Erratic Switching on multiple ATLAS 6924 boards

Check with Atlas, I think there is a fix, done by adding diodes to the two isolated rail inputs.  The boards appear to be sensitive enough to pick chatter from the rails, enough to trip the boards.  Of my 20 some boards I only have two that exibit this, and only in conventional use, of the TMCC or DCS system.  If you get no response from Atlas, contact Steve Horvath, forum member.   Underlined phrase should link Steve's forum profile.

Thanks Mike CT.

I will try to connect with Steve. He helped helped me with several problems I was having with some of the switches. 

It had not occurred to me that signal on the rails could trigger the switching when there is no active power draw from an engine.  

Joe

My two cents from EE and quality control perspective: not enough engineering thought or testing went into what otherwise is a good design idea.  Sounds like sensitive analog parts, insufficient bypassing, not separating grounds to one point, or who knows.  Maybe used minimum parts to get working to get price point?   Reminiscent of Muntz TVs (if you don't know what I am talking about, contact me off line).

I am befuddled by this one. I have these boards, but have not yet put one into service. As far as I know there are no semi-conductors on these boards. Again, I have not looked. So, if true, I would suggest powering these boards from pure DC, not FWR AC. Some DC relays are not intended for use on FWR AC (aka lumpy DC).

Thanks RRMAN and Keith.  I have contacted Atlas and am waiting to hear back.  Keith, I don't know how or if I can bypass the FWR on the board or if I could damage the board doing it.  I will also ask Atlas about DC direct to the board.

Joe

JVM-51 posted:

Thanks RRMAN and Keith.  I have contacted Atlas and am waiting to hear back.  Keith, I don't know how or if I can bypass the FWR on the board or if I could damage the board doing it.  I will also ask Atlas about DC direct to the board.

Joe

Well, I was way off on that reply. The board has a 5 volt regulator and a Microchip microcontroller. I thought it was just a bunch of relays. It still may perform better on pure DC. You do not need to bypass anything on the board, just apply pure DC instead of AC.

Guessing that the microcontroller inputs are not sufficiently bypasses or conditioned and respond to noise or surges that interpreted as valid.  Usually can be fixed in software by delay loops so input has to be held on or off sufficiently long time, relative to noise impulses, before considered valid.

rrman posted:

My two cents from EE and quality control perspective: not enough engineering thought or testing went into what otherwise is a good design idea.  Sounds like sensitive analog parts, insufficient bypassing, not separating grounds to one point, or who knows.  Maybe used minimum parts to get working to get price point?   Reminiscent of Muntz TVs (if you don't know what I am talking about, contact me off line).

It seems like a Rube Goldberg design. It takes 3 relays and a microcontroller to accomplish this function? 

Actually, the boards seem amazingly functional when not affected by spurious signal that triggers this odd switching acation.

Non-derail, power routing for two power sources, panel / trackside signals, switch machine burnout protection and maybe more that I have not figured out.

I'm don't know if this condition is unique to me, caused by something I've done with my wiring or if others are experiencing this as well.  

JVM-51 posted:

Actually, the boards seem amazingly functional when not affected by spurious signal that triggers this odd switching acation.

Non-derail, power routing for two power sources, panel / trackside signals, switch machine burnout protection and maybe more that I have not figured out.

I'm don't know if this condition is unique to me, caused by something I've done with my wiring or if others are experiencing this as well.  

 

I have a "test suggestion". If you can completely isolate the board power from the switch machine power, try powering all of the non-derail boards from a 12 VDC battery. I suspect that electrical noise is coming in the input terminals and coupling across to (what should be an isolated circuit) the internal 5 volt circuit. If the power for the internal 5 volt circuit is derived from a battery, you will break the signal path and the problem should cease. If my thinking is correct, a powered locomotive shunts this electrical noise to common or ground.

If you have wired the same power source to both the inputs and the board power terminals, this may be easier said then done.

Here is the wiring diagram.  the board gets internal power from the ZW B terminal.  The Power routing and track power is coming (currently) only the ZW D terminal. All circuit go to the common ZW U terminal. 

JVM-51 posted:

Here is the wiring diagram.  the board gets internal power from the ZW B terminal.  The Power routing and track power is coming (currently) only the ZW D terminal. All circuit go to the common ZW U terminal. 

 

 

Nice diagram. I am suggesting to disconnect terminals 16 & 17 from the ZW. Then connect terminals 16 & 17 to a 12 volt battery. This, of course, is for testing only. If the problem goes away, I can suggest solutions.