LCS Power Limitations Through PDI ?

From MartyE's Legacy User Group meeting summary:

PDI network “coupler”. This small passive device will allow 2 PDI cables to be joined without needing an LCS device. This will come in handy to extend a PDI cable or when a LCS device is removed the cable can be joined up. This is available through Lionel parts. The PN is 691JOINA00.

I recall a a more recent thread (that I can't find) that shows a newer version with provisions to inject 12v DC power.  You have to call Lionel Service to order these.

Barry, I have done a number of posts on this subject in the past. For convenience here are two pictures. One is the single line diagram of my LCS system showing all the devices and cable lengths. The rose colored dots are the amplifier locations. In this system 5 were needed for flawless system operation. Also included is a picture of the amplifier, Lionel calls it a bus joiner. These are only available by calling Lionel.

As Tracker John states a regulated 12VDC power supply is required for the joiners to work as an amplifier. That is provided by the barrel jack connector that plugs in at the top.

The Lionel Parts website shows the current model:

691JOINB00

PCB / PDI BUS JOINER WITH POWER JACK (NO DISCOUNT)

$16.00

Must contact Lionel to order.

AmFlyer, are you using a single 12V regulated supply with sufficient amps or multiple supplies?

That is the version in the picture above. With these bus joiners there is no reason to change the Lionel provided system power supply. At least in my LCS setup the issues were voltage drop in the DB-9 cables, not a power supply current limitation.

I have a 12VDC accessory bus that runs around the layout. As shown in this picture on the left there is an adjustable regulator that connects to that bus and powers the Joiner. There is a regulator for each of the joiner/amplifiers.

Ignore the Joiner in this picture, it is an old handmade prototype Lionel R&D sent us 4 years ago. Works the same but looks different.

Barry, since you may not be able to order the joiner until we all get back to work, you can cut a PDI cable in half and build your own breakout box to inject power.  Pinouts are:

PDI Cable
1. Serial common      (Black)
2. Power ground       (Brown)
3.  PDI TxData           (Orange)
4.  PDI RxData           (Red)
5. +v power               (Yellow)

You will note the wire gauge is not appropriate to maintain voltage over long runs per AmFlyer's comment.

Here is a better picture with the LCS powered on.

The gauge wire used in the Lionel DB-9 cables was only intended for very short runs. 

Good thread!  

@Tracker John posted:

Barry, since you may not be able to order the joiner until we all get back to work, you can cut a PDI cable in half and build your own breakout box to inject power.  Pinouts are:

PDI Cable
1. Serial common      (Black)
2. Power ground       (Brown)
3.  PDI TxData           (Orange)
4.  PDI RxData           (Red)
5. +v power               (Yellow)

You will note the wire gauge is not appropriate to maintain voltage over long runs per AmFlyer's comment.

Tracker John et al:

First, thanks for the replies/advice.  This is my first post to the Forum and I certainly appreciate the sage advice.

My LCS modules are mounted on a pullout panel in three columns with five or six modules each plus the SER2 and WiFI mounted separately on the same mounting panel. Each module is basically connected to the next with a 12" PDI cable.  The longest cable is the 10 ft. cable from the power supply to the WiFi module, then 12" cable to the SER2, then a 3ft. cable to the first ASC2.  where it becomes 12" cables to the top of the column and then a 12" cable to the top of the next column then 12 " cable down the column, etc. See attached photo.

In other words, voltage drop due to PDI cable length is mitigated by the short runs of PDI interconnecting the modules.

While the breakbox approach would certainly help if it were inserted between say between the 8th and 9th module, I'm still wondering why not simply replace the 1A power source with a 2A power source and be done with it?  I understand that small gauge wire  can cause a voltage drop, but given the relatively short overall total PDI length (e.g. no more than 26 ft. total including the initial 10 ft from the Legacy Base) is there a concern that the last module would not receive sufficient current even if the input current source was doubled?  Given that I had no problems prior to adding the WiFi and SER2 modules, the question is whether adding those two modules (totaling 450 ma of additional power consumption) and three feet of additional PDI cable would mean having to use a breakout box rather than simply increasing the total available current. 

Consequently, is there any thought about the number of interconnections between modules as being a voltage drop concern as well as the overall footage of PDI cable that should drive the decision as to whether use a breakout box versus simply using a larger power supply?

Thanks,

Barry

Given the compact arrangement of your LCS modules, I think a 2A power supply may be the solution given your calculations show your total current draw exceeds 1A.  My LCS use has been on modular layouts where the cable snakes around the perimeter - a lot of 20ft cables in use.  Voltage drop is an issue.

If the 2A power supply doesn't solve the problem, inject power at the tail end of the chain.

Tracker John:

Given your comment to inject power at the tail end of the chain, that would be an easy thing to do using your prior post about cutting a PDI cable to gain access to the two power leads (Yellow power plus/Brown power ground) and simply connecting them to the 2A power supply I have that has the traditional red/black binding posts.  Thus, I would not touch the original 1A supply and add the second 2A supply at the tail.  I would expect that the amount of actual current provided by the 2A supply to be the amount needed to compensate for whatever additional current is needed.

Unless you see an issue with this arrangement, I'll work on this in the morning. I have a 3 ft. PDI cable that would be a perfect 'victim' for this application and would avoid having to take what is already installed apart.

Thanks!

Barry 

i would not inject power into the head and tail ends of the lcs chain. i believe the bus joiner interrupts the power supply at that point in the chain. just try the   higher capacity supply by itself.

rudy

also this video:

https://youtu.be/St8FkSYb7nw

Rudy, power ground (brown), +12v (yellow), and serial common (black) are electrically contiguous on the modules (ASC2 and SER2) I tested.   Power injected into either end should propagate through all modules in the chain.  According to Jon Z, the serial i/o in each module re-transmits packets to the next leg - point-to-point comms.  (the original ASC & other modules used multi-drop serial comms).  This diagram was drawn before I learned which pins were Tx and Tx comms.

OK if Jon Z said it is good, it is good!

(editing this) In re-reading your note, it's not clear to me what exactly Jon Z said regarding an additional power supply. In an abundance of caution, I would recommend using two power supplies only in conjunction with the Lionel bus-joiner PCB, or use a single higher-current supply at one end of the chain.

Barry, that is top end work, your LCS setup looks great. A couple of thoughts on solving your issues. First I have had two ASC2's fail. One would loose its programming when the power was turned off and one just stopped working. So one of the LCS components could be bad. I now have some spares of each type just in case. Four years ago Lionel R&D gave us a copy of their development software so they could see why my system was not working correctly. At the time there were 38 LCS devices and all would work but not reliably. The system behavior leads me to believe the stated ma current requirement for the devices is not continuous, just when the device is performing a function. As I recall, the voltage at the end of the line was a bit over 3V. Adding 3 joiner/amps solved it. Last year I added 6 more BPC2's plus interconnecting cable to the system so two more joiner/amps were added. At $19 these are a cheap, proven and tested fix by Lionel.

I agree with Rudy.  If you use two (or more) supplies with out breaking the power buss then I feel you leave yourselves open for a ground loop situation which could inject noise and spurious signals into the communications.  At the price of the joiners as AmFlyer points out this is the way to go in my opinion.  I have found this thread very informative because I too have a large network (16 devices) and a 40' run.  I had to disconnect my serial connection because it was loading down the network and now thanks to you all I am going to go back and take another stab at it.

dick

So does anyone know the correct Lionel Part Number and or nomenclature to place an order?  I searched in parts ordering and I cannot seem to locate it.

thanks

dick

Tracker John posted it above in this thread.

Re: "ground loop situation" ... my approach would be a single power supply of suitable amp output, make a Y at the power supply and run power to as many PDI Bus Connectors as needed for stable operation.  If you have a long run, use lower gauge wire.  The PDI power supply uses a 3.5mm barrel plug.  Don't know what size the bus connector uses.

I got out my spare joiner and took a closeup picture should it be of any help. The circuit board is 1” square. The barrel connector, the black item at the top, is 3.5mm. 

Well, the 2A 12 VDC supply is installed and all  LCS 18 modules are recognized by eTrain, but only 16 are recognized by LCS WiFi Monitor. For some reason, LCS WiFi Monitor only recognizes 10 of 12 ASC2s;  the last two are not recognized though the 4 BPC2s after the two not seen are indeed recognized by LCS WiFi Monitor.  

Is there a limit to the number of modules of a given type that will be recognized by LCS WiFi Monitor? Clearly, the system now has enough current and given that eTrain sees all 18 modules, the problem presumably resides in software, not hardware.

So here's some background on getting my system's power increased....

On Tuesday morning I cut 3 ft. PDI cable and attached Anderson Power Poles to the Yellow (+) and Brown (-) wires as Tracker John suggested.  I connected the PDI cable to the open PDI port on the last module (BPC2) and connected my 2 A 12 VDC power supply.  The 1A wall wart provided by Lionel was still connected through the coaxial connector by the DB-9 connector.  I powered everything up and all 18 modules were online.  So far, so good...

During the course of the next hour or so I did some testing.  I removed the wall wart from the outlet (left the 12 VDC wall wart connected through the coaxial connector but no longer powered) and the system kept running.  However, when I shut down the system and restarted it, all of the modules had their red LEDs on and they did not extinguish and of course CAB2 could not command switches or track relays.  So I started to take things apart again.   The end result was that I now have the 2A power supply connected to the 12 VDC line that once came from the wall wart and everything was stable.

This evening i played with the setup again.  First I did a continuity check on the PDI cable to ensure that the Anderson Power Pole connectors were properly tied to the correct pins on the PDI connector.  Everything checked out.  Then I removed the 2A supply from the coaxial connection  and used a different power supply to power the PDI connection through the last PDI port on the last module (BPC2).   Everything fired up correctly and all 18 modules were recognized by eTrain and 16 modules by LCS WiFi Monitor.  I then shut down the power, waited about five seconds and started power again.  This time only the first four modules were recognized with the remainder having their red LEDs all lighted.  So I  replaced the power supply with the one that I was using previously and end up with all of the modules with soled red LEDs.   I shut down the system again and repowered, getting the same result.

Lastly, I reinstalled everything as before with the 2A supply providing power through the coaxial connector by the DB-9 and everything once again appears to be stable.

Conclusions:

1  The wiring information provided by Tracker John was correct.

2. Powering the LCS system through the last PDI port on the chain does provide power to the entire network.

3. However, reliability of this approach is suspect.  Apparently all modules were powered (e.g. all had their LEDs on when communication was not established) but for whatever reason the system could not function reliably.

4.  The 'best' approach is to replace the 1A 12 VDC wall wart with a more robust power supply, maintaining the power connection at the DB-9 end of the PDI cabling.

5. I decided against having two power supplies connected at the same time (one at both ends).  If a power supply is connected to the LCS system but is not powered, it can potentially act as a load against the LCS power system dragging the other power supply down in the process.  The safer approach is to have one power supply used and then if multiple power drops are necessary, use a feed off the same power supply.  This assumes of course that the external 12 VDC power system has enough current and sufficient wire gauge to provide the proper voltage/current at the feed points.

Last comment:  Tom noted the quality of workmanship of the panel housing the LCS modules.  I must give credit to where credit is due.  The panel was designed/constructed by TrainWorx of Dallas, TX, the outfit that constructed the layout.  Their shop is about an hour from my home.  They did a first class job designing/building my dream layout.  

At my request, they also installed the CTI-Electronics "Model Railroad Control System" that allows design/programming of a prototypical signal system based upon turnout position and track occupancy. Trainworx installed the CTI Electronics modules on the opposite side of that control panel and wired the signal system.  Roy McMillan of Trainworx did the initial programming to demonstrate that the system works and then I spent the last three months expanding on his initial programming to create a signal system that mimics CSX signal rules ("speed signaling").  The programming length doubled from Roy's 4000 lines to over 8700 lines.  Following that, the most recent update to using CTI was the integration of turnout control.  While the  CTI system was initially cognizant of turnout position (normal/reverse), CTI does have the ability to send TMCC commands to engines and  turnouts. So i added the programming where mouse clicking on a turnout on the CTC display will activate the turnout.  It is a lot easier to look at an active layout panel and select the proper turnout instead of going through a list of 39 turnouts to decide which one to activate.  

While I've been involved on-and-off with 3-Rail since I was a kid, I appreciate having a layout that more closely reflects prototypical train operations.   Perhaps not everyone who sees a 3-rail layout in operation will notice the signals and the aspects they display, but as a CSX retiree I wanted the appropriate signals to be displayed based upon route and track occupancy.

Thanks,

Barry Baines

Keller, TX 

I have 26 ASC2's in my LCS system with no problems. There are 7 BPC2's. What you describe sounds similar to our problems before we added the Joiners to the system. The original developer of the LCS, Mike Reagan works for TrainWorx. Have you discussed this with Mike?

Hi Tom:

I'll be calling Mike Regan later today on a different matter and will let him know of this conversation.

At this point I remain convinced that Lionel's LCS WiFi Monitor Program has a bug is because:

1.  CAB2 can use all of the ASC2s and BCP2s without any problems.  Nothing is offline from CAB2's perspective. In addition, the SER2 and WiFi modules are working fine as I can communicate through both of them.

2.  None of the modules Red LEDs on indicating a system problem communicating.

3.  Harvy's eTrain software lists all 18 modules at the bottom of the screen, so the software recognizes all 18 modules including the two that LSC WiFi Monitor does not.  

Again, everything is working with no problems using CAB2 but for some reason LCS WiFi Monitor fails to recognize 2 of 12 modules (the last two of the ASC2 chain) while recognizing the following 4 BPC2s.   I have yet to configure eTrain  to activate the nine uncoupling tracks attached to those two modules, so I can't confirm that they work under eTrain.  

So when you indicate "26 ASC2s in my LCS system with no problems" does this mean:

1.  The LCS WiFi Monitor software recognizes all 26 modules, or

2.  All of the ASC2s are recognized by CAB2, or

3.  Some other program recognizes all 26 ASC2s as being online; or

4.  All of the above?

In other words, are you running LSC Wifi Monitor and it sees all 26 ASC2's plus whatever other modules you have installed?  If this is the case, how many modules do you have being powered by the wall wart from the PDI cable kit and how many are powered by each "Joiner"? What is the amperage of the power supply that is providing power to the Joiner?  

Thanks,

Barry Baines

Keller, TX

Barry, now I understand better what is happening. Great to hear your LCS is fully operational. I have the programs you mention but I am not using them. My system works fine and recently I have spent far more time using the layout than monitoring it. I control my layout with an iPad connected to the LCS wireless module. It sees all the devices, actually all the 91 programmed TMCC addresses plus the 7 preprogrammed route settings.

@Barry Baines posted:

Hi Tom:

I'll be calling Mike Regan later today on a different matter and will let him know of this conversation.

At this point I remain convinced that Lionel's LCS WiFi Monitor Program has a bug is because:

1.  CAB2 can use all of the ASC2s and BCP2s without any problems.  Nothing is offline from CAB2's perspective. In addition, the SER2 and WiFi modules are working fine as I can communicate through both of them.

2.  None of the modules Red LEDs on indicating a system problem communicating.

3.  Harvy's eTrain software lists all 18 modules at the bottom of the screen, so the software recognizes all 18 modules including the two that LSC WiFi Monitor does not.  

Again, everything is working with no problems using CAB2 but for some reason LCS WiFi Monitor fails to recognize 2 of 12 modules (the last two of the ASC2 chain) while recognizing the following 4 BPC2s.   I have yet to configure eTrain  to activate the nine uncoupling tracks attached to those two modules, so I can't confirm that they work under eTrain.  

Interesting comments, I've noticed some oddities in what the LCS WiFi monitor recognizes, but it wasn't really causing a huge problem for me, so I ignored it.  It calls my CSM2 modules STM2 modules, and it gives me a BPC2 module that I don't have.