, resistance. The problem of increasing the temperature was present even before changing all these components.

I ordered an SSR relay of the same brand and model as the one mounted on the oven: Crydom D2425.

Since it was panel mounted I proceeded to modify the panel to put a heatsink in it.

Between the heatsink and the relay I put some Thermalright TFX thermal paste which has a high thermal conductivity power and is also used in computers and can withstand up to 300 degrees.

After having reassembled everything I turned on the oven, this time it reached 100 degrees centigrade after 10-15 minutes the increase in temperature stopped.

I wonder where he went wrong about the SSR relay changing as I think I've done everything possible. Could anyone help me? Thank you in advance for your reply.

I attach some photos to the message.

 

[marconi]

AL1 and AL2 are over- and under- temperature limits to alert you if the oven goes to far above or below the set temperature. Would this be useful to you?

 

[mdkmdk]

Good you have been looking at the controller's settings. Would you go through the steps as in the youtube video and report to me the settings you find? I can then tell you what to change or keep the same.

More immediately please me tell if you have set up 'InP' value as 'k' ( see at 5m57s) to tell the device you are using a thermocouple - and then tell me if the 'd' function is set to on or off (see at 6m57s).

The input signal is set to the type K thermocouple, which is the one actually present and which has been replaced. The variable "d" corresponding to the derivation time is set with the default value "30", in the manual it is written that if it were set to "0" the derivation would be closed. The variable AL1, on the other hand, is set to the default value "10". The AL2 variable is not present. SLL (setting of minimum set value) and SLH (setting of maximum set value) are the variables that, if set, put a minimum and maximum limit on the temperature that can be manually set on the thermoregulator display, I wonder: what is the difference between these variables and AL1? I never understood what AL1 was really for, is it perhaps an audible alarm that sounds when the temperature is reached?

 

[marconi]

I am wondering now whether your controller is not exactly the same as the one shown in the youtube clip but similar in many respects.

How long does the oven take now to reach 800C?

Please set d to 200 and let me know how the oven performs.

I did ask for all the controller settings in my earlier post.

I will have another search for the manual for your controller. All I could find before was a leaflet with brief details.

Are you using the oven now to melt gold?

 

[mdkmdk]

I am wondering now whether your controller is not exactly the same as the one shown in the youtube clip but similar in many respects.

How long does the oven take now to reach 800C?

Please set d to 200 and let me know how the oven performs.

I did ask for all the controller settings in my earlier post.

I will have another search for the manual for your controller. All I could find before was a leaflet with brief details.

Are you using the oven now to melt gold?

I have not yet done the test of setting the variable 'd' to 200, however the auto tuning is still in operation, to get to 800 degrees the oven takes 2 hours and 27 minutes setting the temperature to 1000 degrees in order to make it rise continuously without being slowed down by the overshooting system. An anomalous thing I noticed is that if it reaches for example the temperature of 790 degrees and I set the temperature variable to be reached from 1000 to 790 to keep the temperature stable for ten minutes, I noticed a drastic drop, in 10 minutes it dropped to 736 degrees. The temperature maintenance system did not work, same thing happened to me when I dropped the temperature to 540 degrees and I set the temperature to be reached at 550 degrees, the temperature did not rise, I had to set 600 degrees to make it rise.

 

[mdkmdk]

Here is the manual of the MC-101 series temperature controller.

 

[marconi]

Thank you. I will study the manual and have a think. It really would help to know the settings which have been input since these determine the way the controller responds.

 

[marconi]

I have not yet done the test of setting the variable 'd' to 200, however the auto tuning is still in operation, to get to 800 degrees the oven takes 2 hours and 27 minutes setting the temperature to 1000 degrees in order to make it rise continuously without being slowed down by the overshooting system. An anomalous thing I noticed is that if it reaches for example the temperature of 790 degrees and I set the temperature variable to be reached from 1000 to 790 to keep the temperature stable for ten minutes, I noticed a drastic drop, in 10 minutes it dropped to 736 degrees. The temperature maintenance system did not work, same thing happened to me when I dropped the temperature to 540 degrees and I set the temperature to be reached at 550 degrees, the temperature did not rise, I had to set 600 degrees to make it rise.

I might be able to comment on what you have written if only I had the PID controller's current settings ...................

I have some ideas on how to proceed once I have the settings in front of me.

Are you writing and reading here in your first language?

 

[mdkmdk]

I might be able to comment on what you have written if only I had the PID controller's current settings ...................

I have some ideas on how to proceed once I have the settings in front of me.

Are you writing and reading here in your first language?

Some variables in the manual are not present in the menu of my temperature controller. I list below the variables related to the PID that are present in the menu:

At, Al1, SC, P, I, d, t, rSt, LCy are present.
The variable 'At' is still set to Yes since I have activated auto tuning.
The variable 'P' is set to 20.0 instead of 30.0 as it would be by default.
All other variables are set with factory default values. You can see these values in the central 'Factory Defaults' column of the manual.

Al2, oH, rE, bUFF are not present.

English is not my primary language. I really appreciate your interest and thank you for your help.

 

[marconi]

Thank you. Armed with this information I will be back in touch. Meanwhile could you do a trial for me please? Please set the Control Period 't' value to 2 seconds suitable for an SSR - at the moment it is set to 20s required for a relay. Then disable the Derivative D function by setting it to 0. Because the oven has a high thermal capacity and hence is slow to change in temperature the D function will not be that useful for temperature regulation. How does the oven perform now when the set Value SV is 800C? eg time to first reach 800C, overshoot and undershoot temperatures, time to 'stay at' 800C. Is it 'better' than when D is selected with a value of 20? Keep the AT selected.

The Control Period Value t being set to 20 is wrong because you are using an SSR which can switch more frequently than a relay without damage. I am pondering that your oven only needs PI control not PID. The reason the AT may be taking a long time or even not succeeding is that it assumes the controller has dual action ie it can raise the temperature and also lower the temperature by adding or removing energy. Of course the element can only add energy not remove it. For the oven to cool it needs heat energy to be removed and this must rely on it flowing out of the oven which I assume has reasonably well insulated walls so it is going to take more time to reduce the temperature by say 50C than to raise it by 50C - the controller does nothing but wait with a zero output and element off when the actual temperature PV has to decrease.

I am finding this problem interesting and a nice revision of the control engineering I learned/forgot decades ago.

 

[mdkmdk]

Thank you. Armed with this information I will be back in touch. Meanwhile could you do a trial for me please? Please set the Control Period 't' value to 2 seconds suitable for an SSR - at the moment it is set to 20s required for a relay. Then disable the Derivative D function by setting it to 0. Because the oven has a high thermal capacity and hence is slow to change in temperature the D function will not be that useful for temperature regulation. How does the oven perform now when the set Value SV is 800C? eg time to first reach 800C, overshoot and undershoot temperatures, time to 'stay at' 800C. Is it 'better' than when D is selected with a value of 20? Keep the AT selected.

The Control Period Value t being set to 20 is wrong because you are using an SSR which can switch more frequently than a relay without damage. I am pondering that your oven only needs PI control not PID. The reason the AT may be taking a long time or even not succeeding is that it assumes the controller has dual action ie it can raise the temperature and also lower the temperature by adding or removing energy. Of course the element can only add energy not remove it. For the oven to cool it needs heat energy to be removed and this must rely on it flowing out of the oven which I assume has reasonably well insulated walls so it is going to take more time to reduce the temperature by say 50C than to raise it by 50C - the controller does nothing but wait with a zero output and element off when the actual temperature PV has to decrease.

I am finding this problem interesting and a nice revision of the control engineering I learned/forgot decades ago.

I did the test with the configuration you suggested, what I noticed is that after 2 hours the temperature was at 750 degrees instead of 764 as in the other measurement with the setting at 1000 degrees and without the configuration you suggested. I don't understand why there is this 14 degrees difference. I don't think it could be the resistance that has degraded considering that it has been changed by a specialized technician a couple of months ago. I didn't always watch the oven continuously so I don't know if the heating stopped sometime, but when I took the measurement it was on.

The second thing I noticed was that at 760 degrees the heating turned off, after which it dropped to 753 degrees and stopped there; after a few minutes it began to rise. Then it reached 764 and switched off again, after that the temperature dropped and after a while it started again reaching 763 degrees, then it did the same thing stopping at 766. In 2 hours and 28 minutes it was at 766 degrees, instead in the measurement made days ago (with the temperature set at 1000 degrees) reached 800 degrees in 2 hours and 27 minutes. Since it seemed impossible that the temperature would rise more, I decided to finish the test.

 

[marconi]

Thank you. Tuning the controller is I regret to say largely 'trial and error'.

Next let us reduce the I value from 210 to 180. Reducing this time increases the integral action I and helps to reduce the error quicker as the temperature PV gets closer to the set value SV, but it can also increase oscillations so some more trial and error with this value. The integral action makes up for the reducing drive from the P proportional action since P depends on the error and the error is decreasing.

Please use 800C as the set value from the moment you turn on the oven. We are attempting to tune the system to achieve 800C 'quickly' not 1000C. If you set SV at 1000C the controller does not use proportionate control until close to 1000C and the temperature enters the proportional control band centered around 1000C. We want proportionate control to kick in as the oven nears 800C. The SV value determines the midpoint of the proportional band.

Please understand the oven and controller has two 'behaviours' working at the same time. From cold to set value temperature the oven is operating in its transient mode since there is the biggest change in temperature. As the oven temperature reaches the set value the transient behaviour should reduce and the oven enters its steady state behaviour only making small changes to keep the temperature near to 800C.

Transient response - Wikipedia - https://en.wikipedia.org/wiki/Transient_response#:~:text=In%20electrical%20engineering%20and%20mechanical,the%20equilibrium%20of%20the%20system.

 

[mdkmdk]

Thank you. Tuning the controller is I regret to say largely 'trial and error'.

Next let us reduce the I value from 210 to 180. Reducing this time increases the integral action I and helps to reduce the error quicker as the temperature PV gets closer to the set value SV, but it can also increase oscillations so some more trial and error with this value. The integral action makes up for the reducing drive from the P proportional action since P depends on the error and the error is decreasing.

Please use 800C as the set value from the moment you turn on the oven. We are attempting to tune the system to achieve 800C 'quickly' not 1000C. If you set SV at 1000C the controller does not use proportionate control until close to 1000C and the temperature enters the proportional control band centered around 1000C. We want proportionate control to kick in as the oven nears 800C. The SV value determines the midpoint of the proportional band.

View attachment 99614

Please understand the oven and controller has two 'behaviours' working at the same time. From cold to set value temperature the oven is operating in its transient mode since there is the biggest change in temperature. As the oven temperature reaches the set value the transient behaviour should reduce and the oven enters its steady state behaviour only making small changes to keep the temperature near to 800C.

Transient response - Wikipedia - https://en.wikipedia.org/wiki/Transient_response#:~:text=In%20electrical%20engineering%20and%20mechanical,the%20equilibrium%20of%20the%20system.

Should I leave the other variables "d = 0", "t = 2" and "AT = yes" unchanged and change only the variable "I" from 210 to 180?

Could I do the test by entering a lower temperature like 500 degrees or does it always need to set 800 degrees? Putting 500 degrees in an hour I would have the response and I could do the test faster, otherwise like the previous test I will proceed by setting SV to 800.

 

[marconi]

Should I leave the other variables "d = 0", "t = 2" and "AT = yes" unchanged and change only the variable "I" from 210 to 180?

Could I do the test by entering a lower temperature like 500 degrees or does it always need to set 800 degrees? Putting 500 degrees in an hour I would have the response and I could do the test faster, otherwise like the previous test I will proceed by setting SV to 800.

Yes leave those variables unchanged. Only change I. You can use 500C for the test.

 

[Darkwood]

Tried skipping through quickly so sorry if already mentioned, the colour coding on your thermocouple wires could be from a different global zone thus identified incorrectly, assuming the controller has been configured for the therm' type it is monitoring, not doing so will give the wrong feedback voltages and the wrong temperatures. Type K And J are commonly used but must be correctly input into the PID controller fir it to Interpret what the thermocouple resistance feeds back is.

 

[marconi]

Thermocouple Cable Colour Codes and Tolerances - TC Ltd - https://www.tc.co.uk/thermocouple-information/thermocouple-colour-code.html

The k type thermocouple has red and green wires which would indicate it is using the German colour code red for plus and green for minus. This is how it is connected to the controller. See #49.

mdkmdk - could you send me details of the thermocouple please? Is it made in Germany. Do the red and green wires go all the way to the thermocouple or have you joined the thermocouple wires to the red and green wires which go to the controller?

If you have a small permanent magnet and hold it near the red wire and then the green wire the wire which is most magnetic is the negative lead.

 

[mdkmdk]

Yes leave those variables unchanged. Only change I. You can use 500C for the test.

I did the test you suggested by changing the variable 'I' and setting it to '180':
The oven reached the temperature of 475 degrees and the heating turned off.
It dropped to 486 degrees while the heating was off.
The temperature dropped, and at 474 degrees it started again.
The temperature dropped to 466 degrees, and the heating turned on again.
The heating turned off at 477 degrees.
Then I finished the test as it never seemed to reach 500 degrees.

Thermocouple Cable Colour Codes and Tolerances - TC Ltd - https://www.tc.co.uk/thermocouple-information/thermocouple-colour-code.html

The k type thermocouple has red and green wires which would indicate it is using the German colour code red for plus and green for minus. This is how it is connected to the controller. See #49.

mdkmdk - could you send me details of the thermocouple please? Is it made in Germany. Do the red and green wires go all the way to the thermocouple or have you joined the thermocouple wires to the red and green wires which go to the controller?

If you have a small permanent magnet and hold it near the red wire and then the green wire the wire which is most magnetic is the negative lead.

Tried skipping through quickly so sorry if already mentioned, the colour coding on your thermocouple wires could be from a different global zone thus identified incorrectly, assuming the controller has been configured for the therm' type it is monitoring, not doing so will give the wrong feedback voltages and the wrong temperatures. Type K And J are commonly used but must be correctly input into the PID controller fir it to Interpret what the thermocouple resistance feeds back.

A few months ago I also changed the thermocouple. Before replacing it I did the test with the magnet and it showed that it was a type K thermocouple. Later I replaced it with a new one of the same type, obviously K. I remember that before changing it I inquired about the colors of the wires and it emerged that they probably refer to the German colour code.

 

[marconi]

Please could you clarify ‘salt from off… ‘?

 

[mdkmdk]

Please could you clarify ‘salt from off… ‘?

It dropped to 486 degrees while the heating was off. Sorry for the accidental mistake, I made the change to the reply and it will be visible shortly after being checked by the forum moderators.

 

[marconi]

First it seems to me the controller is working and that 25C or so away from 500C seems to me quite good after the initial period of heat up. Would you please leave for longer and find out how close the oven temperature settles to 500C?

Jogged by Westward10's post I am thinking about the error in measuring the oven temperature PV introduced by the connection of the thermocouple's red and green leads to the metal terminals on the back of the controller. These two connections are the other half of the thermocouple circuit. A thermoelectric circuit has a hot sensing end thermocouple and a cold reference end thermocouple. The combined circuit produces a voltage/current proportional to the difference between the temperatures of each end. See:

Thermocouple Cold (Reference) Junction Compensation - https://blog.beamex.com/thermocouple-cold-junction-compensation

The reference thermocouple made by the terminal connections will reduce the voltage produced by the hot end. Thus temperature measurements by the hot end thermocouple will be seen by the controller as lower than they actually are. The controller has a an SC value for correcting sensor measurement errors so I will come back to you later about entering a value other than 0.

The problem is that the reference temperature 'cold junction' is actually getting warmer and warmer as the oven does because the terminals are enclosed in a space near the oven and ssr.

How accurate does the temperature need to be for your purpose?

 

[mdkmdk]

First it seems to me the controller is working and that 25C or so away from 500C seems to me quite good after the initial period of heat up. Would you please leave for longer and find out how close the oven temperature settles to 500C?

Jogged by Westward10's post I am thinking about the error in measuring the oven temperature PV introduced by the connection of the thermocouple's red and green leads to the metal terminals on the back of the controller. These two connections are the other half of the thermocouple circuit. A thermoelectric circuit has a hot sensing end thermocouple and a cold reference end thermocouple. The combined circuit produces a voltage/current proportional to the difference between the temperatures of each end. See:

Thermocouple Cold (Reference) Junction Compensation - https://blog.beamex.com/thermocouple-cold-junction-compensation

The reference thermocouple made by the terminal connections will reduce the voltage produced by the hot end. Thus temperature measurements by the hot end thermocouple will be seen by the controller as lower than they actually are. The controller has a an SC value for correcting sensor measurement errors so I will come back to you later about entering a value other than 0.

The problem is that the reference temperature 'cold junction' is actually getting warmer and warmer as the oven does because the terminals are enclosed in a space near the oven and ssr.

How accurate does the temperature need to be for your purpose?

From 475 to 500 degrees it would take a long time since the heating stops constantly.

I completely redone the panel when the resistance was changed and I made the preparation for the new thermocouple since it did not have a bolt on which to stand. I soldered it at a slight angle to prevent the thermocouple wires from getting close to the resistor wires and the oven exhaust pipe.

It is important for me that the temperature is accurate and real and that it reaches 800 degrees as soon as possible.