Railpower 1300 testing

[mmagliaro]

Ah, so simulators are just that: simulators. Not perfect. Still, it is weird that the simulation of a bridge rectifier is an accurate representation of its internal components.

Well, no, they aren't perfect.  For example, you can put a component into the circuit that is horribly overloaded, and the simulator will not simulate things like thermal runaway, or the component overheating and causing an open or short circuit. 
The circuit will appear to function just fine.  In this case, one of those diodes would have "infinite" current flowing through it, but since nothing "burns out" in a simulator, it would still do its job.

There are meter and probe "tools" that you can touch to components to see what their voltage drop, current, and power are, and you can also snap permanent little meter tools directly into the circuit anywhere you want so you can monitor what happens at different test points.  But it's up to you to say, "Yikes... that diode's gonna blow!"

[peteski]

Ah, thanks for the explanation Max.  Simulator seems to be a very useful tool, as long as you remember its limitations.

[mmagliaro]

Ah, thanks for the explanation Max.  Simulator seems to be a very useful tool, as long as you remember its limitations.

It's incredibly useful.  It saves a ton of time getting a circuit "almost" perfectly right before breadboarding it or building it.
And of course, you can change components and circuit design as quickly as pointing and clicking your mouse.

[mmagliaro]

I know it's been a while, but I finally got back to this and implemented Peteski's suggested wiring modification.  It works as well in real life as it did on the simulator.

Here are photos of the mod on the actual board, so you can do this at home if you wish.

You will notice in the photos that one end of the diode and one end of the resistor are soldered right to one of the traces on the board.  Make sure you scratch the green coating off the trace at the point and expose some clean copper so you can solder to it.  Also be sure to orient the diode with the banded end in the direction shown.

The performance specs are:

At No Load:
21v maximum peaks
RMS DC 14.5v maximum

At Full Load (about 400 mA):
19v maximum peaks
RMS DC 11.98v maximum

Compare this to the stock unit which has 25v peaks, even under load. 
You could use a 16v zener and the peaks would be only 15v maximum, but your RMS DC at full load
would be limited to 10.5v.   That is not bad, but for running some DCC engines on DC, it could be
a problem.  10.5v is just not enough to get over the "decoder drop" and run a DCC engine on DC
sometimes, and after all, the whole thing that prompted this topic was the issue of running DCC
engines on this power pack.

I think the 20v zener, and 21v peaks, are the best compromise, and I'm pretty confident that knocking 4 volts off those peaks from the stock unit should do the trick.

Also, as we had hoped, the heat load is nil.  The diode no longer even gets warm even at full load.

[Point353]

Is there any confirmation that these modifications will prevent failures of the Rapido dual-mode decoders?
Was the failure mode of the Rapido decoders ever determined (and, if so, made public)?
Has @rapidotrains ever responded with that information?

[MK]

Thanks Max!  Mini summer project for me!

[mmagliaro]

Is there any confirmation that these modifications will prevent failures of the Rapido dual-mode decoders?
Was the failure mode of the Rapido decoders ever determined (and, if so, made public)?
Has @rapidotrains ever responded with that information?

No, no, and no.   No proof and no response or confirmation from anyone.   As far as I'm concerned, until  MRC or Rapido conducts laboratory tests and publishes the results to prove their point one way or the other, the only thing we have is anecdotal evidence that the decoders have failed a lot when powered by the 1300 or 1370 power packs.  And upon looking at all the scope traces of all the power packs we as a group have been able to publish in this thread, the 1300 has one stand-out feature: that it's voltage peaks hit 25v which is higher than any other pack we've seen.

The 25v peak is just my best guess at what the problem could be, given the limited testing we can do.
And since reducing that peak can be done easily, I see no reason not to do it.

If MK decides to make this mod to his power pack, I think it would very helpful if he could report back and say how he feels about the behavior of the power pack after making the mod.   i.e. whether he sees any difference in the behavior of his engines, whether he likes or dislikes the performance with the mod, or whether he really sees no difference at all (which is what I'm hoping for).

[MK]

Uh oh, Max just put me on notice!   . I'll definitely report back once I get around to it but two caveats.

A) It's not on the top of my priority list and it is summer with nice weather out so MRR takes a back seat.

B) I don't have any sound decoders (yet).  All my DCC locos are Digitrax with a handful of TCS.

Would (B) invalidate the test?

[CBQ Fan]

My new decoder arrived from BLI and it fried again in about 30 seconds at low power.  So the issue has to the entire unit.  It never ran right before in the short time it took to fry.  Email sent to BLI asking them to look at the whole unit and sort it out.  My new E6 runs sweet, and the rest of older run E units run much better now that traction tires are gone. Don’t think it is the power pack. 

[mmagliaro]

My new decoder arrived from BLI and it fried again in about 30 seconds at low power.  So the issue has to the entire unit.  It never ran right before in the short time it took to fry.  Email sent to BLI asking them to look at the whole unit and sort it out.  My new E6 runs sweet, and the rest of older run E units run much better now that traction tires are gone. Don’t think it is the power pack.

What power pack are you using when this happens?
No, I agree.  I would say that in the case you are describing, it's not the power pack and something else is going on.

[CBQ Fan]

What power pack are you using when this happens?
No, I agree.  I would say that in the case you are describing, it's not the power pack and something else is going on.

Tech 4 260

[mmagliaro]

Tech 4 260

Okay, and that one does not have 25v peaks.   It has peaks superimposed on the smooth DC, and decreases the amplitude of the peaks as you turn up the throttle so that at full speed, you have a smooth 19v DC (rather high for a the motor, but not over the DCC decoder spec).

[Point353]

Okay, and that one does not have 25v peaks.   It has peaks superimposed on the smooth DC, and decreases the amplitude of the peaks as you turn up the throttle so that at full speed, you have a smooth 19v DC (rather high for a the motor, but not over the DCC decoder spec).

MRC specifies the outputs for the Tech 4 220 and 260 models as 23 VDC.
For comparison, the Tech 4 200 is specified as 15.5 VDC and the Tech 4 280 (dual throttle) is specified as 14.5 VDC.

[mmagliaro]

MRC specifies the outputs for the Tech 4 220 and 260 models as 23 VDC.
For comparison, the Tech 4 200 is specified as 15.5 VDC and the Tech 4 280 (dual throttle) is specified as 14.5 VDC.

That might be what it says in the MRC specs, but this fellow scoped the 220, and it actually has a flat output of 19.3v DC at full throttle:   http://www.sumidacrossing.org/LayoutElectricity/ModelTrainPower/PowerPackTesting/
(scroll down and find the discussion of the 220)

Even so, 23 isn't 25.  And since decoders aren't dropping dead by the thousands every time they come near one of these power packs, whatever is going on is probably something that is right on the edge of the decoder's tolerance.  It wouldn't surprise me at all to discover that a couple more volts in those pulses is enough to occasionally make these decoders fail (at the 25v mark).

[Point353]

Since raising the initial alarm, the utter silence on this subject from @rapidotrains continues to be somewhat suspicious.