This is with scope ground wire clipped on the power supply V- terminal. Test lead is on either wire on the limit switch and the signal never changes with the switch pressed. The average voltage agrees with the volt meter set to DC reading 7.12V. The noise is always present. Both switches show the same behavior, nothing changes when pressed.
Well, you've got rid of the 60 Hz hum at least. That's good. But the limit switches are already doing their jobs so you're not really chasing the problem as I see it.
However, what happens if you adjust the trigger level? Does the fuzz move up and down with it? Try setting trigger to an unused channel.
BTW: The fuzz isn't necessarily a problem. The scope is showing noise that is ignored/unseen/filtered by the controller.
Hello Evan. Thanks for the info. I did adjust the trigger up and down and park it dead center of noise. No change. It sits at that range at all times no matter what the switches do. Maybe this is a current type detector after all.
More than likely it is some noise on your line or system. The simple micro switch setup is just a current loop. Maybe driving some opto's on the main board. Opto's normally 10 Ma for full on. Like a 4n25 or CNY17-4, etc.
I found out that the lid on the plasma CNC controller did not have any path to chassis ground as it was painted. I removed some paint where a screw make contact and shielded the lid. After running for a day it had no more of the random problems. The tech support said there is no method to reset the machine 0/0 without some commands and those commands take seconds to update. The agreed with me that noise must be hitting the microprocessor. How noise affects one set of parameters like machine coordinates but not others? No idea but so far so good.
I like to try to understand how people do certain things that I haven’t seen before. In this case it’s a simple limit switch lever type. In this case, there are two wires that go to the switch and the switch is wired in normally closed. So that if you press the button to switch opens. I was trying to figure out how it is that you have the same voltage 7.12 V DC on each wire whether the switch is pressed or not. I was trying to understand how something could signal the microprocessor the status of the switch while the voltages are the same. I still am not clear How this works. Just making a guess I assume that if you have two voltages combining, then you have a certain amount of current. Whereas if you remove one of the voltages by breaking the switch, then you have less current and possibly there is a current detector IC. But this is just guessing, but I assume they are doing something because it’s high noise environment and the circuit they chose must be due to noisy immunity, which would be great to understand.
@"T Chap" said:
I was trying to figure out how it is that you have the same voltage 7.12 V DC on each wire whether the switch is pressed or not. I was trying to understand how something could signal the microprocessor the status of the switch while the voltages are the same.
If you measure the voltage from one of the pins of the switch to ground that may be just some random noise. Both pins could have some reference to ground but they don't have to. They can be floating. What's important is the differential voltage.
The switch is nothing but a variable resistor. It's resistance is near zero when closed and nearly infinite when open. So the differential voltage must be close to zero when the switch is closed and the current must be close to zero when it's open. The other two parameters, current when closed and voltage when open, are arbitrary and depend on the circuit and method to detect the state of the switch. But they somehow have to be different from zero.
Also, remember Kirchoff's law: The sum of all voltages measured in a closed loop must be zero. So if you measure (1) the voltage between pin A and B of the switch, (2) between Pin B and grond, and finally (3) between ground and pin A, they (should) all sum up to zero. BUT, this holds true only if all voltages are measured simultanously and with an ideal instrument with infinite resistance.
In the real world the instrument represents a load so the voltages drift depending on where you put the probes, especially if the pins are floating relative to ground and the voltage source is high impedance (noise picked up from rectified HF and leakage currents). So the readings can be false and don't sum up to zero.
Comments
Maybe the limit switch circuit is not that sophisticated and you just have a bad or noisy switching power supply
This is with scope ground wire clipped on the power supply V- terminal. Test lead is on either wire on the limit switch and the signal never changes with the switch pressed. The average voltage agrees with the volt meter set to DC reading 7.12V. The noise is always present. Both switches show the same behavior, nothing changes when pressed.
Well, you've got rid of the 60 Hz hum at least. That's good. But the limit switches are already doing their jobs so you're not really chasing the problem as I see it.
However, what happens if you adjust the trigger level? Does the fuzz move up and down with it? Try setting trigger to an unused channel.
BTW: The fuzz isn't necessarily a problem. The scope is showing noise that is ignored/unseen/filtered by the controller.
Hello Evan. Thanks for the info. I did adjust the trigger up and down and park it dead center of noise. No change. It sits at that range at all times no matter what the switches do. Maybe this is a current type detector after all.
Limit switches are simply beasts. Just need to know where the wiring goes is all. Without a wiring diagram you're doing far too much guessing.
One answer is to open up the boxes, meter it out and do a rough hand drawn diagram.
Sounds like a lot of screwing around for the sake of EMI. Just stick a differential xceiver on the line.
More than likely it is some noise on your line or system. The simple micro switch setup is just a current loop. Maybe driving some opto's on the main board. Opto's normally 10 Ma for full on. Like a 4n25 or CNY17-4, etc.
I found out that the lid on the plasma CNC controller did not have any path to chassis ground as it was painted. I removed some paint where a screw make contact and shielded the lid. After running for a day it had no more of the random problems. The tech support said there is no method to reset the machine 0/0 without some commands and those commands take seconds to update. The agreed with me that noise must be hitting the microprocessor. How noise affects one set of parameters like machine coordinates but not others? No idea but so far so good.
Thanks for the Update!
While I had not been able to provide good advice I had tried to think about it.
Good luck!
I like to try to understand how people do certain things that I haven’t seen before. In this case it’s a simple limit switch lever type. In this case, there are two wires that go to the switch and the switch is wired in normally closed. So that if you press the button to switch opens. I was trying to figure out how it is that you have the same voltage 7.12 V DC on each wire whether the switch is pressed or not. I was trying to understand how something could signal the microprocessor the status of the switch while the voltages are the same. I still am not clear How this works. Just making a guess I assume that if you have two voltages combining, then you have a certain amount of current. Whereas if you remove one of the voltages by breaking the switch, then you have less current and possibly there is a current detector IC. But this is just guessing, but I assume they are doing something because it’s high noise environment and the circuit they chose must be due to noisy immunity, which would be great to understand.
If you measure the voltage from one of the pins of the switch to ground that may be just some random noise. Both pins could have some reference to ground but they don't have to. They can be floating. What's important is the differential voltage.
The switch is nothing but a variable resistor. It's resistance is near zero when closed and nearly infinite when open. So the differential voltage must be close to zero when the switch is closed and the current must be close to zero when it's open. The other two parameters, current when closed and voltage when open, are arbitrary and depend on the circuit and method to detect the state of the switch. But they somehow have to be different from zero.
Also, remember Kirchoff's law: The sum of all voltages measured in a closed loop must be zero. So if you measure (1) the voltage between pin A and B of the switch, (2) between Pin B and grond, and finally (3) between ground and pin A, they (should) all sum up to zero. BUT, this holds true only if all voltages are measured simultanously and with an ideal instrument with infinite resistance.
In the real world the instrument represents a load so the voltages drift depending on where you put the probes, especially if the pins are floating relative to ground and the voltage source is high impedance (noise picked up from rectified HF and leakage currents). So the readings can be false and don't sum up to zero.