rrinker The cap smooths out the pulse, and provides a bit of hysteresis so dirty wheels don;t make the detection drop in and out. The rest of it - Rob just takes advantageof the fact that a 555 can act as more or less a simple flip-flop, and at the same time provide both an open collector and bipolar output.
pins 2 and 6 on the 555 are the trigger and threshold. The 555 toggles when the voltage on the one is > 2/3 Vcc and when < 1/3 Vcc on the other. It looks like the cap (C1) charges thru a 10K resistor (R1) and has an RC time constant of ~20 ms, and that the transistor discharges the cap when a pulse occurs when the track polarity reverses.
DCC pulse are 58 or 100 usec wide. I don't know how much time there is between packets. I think 20 ms is long enough to keep the output latched between DCC packets, but I would think it should be longer for dirty wheels or > 1 sec to provide some delay before indicating the block is clear.
the 555 output becomes active as soon as a pulse is detected. looks like there is no delay.
i ordered some of the Vitec current sense transformers.
greg - Philadelphia & Reading / Reading
In Rob's circuit, the transistor portion, with the capacitor, indeed only captures every other pulse, just the positive half of the DCC signal, or the negative half, dependign on which way you wind the wire through the coil. Let's just say, the positive half relative to the diode polarity. The cap smooths out the pulse, and provides a bit of hysteresis so dirty wheels don;t make the detection drop in and out. The rest of it - Rob just takes advantageof the fact that a 555 can act as more or less a simple flip-flop, and at the same time provide both an open collector and bipolar output. For more specialized applications, it's probably possible to replace the 555 part of the circuit and use something 'simpler - although I don;t think you can get much simpler than one small cip for TWO blocks. Last time I bought any, the 556 was actually CHEAPER than a 555. The whole thing is pretty cheap. What complicates the matter is getting that output to drive something - if all you want to do it turn on an indicator when a train is in block, like for staging or progress up a helix, all well and good. But unlike some other options (for coil type detectors, something like what Dick Bronson at RR-CirKits has) which already have an interface tosome sort of cab bus, mostly all Digitrax Loconet, you need to add that part to complete a signal system. A standalone Loconet if the DCC system is somethign else is one of the most economical, far less expensive than C/MRI boards. There's always the AIU for NCE systems. I don;t think Lenz or MRC have anything conceptually like a multi-port I/O board. Or maybe the new NMRA Bus, I'll believe it when I see it, but as of today, Loconet has been around a long time, has the most 3rd party support DESPITE being a proprietary product, and has the most DIY projects
--Randy
Modeling the Reading Railroad in the 1950's
Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.
rrinkerYou can;t just repalce the bridge rectifier with the current transformer, the output of the current transformer will be AC and you'll have a high speed flashing output (assuming the optiocal coupler is fast enough for DCC frequency).
I think the transistor captures only every other pulse from the transformer. The 555 timer and the RC circuit driving it effectively latches the output in between pulses. While I'm sure the drive capability of the timer may be important for some applications, I think the delay the timer provides is key.
While the bridge circuit is simpler than the transformer circuit, there's obviously less voltage drop with the transformer.
You can;t just repalce the bridge rectifier with the current transformer, the output of the current transformer will be AC and you'll have a high speed flashing output (assuming the optiocal coupler is fast enough for DCC frequency). Rob's circuit already generates an open collector output from the 556. He has some other circuits that utilize the 55/556 as well, not just as timing circuits. Pretty ingenius.
I did build up one of Rob's detectors using a transformer that was elsewhere suggested as a suitable replacement for the one he originally used, which is no longer available. In never did work, and when finally someone dug up the specs for the VT-5, it was obvious why - the so-called 'suitable' replacement wasn't even close in rating. With the proper components, this should work well though - the sme or a similar circuit is used in any number of commercial products and I've seen those in action.
for me, building the electronics is just as interesting as building the layout and scenery. less expensive and only what I need.
i see that the Vitec current sense transformers aren't that expensive (<$2) and will look into building a detector circuit based on paisley's design. Looks like they could simply replace the bridge in teh NMRA circuit.
I use NCE BD-20's for signal detection and RR crossing flashers. The feeder for the block passes through a pickup coil and is activated by the current drawn by any load on the feeder wire.
Two diode drops is going to be about 1.4V with silicon diodes. That will probably be noticeable. A bridge wired the same way but without the detection circuit part will make the drop equal between detected and non-detected blocks so there is no speed variation.
Is this for DCC or DC? If for DCC, why not a coil type detector? The ones on Rob Paisley's site are extremely inexpensive, and there is no voltage drop, so nothing needs to be done for undetected intermediate blocks. Those old style circuits had their place, but I think the coil transformer type are much better suited. You may hav enoticed I don;t recommend people get the Digitrax BDL-168, even when they have Digitrax systems - same thing, it uses the drop across a diode - although I think it only uses 2 total diodes per block, not 4 - in fact, add a little extra to aht circuit you posted and for DCC you'd only need 2 diodes - one bridge for 2 blocks - the DCC frequency is fast enough that a little capacitance can hold up during the negative half of the signal. For DC, of course, you need detection in both directions.
the NMRA block detection circuit uses the voltage drop (~1.4V) across a brdige rectifier to detect the current flow through a block.
is this voltage drop significant enough to affect motor speed when moving from a block with a detector circuit and one without?would it make sense to put a brdige rectifer in series with the circuit(s) for blocks without detection circuits?