484ffcf8010a559ffdc7da27324df8c3cccb2ddf
Board-new/ESP32-ISO-dat/ESP32-ISO-dat.md
... | ... | @@ -0,0 +1,6 @@ |
1 | + |
|
2 | + |
|
3 | + |
|
4 | +## ESP32 ISO board |
|
5 | + |
|
6 | +- [[inductive-load]] |
|
... | ... | \ No newline at end of file |
Tech/actuator/inductive-load-dat/inductive-load-dat.md
... | ... | @@ -0,0 +1,27 @@ |
1 | + |
|
2 | +# inductive load dat |
|
3 | + |
|
4 | +## Cause analysis |
|
5 | + |
|
6 | +### ground bounce |
|
7 | + |
|
8 | +Almost always, RS232 disconnects are from something called ground bounce, which means that some inductive device like a motor or a solenoid puts a lot of power onto the ground. When that happens, the signaling, which is relative to ground, can end up being high or low. |
|
9 | + |
|
10 | + |
|
11 | +### float ground by seperated power supply |
|
12 | + |
|
13 | +The problem is worse when the systems all share a ground. If you powered the new board with a battery, not a power supply, this can help to float the ground. Or, if the PC is a laptop, disconnect from the wall and run it on battery. Try that first if you haven’t. if you haven’t tried, change the cycle time, where maybe the motor is on 5 seconds and then off 5 second. That makes it easier to see when the error happens- at the start or the stop of the motor. If the start, it is going to be a voltage drop on the RS232 chip, and a local capacitor of 5uf or so can make a huge difference. If this happens with the motor turning off, this is almost certainly inductive spike causing a ground bounce. |
|
14 | + |
|
15 | +I usually put some extra capacitance on CPU, comms chips, and any logic chips and make sure there is a good ground between them- either a ground plane, or at least a 20mil trace. Same with power. The caps will handle all short term power changes- but you need the larger power traces to get that power to or from the chip before the cap is no longer able to keep voltage in limits. |
|
16 | + |
|
17 | + |
|
18 | +### oscilloscope check |
|
19 | + |
|
20 | +Do you have a DSO oscilloscope? You will never see this with a power supply meter- you need to check for voltage spikes that are much shorter than 0.1 second. Often, they are more like 0.00001s. If you don’t have that test equipment, I do, and maybe you can send the proto board over with just the bare minimum parts- RS232 chip, motor drive chip, controller. I have motors and all that stuff. |
|
21 | + |
|
22 | + |
|
23 | +## Solution Tried |
|
24 | + |
|
25 | +- add 100uf capacitors to 5V/GND and 3V3/GND |
|
26 | +- use a seperated 12V/2A battery, USB-RS232 board does not keeping reboot, can still can not be opened COM port to use |
|
27 | + |