LM35 sensor and direct heat

Discussion in 'The Projects Forum' started by telerian, Jan 19, 2011.

  1. telerian

    Thread Starter Active Member

    Jul 19, 2010
    57
    0
    I am using the circuit on this site http://wingzero.ath.cx/index.php?page=electronic-thermostat and I have 2 questions

    1. The relay keeps switching on/off fast at the point when the preset temperature is reached until it stabilizes
    2. I am using to control a heating coil (60W) which inserted inside an aluminum tube, is it ok to fix the sensor directly on the tube?
     
  2. soda

    Active Member

    Dec 7, 2008
    174
    13
    The LM35 temperature range goes from -55 to 150c so first test the temp of the tube
     
  3. telerian

    Thread Starter Active Member

    Jul 19, 2010
    57
    0
    I do not have the necessary equipment to test high temp. and I need to keep the tube temp. to 30-35C so theoretically if I fix the sensor on the tube with the circuit controlling the heating element it shouldn't go above the 30-35C. right ?
     
  4. tracecom

    AAC Fanatic!

    Apr 16, 2010
    3,869
    1,393
    1. It sounds like the hysteresis is set too high. First make sure the 10M resistor is correctly connected to the LM311. If it is correct, then try reducing the value to about 1M.

    2. If the temperature of the tube does not go above the maximum of the LM35, it will be okay, but if there is a malfunction, the LM35 could be destroyed. In addition, if you are trying to control the temperature of a medium (air or water for example) your LM35 should be exposed to the medium, not the heater. After all, it's the temperature of the medium that is to be controlled, not the temperature of the heating element.
     
  5. telerian

    Thread Starter Active Member

    Jul 19, 2010
    57
    0
    I will try a smaller resistor and see how it goes (or maybe use a Pot)

    I agree with you about the medium but I want to make sure the pipe does not get very hot and burn someone's hand
     
  6. bertus

    Administrator

    Apr 5, 2008
    15,648
    2,347
    Hello,

    As tracecom already told you try a lower value for the 10M resistor to make the delta temperature switching larger.
    Also the capacitor just afther the rectifier bridge is very small.
    Also put a o,1 uF capacitor at the entrance and the output of the stabelizer chip.

    [​IMG]

    Bertus
     
  7. wayneh

    Expert

    Sep 9, 2010
    12,127
    3,048
    I have fought with hysteresis on a temperature controller. The challenge is to get a narrow hysteresis band - tight control at your setpoint - but a solid switching without chatter. Those are competing goals.

    I was using a LM339 comparator and the 10M resistor recommended in the datasheet for hysteresis is the extreme. It CAN work, but I got much better switching (less chatter) by using a lower resistance (3.3M, I think) to increase the hysteresis.

    In my case the deadband was too large though, so I used an op-amp with a gain of ~20 on the LM35 output. That meant that hysteresis of say 5mV at the comparator gave a deadband of 0.5°C without the op-amp, and only 0.025°C with the op-amp.

    The LM35 is pretty robust and can tolerate a surprisingly high temp. Take a look at the datasheet. I think you'll be fine attaching it to your tube. But don't confuse the temp. of your heater tube with the temp. of whatever you're heating.

    PS: I think the resistor to ground on the LM35 output is too low. The datasheet recommends 220Ω.
     
  8. telerian

    Thread Starter Active Member

    Jul 19, 2010
    57
    0
    Do you mean high ?, it is 4.7K
     
  9. wayneh

    Expert

    Sep 9, 2010
    12,127
    3,048
    Yes, I keep interchanging load and resistance which are inverses.
     
  10. tracecom

    AAC Fanatic!

    Apr 16, 2010
    3,869
    1,393
    Many DMM's have a temperature measurement function that uses a thermocouple wire. My cheap one has it, but I don't use it very often and sometimes forget that it's there. Maybe yours does? And yes, the heating element may go much higher. In a similar situation, I have seen the surface temp of the heating elements at almost 100 degrees C even with intermittent duty.
     
    Last edited: Jan 20, 2011
  11. telerian

    Thread Starter Active Member

    Jul 19, 2010
    57
    0
    Mine doesn't have temp. sensor. How could the element surface go high if it is controlled, maybe with a high power one which goes hot fast but mine is only 60W coil of soldering iron.
     
  12. tracecom

    AAC Fanatic!

    Apr 16, 2010
    3,869
    1,393
    If the thermostat does its job, then you are correct. But if the thermostat were to fail (for example if the relay were to stick closed), then the heating element could get very hot - hot enough to melt solder - maybe 450 degrees C.

    You probably already know that you can easily use your DMM to read the millivolt output from the LM35 and know its temperature. For example, if the output from the LM35 reads 355 millivolts, the temperature is 35.5 degrees C.
     
  13. telerian

    Thread Starter Active Member

    Jul 19, 2010
    57
    0
    I am not sure I understand what you mean by the stabilizer chip and where I should add a capacitor
     
  14. telerian

    Thread Starter Active Member

    Jul 19, 2010
    57
    0
    Yes, I know about the voltage reading and temp. 10 mv for 1C. How do you suggest I put a fail safe system for the heating ?
     
  15. tracecom

    AAC Fanatic!

    Apr 16, 2010
    3,869
    1,393
    Bertus means on the input and output of the LM78L05; he is referring to the LM78L05 as a stabilizer.
     
    Last edited: Jan 20, 2011
  16. tracecom

    AAC Fanatic!

    Apr 16, 2010
    3,869
    1,393
    Probably with a thermal fuse in series with the power lead to the heater. The thermal fuse should be selected so that it opens and interrupts the power to the heating element at a safe temperature. Of course the thermal fuse must be located physically close to the heating element.

    http://en.wikipedia.org/wiki/Thermal_cutoff

    Such devices are often used in hair dryers and curling irons to prevent burns.
     
  17. telerian

    Thread Starter Active Member

    Jul 19, 2010
    57
    0
    I don't think so because there is a 10uF capacitor there and he asked me to replace one of them with a 220uF
     
  18. tracecom

    AAC Fanatic!

    Apr 16, 2010
    3,869
    1,393
    Normally, the LM series of voltage regulators need a small ceramic cap on the input and output to avoid oscillation of the IC. These caps are in addition to larger electrolytics that are used as filters. My practice based on the datasheet for the LM regulators and on the advise of other forum members with lots of experience is to put a .33uf on the input and a .1uf on the output.

    But Bertus will be along shortly and answer for himself. In fact, I shouldn't have interfered. ;)
     
  19. bertus

    Administrator

    Apr 5, 2008
    15,648
    2,347
    Hello,

    To get back at the capacitors at the regulator (LM78XX).
    Tracecom is right about preventing oscillations.

    [​IMG]

    Also it is good to have decoupling capacitors on the other active chips:
    Decoupling or Bypass Capacitors, Why?

    Bertus
     
  20. telerian

    Thread Starter Active Member

    Jul 19, 2010
    57
    0
    So to make sure I understand you correctly, I should put those capacitors in parallel with the existing 10uF ?
     
Loading...