Thermistor Circuit Difficulties

Thread Starter

Deve

Joined Dec 28, 2015
95
Hello. This is my first post but not exactly my first rodeo. I have designed a circuit that I want to expand upon, but I am baffled as to why it is working backwards. The comparator works fine, the LED comes on as the temperature rises, I can set the temperature via the potentiometer just fine. Life is good so far. But when I try to add a relay so I can fire the Piezo, I cannot understand why the relay coils are energized when the LED is OFF and de-energized when the LED is ON. I want that part to be the other way around. I am sure it is something simple, but I have been playing with this circuit for days and cannot seem to grasp why it is backwards. Also, the LED is very dim but on a little when it is supposed to be dead off. I am always appreciative of direction and advice, so please do not hesitate to tell me what ELSE this circuit needs to be solid. Thank you!
 

Attachments

Kermit2

Joined Feb 5, 2010
4,162
A quick look and I see that you need a low on pin 1 to light your LED. You are using an NPN transistor which requires a high on its base to operate. This is causing the relay to operate when the LED is off.
Simplest fix is to change which contacts you use with relay off and on.
 

bertus

Joined Apr 5, 2008
22,270
Hello,

You have the back-emf diode connected on the wrong place.
The transistor could already have been fried.
Also the remark of Kermit must be taken in account.

deve_watertemp_detector.png

Bertus
 

Thread Starter

Deve

Joined Dec 28, 2015
95
Wow. Thanks! Umm. I drew the schematic wrong.. thanks for that correction. That IS how the circuit is really done. Kermit.. what do you mean by switch contacts? I have PNP transistors, I just do not know enough how to hook it up so it works correctly. I really do not want power to the relay for no good reason, otherwise, just going to NC instead of NO on the relay would work. Great resource!
 

Thread Starter

Deve

Joined Dec 28, 2015
95
Ok, thanks to you guys, I have it working. To be honest, I was thinking PNP but just couldn't connect the dots. Thanks for the great resource. If anyone has comments on how to improve this circuit, please say. I want to build these for others and would like the circuit as solid as possible. I have no capacitors in the circuit which seems strange and the NTC Thermistor being 100K wasn't sure what values to use for the voltage divider so that was my best guess. Its for a critical systems warning circuit for vintage cars/trucks. Thanks so much for the help. Brain Freeze can last awhile with me!
 

MikeML

Joined Oct 2, 2009
5,444
Why do you need a 12V relay to switch 12V to a 12V Piezo? Why not just switch the Piezo directly with the collector of the NPN?
 

Thread Starter

Deve

Joined Dec 28, 2015
95
Its a PNP, and there isnt enough current at the LED to fire the relay. There may be to fire the Piezo. Haven't tried that. You will find many holes in my knowledge. The idea of firing the Piezo direct off the transistor and skipping the relay makes sense too. The fewer parts and a more stable circuit is what I am after no matter how long it takes. All of the similar solutions I have seen, ordered in, tested, etc use a relay so I have relay stuck in my head.
 

Alec_t

Joined Sep 17, 2013
14,280
The 393 has open-collector outputs, so will need a pull-up resistor to replace the LED+ resistor if those are moved to be in parallel with the relay coil.
 

R!f@@

Joined Apr 2, 2009
9,918
Is it a 12V Piezo ?

What is the part number or show a picture of it.
If it is a small one, remover relay and put the Piezo where the relay was. Keep the diode in place, just swap the relay with the Piezo
 

Kermit2

Joined Feb 5, 2010
4,162
Or take the led from pins 8 and 1 and connect from 1 to ground.
Take the thermistor and the 100k resistor and swap their positions in the divider. That should get your output voltage states swapped so the NPN gets turned on. That 10k resistor will be to large I think. Drop to 1k and see
EDIT
Strike all that. Is the LED coming on when you want?
Is the relay supposed to operate at the same time?
Your led will come on when the voltage at the divider goes higher than your reference. Since the divider is hooked to a neg. Input the output at 1 will go low.
Swap the input pins between the reference and the divider. You still have unused comparator pins to drive led
 
Last edited:

Kermit2

Joined Feb 5, 2010
4,162
Look at the middle letter of a device to see how it operates.
NPN needs the base to be more POSITIVE than the emitter.
Not Pointing N. Describes the arrow direction of the schematic symbol.

For PNP just reverse everything.
POINT N PLEASE describes the schematic symbol.
 

Thread Starter

Deve

Joined Dec 28, 2015
95
Here is the updated schematic. I want to be able to market this, so any help in refining the circuit.. I just do not want an unstable circuit out there. The other half of the LM393 will be used for a Low Fuel Warning using the existing fuel sending unit (approx 0-100 ohms) but that is another issue. The nice part of how this has transpired so far is that I CAN use a relay if the circuit evolves further.

Kermit, I thought of swapping pins 2 and 3 of the Comparator, but then the LED goes OFF when the temperature point is reached (via the pot), so I decided to go with what you suggested and use a PNP. Works flawlessly now. LED comes on, Relay Engages. Perfect!

So I run this website on vintage Chevy Trucks and the whole point of the website is to help others figure out the nuances of Restoration. I design many things and have a policy of sharing everything for free, but if someone wants me to build something for them, I will do it for a fee. The instructions are always posted first in complete detail. So, this circuit will be out there and used alot over time. Thus the need for getting this right. I came to the right place! THANKS and if you see a need to change something, please speak out. Here is where this project is on the web site and the schematic attached below.

http://devestechnet.com/Home/WarningSystem
 

Attachments

marcf

Joined Dec 29, 2014
288
Just a thought, but how critical is this circuit, and I assume that the temp sensing (100k@25C NTC thermistor) is somewhat remotely located from the rest of the circuit. You could use the unused LM393 to supervise the connection. If the circuit goes open, you will never know it.
 

Thread Starter

Deve

Joined Dec 28, 2015
95
I hate to ask because I have already been helped so much here, but how would that circuit for supervising it look? The Thermistor is put inside of a tin housing and bolted to the thermostat housing on an old engine. The wire lengths are about 4 feet. I have tested my circuit already with this length. Supervising is a good idea. I would want to use the same piezo.
 

marcf

Joined Dec 29, 2014
288
Hello.

I am not sure of the PN for you thermistor, but I suspect it might be something like this:

http://www.vishay.com/thermistors/list/product-29049/

The device has a negative temp constant, and the 100k @25C part has a resistance of 340k at 0C. Open would be infinite.

By using the voltage divider equation, you should be able to calculate what the voltage should be at the junction of the Thermistor and the fixed 100k resistor. From there just set the voltage on pin6, again by using the voltage divider equation. (R1/(R1+R2)) * 12V. Of course you do not need to use 0C as a 'trouble' point.

I would also reference the Thermistor on the ground side rather than on the +12V side. It would eliminate the chance of having a hot wire land on something it should not.

I would also use NPN resistors. They would be 'off' when the comparator is low and 'on' when the comparator is high. (They would turn on via the pull up resistor).

I probably have a bias against PNP transistors as they are 'up side down', and I had to go back and learn transistors after I learned about tubes. I mean every one knew that the plate (collector) had to be tied positive with respect to the cathode (emitter). After all electrons come out of the ground don't they?
 

Thread Starter

Deve

Joined Dec 28, 2015
95
Thanks for the tips. I do not like PNP either. It's like right brain vs left brain. Totally unnatural. But, I need the relay to turn on when the output is low, so I couldn't figure out any other way. Like I said, lots of holes in my knowledge but doing the very best I can do. I really want to use the other side of the comparator for a low fuel warning, taking the sending units output to control the low fuel warning light. Not sure yet, LM393s are cheap. It IS a good idea to move the Thermistor to the ground side. Since the values are the same it won't change anything else.
 

marcf

Joined Dec 29, 2014
288
Take a look at this. This circuit is from the 'top' of my head and has never been actually bread boarded.

The term 'NO' and 'NC' for relays means (to me any way) that the 'NO' contacts will make when the relay is energized and the 'Normally Closed' contacts make when the relay is not energized. Another problem you may see with a relay rated at 10 amps is that some relays have a minimum current that they will work at. Trying to use a 10Amp rated contact to switch 10mA may not work as expected.

The spec sheet says that the minimum current gain for the 2n4401 is 100 if the collector current is 150mA, so 2ma of base current should result in 200ma of collector current. This should be more than enough to drive your 10A relay.

Ignore "Input Pin-2 Non Inverting" over the really contacts. I should have deleted it.
 

Attachments

Top