PSXX - AR Current Limit/Trip Value

[mmagliaro]

Experiment I'd like to see:
Forget the engine.  Make up a little circuit consisting of a 2000 uF cap with a diode and 100 ohm resistor in series with one leg of the cap.  The diode is to make sure you don't hit the cap with the wrong polarity and blow it up.  The resistor is to limit the current when the cap charges.  Essentially, this is your "keep alive" all by itself.

wire 1 --- + CAP -  --->|-- --- resistor ---  wire 2
                                diode

Make sure the cap is fully discharged by shorting its leads with a wire briefly.  Then touch the two wires to your rails and see if this same phenomenon occurs.  That will eliminate the engines, motors, the decoders, etc.   I have a feeling this will cause the problem all on its own.

[peteski]

Experiment I'd like to see:
Forget the engine.  Make up a little circuit consisting of a 2000 uF cap with a diode and 100 ohm resistor in series with one leg of the cap.  The diode is to make sure you don't hit the cap with the wrong polarity and blow it up.  The resistor is to limit the current when the cap charges.  Essentially, this is your "keep alive" all by itself.

wire 1 --- + CAP -  --->|-- --- resistor ---  wire 2
                                diode

Make sure the cap is fully discharged by shorting its leads with a wire briefly.  Then touch the two wires to your rails and see if this same phenomenon occurs.  That will eliminate the engines, motors, the decoders, etc.   I have a feeling this will cause the problem all on its own.

Please note that Max's circuit is different than the stay-alive circuit you use.

I was going to suggest just placing a 100 ohm resistor across the track.  It is not polarity dependent, and the current passing through it will be just as great as if it was in the capacitor circuit.  But do not leave the resistor connected for anything but brief period  It will get rather hot quickly.

[Chinapig]

Luckily, I make my stay-alives easily detachable from the locos by a small two-pin plug and socket.

Now here's the surprising thing - to me anyway.  The loco and decoder (with stay-alive removed) still trips the PSXX.

I've also tried to trip the PSXX with (1) a wagon equipped with a resistor wheelset and (2) a track current detector (basically an LED plus resistor) and neither tripped the circuit.

What does one make of that?

[peteski]

It really makes no sense to me, but while I'm familiar with electronic circuits, I am not intimately familiar with DCC auto-reversers.   Almost seems like this one is too sensitive (defective?)   The trip current stated in its instructions and the actual trip current are *WAY* different.

You also didn't expand on the behavior.  When you place the loco on the reversing section, what exactly happens?  Does it just flip polarity, or does the breaker activate and totally cuts all power to the section?  If it does cut power, does it try to restore power after whatever time delay is  set, and then cycles through that behavior?

I'm also curious why are you placing the loco directly on the reversing section?  That is not the intended way for the reverser to operate.  How does it behave when you operate the setup in the way it is intended (simple drive the model to enter and exit the loop).

[Chinapig]

The loops form a storage section for our club's modular layout hence the need to place locos directly onto it.  We've used the previous PSX-AR before and I don't recollect this problem.

When a loco is initially placed on a loop the electrics act as if it had shorted with the adjoining section - the rail polarity changes and the turnout throws.  Then it acts as normal.  Interestingly, if you put another loco on afterwards exactly the same happens.

I'm just glad to have found a solution.

[John]

If you have a little time - lets see if we can figure it out ..

1) disconnect the PSXX completely, remove all the cars and locos - then measure the resistance across the tracks .. if all is correct you should read infinity
2) measure the resistance between each track and the end of the feeder wires.  Also cross measure - you should get close to 0 ohms on the wire connected to the rail, and infinity from the other rail .. Basically just want to eliminate any short or vampires
3) If you have another PSXX, try it - if the problem goes away - revisit the first PSXX
4) Make sure all the gaps are isolated .. then measure resistance across the gaps and cross measure across the gaps .. ie Rail A to Rail B

5) If you can share an image with us how this section of track is wired - it might be helpful ..

6) Is it possible to read backthe CVs?

[mmagliaro]

Two things I just read make me think...
Peteski suggests that the trip of the unit is reacting at a point way lower than it should be, and ....
putting a car with a resistive wheelset on the track does not cause the track.


So... maybe it is a defect in the unit itself and it really is tripping too low.
The resistive wheelset is probably something like a 10k ohm resistor, and the 100 ohm keep-alive circuit by itself does not trip.
So to test if the issue is just that it trips at much too high a resistance than it should, you could try successively smaller
resistances 100, 82, 68... etc.  And only hold them on there for an instant so they don't burn up.

A 5 amp trip point at a nominal 12v track voltage means it should no trip at anything higher than about 12/5 = 2.4 ohm  But if we find that it trips well above that, like say, 22 or 33 ohm, I'd really start to suspect that unit, especially since you could "fix" it by raising the current threshold.

[Chinapig]

I've just sent an email to Tony's Trains, the manufacturer of the PSXX units asking if they could explain what's going on.

I've got two of the units and both seem to be acting in the same fashion so perhaps there is a simple explanation.

I'll keep you posted with any updates I get.

[mmagliaro]

Should we presume from the time that has passed that you aren't getting an answer from Tony's ?

[Chinapig]

Yeah, no reply as yet I'm afraid.

[Chinapig]

Tony's Trains got back to me pretty much straight away but I've been too busy on other projects to follow it through until now.

They have said:

    The PSXX-AR current limit threshold is exceeded as shown by the flashing D24 blue LED.

    Your cap circuit creates a surge of power exceeding the Current Limit of the PSXX-AR, then D23 RED LED turns on, and the PSXX-AR cuts power, doing   its job of protecting against a current spike.

    You can program CV58 to delay the PSXX-AR reaction to a current spike

    Note the current spike still exists, the PSXX-AR has been programmed to not react for a certain period of time determined by CV58

Programming CV58 to 3 works on all of my locos except for a couple which they suggested I look at my keep-alive circuit to make sure that I've wired the current limiting resistor and diode in properly.  I'll be looking at this later.  This seems a fair comment as one of my Cab Forwards trips the system and the other doesn't - they both have the same decoder.

Another thing they suggested was setting CV49 to 3 (as long as your booster is rated to 3 amps) to increase the trip current from the default of 1.8 to 2.7 amps.

As far as I'm concerned the problem is solved and I understand enough detail to get by.

My thanks to Tony's Trains.

[mmagliaro]

Okay, something is not adding up here.  You have at most 880 x 2 = 1760 uF of keep-alive capacitance, with a 100 ohm
resistor in there.  Way back at the start of this thread, Peteski and I were discussing the surge current to charge the cap as being the thing that's tripping your PSXX.   He correctly pointed out that 100 ohms in series with that cap would limit the current
to v/100 (something like 110mA or 2x that at most), which would not trip the PSXX.  Plus, you mentioned that it tripped even when the keep-alive was removed from the engine.   So their answer isn't really explaining this, and it just boils down to "increase the trip time", and that fixes it, but we do not know why.  Maybe people will say I'm being "fussy", but that thing is tripping because there's a current spike occurring - a big one - and current spikes burn things out.

[Chinapig]

I think I understand what you’re saying but typically these problems come down to me doing something wrong – I still don’t know what it is/was and maybe I never will but at least it’s working.
I’m still going to look at it again but now I’ve got them working I must finish the remainder of the loops as I need them for the club layout.
If I find out any more on the subject I’ll certainly update this thread.