Well, it wasn't that quick, I screwed up a header footprint and had to order the boards again! 
I was more careful this time.
Given a design of this level of simplicity, I can create the schematic and board layout in less than an hour, it's a pretty simple task with the right tools.
hi john and Stan great job ill take 5 boards and components, i'll put in a miniature pot for r1 side mounted for adjustment. if there are any left save e if not ill purchase on your next order!
Alan Mancus
All gone for this run, however the Gerber files were posted previously, so you can just visit OSH Park and order the blank boards.
Still seemed quick to me, even with the header footprint problems. 
Well, they should be done next week, the first one worked fine, and indicated that Stan was right on with his suggestion of having a selected resistor.
Guys, what is wrong with buss the grounds and toggle (or rotary) the sidings. It will only give you perfect 10 signal. Why make it complicated.
Nothing wrong with it Marty, we got launched into this other project and I ended up making the boards. It's slightly more automated as far as the WD is concerned, but it's not a major issue. In any case, I haven't installed anything yet, we tore the yard apart to put in home brew uncoupler tracks on each siding. We also had to fix some switches, so it's not ready to wire yet.
John, keep us posted. Good to say hello at York.
Almost home on the boards. They're all assembled and when calibrating, I found that the 2.2uf capacitors I used are right on, but Stan must have used one that had a higher value. I have to get some higher value resistors to get the time closer to the 1 second target, they're running in the lower with the assortment I have. With a 1meg resistor, I only get around 800ms between resets, and even with the 1.2m it's around 925ms and sometimes lower. I have to get some 1.3M and 1.5M resistors to bring them in closer.
I guess there's nothing to be gained by beating the "R1 should have been a trimpot" drum.
I think the R1-C1 time-constant should be linear. So if initially measuring 800 msec using R1=1.0M, then targeting 950 msec would suggest R1=950/800=1.19 M, so 1.2M should do the trick.
stan2004 posted:I guess there's nothing to be gained by beating the "R1 should have been a trimpot" drum.
Not unless it makes you feel good, I'm not doing the boards again!
stan2004 posted:I think the R1-C1 time-constant should be linear. So if initially measuring 800 msec using R1=1.0M, then targeting 950 msec would suggest R1=950/800=1.19 M, so 1.2M should do the trick.
Yep, it was a bit less than 800, and it shifts lower a bit when I actually solder the resistor in, obviously holding them in the holes isn't perfect. I actually have one that's right on 1.0 seconds with a 1.5M resistor soldered in, didn't try that with lower values. I also didn't have that many 1.5M resistors, so I'll get more this time. It looks like they're going to want either the 1.3M or 1.5M resistors to get close, seems there aren't any values between those.
Shifting gears. It seems like, say, once a year there's a report that a shut-down command engine silently sitting by itself on a powered siding or spur...suddenly and mysteriously starts up and, in some cases, takes off at full speed. My interpretation of the consensus conclusion is that some kind of power glitch or the like has reset the engine electronics - and it starts up in conventional since there was no watchdog (which has come and gone from the initial powering of that siding or spur).
So....in addition to the primary function of keeping a newly powered engine in command mode...is there a secondary benefit of keeping a newly reset engine in command mode? 
Well, time will tell, as soon as I get the resistors, I'll ship them out and they'll get some field tests. The one I completed, looks like it works fine in my DCS-RC, and it fits with gobs of room to spare. It even happily pulses the reset every second, just need the resistors to do the rest.
You will have to solder the three-pin connector to the DCS-RC motherboard, that connects the reset board.
I was able to duplicate a engine take of by .... trying to re-rail a engine on the turntable in tiu #1s zone with power "on" and the engine in tiu #4s zone took off......
Lets try that again but this time with a screw driver slightly shorting out the table (center to outside rail) Engine took off again....
The question.... Did I do a double short ? first one puts the engine in neutral, and second one put the engine in forward
Or did the I create some type of unwanted DCS command. I'm leaning towards this one.
Kiss.... one momentary off toggle will also create the watch dog if installed on the input side if the channel...
Almost perfect but not quite because there may be other tracks that you don't want a power interruption.......
Besides I know you electronic nuts like this type of challenge.. that's what you guys do....
stan2004 posted:Shifting gears. It seems like, say, once a year there's a report that a shut-down command engine silently sitting by itself on a powered siding or spur...suddenly and mysteriously starts up and, in some cases, takes off at full speed. My interpretation of the consensus conclusion is that some kind of power glitch or the like has reset the engine electronics - and it starts up in conventional since there was no watchdog (which has come and gone from the initial powering of that siding or spur).
So....in addition to the primary function of keeping a newly powered engine in command mode...is there a secondary benefit of keeping a newly reset engine in command mode?
Do see a "creeper" on the CL&W once in a while. Sometimes it has the same engine id as a different one in a different remote when the two get out of sycn on engine id numbers ..
Well, a WD signal won't fix those. 
Good afternoon gentleman
I've been following this thread since it started, very interesting. And I'm amazed at the knowledge on this forum. Thank you all for the help you give me. I'm still having DCS issues to the point that its not much fun anymore. Anyhow as I was cleaning out my closet I came across some old technical manuals of mine.(my job for uncle sam) I picked up the operators manual out of the group and noticed it seemed like a familiar size and weight. Something wrong here
Clem 67N20
FYI, the DCS-RC Watchdog Timer Modules are going out tomorrow, you should get them sometime next week. I also have one assembled module left, so if anyone missed out and wants one, let me know. 
All the DCS-RC WD Modules are gone and shipped. I hope you all have everlasting watchdog signals.
On the OSH Park web site, is there a way to find out the latest version of the PBW board in the "Shared Projects" ?
Or could you post the final Gerber file here ?
BTW, Great Post and Project ! Thanks !
I added the Gerber files necessary for OSH Park to the first post in the thread and updated the title of the thread.
Good Afternoon Stan2004:
The man asked me to involve you with this question.
The little watchdog timer board for the MTM DCS-DC controller seems to cut the signal at a point when it is just turning on.
If my scope is telling the truth, at the peak of the sine wave, is it the watchdog signal coming on at about one second after power up? A then turning of one second later?
Let the man know your thoughts.
Here's a Z-4000 at command voltage powering my Perpetual Barking Watchdog.
About 3/4 sec after the green LED blinks, the watchdog is a relatively small burst of digital activity around the peaks of 5 (x marks the spot) consecutive AC cycles...so for just under 1/10th of a second. Then there is no further watchdog activity until the next green LED blink. This is admittedly difficult to capture on a scope unless you know what you're looking for!
Given that fact, I would have probably been better off extending the time to around 1.2 seconds to insure it made it in. I made sure the reset time was at least 900ms, so all the boards should work, but there's probably no reason to make the window that small.
I saw crystal-clock stable delays and the 0.9 sec even accounts for the slightly different delay timing when track power is DC. So the watchdog should not be getting pre-maturely cut off. Since there are no published spec I am aware of on how often or how soon after applying power a PS engine must see a watchdog to stay silent, I figure the more often the better.
Presumably ChooChooPaul will clarify if his inquiry is because he is trouble-shooting a problem or was messing around under-the-hood.
Paul is having a problem, I got all of his modules back and tested them here, and they worked for me. We're trying to figure out what's different for his situation than what I tested.
I'm willing to contribute my 2 cents if you/he describe the problem. Or if you're still in the discovery phase then I'll stand by...
I'll let Paul take over and walk you through it. I couldn't make it fail, so I don't have a good idea of what might be happening.
Hey Stan2004 and GRJ:
After GRJ could not make my six modules fail to produce my results on my home layout, which was with the extra board installed in a stock DCS-RC controller, after applying power, both the PS2 and a PS3 engine would start up in conventional mode.
As I understand the project, it is to apply to the track a watchdog signal that cycles on and off once a second.
This morning after GRJ success in testing, I wanted to see what the barking dog signal looked like on my scope. So I connected the power supply to the inputs of the scope, and saw a clean sine wave. I then connected the stock DCS-RC controller to the network. The first thing the showed up was a modified tick in the peak of the sine wave, about one second later the watchdog signal appeared over the crest of the sine wave. After another second it was gone. So in about two seconds the event was over, as far as I could tell. If the addition of the board to the DCS-RC is resetting the controller 900ms after power is applied, it would seem that the watchdog signal never gets to the engine to keep it dark and silent waiting for the startup signal.
It also seems that the watchdog signal from a MTH TIU rev.L starts at applying power to the output channel of Fixed 1, and continues until the red indicator light in the unit flashes the number of the TIU.
With an engine sitting on the test track and applying power to the modified DCS-RC-WD controller the engine sees that first tick in the sine wave, a then waits for the watchdog signal to figure out which mode to start in, conventional or command. If the engine does not see a signal in about five seconds it starts up in conventional mode.
Some of us have the continuing problem in yards on club layouts, we apply power to the supply feed of the yard with the channel of the TIU connected passively to it. The watchdog signal is now over, the switch to now thrown to the yard spur, and the engine starts up in conventional mode. I have three applications for the fix to the DCS-DC.
If we change R1 in the circuit to provide more time before reset after power up, would that let the dog do his job, no plastic bag needed.
I'm not sure what you mean by a "tick" but the watchdog signal is as shown in my oscilloscope photo. It is relatively small in voltage relative to the track voltage and occurs over 5 cycles or for just under 1/10th of a second. The repeat rate is on each green LED blink.
I'm not sure what GRJ verified. Are you seeing the watchdog burst (at the top peak of 5 consecutive AC cycles) every 900 msec or once per LED blink?
The "tick" happens when we connect the DCS-RC to the power supply. It is a small change in the peak of the sine wave. The burst of watchdog signal starts sometime in ms after power up. Then after your 5 cycles it goes to the house. If the total time for reset is from power up then 900ms, it seems the timing is on the knife edge of not being seen by an engine on the track.
That "tick" appears to be one-time event when you first power up the DCS-RC so I think we can safely ignore it. The add-on PCB re-sets the microprocessor code in the DCS-RC every 900 msec (or whatever it is set to presumably within my "spec" of between 0.9 and 1.0 sec). The watchdog burst of just under 100 msec occurs less than 800 msec after each re-set. This delay time is governed by a crystal oscillator so it is rock solid. So, yes, it may be on the knife's edge but you should always get the full 100 msec of watchdog (5 AC cycles) activity as shown in the scope photo.
That's why I ask what exactly GRJ is seeing that he says it is working.
So do I understand it that you kept one of the boards? If so can you see the burst occuring on 5 cycles? If so, then there's something else going on whereby the engine is not receiving the bursts.
Just tossing it out there, is it possible that there is a variation in when the watchdog occurs between different production runs of the DCS-RC? maybe all the previously tested ones send a watchdog at 800ms, but some other variation sends it at 900 or 1000ms? Just guessing at things, but figured I'd throw it out there.
JGL
I will not have any boards until Friday when the six get back from GRJ. The wave forms I saw this morning on the scope were from a factory unmodified DCS-Remote Commander. The results I found in my testing of the DCS-RC-WD unit, were in Lab conditions, clean track, short wire runs, the engine on the track still started in conventional mode.
We must remember the the control is only half of the system. Under DCS command the engine is the other half of the solution. It seems the engine sees the "tick" on power up, then looks for the watchdog before making it's choice as to which mode to start in.
I misunderstood the experiment. It appears you applied power to a factory-stock DCS-RC and your PS engine started up in conventional. If that's the case, then there's a separate problem. Ignoring that "tick" which could just be a power turn-on blip from current-inrush, the watchdog activity you saw on the scope is as expected. For the DCS-RC you get that one short 1/10th sec burst of activity and that's it. And that's the purpose of the add-on board...to re-set the DCS-RC every second so it perpetually generates those bursts.
This is completely different behavior than the TIU which generates watchdog activity for many seconds following initial power-up.
Stated differently, if you can't get your engine(s) to power-up silent in command-mode with a factory-stock DCS-RC attached, then that needs to be addressed first. In my opinion of course!
When I built the boards, I insured that all of them had a reset window greater than 900ms. I set my top limit at 1100ms, most of the boards are centered around one second.
After further discussions with Paul, I decided on one more test. This was only with one module, I assumed the others would perform in a similar manner. Paul had mentioned that the engine stayed dark for about 45-60 seconds and then started up. I can't explain why this would happen short of a power interruption as once they recognize the WD, they should stay dark until started.
I connected DCS-RC output terminals with the WD board installed directly across the TIU output of Fixed #1. That TIU channel is connected to a roughly 40 foot loop of track with an 8 foot siding with a switch on the center rail to power it. I placed a PS/2 engine on the siding with power off. I left the TIU and DCS-RC running for about 5 minutes to insure the TIU WD had expired, then flipped the power on to the siding. After 5 more minutes, and the engine was still dark.
I turned the siding off, and unplugged the DCS-RC from the TIU channel connection. I flipped the siding power on and the engine came up in conventional in a couple of seconds.
I repeated this test three times to insure it was a repeatable result. In all cases, the DCS-RC with the reset board installed kept the engine dark for at least 5 minutes. I'm assuming it would have stayed dark until addressed.
If there's something I missed, let me know.
Good Morning out there in Radio Land!
First of all let's knock down some of the cob webs in this thread.
In all my operating experience with the DCS-RC system and the full DCS-TIU Rev L system, the MTH equipment never started up an engine without a command from a handheld controller. All the problems I have had always came down to a malfunction of boards or battery problems in the engines I was working on for repair.
When this modification to the DCS-RC came to the light of day, it seemed like the answer to the problem we have with operating a yard in a club layout. Members come to run trains, they do not have the time or knowledge to follow a long list of do this before that, then put a foot in and shake it all about.
So when the boards came to install, which GRJ can say were done correctly, it was not a cause for joy when three different engines, both PS2 and PS3, all started in conventional mode with the modified DCS-RC-WD controller connected passively across the power source to the track.
I realize that the functions of control with a handheld are lost with the addition of the board into the DCS-RC, but with a longer timed reset cycle could you send a data string to a engine between resets?
Now that the pot is boiling again, I will wait and see what cooks up today.
clem k posted:Good afternoon gentleman
I've been following this thread since it started, very interesting. And I'm amazed at the knowledge on this forum. Thank you all for the help you give me. I'm still having DCS issues to the point that its not much fun anymore. Anyhow as I was cleaning out my closet I came across some old technical manuals of mine.(my job for uncle sam) I picked up the operators manual out of the group and noticed it seemed like a familiar size and weight. Something wrong here
Clem 67N20
That Helicopter manual is clearly lacking some critical information pages huh?
(Or is it because the DCS manual has stuff for the "2Rails or NoRails" crowd?)
choochoopaul posted:I realize that the functions of control with a handheld are lost with the addition of the board into the DCS-RC, but with a longer timed reset cycle could you send a data string to a engine between resets?
I'd just like to figure out why you get different results than I do when testing the same hardware!
Access to this requires an OGR Forum Supporting Membership
