ESP Fan Controller to drive a relay

[BobEllis]

I am trying to use Rod's fan controller circuit Thermo-Fan to keep your amplifier cool to drive a relay. The ultimate application is a fridge controller to keep it a cheese making temperatures. I'm having an issue with the relay closing and opening slowly.

It is a parts bin build as shown in Figure 2 with the following exceptions:

U1 is half an OPA2134, the other half has both inputs tied to ground.
D2 is a 9.1V zener.
Q1 is BC556C
Q2 is MJE15030
The fan is a 12V 160 ohm relay that draws 75 mA.
R8 is 221R
R9 is 27R
Temperature sensor diode is a 1N4007 wrapped in heat shrink
Power is a wall wart delivering 15.9V with just the circuit, dropping to 14.5V with the relay powered.

Relay closing with temperature change took almost a second of buzzing, opening a bit longer. In an attempt to make switching faster, I reduced R7 to 1K. Some improvement.

I eliminated the feedback resistor, R4 to make it a comparator hoping to make the circuit switch faster. It did, but the relay still buzzes.

The voltage across the zener drops from 9.5 to 9.3V when the relay is powered. although the voltage across R8 indicates 27 and 22 mA relay off/on, so the opamp circuit seems to be working properly. The OPA2134 should draw about 10 mA, another couple mA for the rest of the regulated circuit, leaving at least 5 mA for the zener. I tried reducing R8 to 150R, with no change in operation.

I haven't broken out my scope yet, but the OPA output seems to swing properly, albeit slowly, but that may be the response time of my bargain DVM. Off it sits high at .6V below the rail, Q1 base just a few mV below the rail. On, opamp output is .1V above ground, Q1 base .6V below the rail, as expected.

Does anyone have any ideas to make the relay switch on and off faster? Would 100R emitter resistor for Q1 help shutoff?

Thanks for your help.

 

[will_raymo2000]

Not totally answering your question but this seems a simple circuit: Simple Temperature Controller

Possibly worth trying if you have no luck with the present circuit.

 

[BobEllis]

Thanks, that's basically the same circuit that I have now by removing R4 - a comparator, although the output drive circuit is different and I don't have thermistors other than CL30s in my parts bin.

 

[Johno]

The comparitor has no hysteresis and will amplify noise on the input when near the switching threshold which will swing the output rail to rail and make the relay chatter.

Give it some positive feedback (not negative feedback) and add Rin for the +ve input to be the feedback voltage divider.

 

[djQUAN]

Rod has another project suitable for that.

Project 131

He mentions replacing the LDR with a thermistor if the application would be to sense temp rather than light.

 

[BobEllis]

Thanks.

From Project 131

Hysteresis is controlled by R4, and is essential to prevent relay chatter when the light or temperature is right at the reference level. The amount needed depends on your application, so adjust R4 )or select a likely candidate from the table below) until there is no relay chatter even if the input voltage is right at the trigger level. The value shown (1 Meg) is the bare minimum - most applications will use a lower value. For probably 99% of all uses, R4 will be between 100k and 1Meg.

That's odd - I got less chatter removing R4 from the circuit, but did notice a difference in the trigger point. Guess I'll try a few different values.

 

[BobEllis]

I guess I did something else at the same time as I clipped R4. I'm in one of the oddball applications. with R4=100K I got three or four vibrations, but R4=68K it snaps on and snaps off perfectly.

Thanks for pointing me to the project 131. Now to noodle on how R4 creates the hysteresis.

 

[djQUAN]

No problem.

The lower resistance produces higher amount of positive feedback therefore a larger hysteresis.

 

[BobEllis]

Oddly, Project 42's R4 provides negative feedback. I don't get it, but the circuit works.

 

[djQUAN]

in P42, R4 provides negative feedback (it is connected to the inverting input) so that the output varies linearly with regards to the input.

in P131, Positive feedback (connected to the non inverting input which results to hysteresis) allows the output to stay at only two states... ON and OFF.

 

[BobEllis]

Right. Which is why I am confused - lower R4 in P42 means lower gain. Without R4, the opamp is acting like a comparator the output level changes slowly enough to cause chatter. With low gain, it changes quickly enough not to chatter.

 

[Johno]

The rate of change of the op amp/comparitor is not in question. Its the rate of change of the input signal that the op amp is reacting to.

With no hysteresis, noise on the input (which is unavoidable) swings the input either side of threshold and causes the output to swing.

With hysteresis, the threshold is moved by the positive feedback so that an input swing larger than the noise level is needed to return the comparitor to the off state.

You could design the hysteresis to require a temperature decrease of 10'C for example, before the fan turns off.

 

[BobEllis]

I see how positive feedback makes the comparator have only two states, high and low. In P42, applying negative feedback kept the relay from chattering. Without no feedback watching the armature of the relay I could watch it slowly disengage with some chatter on first the NO contact and then the NC contact as the diode cooled. Increasing negative feedback resulted in snap action one would expect with hysteresis.

 

[djQUAN]

ah I see what you mean.

Relays have built in hysteresis. Try powering one off an adjustable supply and slowly adjust the voltage. At some point, the contact will close and stay closed. To open the contacts, you have to decrease the voltage to a much lower value hence relays have 'pull in' current and 'holding' current.

When you configured P42 as a comparator, The output of the op amp changes state unpredictably due to noise causing chatter.