The Strangest Problem :)

idbruce

As many of you know, I am currently working on building a 3D printer, and I have come across the strangest problem. Before I tell you what the problem is, please allow me to set the scene.

I normally do most of my work in the basement, but during the winter months, the basement gets unbearably cold, so I have moved the project upstairs to a much warmer and dry environment. In fact, my current work area is so dry that electrostatic discharge is common. Anyhow I have a table and a cabinet top where I am performing most of my work and testing. The cabinet top is the test area, and on this top I have placed and wired my controller board, (2) stepper drivers, a user interface board, an LCD display, (2) linear actuators, and a power switcher board. The inside of the cabinet is where my power supply currently resides. All devices are wired up and ready for power. So that is the scene.

Now for the problem. As mentioned, electrostatic discharge is common and I know that I really should be wearing protection, but some strange stuff is happening. With the discharge of any static, all the equipment powers up from an OFF state.

For instance, imagine a linear actuator laying on a wooden top. The actuator is an aluminum tube with a stepper motor fastened to it. The stepper motor has four wires which go to a stepper driver. The stepper driver has four wires which go to the Propeller controller. The Propeller controller power switch is always in the 5V servo position. The power going to the controller is routed through a power switcher board but is controlled by a tact switch on the user interface.

If I touch the aluminum tube with a static discharge, the whole system fires up or if I discharge static anywhere else on the attached equipment for that matter That discharge is traveling a long distance to enable this kind of activity Basically it is somehow triggering the the 4011N shown below.



I have not tried to troubleshoot the problem yet, I am just throwing it out there for a smile or two.

EDIT:The schottky diode has been temporarily removed from the power switcher board, due to prior troubleshooting.

tonyp12

1M and 100K ohm are pretty high values and any static voltages will "win" over it.
I would 1/10th it, so 100K and 10K instead as you are not running of a battery and trying to save current.

frida

Will a capacitor in parallel to R1 solve it?

idbruce

Tony

The 1M resistor was chosen as a time delay, to prevent accidental power downs. There is more info in the following quote. Any other ideas?

Here is another schematic to accomplish the same task: http://www.mosaic-industries.com/emb...ch-circuit.png

EDIT: After desoldering some components from some old electrical junk, I finally had enough components to test the circuit linked to above, and I must say it works very well. I did not have any 5M resistors lying around, so I slapped in a 10M resistor and tested it with a 12VDC nominal power supply. One quick button press and voltage appeared on my DVM and to shut the circuit off, you must hold the tact switch down for about 3-1/2 seconds. Taking it a step further, I replaced the 10M resistor with a 1M resistor and tested again. One quick button press and voltage appeared on my DVM and to shut the circuit off, you must hold the tact switch down for about 1-1/2 seconds. So the lower the resistance, the faster the circuit can shut down and vice-versa. 1-1/2 second time delay is just about perfect for my needs. All I need now is a worthy MOSFET or transistor to add onto the end of it for controlling the main supply.

The full discussion pertaining to this circuit is here: http://forums.parallax.com/showthread.php/156184-Discussion-Of-Latching-A-Momentary-Switch-To-Control-A-Power-Supply

idbruce

As I see it, considering that all peripherals go through the Propeller controller, the static must be traveling through VIN, because the GND is switched by Q1 and going out of X1-1. I believe it might be back tracing through Q1 and R4, if that is even possible.

EDIT: I take that back, because VIN is a straight shot to the power supply.
EDIT: Static must be traveling through X1-1 and triggering Q1 somehow, because I do not see any other way for it to go from the peripherals to the power switcher board.

idbruce

I have added a colored and labeled schematic for easier tracing and understanding.

Perhaps a diode on the X1-1 line??? That way the ground only has one way to go????

jmg

idbruce wrote: »

EDIT: Static must be traveling through X1-1 and triggering Q1 somehow, because I do not see any other way for it to go from the peripherals to the power switcher board.

I do not see any Chip decoupling?
A series R before Vcc decoupling, also drops the currents from any spike. Make it as large as practical 1~10k may work on your design ?

With ESD or contact triggering, I find it useful to run up Spice.

Enter a generator with a couple of kV and rise time of 1ns region, and feed that info a few Femtofarads and onto a few nanohenries,
then see what amplitude and duration pulses you get.
Vary the Rise time and fF coupling to get a feel for the effects.

Usually, I would suggest 4000 series CMOS, as that is inherently slower and fails to 'see' narrow spikes, but you seem to already be using that ?

idbruce

Thanks for the input jmg

Nope, no decoupling, I keep forgetting that important necessity.

As for Spice, I think I tried using that about 15 years ago, but I will give it another try, just to get familiar with it and perhaps find out more about this circuit.

Usually, I would suggest 4000 series CMOS, as that is inherently slower and fails to 'see' narrow spikes, but you seem to already be using that ?

Yea, I had an old NTE4011B laying around, so I used that for the circuit.

With ESD or contact triggering, I find it useful to run up Spice.

Apparently this problem is more common than a novice would think.

jmg

idbruce wrote: »

As for Spice, I think I tried using that about 15 years ago, but I will give it another try, just to get familiar with it and perhaps find out more about this circuit.

Download
www.linear.com/ltspice‎

Free and easy to use.

idbruce

Thanks for the link jmg, it is installing now.