Hello all,
I have this project in mind for a spot welder where i need to measure the resistance of the attached weld cable. The idea here is to have constant current source using an opamp with its feedback connected to the weld cable to be tested. Initially wanted to do it without re-attaching the cable to another terminal could not figure it out so made it separate.
Here is what i hope to accomplish or could be wrong completely. Once the electrodes are shorted the base of Q7(iSet_EN) is set high by the microcontroller. A voltage drop at the cable would be read at 1 & 1 & 2 of J1 which would be the total resistance of the cable.
The design and function is fairly simple during a weld. MOSFETs Q1-Q6 function as a switch and are driven by their corresponding drivers U1-U6. They switch the stored charge from the supercapacitors C1-C6. These capacitors are in 3s2p configuration with a max voltage of 8.1v. The trigger pulse is send by U8, the Atmega328 from pin 10.
The constant current source is comprised of the opamp U1, U8 send a pulse to Q7 which in turn switches on the power to U7, Q8 base current is controlled by U1's feedback.
Test.pdf (72.9 KB)
Forgot what i wanted to ask,
To connect this to an arduino, it is by connecting pin 1 and 2 for J1 to ADC pins 25 and 26. How is the voltage drop measured ? With a DM it would be measured on pins 1 and 2. I can understand if it was one pin then it is reference to gnd but not sure how this would be done here.
Any suggestions ?
-
That’s a sure fire way to blow up an atMega328. No contact will result in 10 volts on pin 25. Device VCC is the limit.
-
U7-Q8 doesn’t look like a constant current source to me.
What kind of contact resistance are you trying to measure? Do you understand how four wire ohmic measurement works?
I would use a controller that is galvanically separated from the cable being measured. Push a known current through the cable. Connect the analog input to the plus side of the cable and gnd to the other end of the cable.
Use Ohms law and do some calculations how much current You will need to get a reasobale amount of millivolts for the controller to read. You will probaly need quite high current because a welding cable is a very low ohm thing.
Using precision measurement amplifiers, in order to use a lower test current, will call for a certain size of the wallet, or scills in that science.
WattsThat:
-
That’s a sure fire way to blow up an atMega328. No contact will result in 10 volts on pin 25. Device VCC is the limit.
-
U7-Q8 doesn’t look like a constant current source to me.
What kind of contact resistance are you trying to measure? Do you understand how four wire ohmic measurement works?
Yea without a lot that would be stupid to do.
Simulating U7-Q8 i seem to get a constant current.
My idea was to use Kelvin measurement to measure the voltage drop on the load. The load here is a 4-AWG cable. When laying out the board the trace from pins 1 & 2 from J1 would be thinner going to pin 25 of the atmega and thicker between pins 2 and 1 from Q8 and J1 respectively.
Railroader:
I would use a controller that is galvanically separated from the cable being measured. Push a known current through the cable. Connect the analog input to the plus side of the cable and gnd to the other end of the cable.
Use Ohms law and do some calculations how much current You will need to get a reasobale amount of millivolts for the controller to read. You will probaly need quite high current because a welding cable is a very low ohm thing.
Using precision measurement amplifiers, in order to use a lower test current, will call for a certain size of the wallet, or scills in that science.
I've managed to achieve it. i can see a fair amount of voltage drop at the weld cable about 400mV. How much can the Atmega328 read on and ADC pin ?
A 5 volt controller, max 5 volt to any I/O pin and max 3.3 volt for a 3.3 volr powered controller.
Thanks, I know the max but is there is minimum.
anishkgt:
Thanks, I know the max but is there is minimum.
Yes, it's in the processor data sheet. It will be a voltage slightly less than VSS (usually ground).
