Joe, For several decades I've had fantastic memories of growing up in Bronx, NY and traveling through all 5 Burroughs by Subway and the elevated sections. I wish you would have done a complete How-to Videos from start to finish. 

Do you think you can start from where you left off to finish the rest of the project in video Please. 

Sincerely 

Allan Martinez 

That's a serious lot of hand wiring!  

If you ever go this route again, I'd consider ebay item: 251523663500, $2.13/ea

These could be used for each sense station and wired to a uP, you'd have sensing for each location with a lot less labor.

John,

Could you maybe explain how  they get wired up to the Track ?

Maybe I would like order one  and play around with it. 

Thanks, John NYC Subway Transit Signal 

Kb2agpjohn@aol.com  

The are put in series with the track power, they sense any current flow, AC or DC.

John,

Try to find them e bay the item number is not coming up.

Any other way to find them ?  

That could  work for me.  I think I will order a few and play around with them.

Thanks, again,  John NYC Transit Signals 

Don't know what you typed, but when I simply paste the number from my previous post, I get this.

I found it using the # in the ebay search box

or search by the tittle

New design 20A range Current Sensor Module ACS712 Arduino module ACS712T GOOD

John,

I'm ordering One of them. 

I will Play around with it.

I hope it works out. 

I will everyone posted . 

Thanks, John NYC Subway Transit Signal 

So basically this senses current draw when a loco is near the drop in the track? Gap the center rail and it senses the current draw in that block?

Note that GRJ is simply referring to the current sensing function of the OP's circuit.

From my understanding of the OP's circuit, the eBay current-sensing module replaces the bridge-rectifier.  As OP says, his 8-Amp bridge rectifier drops about 1.4V.  This is quite a burden; in other words if a subway car is pulling, say, 2 Amps, then the current sensor (i.e., the bridge rectifier) is burning 1.4V x 2A = 2.8 Watts!  The eBay module uses a different current sensing technology which burns less than 1% of that (i.e., only milliWatts) and does not impose a 1.4V "tax".

As I understand GRJ's comment, the eBay current sensors would be hooked to a microprocessor (Arduino) which would perform the additional signal-conditioning and timing functions presently performed by the other components of the OP's circuit.

What's nice about the "Hall sensor" sensing technology in the eBay module is it senses both DC and relatively high-frequency AC.  This means it can be used with DC track systems like in HO/G-gauge, 60 Hz AC O-gauge, and high-frequency DCC systems.  The current-transformers (doughnut-type current sensors) used in some O-gauge AC systems cannot sense DC currents and may not adequately sense the high-frequency AC currents in DCC systems.  Obviously an end-user doesn't care about a universal sensing solution - just one that works for the layout at hand - in which case if this is an O-gauge 60 Hz AC system I'd suggest looking at ~$1 eBay current transformer modules which would also burn less than 1% of the power (and no 1.4V voltage-drop tax) relative to the bridge-rectifier sensor...for example eBay 201614919336:

Stan is correct, and either suggested solution would work fine.  I have a couple of the current transformers that I picked up on eBay to tinker with, they work as advertised.  I have yet to use them in a real project, but maybe soon...

christhetrainguy09 posted:

So basically this senses current draw when a loco is near the drop in the track? Gap the center rail and it senses the current draw in that block?

Yes.  What nice about the OP's bridge-rectifier "current sensor" is that it generates a 1.4V (or so) output voltage for a wide-range of currents.  1.4V is plenty of voltage to trigger down-stream circuits for signaling or whatever.  An 8 Amp bridge rectifier is less than 50 cents on eBay:

OTOH, the current-sensor in GRJ's eBay module or the current-transformer doughnut sensor generate an output voltage proportional to current flowing in the circuit.  So if your consist consists of un-powered cars, you need to figure out a way to draw current for non-engine rolling stock which do not draw current.  Otherwise you have an occupancy detector that only detects occupancy of current-consumers (engines, lighted cars, etc.).  In HO systems this is routinely done by replacing wheel-axles in non-powered cars with wheel-axles that have a resistor built-in to draw current between the 2 rails! 

This current-sensing approach to occupancy detection is applicable to the OP's layout because it is a subway where consists (typically) have all lighted/powered cars so you are "guaranteed" to draw some current from end-to-end.  As mentioned, in HO you replace insulated axles with ones with a resistor built-in to draw some current.  In O-gauge I suppose you could replace an un-powered wheel truck with a powered wheel truck (i.e, one that has a center-rail roller/pickup) and insert a resistor, lamp, or some other electrical load that draws current.

Stan thanks, we at the club run a caboose on every train, so i think i have that covered. But i am open to ideas for the non caboose trains.

using GRJ module, simply run track power to one terminal and the power drop to the other terminal and plug the rest of the pins up to Arduino and write a little code and bazinga you have a block detection   

You can go with the old standby for block detection, the isolated rail.  Absent the isolated rail, you have to figure some way to sense cars are on the track, and current seems the only logical way except for stuff like optical sensors sprinkled around.

GRJ, i would love to do the isolated blocks but the layout was wired by a wire happy guru 20years ago and grounds are tied together all over the place. Would take too much time to redo the wiring as it stands now. So i created a photocell method works fine at evening times but during the day its haywire (environmental light) o i thought this might be a better solution. 

Well, if you can't do the insulated tracks, the optical solution is the other method that doesn't involve current sensing.  One thing you can do to improve the reliability of optical sensing is to use modulated beams and a smarter detector to sense the beam.  If you simply drive the beam with an 8khz signal, for instance, you can then have the receiver look only for a signal that's modulated at 8khz.  That eliminates the normal issues with stray light triggering the sensor.

christhetrainguy09 posted:

...using GRJ module, simply run track power to one terminal and the power drop to the other terminal and plug the rest of the pins up to Arduino and write a little code and bazinga you have a block detection   

Well, it depends on your definition of "a little code" and your definition of "bazinga"!

The IC sensor in GRJ's module uses the relatively recent ACS712 IC chip manufactured by Allego Microsystems which generates a voltage proportional to sensed current.  The scale for the 20 Amp sensor chip is 100mV per Amp.

So even if you have a caboose at the end of every consist, let's say its lamp draws 5 Watts.  At 18V command voltage, that's less than 300 mA (0.3 Amps).  The eBay ACS712 module generates "only" 100mV per Amp...so when the caboose rolls by, the output voltage to the Arduino would be a modest 100mV/Amp x 0.3 Amps = 30 millivolts!  I'm not saying 30mV is difficult to detect but to quote GRJ, no job is easier than the one you imagine someone else doing!!! 

As mentioned, the "beauty" of the bridge-rectifier detector method is that even if the current is "only" 0.3 Amps, the output of the detector will be about 1.4 Volts...or some 50 times greater than the eBay sensor!  In the OP's case, this 1.4 Volts is used to trigger an opto-isolator (which trips at about 1 Volt).

So, as GRJ points out, if you can't or don't want to use the isolated outer-rail method, you need to consider something like an IR/ITAD detector if occupancy detection is the objective.

The current sensing sounds fun and i think would work for the club application.

Thats a great idea, better than photocells, but lets not highjack the thread and steer it off topic.

Well, we're still sensing block occupancy.

This is true! Well i ordered 3 of those boards figured why not give it a try! 

 In line with the modulated optical sensor method GRJ mentions you can buy an eBay module for about $2 that consists of a modulated/pulsed infrared LED at about 40 kHz and a mating infrared detector.  This way you can operate without issue during the daytime or with bright train room lighting.

The spectre of replacing un-powered wheel trucks with powered trucks loaded by a light-bulb or resistor or whatever (to create a current load) seems like a non-starter in the O-gauge world.  Again, I think the current sensing approach is fine for the OP's subway application because all cars are lighted and/or motorized and hence draw current.

I agree Stan, the only reason i am perusing this is we run a lighted caboose at the end of every train, and the passenger trains are all lighted (being converted to LEDS).

that module looks interesting, wonder how i could get it between the rails, maybe unmount them from the PCB and sink them into the board?

Well, if your passenger cars have been converted to LEDs, it's likely they each draw far less than 5 Watts of an incandescent caboose.  For example, the Hennings LED passenger car module might be set to as little as 10 mA.  So using the ACS712 eBay current sensor module, that would generate a voltage output of only 0.1 V/A x 0.01 A = 10 milliVolts when a passenger car is running by that block sensor.  10mV is arguably difficult to easily detect.   I understand a block can be longer than 1 car length so you'd be detecting multiple cars and hence the current to be measured would be higher, but I think you get my point.

Anyway, the emitting LED and detecting sensor on the modulated eBay module are of the 5mm ilk.  So, yes, you could drill 5mm holes in your track bed or something like that.

Note that the "trick" to this modulated IR infrared occupancy detection is the integrated IR detector chip tuned to 40 kHz or whatever.

These are 3-terminal chips with the photodetector and all the electronics to detect just a pulsed/modulated IR reflection.  So if you are configuring a system/layout you only need one generator of 38/40 kHz driving multiple LEDs all pulsing at a common frequency...and one of these 15 cent IR receiver modules detecting the reflections when a block is occupied.  Note that the current detection method might generate a somewhat difficult to detect millivolt range output whereas this IR Receiver chip swings a full 5V or so between detect and no-detect!

Sounds like it might be worth my wild to rewire for isolated blocks! The photocell logic was time consuming, and the IR seems like too much tech for most of the club members. My saying "K.I.S.S"   

Agreed. I'd think the effort to create isolated outer-rail blocks vs. isolated center-rail blocks would be similar.  That is, you'll be creating gaps between track sections and running additional wires to the newly formed electrical "islands".  An obvious advantage of the isolated outer-rail method is the amount of current flowing when the block is occupied is negligible...measured in the, say, milliAmps.  The isolated center-rail method using current-sensing carries the full current of the train which can be in the Amps.  This has implications on the type of wiring used.  And has been discussed there are tradeoffs in current-sensing technique.  The outer-rail method has negligible current flow as you are simply detecting the presence/absence of the common outer-rail voltage supplied via one or more wheel axles - a relatively simple circuit.

Agreed. It's the same effort as center rail gaps. The relays we had under the layout for the first test of isolated blocks were from Azatrax, he has great products but don't think these were the right thing for this layout. 

Ive seen you're previous threads with the one channel boards for about 2$ 

I have a 8 channel Arduino relay module. Just a little confused at where to hook the wires up to. (Coffee hasn't kicked in yet) 

christhetrainguy09 posted:

...I have a 8 channel Arduino relay module. Just a little confused at where to hook the wires up to. (Coffee hasn't kicked in yet) 

Is this a question or a statement?   I seem to recall drawing up a diagram for another thread on how to use these $1/channel multi-relay modules but can't seem to quickly find it.

Anyway, if you're seeking suggestions on how to hook it up for isolated outer-rail, I suggest starting another thread to keep the hi-jacking police at bay.   What I'd want to know is:

1. What voltage is your 8-channel module (5V, 12V, etc.)?

2. Is it "hi" or "low" level triggered?  Or what is the eBay listing #.

3. These modules do not have any timing or delay to demote or mask intermittent axle contact (relay chatter) when the first few axles enter the block or when the last few axles exit the block.  Is this feature of interest - which would add, say, 25 cents per channel and assembling a few components (capacitor, etc.).

I will start another thread later tonight. That should keep the hijack police happy.  

New thread.

Hi. I see there has been an interesting discussion here. With the ACS712 I guess I can eliminate the optoislator, but what about the filters for compensating for "dirty track" that may cause flickering in my output signal?  Would I then need an Arduino to capture the output and then perform my signal logic via software?  I was once a software engineer so this is not a daunting task for me.  This was something I considered but decided against given the complexity of my signal modules and the very fact that Arduino cannot handle the number of inputs and output needed for some of my modules, let alone issues when wiring up an Arduino (i.e. solder bridges destroying inputs/outputs). What I see here is a possible software solution which I ruled out, however will the ACS712 provide the needed output voltage to drive my signal logic if I go with the hardware solution (i.e. NAND gates) that will be part of my design.  My guess is that it does not.

I know that the 1.4 Volt "tax" is a downside.  I am reluctantly willing to live with that.

Actually, the ACS712 is not a good match to your application.  It is a linear sensor which means you get a voltage output proportional to the measured current.  This comes at a cost - both the IC chip (relative to a cost of a bridge-rectifier) and, to your point, additional circuitry to convert the output to a yes/no digital signal as needed by the downstream circuitry.  It appears the 1.4V "tax" and arguably punitive power dissipation are acceptable so really all I can say is thanks for sharing! 

One observation about the thread topic of "current detectors".  From what I've seen the big effort in block current detectors is for 2-rail applications like HO where they don't have the luxury of the extra isolated rail for "free" occupancy detection.  2-rail being generally DC, the bridge-rectifier and ACS712 operate/detect DC currents as well as 60 Hz AC and high-frequency DCC.  I suspect if 2-rail had historically been 60 Hz AC, there would be more published circuits taking advantage of doughnut current-transformers as occupancy detectors without the 1.4V "tax".  Not quite apples-apples but we all have current detector circuits protecting our safety - the GFCI (Ground Fault Circuit Interrupter).  Generally implemented with a low-cost application-specific IC chip, it is a clever circuit which can detect remarkably small AC currents generating what amounts to a yes/no output.

If I was to use an ACS712, software would need to sample the output. That sample would be in the form of a loop that would check the output say 1000 times. That would replace the filter. In order to achieve the voltage needed to drive the logic, that resulting output from the loop could be fed into another softwate program that drives an Arduino output signal which could be 5 volts DC.  That output could then drive the signal logic via hardware or software. Just another way of achieving the same goal. The advantage here is that there would be no 1.4V drop. I may have considered this, but all 123 detectors are now wired and tested.