Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

How to prevent floating input for this on-off control circuit driving an SSR?

Status
Not open for further replies.

doncarlosalbatros

Full Member level 2
Joined
Nov 26, 2016
Messages
143
Helped
1
Reputation
2
Reaction score
1
Trophy points
18
Activity points
1,758
I was doing on-off temperature control for around 40‎°C for an enclosed big enclosure manually by switching on and off a 1100W heater(with a fan) with my hand. The accuracy is not very important roughly 40‎°C is fine. Now instead me doing this, I want to automate this on-off process. Basically 40‎°C corresponds to 8V output of the temperature transducer.

Here below is raw voltage output of the temperature transducer when it is heating and cooling around 40‎°C(sampled at 4kz):

Capture.PNG

Here is a close view of the raw data:

regul3.png

Now I tried the following comparator circuit(with LM2903) where I set the threshold to 8V with a hysteresis of around 35mV, and fed the 4kHz above sampled input to the circuit as PWL.

Here the comparator triggers **broken link removed**.

jHPn5.png

Below shows the input as the sampled voltage from the transducer, the non-inverting input which shows the hysteresis level, and the heater current in order.

MXE6a.png

Here is my question:

Here is the datasheet for the transducer.
Now this sensor is set to 0-10V output and I wondering how would floating input turns off heater for this circuit. Is connecting 560k or 1Meg to Vcc load the transducer or damage it? Transducer output can be anything between 0 to 10V.


If there's no fundamental problem with this circuit, I just need to control something better than or same as a human doing. Normally when I see the temperature exceeds 40‎°C I turn off off the heater with my hands and when lower than 40‎°C I turn on on heater. I act like a comparator and SSR. So I need to automatize it instead of staying there for one hour. And it is fine for my purpose. Then I turn off the heater after making a measurement .

How can I obtain heater off when the input is floating? I'm not sure if the R5 connected to Vcc is a good idea. Because if the input floats the heater will run forever.
 

I do not think the 3102 you called out is isolated.
I think the return for the current output and the return for the voltage output (both pin 6) are both tied to the 24V supply return (pin 8)

the 3112 is isolated and the documentation shows that

if you adjust a few resistor values, you could run the whole thing on one 24V supply
and is you use two supplies, 24V for the sensor and 12 V for the rest, the only common connection is the ground.
 

I do not think the 3102 you called out is isolated.
I think the return for the current output and the return for the voltage output (both pin 6) are both tied to the 24V supply return (pin 8)

the 3112 is isolated and the documentation shows that

if you adjust a few resistor values, you could run the whole thing on one 24V supply
and is you use two supplies, 24V for the sensor and 12 V for the rest, the only common connection is the ground.

I think you misunderstand what I'm asking.

My worry is if the temperature input is unplugged the heater will be on forever.
So I need the comparator's inverting input to go high(which turns off the heater) when the temperature input is unplugged.

I don't want to use the same power supply for practical reasons.
 

Hi,

Reading the transducer datasheet I see a couple of relevant values:
* supply voltage: 16.8V ... 31.2V
* load (@ voltage output) >= 10 k Ohms
* open output: < 18V

--> So from this ...I see no problem for the transducer. You are within specifications on all three specifications.
But for accuracy reasons I'd avoid the 100 Ohms series resistor (at least at the position you use it)

But that's only half (if ever) of the truth!
There are a couple of things to consider.
First of all: I see no reason why the output of the transducer should be high impedance.
Please tell us why you consider this situation at all.

If you consider to disconnect the wires, then:
* Don't do this when the whole circuit is powered,
* if you do this with the circuit powered, then you need additionl protection circuitry. For both the transducer and the comparator circuit.
* the "pull up voltage" (12V in your case) is not relevant, because the transducer is not connected...

The biggest problem I see:
Is the comparator. It is supplied with 12V ... and your pullup is to 12V, too.
But the comparator's datasheet tells "common mode input voltage" is up to "VCC - 1.5V". Now your circuit violates this specification. In this case the comparator output state is unknown ... and this is exactly what you want to avoid.

Thus a more valid solution is to use two resistors: maybe use
* 200k as pullup (to VCC = 12V)
* and a 1M to GND.
With this the resulting "unconnected" node voltage will be 10V.

*****
Here a couple of additional hints.
* use the series resistor after the pullup/pulldown resistor. And use 10k.
* use two protection diodes. Maybe 1N4148 each one in parallel of the pullup and pulldown respectively.
* use a (foil) capacitor as LPF noise filter in the range of 1uF ... 10uF directly at the comparator input
* use a big bulk capacitor to stabilize the 12V supply. (Any weak supply may cause oscillations, because the relay current will drop the supply voltage...and thus the voltage at -IN, which causes negative feedback). Hopefully your hysteresis is higher than the negative feedback.

Klaus

Btw: I'd use only one power supply. This avoids power sequencing problems. But needs special care on circuit design.
 

Hi,

Reading the transducer datasheet I see a couple of relevant values:
* supply voltage: 16.8V ... 31.2V
* load (@ voltage output) >= 10 k Ohms
* open output: < 18V

--> So from this ...I see no problem for the transducer. You are within specifications on all three specifications.
But for accuracy reasons I'd avoid the 100 Ohms series resistor (at least at the position you use it)

But that's only half (if ever) of the truth!
There are a couple of things to consider.
First of all: I see no reason why the output of the transducer should be high impedance.
Please tell us why you consider this situation at all.

If you consider to disconnect the wires, then:
* Don't do this when the whole circuit is powered,
* if you do this with the circuit powered, then you need additionl protection circuitry. For both the transducer and the comparator circuit.
* the "pull up voltage" (12V in your case) is not relevant, because the transducer is not connected...

The biggest problem I see:
Is the comparator. It is supplied with 12V ... and your pullup is to 12V, too.
But the comparator's datasheet tells "common mode input voltage" is up to "VCC - 1.5V". Now your circuit violates this specification. In this case the comparator output state is unknown ... and this is exactly what you want to avoid.

Thus a more valid solution is to use two resistors: maybe use
* 200k as pullup (to VCC = 12V)
* and a 1M to GND.
With this the resulting "unconnected" node voltage will be 10V.

*****
Here a couple of additional hints.
* use the series resistor after the pullup/pulldown resistor. And use 10k.
* use two protection diodes. Maybe 1N4148 each one in parallel of the pullup and pulldown respectively.
* use a (foil) capacitor as LPF noise filter in the range of 1uF ... 10uF directly at the comparator input
* use a big bulk capacitor to stabilize the 12V supply. (Any weak supply may cause oscillations, because the relay current will drop the supply voltage...and thus the voltage at -IN, which causes negative feedback). Hopefully your hysteresis is higher than the negative feedback.

Klaus

Btw: I'd use only one power supply. This avoids power sequencing problems. But needs special care on circuit design.


Hello and thanks for the answer,

For practical reasons I will use separate power supplies, 24VDC for transducer and 12VDC for the comparator circuit.

I want to go through your suggestions and try to clarify what worries me.

First here below is the circuit I modified after your suggestions:
ccctest.png

* The 100 Ohm Rout in my first circuit was to mimic the output impedance of the transducer which is not given. I estimated it but since not given now I deleted it in the new circuit. If you can refer the output impedance of the transducer from the datasheet let me know. I only needed it for simulations.

* Now following your suggestions;

I removed 100 Ohm Rout.(Transducer output resistance is unknown to me)
I added 10u and 100n capacitors for Vcc.
Added pull up and pull down R5 and R4 as 200k and 1Meg respectively to lower the input voltage not to violate the "VCC - 1.5V" spec.
Added two parallel diodes 1N4148. I added R8 as series 10k and C2 to form a LP capacitor.

[
***What do you think about the circuit now?(Hope I modified it accordingly with your suggestions)
[

***But my main worry is when when the transducer output is seeing 10V applied to its output port. Transducer might be 0V up to 10v. In this case it will be like we are connecting a 9V battery to this transducer output through a 200k. So we are applying voltage to its output port where the transducer will sink around 100u. Now this reverse current will always exist. I hope I made where Im stuck better.
This confuses me because it is like connecting 9V battery to the output terminals of a function generator. An output port outputs voltage and sources current. How come we can apply voltage to an output port. So in this case the 10V at the node of R4/R5 is continuously applied to the transducer output port. Isnt this a problem?

Thanks!
 

I don't yet understand why disconnected transducer is your "favorite" failure mechanism. May be the reason is in the specific wiring of your apparatus.

In any case I see the problem that you are focusing on one of several possible fault mechanisms that might cause uncontrolled heater operation and arbitrarily disregard others, e.g. transducer power supply failure, RTD cable short. Also defective comparator circuit and shorted SSR should be considered for a fail-safe heater.

At the end of the day, an independent latching mechanical overtemperature switch would be the best solution.

I have a minor comment regarding LM2903 common mode range. By device design, there's no gain reversal as long one of the inputs stays within the common range. That's guaranteed in your circuit by noninverting input bias condition, don't need to keep the inverting input within common mode range.
 

Hi,

***But my main worry is when when the transducer output is seeing 10V applied to its output port. Transducer might be 0V up to 10v. In this case it will be like we are connecting a 9V battery to this transducer output through a 200k. So we are applying voltage to its output port where the transducer will sink around 100u. Now this reverse current will always exist. I hope I made where Im stuck better.
This confuses me because it is like connecting 9V battery to the output terminals of a function generator. An output port outputs voltage and sources current. How come we can apply voltage to an output port. So in this case the 10V at the node of R4/R5 is continuously applied to the transducer output port. Isnt this a problem?

A good transducer output shoukd be push-pull. It can handle currents in both directions.
As long as we are within specifications...these are: "load resistance" and "max applied voltage" ...we are safe.

And a good transducer should be feedbacked. ..in a way that any output current that tries to pull the output voltage higher or lower should be regulated out. This results in very low output impedance (up to a frequency limit and current limit). Maybe the output impedance is less than 1 Ohms. At least it needs to be low enough to keep specified precision within specified load resistance range.

Klaus
 

Hi,



A good transducer output shoukd be push-pull. It can handle currents in both directions.
As long as we are within specifications...these are: "load resistance" and "max applied voltage" ...we are safe.

And a good transducer should be feedbacked. ..in a way that any output current that tries to pull the output voltage higher or lower should be regulated out. This results in very low output impedance (up to a frequency limit and current limit). Maybe the output impedance is less than 1 Ohms. At least it needs to be low enough to keep specified precision within specified load resistance range.

Klaus

Hi,

You wrote "A good transducer output should be push-pull. It can handle currents in both directions." I really never heard of it. Interesting. I thought a transducer can/should only output current one direction.

Thanks for this, I made the circuit but now I will also add the pull up and down resistors.

One more thing I think in my circuit diodes should be after R8 (10k) and before the cap correct?
I draw it wrong I guess..
https://www.edaboard.com/attachment.php?attachmentid=153476&d=1559641247
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top