Assistance on connections to Power Supply

[Les Jones]

Hi zexclo,
Can you confirm the type of output on your PID controller as the link you provided in your first post is for a different model to the one you have ? Is the output an on/off switching type or is it a proportional signal. (All my comments have been made on the assumption it is an on/off switching type.) I am confused when you said that you measured the output as 27 volts and then you used a potential divider consisting of a one ohm and a two ohm resistor. I do not believe either type of PID would be capable of driving such a low resistance potential divider. (27 volts across 3 ohms is 9 amps, 9 amps x 27 volts is 243 watts) If the PID is really capable of such a high power out then it will drive the heater directly.

Les.

 

[Tony Stewart]

It would help to define what you need to regulate T for.

Lower mass , the easier it gets.

e.g when I wanted to test Xtal cans for TCXO application, we wanted to regulate T to 40C and 70C with 15 seconds and within 1 deg, so all we needed was 1/2 W SMD on flex cct wrapped inside styrafoam socket for low mass to achieve with only fast On/off control. , essentially PWM,PFM control.

e.g when I wanted to test circuit board from -40C to +70C , I bought an insulated PVC picnic box and used dry ice with 5W muffin fan , for cold after 15 minutes then remived dry ice and used 35W solder iron suspended in air inside box with fan and reached in 30 min. Without regulation. Just monitoring T.

If high mass, and rate change is fast, then overshoot is possible so PID is needed, otherwise not.

 

[KeepItSimpleStupid]

Let's go over some of the I/O options, so you can understand them better. A reminder that the outputs could be something like heat/cool or even open/close for a valve. Others could be used as alarms or control (e.g. RS232) There is only a set number of pins.

Output Options
A = None
(MEANS NOT FITTED)
B = Solid-state relay, Form A, 0.5A, with RC suppression
(An AC relay used to turn on a very small load)
C = Switched dc or open collector, isolated
(This is one of the outputs you have, You need to supply a low current
DC source or the device does)
(It may have to be current limited by you)
(It could be used to turn on a 3-32 V input SSR with an EXTERNAL 3-32 V supply)
D = Electromechanical relay, Form C, 5A with RC suppression
(Form C is, I believe SPDT), suppression for AC loads I would guess)
E = Electromechanical relay, Form C, 5A without contact suppression
(See D)
F = Universal process 0-5V(dc), 1-5V (dc), 0-10V(dc), 0-20mA(dc), 4-20mA(dc), isolated
(These are typical process signals. The voltage outputs will have an output
(resistance associated with them). 0-20 mA is easy to turn into a 0-5 or 0-10 V control
(signal. Using current removes the distance limitation when using voltage)
K = Solid-state relay, Form A, 0.5A, without contact suppression
(An AC SSR used for "pilot duty" to usually turn on another TRIAC/SSR)
T = External signal conditioner power supply, 5, 12 or 20V(dc) @ 30mA
(This is just a small power supply)

 

[Tony Stewart]

The only requirement for PID loop is to prevent overshoot.
If heat transfer is not equal to all internal parts, then you need multiple sensors.

I used variable speed fan on a 180W power supply to keep the Fan speed lowest and temperature <55'C using only P or proportional control. Often this is all you need.

Start by defining your mass to be heated , temperature rise, Rth thermal resistance of enclosure and T error for regulation and max response time. Do this before any implementation thoughts.

 

[KeepItSimpleStupid]

The only requirement for PID loop is to prevent overshoot.

The use of PID also makes the setpoint and measured value agree.
With P, they don't at all
With PI, they do, but might have overshoot
With PID, it;s likely overshoot can be controlled.

 

[Tony Stewart]

The use of PID also makes the setpoint and measured value agree.
With P, they don't at all
With PI, they do, but might have overshoot
With PID, it;s likely overshoot can be controlled.

Actually all PID controllers have limits and thus are nonlinear meaning full power is applied if error >X

In my examples , the limits for P are tighter, so it is only linear over a narrow range below setpoint.
As in all cases of PID control theory, the error is inversely related to proportional gain.

Thus it always depends on the previous specs for tolerance I stated. Which relate to conductance and thermal insulation and the heat source and rate if change with error limits.

 

[KeepItSimpleStupid]

Your correct on all counts.

Actually all PID controllers have limits and thus are nonlinear meaning full power is applied if error >X

In an analog controller P out is limited by the power supply rails or a control. Eurotherm implemented the Maximum output of the output of the error amplifier".

In software, it's very wise to limit the integral of E to avoid Reset windup" or long recovery time.

Yea, your right, it's linearly proportional between the proportional band of 1/P

I did one wierd controller in an environmental chamber where I controlled the the oven setpoint using the proportional band as my -100+100% and implemented my own PID in software to control the temperature of a surface. Quite wierd.

The first PID controller I did in software was back in the 80's. We had voltage, current, power, current, temperature control all going at the same time, It was actually implemented as one as control and the others were limits. dP/dt was also implemented. Reset windup was implemented. We even implemented an energy limits on heat-up to detect a shorted or misplaced thermocouple. To top it all off, it had recipies and 7 loops and stability criteria. This was before the "IBM PC", Probably early 80's. Written in Fortran under the DEC RT11 F/B OS. We did need one assembly language instruction for whatever reason. Move From Previous Instruction or Data Space.

 

[Tony Stewart]

When I had a waterbed and heater, the mass is large with a 2 day time constant from 22'C but a few degrees was noticeable so I used a simple OA comparator and bridge to drive the heater relay and it would chatter from EMI during transition (regulation) ... since the regulation was proportional but over a very small range, I added 0.5 deg pp hysteresis just above the noise level and then it worked fine. Except if the heater was on, and I jumped into bed it would have closer contact with the bed from added pressure and was must enough to shut off thermostat for 10 minutes or more until temp dropped 0.5 deg. So thermal sense accuracy is also important with high gain or bang bang comparator oops I mean hysteresis....

All heating thermostats work this way for residential heating. Commercial uses PID controllers for heat when available.

 

[KeepItSimpleStupid]

We haven't been helping the OP, so this time, I'm going to try harder.

Look here https://www.watlow.com/downloads/en/manuals/series 982 rev.r 01-12-07.pdf at PDF page 21 and Figure 2.11b. This is what I think you have.
An open collector output. This is basically designed to drive those 0-32 V SSR relays with pin #12 going to the (+) of the inherent LED in the SSR and pin #13 going to the (-) side of the inherent LED in the SSR.

If you would like, you can take a standard LED (50-100 mA max current) and connect between pins #12 and #13 and play. The LED should turn on/off when your significantly above and below the setpoint providing everything else is right.

If this is a standard Thermocouple controller, a short at the TC input basically sets the measured temperature to the room temperature.

So take your pick if you want to use AC or DC to operate your load and select an SSR appropriately that accepts a digital signal between 18-32V. The 3 to 32V SSR's would do that. I'm in favor of using DC.

Other ideas are possible, but lets not go there for now. You would have to build a box but you could eliminate the SSR.

The DC supply offers some advantages. You would be just turning on and off the output on the output side with a DC rated SSR. The meters on the power supply will easily show you voltage and current, so you'll know if there are any other issues. The power supply will also provide soft start, so light bulb operation is still possible. It also would provide current limiting. SO, I'm in favor of the DC supply.

So, plan would be:
1) test controller using a simple LED
2) Use your power supply/heating element to figure out if it is under/oversized. You don;t want to be operating at say 1V at the desired setpoint.
3) Order a DC SSR of the appropriate size with a heatsink and thermal compound. e.g. https://www.newark.com/wakefield-so...p/00Z1245?MER=TSTSO_S_C_HeatSinkCompound_None

Asides:
Yes, you could make something containing an adjustable reference voltage that could be turned on an off by an optoelectronic signal. The reference voltage would, in effect, set the max output of your power supply using the analog programming input. e.g. toggle from say 12 to 0V programming or anywhere in between. This isn't recommended because of your novice level. This would eliminate the SSR.

The correct controller would have had a process output possibly the ability to limit that process output.

Again, sorry I didn't pay more attention to what you needed. I agree, that the light bulb may or may not be a good solution. IR heating which a light bulb can do is a good way of heating stuff.

With IR heating, https://www.aceheattech.com/short-wave-ir-lamp.html , heating is done by radiation, with resistive heating, it's mainly by conduction.

 

[Tony Stewart]

There are ways to solve this. None of which can be verified without design specs.

This is important for proper design.

Please try to define. or describe for others to define
1) mass to be heated ,
2) temperature rise, or setpoints
3) Rth thermal resistance of enclosure
4) allowed Temp error for regulation
5) max response time.
6) sensor location or estimated response time

Output is then defined by available power , temperature slew rate , sensor latency and loop response characteristics then implementation is trivial.

 

[Les Jones]

Well done "KISS" in finding the correct manual for the PID controller. (I had tried by shortening the URL so it finished at manuals/ but it would not allow access to that directory level) Considering the size of the controlled environment (120cm x 40cm x 40cm) and the small temperature difference between the environment and set point (3 Deg, C) (See post 14) I think a small fan blowing air past a few wire wound resistors will be the best heat source. I also think the controlled environment will need to be well insulated to avoid large temperature gradients.

Les.

 

[KeepItSimpleStupid]

LES:

reality suggests something entirely different like a thermoelectric cooler. https://www.tec-microsystems.com/EN/Home.html For the uninitiated. it's a solid state heat pump. These https://www.eicsolutions.com/produc...ioners/200btu-thermoelectric-air-conditioner/ would probably be more appropriate.

I made an ersatz one from a cooling block that we had lying around at work and used it to "cool" a space. This could also be a job for a small environmental chamber. The one I hep spec went from -80 to 200 C. Heating and cooling around room temperature is always a pain. In one case we just use a muffin fan connected to a variac to help control the speed. There was no available money to actually try to "control" the temperature. We had a microscope stage with TEC cooler with water as the sink. Too cold means condensation, so the e-chamber had the low cost (nitrogen purge) rather than humidity control.

The OP might actually do better, with a fixed heat source outside the box and control the fan speed of the air entering the box.

The TEC cooler and/or a water circulator might be worth looking into. All of this goes back to defining the problem like Tony said.

For all we know, is there is little money available and the stuff was lying around the lab and the OP wants to know if it could work.

I've dealt with temperature control from nearly 0 K to the melting point of copper and control around room temperature. One thing I never had the chance to play with was a Vortec cooler.

 

[Les Jones]

Hi KISS,
It sounds like you have far more experience at temperature control than I have. I have not played with Peltier effect devices but I can see it could be an advantage to be able to cool as well as heat just by reversing the polarity with the set point only about 3 Deg. C above ambient. I initially thought it might be possible to use one of the small DC to DC converters that are designed to provide an isolated supply as a means of using the salve input on the DC power supply to switch the output on and off but I don't think the open collector would be capable of driving even a 1 watt version. (I think the output is limited to about 20 mA as the PID unit has an internal 790 ohm current limiting resistor and recommends a minimum load of 500 ohms. With the 27 volts open circuit voltage that zexclo measured the current works out to about 20 mA.) I think zexclo needs to specify the tolerance of the target temperature. It seems on this sort of forum one of the main problems is defining the question accurately.

Les.

 

[KeepItSimpleStupid]

Hi KISS,
It sounds like you have far more experience at temperature control than I have. I have not played with Peltier effect devices but I can see it could be an advantage to be able to cool as well as heat just by reversing the polarity with the set point only about 3 Deg. C above ambient.

I worked in a semiconductor lab. Lots of toys to play with and fix. One of our early Scanning Electron Microscopes was mostly 741 OP amps. The Peltier devices are basically now controlled with an H-bridge and PWM. They are very inefficient. The modules I used modules were like 12 V @ 8A or so. The max delta T is about 60 deg. C.

I initially thought it might be possible to use one of the small DC to DC converters that are designed to provide an isolated supply as a means of using the salve input on the DC power supply to switch the output on and off but I don't think the open collector would be capable of driving even a 1 watt version. (I think the output is limited to about 20 mA as the PID unit has an internal 790 ohm current limiting resistor and recommends a minimum load of 500 ohms. With the 27 volts open circuit voltage that zexclo measured the current works out to about 20 mA.)

Yep. 20 mA is fishy. i.e. Typical LED current

1. But say take a 5V reference IC (Assuming 0-5V control for now, not gonna go back and verify)
2. Then a 10T pot to get 0-5V for programming the supply from 0-100%
3. Now grab an OPTOMOS SPDT relay: one side to the 0-5 and the other to ground. (i.e. Selects the 0-5V signal or ground)
4. Buffer and current limit the output
5. The LED in the OPTOMOS part would function the same way as a SSR.
6. Powered from a mains supply or maybe a linear wall wart.
7. Still details to work out.

I think zexclo needs to specify the tolerance of the target temperature. It seems on this sort of forum one of the main problems is defining the question accurately.

Controlling close to ambient is tough, honestly it's anyone's guess if the environment is for fruit flies or a laser diode

 

[Les Jones]

I was trying to think of a way to avoid finding a floating supply for the control voltage to the DC power supply. The wall wart would solve that problem and is cheap. The control voltage is 0 to 10 volts and the positive is connected to the positive output of the DC power supply and the negative to the "PAR" input of the DC power supply. See page 16 of the manual. This is the URL to download the manual
**broken link removed**
I think just using a potentiometer on the wall wart output would be good enough for setting the output voltage.
I think your above solution should sort out the electrical design.
Edit I have just noticed your comment that the Peltier device requires 8 amps. The DC power supply is only rated at 3 amps.
Next to the thermal design.

Les.

 

[Tony Stewart]

Hi,

I would like to seek some assistance on connecting my PID controller to my DC power supply.

I am currently making a temperature regulator using

1. WATLOW controller model ( 982C-25CC-KRGG)
2. Lightbulbs (Philip Halogen 50W, 12V lightbulb)
3. GW DC power supply (Model GPS-3030)

I am encountering some difficulties connecting the controller to the DC power supply.
From my learning,I understand that the controller is a PID controller, which sends pulses to the DC supply which in turn will power up the light bulb should the temperature drops what is needed to maintain.

However, I am having difficulty understanding and connecting the PID controller. As to should the setting of the power supply be set to master output? and how should the circuitry be connected.

For any kind assistance!

a) GPS-3030 is not pulse controlled but linear control 0~10V for 0~100% of Vmax for the supply.

BUT if using a halogen light bulb power out is not linear. with 5:1 R ratio for hot:cold at fullff power
As V/R rises almost the same with temperature it is a crude constant current device for 0~30% V and thus linear power with V, but above 70% Resistance rises near max and thus P=V^2/R and interpolates in between.​

b) What temp sensor? linear digital?
c) 982C-25CC-KRGG controller must use 0~10V output for power control

All other physics variable I mentioned are still more important to realize low setpoint error.
Is there a Device being tested with self heating involved?

​

 

[zexclo]

First off, thanks for the exchange and sharing of information.

Hey Kiss, thanks for the link and that is definitely lots of information, thanks! Going to take some time for me to read and digest the informations and read and understand all the replies here.

Really sorry for my slow response, wasn’t able to post a reply due to the equipment was not with me earlier.

====

Hi Les,

In response to post # 21 I was measuring the output from the model 982C-25CC-KRGG earlier and the reading I got 27V, however I think I seemed to be have blown the fuse. Give a moment as I re-check the readings and to see if I have fused it.

====
Hi Tony,

In response to post#22 I am currently making an incubator for some running tubes. This temperature regulator will be used to keep the air in the tube to be 25 degree Celsius, as the surrounding outside(environment is around the box is about 22 degree Celsius). My project supervisor wants to use the PID in the setup, as the results are used for research purposes.
Hence, the heating mass is generally air in the box.

I am considering using a wire wound resistor of 50-100w to help heating up the air and the circulation of the air will be done by a CPU fan placed beside it and another place near the top. A thermocouple will be placed at the other top end of the box to measure the temperature of the air in the box.

====

Still digesting informations till post#28.
Give me a moment to read, and repl but keep the exchange going!
Thanks all and really appreciate the help!

Cheers,
Zexclo

 

[zexclo]

Hey Kiss,


Highly doubt I would be making an adjustable reference voltage that could be turn on and off by an optoelectronic signal at this point of time. That would probably be an improvement from this project when I am better with electronics.


The idea of fix heat source outside the box and controlling of the fan speed of the air entering the box sound cool, but it does seems more complicated. Might explore that idea after I manage to get this system set up.


Yea, I am running quite tight on budget with little money to spare.

====


Hi Les,


Have checked, the PID is working and pulsing at appr. 27 VDC from pin #12 and #13 when connected to a multi-meter.


In response to post on #33, I aim to keep the temperature tolerance to app 1 degrees. Do feel free to ask me questions on my setup and tolerances/information that is needed, and I will answer to my fullest (and very limited yet growing) knowledge!

=====


Hey Tony,


In response to the post on #30


1. The mass to be heat is air measuring 120cm x 40cm x 40cm.


2. Temperature rise should be from 22 to 25 degree Celsius. The set temperature will be 25 degree and heating will occur when temperature drops below 25 degree.


3. The enclosure (the box containing the 120cm x 40cm x 40cm) is made of acrylic thickness of 0.5cm.


4. By allowed temperature error for regulation, I believe you meant by time for the temperature to reach set temperature of 25 degree Celsius. For now, as I am quite new to this circuitry system, I aim to keep it under 30minutes and hopefully later under 10minutes when I am better.


5. By max response time, I inferred that the response of the PID upon detection of temperature difference and action is taken?


6. Sensor will be located at the other side, top end of the box. The heater near one end where the fan will be blowing out the heat, and the sensor will be at the other end, top corner.

----

In response to #36

b. am currently using a thermocouple attached to the PID as the temperature sensor


Do let me know if you need other design information and also feel free to voice out your recommendations. Cheers.


=====


I noticed that there is a mention of wall wart as alternative to the DC SSR setup.


Do correct me if I understand it wrongly,


1. A wall wart with volt control of 0 to 10v will be used.

2. The positive end of the wall wart is connected to the positive output of the DC power supply and the negative end of the wall wart is connected to the “PAR” input of the DC power supply (GPS-3030)

How do I connect the potentiometer and a 75watt wire wound resistor to this setup?

on the other hand
If I use a DC SSR, how do i connect it to the system with the PID and 75 watt wire wound resistor?

 

[KeepItSimpleStupid]

In response to post#22 I am currently making an incubator for some running tubes. This temperature regulator will be used to keep the air in the tube to be 25 degree Celsius, as the surrounding outside(environment is around the box is about 22 degree Celsius). My project supervisor wants to use the PID in the setup, as the results are used for research purposes.
Hence, the heating mass is generally air in the box.

Very Confused! But, maybe it doesn't matter. The boss is the boss. It's HIS problem.

Rule #1. The boss is right
Rule #2. If the boss is wrong, see rule #1.

"running tubes" = worms? I have no idea what a running tube is,

The air in the tubes is supposed to be 25 deg, yet your heating the air in the box.

==

Remember the LED test I talked about? You initially just need to measure the voltage at the pin with the resistor to the emitter of the transistor.

 

[zexclo]

Yeah, boss is boss, his word is law.

Oh,by running tubes I meant, tubes from one end of the box to the other end. The tubes are small carrying liquid.
ie
Imagine you are in a washroom in an apartment, there are pipes running from the top of the ceiling to the floor.

==

Yup, will be getting a LED to conduct the test probably later in the day, as currently there is no LEDs around.