Huge current affecting the output of sensor

Hello,

I'm working on a water heater circuit.

I use PIC10f322 as microcontroller and LM35CH(metal can package) as temperature sensor.

My heating coil pulls around 9A and in the future we might need to use 2 of this coil.

Everything works without any problem till coil starts working. When it starts working, the output of 7805 goes up to 5.67V which is normally is 5.04V. Output of LM35 goes up to around 800-1000mA which is normally 250-350mA (that means 25-35 degrees).

What do you think cause the boosting up the voltages? How can i solve this issue?

You can find my schematic and layout below;

Schematic https://imgur.com/80znoKY

Layout top https://imgur.com/JpB9jeM
Left polygon is GND, right top polygon is +12V from power supply, right bottom polygon is +12V coming from top relay to the 2nd coil

layout bottom https://imgur.com/MiRvLiv Left polygon is GND

Thank you everyone who reads.

I don' understand the layout. Gnd isn't even connected between power input and outputs. In any case there will be huge ground voltage drops spreading over to the controller side.

With small changes, e.g. rotating the relays by 180 degrees, the current pathes can be considerably improved. You need extra wide traces for 9 A and should use wider thermal spokes at the high current pads. I guess, the connectors are neither rated for 9A.

I would put a diode between A2 (cathode) and +5V (anode) else when the transistor is off Vbatt is fed to IC.
Correction, what value is R8, its feeding Vbatt straight into +5V line
Frank

I would put a diode between A2 (cathode) and +5V (anode) else when the transistor is off Vbatt is fed to IC.

Click to expand...

Seriously speaking, I can't even see what's the intended purpose of this circuit part, I guess it's somehow erroneously designed.

Presumed R8 is large enough, Vbat won't at least hurt the microprocessor.

No diodes across your relay coils.
That is going to cause you problems when the relays are turned off!!
Transistor base drive resistors are too small. Use 1K.

Hello again,

Thanks for all the answers and critics, with your help hopefully its going to be much better.

@FVM
What i tried about polygons is instead of making huge tracks, I tried to make 2 different power area with polygons, is that a bad idea? Is this what you mean there will be huge voltage drop? Bottom plane is whole gnd its connected to main input connector. And you are absolutely right I've should spend a little more time about component placements.

@chuckey and @FVM
That part is for reading the battery level, r8 is 100k and r9 is 300k. Maybe i should use a pnp transistor to switch the power instead of gnd.

@neddie
thanks for advices, i'm going to add diodes. I already used 1k ohm on transistor, just forgot to erase those values from the design.

You didn't yet comment my statement that the ground node isn't connected at all. May be you posted the wrong layout printout?

That part is for reading the battery level, r8 is 100k and r9 is 300k. Maybe i should use a pnp transistor to switch the power instead of gnd.

Click to expand...

You would do that to disconnect the measurement. But I don't see why, the battery power supply is permanently connected, so why should you disconnect just the measurement?

https://imgur.com/R0TkCHa
Here is the bottom layout, you can see the gnd connection of the circuit.

I just realized that with npn transistor the situation of the current will be flow through +5V --> R8 -->mcu-->Gnd. That's true right?

https://imgur.com/wPtELAr
What i wanna do with pnp transistor as you see above; I will put the transistor to saturation just for reading then put it in cut-off mode again. So when its in saturation i will read the 12v battery around 3-4V and compared it with 5V to see the battery level.