Spacing between NCE BD20 detectors


Bill Wilken
 

I want to keep my Atlas signal blocks coterminous with my railroad's four power districts, each of which is approximately 25 feet in length.  Initial experimentation indicates that placing one NCE BD20 in the middle of a block will not detect a train. Can anyone in this group offer any guidance on using multiple detectors in a block?


wirefordcc
 

Hi Bill,

If you haven't tried it already, try putting more than one turn through your BD20.  Literally just loop it around and go through it again.  This increases sensitivity of the BD20.  I don't think you need multiple detectors in a block.  Check the BD20 instructions for multiple turns through it.

Allan Gartner
Wiring for DCC


Bill Wilken
 

Allan,

I've tried a variety of "winding" strategies, but without success.  When I talked with NCE's technical support fellow the other day, he was pretty skeptical about the feasibility of 25' blocks.  He suggested that the only viable option was installing a POT on the BD20, which I haven't yet attempted.  In the meantime, I'm double-checking all my wiring.  My experience (successfully) building a test-bed tells me that it's all too easy to goober up connections.

Bill


On 9/26/21 10:45 AM, wirefordcc via groups.io wrote:
Hi Bill,

If you haven't tried it already, try putting more than one turn through your BD20.  Literally just loop it around and go through it again.  This increases sensitivity of the BD20.  I don't think you need multiple detectors in a block.  Check the BD20 instructions for multiple turns through it.

Allan Gartner
Wiring for DCC


David McBrayer
 

Bill, 
I don’t know all of the in’s and out’s of your layout.  NCE has an article on their website that discusses a BD20 detection challenge that reads similar to yours at this link, https://ncedcc.zendesk.com/hc/en-us/article_attachments/200585105/DetectionSensitivityandNECBD20sonCMRI.docx. Hope this helps.  

Dave McBrayer 
Auburn, CA 
———————

On Sep 26, 2021, at 07:54, Bill Wilken <bill.wilken@...> wrote:

Allan,

I've tried a variety of "winding" strategies, but without success.  When I talked with NCE's technical support fellow the other day, he was pretty skeptical about the feasibility of 25' blocks.  He suggested that the only viable option was installing a POT on the BD20, which I haven't yet attempted.  In the meantime, I'm double-checking all my wiring.  My experience (successfully) building a test-bed tells me that it's all too easy to goober up connections.

Bill

On 9/26/21 10:45 AM, wirefordcc via groups.io wrote:

Hi Bill,
If you haven't tried it already, try putting more than one turn through your BD20.  Literally just loop it around and go through it again.  This increases sensitivity of the BD20.  I don't think you need multiple detectors in a block.  Check the BD20 instructions for multiple turns through it.

Allan Gartner
Wiring for DCC

--
Dave McBrayer
Auburn, CA


wirefordcc
 

Bill,

I believe you said that your BD20 was not detecting a train in the block.  The pot is for desensitizing the BD20.  So that may not help you.

I did notice in the BD20 instructions where it says it may not work on long blocks.  But it referenced false detections due to couple capacitance and glue and ballast.

Is your long block fed every section or two of track to your bus wires.  Note:  One of your bus wires should be connected to a wire going through your BD20.  Make sure you don't have another connection (called the goober connection!) going to your booster that could be bypassing your BD20.

Those are the obvious things I can think of.

Allan


Bill Wilken
 

Interesting.  Thanks for forwarding. 

On 9/26/21 11:23 AM, David McBrayer via groups.io wrote:
Bill, 
I don’t know all of the in’s and out’s of your layout.  NCE has an article on their website that discusses a BD20 detection challenge that reads similar to yours at this link, https://ncedcc.zendesk.com/hc/en-us/article_attachments/200585105/DetectionSensitivityandNECBD20sonCMRI.docx. Hope this helps.  

Dave McBrayer 
Auburn, CA 
———————
On Sep 26, 2021, at 07:54, Bill Wilken <bill.wilken@...> wrote:

Allan,

I've tried a variety of "winding" strategies, but without success.  When I talked with NCE's technical support fellow the other day, he was pretty skeptical about the feasibility of 25' blocks.  He suggested that the only viable option was installing a POT on the BD20, which I haven't yet attempted.  In the meantime, I'm double-checking all my wiring.  My experience (successfully) building a test-bed tells me that it's all too easy to goober up connections.

Bill

On 9/26/21 10:45 AM, wirefordcc via groups.io wrote:

Hi Bill,
If you haven't tried it already, try putting more than one turn through your BD20.  Literally just loop it around and go through it again.  This increases sensitivity of the BD20.  I don't think you need multiple detectors in a block.  Check the BD20 instructions for multiple turns through it.

Allan Gartner
Wiring for DCC

--
Dave McBrayer
Auburn, CA


Bill Wilken
 

I've been pretty scrupulous about having feeders attached to every section of track.  For my amusement this afternoon, I'll be taking a close look at every wire connection.  Unfortunately, I wasn't sufficiently prescient to build my layout 56" off the floor.  Thinking about kids, I opted for 48", which makes every venture beneath my track work decidedly unpleasant.  Indeed, "build your benchwork 56" high" ought to be the first commandment in any DCC how-to-do-it manual!

On 9/26/21 12:47 PM, wirefordcc via groups.io wrote:
Bill,

I believe you said that your BD20 was not detecting a train in the block.  The pot is for desensitizing the BD20.  So that may not help you.

I did notice in the BD20 instructions where it says it may not work on long blocks.  But it referenced false detections due to couple capacitance and glue and ballast.

Is your long block fed every section or two of track to your bus wires.  Note:  One of your bus wires should be connected to a wire going through your BD20.  Make sure you don't have another connection (called the goober connection!) going to your booster that could be bypassing your BD20.

Those are the obvious things I can think of.

Allan


Don Weigt
 

Bill,

There's often confusion between the terms wire, cable, and pairs. Many people call a multi-conductor cable a "wire", when it's a group of wires. Also, "speaker wire" is really a pair, as is "zip cord." They would usually carry the power and the return; currents in opposite directions, whose magnetic fields approximately cancel.

The BD20 works by detecting the magnetic field created by train current going through the wire threaded through the hole in the detector. That wire becomes part of a transformer. If you put both the power and return wires through the BD20, the currents, going in opposite directions, will cancel each other's magnetic fields. You want only one or the other wires, not both, wound on the BD20. Perhaps that's your problem? Note in the wiring diagram it shows only the power feeder wire, not the return wire, routed through the sensor hole.

Nothing is said about needing a power supply for basic detection. But, I wonder if one might be needed. I've never used a BD20, so I don't know. It would hurt nothing to do this test: Connect a 5V power supply between terminals 1 and 4, as called out for using the onboard LED or an external relay. Make no other changes. With the 5V power on, does the detector work, and not without it?

Don Weigt
Connecticut

--
Don Weigt
Connecticut


Bill Wilken
 

Don,

I know that adding a 5-volt connection to posts 1 and 4 on a BD20 will permit the onboard LED to light or not depending on block traffic.  I've also read, as you suggest, that it could impact BD20 sensitivity, but I've never found anything to that effect in NCE's documentation.  In any event, it's worth a shot.

Thanks for your feedback.

Bill


On 9/27/21 9:10 AM, Don Weigt wrote:

Bill,

There's often confusion between the terms wire, cable, and pairs. Many people call a multi-conductor cable a "wire", when it's a group of wires. Also, "speaker wire" is really a pair, as is "zip cord." They would usually carry the power and the return; currents in opposite directions, whose magnetic fields approximately cancel.

The BD20 works by detecting the magnetic field created by train current going through the wire threaded through the hole in the detector. That wire becomes part of a transformer. If you put both the power and return wires through the BD20, the currents, going in opposite directions, will cancel each other's magnetic fields. You want only one or the other wires, not both, wound on the BD20. Perhaps that's your problem? Note in the wiring diagram it shows only the power feeder wire, not the return wire, routed through the sensor hole.

Nothing is said about needing a power supply for basic detection. But, I wonder if one might be needed. I've never used a BD20, so I don't know. It would hurt nothing to do this test: Connect a 5V power supply between terminals 1 and 4, as called out for using the onboard LED or an external relay. Make no other changes. With the 5V power on, does the detector work, and not without it?

Don Weigt
Connecticut

--
Don Weigt
Connecticut


Martin Piech
 

I have heard that adding a resistor to a freight car axle  and/or caboose also helps with detection. Train length will dictate the number of resistors needed depending upon block length and applications (signals, grade crossings, etc).


On Sep 27, 2021, at 9:23 AM, Bill Wilken <bill.wilken@...> wrote:



Don,

I know that adding a 5-volt connection to posts 1 and 4 on a BD20 will permit the onboard LED to light or not depending on block traffic.  I've also read, as you suggest, that it could impact BD20 sensitivity, but I've never found anything to that effect in NCE's documentation.  In any event, it's worth a shot.

Thanks for your feedback.

Bill


On 9/27/21 9:10 AM, Don Weigt wrote:

Bill,

There's often confusion between the terms wire, cable, and pairs. Many people call a multi-conductor cable a "wire", when it's a group of wires. Also, "speaker wire" is really a pair, as is "zip cord." They would usually carry the power and the return; currents in opposite directions, whose magnetic fields approximately cancel.

The BD20 works by detecting the magnetic field created by train current going through the wire threaded through the hole in the detector. That wire becomes part of a transformer. If you put both the power and return wires through the BD20, the currents, going in opposite directions, will cancel each other's magnetic fields. You want only one or the other wires, not both, wound on the BD20. Perhaps that's your problem? Note in the wiring diagram it shows only the power feeder wire, not the return wire, routed through the sensor hole.

Nothing is said about needing a power supply for basic detection. But, I wonder if one might be needed. I've never used a BD20, so I don't know. It would hurt nothing to do this test: Connect a 5V power supply between terminals 1 and 4, as called out for using the onboard LED or an external relay. Make no other changes. With the 5V power on, does the detector work, and not without it?

Don Weigt
Connecticut

--
Don Weigt
Connecticut


John Eldridge
 

Hi Bill, 
 
Have been following this topic of your “Spacing between NCE BD20 detectors” question, and have some questions that I have not seen
asked or answered. 
 
1. Are you using a NCE AIU to receive the signal from the BD20 to encode the Block Occupancy to your DCC system? If not, what is your interface?
 
2. If you are using AIU’s to read your BD20's, does the LED on the AIU for the BD20 in question NOT light, while LED’s for other BD20's do light? 
 
2. Based on your reply to Don Weigt, am I correct that you are using only Pin 2 “Logic +” terminal + Pin 1 “Ground” on the BD20 to your interface? 
 
Please advise, as I have been using BD20's connected with pins 2 & 1 only toAIU interfaces quite successfully for over a year.
 
John Eldridge
Texas


Bill Wilken
 

John,

I'm connecting the BD20s to Atlas' universal signal control board via three wires, one logic, one common, one power.  the Atlas boards in turn have two other posts that use 5 volt power to light the Atlas signal itself.

BIll


On 9/27/21 9:52 AM, John Eldridge wrote:
Hi Bill, 
 
Have been following this topic of your “Spacing between NCE BD20 detectors” question, and have some questions that I have not seen
asked or answered. 
 
1. Are you using a NCE AIU to receive the signal from the BD20 to encode the Block Occupancy to your DCC system? If not, what is your interface?
 
2. If you are using AIU’s to read your BD20's, does the LED on the AIU for the BD20 in question NOT light, while LED’s for other BD20's do light? 
 
2. Based on your reply to Don Weigt, am I correct that you are using only Pin 2 “Logic +” terminal + Pin 1 “Ground” on the BD20 to your interface? 
 
Please advise, as I have been using BD20's connected with pins 2 & 1 only toAIU interfaces quite successfully for over a year.
 
John Eldridge
Texas


John Eldridge
 

Thanks Bill, 
 
Based on your reply, confirming how this BD20 is connected, you said 3 wires: “one logic, one common, one power”.  
 
Have you measured the “power” voltage @ the BD20 terminals 1 to 4 to confirm you actually have 5V? 
 
If you do, you should connect your “logic” wire to pin 3 to take advantage of the power provided.
 
Also, are you saying that with 5V power @ the BD20 terminals, the LED on the BD20 is NOT lighting with a train in the block?

John Eldridge
Texas


Bill Wilken
 

Definitely have 5 volts.

On 9/27/21 10:24 AM, John Eldridge via groups.io wrote:
Thanks Bill, 
 
Based on your reply, confirming how this BD20 is connected, you said 3 wires: “one logic, one common, one power”.  
 
Have you measured the “power” voltage @ the BD20 terminals 1 to 4 to confirm you actually have 5V? 
 
If you do, you should connect your “logic” wire to pin 3 to take advantage of the power provided.
 
Also, are you saying that with 5V power @ the BD20 terminals, the LED on the BD20 is NOT lighting with a train in the block?

John Eldridge
Texas


Jim Betz
 

Bill,

  The typical method for detection placement is for it to be on "the power to the block"
rather than on "the power to one section of the block".  I'll take the time to describe
this further.
  Let's say you have a section of track that has one "power district" that is three
sections of track (with their own feeders) long ... we'll call them A, B, and C.
And each of them has it's own feeders all fed from the same track bus (segment).

  The proper placement of the detector (the "doughnut") is between the booster
and the 'feed' to all three segments ... not in the feed to the "B" (middle) block.
  The idea is to detect all of the power use in the entire track section.  

  Now, if your wiring is too "wimpy" (wire gauge is too small) that can also make
an entire segment (power district) undetectable because the wire is acting as a
"resistor" and limiting the current flow below the level of the BD20 to detect.
  The other problems related to the wiring itself (as have already been hinted at)
are those caused by not twisting the track bus (NOT the feeders) and getting
cross talk between parallel runs of power running to different detection districts.
  The common wisdom is to twist the wire before the detector and not after,
to use larger wire, and to -never- tie wrap or otherwise bring the wires for 
different track segments (detection blocks) running parallel to each other for
even a few inches.  The separation between track bus wires (pairs) should be
at least 3".

  Wiring for signaling is/can be "tricky stuff" and often needs to be 'adjusted'
in each situation and wiring that is perfectly good for a layout without detection
can be unusable when you start adding detection.
   If you suspect your wiring then take just one block and "make it as good as it
can be" with respect to recommended wiring practices (including wire gauge).
If that cures your problem then you know what to do for the rest of the
layout.  When the suspected problem is cross talk (aka "coupling") then
simply unpower all the wires in a bundle except one and see if that one
now works.  However, cross talk can be "cumulative" and elusive to find.
                                                                                                                   - Jim


John Eldridge
 

Do you have other BD20's connected to Atlas' universal signal control board in other areas of your layout that work?

Just this one BD20 is not working?


General
 

Buy a creeper!


thomasmclae
 

Notice what Alan said, ALL power for the track in the block must go through the BD20.
I have three BD20 on the same module, about 3" apart on separate mainlines.
Two work like a charm, one is always triggered.
I checked the wiring, and the two working BD20 are on track with a single track drop (about 6').
The not working one is connected to track with multiple track drops.
I will be investigating how to have the feed through a single drop.

Thomas
DeSoto, TX


Bill Wilken
 

Before starting my project, I set up a small test bed to make certain that I understood how to wire BD20s together with Atlas universal signal boards.  The first thing it taught me was the importance of checking and double-checking wiring connections. Unfortunately, the test bed was too small to address factors such as distance, ballasting, drops, wiring twists, etc.

On 9/27/21 1:25 PM, thomasmclae via groups.io wrote:
Notice what Alan said, ALL power for the track in the block must go through the BD20.
I have three BD20 on the same module, about 3" apart on separate mainlines.
Two work like a charm, one is always triggered.
I checked the wiring, and the two working BD20 are on track with a single track drop (about 6').
The not working one is connected to track with multiple track drops.
I will be investigating how to have the feed through a single drop.

Thomas
DeSoto, TX


Martin Piech
 

Thomas,
I’m new to all of this. The terminology I’m familiar with for powering track  is main bus and feeds. Will the BD20 work properly with a feed where there are multiple feeds off of a main bus? Further, My understanding is to have separate power districts that are isolated from each other and then blocks within a district that are also isolated. Does that have any bearing on this discussion?



On Sep 27, 2021, at 3:20 PM, Bill Wilken <bill.wilken@...> wrote:



Before starting my project, I set up a small test bed to make certain that I understood how to wire BD20s together with Atlas universal signal boards.  The first thing it taught me was the importance of checking and double-checking wiring connections. Unfortunately, the test bed was too small to address factors such as distance, ballasting, drops, wiring twists, etc.

On 9/27/21 1:25 PM, thomasmclae via groups.io wrote:
Notice what Alan said, ALL power for the track in the block must go through the BD20.
I have three BD20 on the same module, about 3" apart on separate mainlines.
Two work like a charm, one is always triggered.
I checked the wiring, and the two working BD20 are on track with a single track drop (about 6').
The not working one is connected to track with multiple track drops.
I will be investigating how to have the feed through a single drop.

Thomas
DeSoto, TX