Unknown issues in circut design

I am working on a project where I need to measure the current drawn by multiple loads from a single power supply using a 8-bit PIC micro-controller. Also, when an overload condition is detected in any of the loads, I need to cut it off using a MOSFET. Here, each port is limited to a specific current and the sum of all the port current is the total rating of the power supply.
To do this I have setup the circuit as shown below.

Here I have 4 output ports of the power supply and 4 loads are connected to them. I am using a MOSFET and a 1-Ohm shunt resistor on each port to measure the current drawn.
Firstly, when I connect 1 load (LOAD1) the circuit works as expected. Here the load is variable dummy load. For this configuration I am getting good readings. I verified the reading with a multi-meter.
When I connect a second load the problem starts. I keep the 2nd load’s current at a constant level by using an LED as LOAD2. Now when I increase/decrease the 1st load, I find changes in reading of the 2nd load. The value increases with increase in current from the 1st load and vice versa when the current decreases.
Example:
- LOAD1 = 200mA, LOAD2 = 20mA
- LOAD1 = 350mA, LOAD2 = 40mA
- LOAD1 = 600mA, LOAD2 = 100mA
I find the same readings in the multi-meter also. I am measuring voltage across the 1 Ohm resistor with the multi-meter. But I found proper readings, i.e. 1st load current changes and 2nd load remains constant, when I connected the multi-meter as an ammeter, i.e. in series with the circuit.
I think the problem is in the way I am measuring the current. Can anyone provide some insights on how this should actually be done?
Any comment or help is highly appreciated.
Thanks for reading.

Hi,

please post your schematic, and your PCB layout.

The code also could help.

(My first assumption: PCB layout problems. Especially ground bounce. Is it a breadboard test?)

Klaus

Hi, this is the schematic or Block Diagram

Hi..
Thanks for your reply... It is not a breadboard test. I already made pcb and tested on it. Issue not solved.

Hi,

please post your schematic, and your PCB layout.

Click to expand...

What is your "load"? are these bought resistors?


Klaus

Hi,

I am using a variable dummy load using MOSFET. Image attached. I made this on a veroboard.

I am also using a PIC development board for mcu purpose and made an extra pcb for op-amp + Mosfet and shunt resistor for switching and current sensing purpose. I have the .sch and .brd file. Should I attach that ?

The schematic is same as this one- ...

Thanks again.

Hi,

are you sure your dummy load circuit doesn´t oscillate?
Is the 9V stable?

I have the .sch and .brd file. Should I attach that ?

Click to expand...

I already asked twice for it.
... Maybe it´s too secret to show us...

Klaus

Hi,

I have attached the sch and brd files. The PCB has a provision to connect 16 loads, but I am using 2 ports as of now. Also, the PCB is a home made etched one, it does not have any solder masks.

The 9V power supply is stable.
I am not sure whether the dummy load circuit oscillates or not. How can I check it?

The hardware setup is connected in this way. All the boards are connected using burg wires.





Thanks..

Hi,

now i´m relatively sure, that my first assumpion is true: Layout proble: Ground bounce.

--> You don´t have a GND plane! you may have some copper poured areas ..but still extremely bad GND wiring.

See the GND path in pink:
1) it´s a long way to go
2) there are several paths with extremely thin wires
3) the GND current on it´s way to the connector is "added" = influenced by the current of other channels. (current of Q16 ...+Q14 ....+Q15)
4) every voltage drop on a wire/trace is added to the voltage drop of the resistor --> causign wrong current readings



--> make the top side a solid GND plane.

If you want to make it with individual traces (worse than GND plane!) then use star connection to GND connector with wide traces to avoid influence of one channel to the other.
Also use individual star connection for each amplifier connection.

***
On top right corner I did some improvements. With a little description.
You see the GND plane. Solid!



Klaus

Hi Klaus,

Thanks for the explanation. It seems to make sense now.
I will make a new PCB considering your suggestions and test it.

Do you think the Dummy Load is also affecting the readings?

Thanks once again..