Railpower 1300 testing

[mmagliaro]

I wanted to respond to this comment from jagged ben, which I missed in all the roiling over this subject.

I do not agree that the 1300 violates the NMRA standard.  Here's the actual sentence at issue:

    Digital Decoders intended for "N" and smaller scales shall be designed to withstand a DC voltage of at least 24 volts as measured at
    the track.

The point I've been making since I chimed in on this thread is that that kind of statement must refer to the RMS voltage, not the peak voltage.

Yes, the statement in the NMRA spec refers to RMS voltage, and I'm sure that's how they intended it - that a modeler could simply measure the voltage on the rails with a meter.

The problem, however, is that DCC power supplies use square waves (as Peteski altered me to earlier).  As such, the peak voltage = the rms voltage.   So when a user puts a meter on the track on a DCC system and measures, say, 16 v, the spec would say "that's okay" because on a DCC system, that 14v would also be the peak (whether the user realizes that or not).  But in the case of the 1300, the meter would read 16v and there would be peaks of 25v.

I think this is a situation that the spec did not take into account.

[Maletrain]

Getting back to the word "spikes", I wonder how these various DC power packs behave during short-period transients associated with things like intermittent losses and reconnections of electrical contacts between track and pick-up wheels, and short-period short-circuits like occur on Atlas Code 80 switch frogs with wide wheels passing over them.

[mmagliaro]

Getting back to the word "spikes", I wonder how these various DC power packs behave during short-period transients associated with things like intermittent losses and reconnections of electrical contacts between track and pick-up wheels, and short-period short-circuits like occur on Atlas Code 80 switch frogs with wide wheels passing over them.

You don't get an inductive kick off a transformer secondary when you interrupt the load on it for an instant.  Unlike a motor winding, something about a transformer causes it to not store energy that would be released from a collapsing magnetic field that way it does in a motor.  I admit, I'm on shakey ground here as my AC and magnetics theory is not sharp, but have read this in more than one place over the years and I accept it as true.

Given that, I don't see how a spike would come out of the power pack just because wheel pickup goes intermittent.
And since the transformer secondary can't deliver more than 25 volts and there are no capacitors in the circuit, so I do not see anywhere that a "spike" can come from.

I'll take this opportunity to repeat my request for a 1370 test subject to borrow and open.  I want to investigate some things in that unit that may shed light on the 1300.

[peteski]

You don't get an inductive kick off a transformer secondary when you interrupt the load on it for an instant.  Unlike a motor winding, something about a transformer causes it to not store energy that would be released from a collapsing magnetic field that way it does in a motor.  I admit, I'm on shakey ground here as my AC and magnetics theory is not sharp, but have read this in more than one place over the years and I accept it as true.

Given that, I don't see how a spike would come out of the power pack just because wheel pickup goes intermittent.
And since the transformer secondary can't deliver more than 25 volts and there are no capacitors in the circuit, so I do not see anywhere that a "spike" can come from.

I'll take this opportunity to repeat my request for a 1370 test subject to borrow and open.  I want to investigate some things in that unit that may shed light on the 1300.

There are also components like diodes and a transistor in series with the transformer's secondary winding (coil) which also seem like they would prevent any inductive kickback voltage.  And on the DCC-equipped locomotive side of the equation, we also have rectifier diodes and couple of transistors in series with the motor's coils to minimize any inductive kickback from the motor's coils.  I'm also slightly rusty on my AC theoretical knowledge, but I don't believe that momentary loss of contact between the loco wheels and the track would generate any inductive-based spikes.

[up1950s]

I'll take this opportunity to repeat my request for a 1370 test subject to borrow and open.  I want to investigate some things in that unit that may shed light on the 1300.

Get you one and I will Paypal you what you dished out in full . When done keep it . Least I can do for all you have done for us Max .

[jagged ben]

The problem, however, is that DCC power supplies use square waves (as Peteski altered me to earlier).  As such, the peak voltage = the rms voltage.   So when a user puts a meter on the track on a DCC system and measures, say, 16 v, the spec would say "that's okay" because on a DCC system, that 14v would also be the peak (whether the user realizes that or not).  But in the case of the 1300, the meter would read 16v and there would be peaks of 25v.

I think this is a situation that the spec did not take into account.

Max, any damage that the railpower 1300 or any other DC powerpack might cause a decoder has nothing to do with a DCC waveform.  Measuring DCC waveform RMS voltage accurately is a whole other discussion. 

The way I read it, that 24VDC in the standard applies to analog mode.  If that's not what they intended, they did a very poor job communicating their intent.   But maybe there's nothing surprising about that.  All I'm saying is if they meant that 24VDC to apply to high frequency waveform peaks, they darn well should have said so.  Because only a more-than-typically-knowledgeable hobbyist would even glance  upon the thought that they might have meant that, instead of meaning an RMS i.e. 'what i measure with my non-scope meter' voltage.

[mmagliaro]

Max, any damage that the railpower 1300 or any other DC powerpack might cause a decoder has nothing to do with a DCC waveform.  Measuring DCC waveform RMS voltage accurately is a whole other discussion. 

The way I read it, that 24VDC in the standard applies to analog mode.  If that's not what they intended, they did a very poor job communicating their intent.   But maybe there's nothing surprising about that.  All I'm saying is if they meant that 24VDC to apply to high frequency waveform peaks, they darn well should have said so.  Because only a more-than-typically-knowledgeable hobbyist would even glance  upon the thought that they might have meant that, instead of meaning an RMS i.e. 'what i measure with my non-scope meter' voltage.

I didn't mean the the waveform shape had anything to do with the damage.  What I meant was, when they wrote a DCC standard stating what the maximum DC voltage should be when measured at the track, maybe they were thinking about the voltage you would see from the DCC booster.  After all, this is a DCC standard we are reading.  It would be logical to think that they were telling the manufacturers what maximum voltage the boosters should be putting on the rails.

The two lines read:

Digital Decoders intended for "N" and smaller scales shall be designed to withstand a DC voltage of at least 24 volts as measured at the track.  Digital Decoders intended for scales larger than "N" shall be designed to withstand a DC voltage of at least 27 volts as measured at the track.

You seem to feel that when they said "withstand a DC voltage", that meant analog mode (i.e. a "power pack" connected to the rails).   That could be.  But this is a DCC standard, so I would expect them to be talking about what the boosters should be putting out and what the decoders should be able to withstand.  They aren't talking about peaks or frequency.  All I read it to mean is that when you hook up your booster and put a meter on the track, you better not see more than 24 volts.

That way, the boosters and the decoders play together nice and don't damage each other. 

This whole section of the NMRA spec begins with:
"The baseline method for providing the power to operate locomotives and accessories, which shall be supported by all Digital Command Stations and Digital Decoders"

And in fact, nowhere in this spec does it say anything about what happens if you put a decoder on a non-command-station power supply, or if that is even required to be tolerated by the decoder.  I really think they were just thinking about command stations (boosters) and decoders.

So ...
When the power source is a booster, you'll see 24 volts (RMS) and that will also happen to be the peak.
When the power source is NOT a booster, but is a DC power pack, you could read 24 RMS and have peaks far higher than that.   

I think I see where you're coming from.  You are thinking about the practical application where somebody plunks a DCC engine on a track and needs to be sure that it's safe no matter what the power source is, if they measure the track and see less than 24v.  I'm not convinced that this NMRA document addresses that question.

Maybe that is what they meant.  I don't know.  I give up.   

[Doug G.]

I know how you are interpreting that sentence, but I think it's clear what they meant.  They did not mean it had to withstand "any voltage of 24 or more".   The modifier "or higher" is clumsily placed in the sentence.

It means the decoder must withstand a DC voltage of 24 or more - or in other words,  at least 24v.    If it withstands 24, it meets the spec.  If it can withstand 50, it meets the spec.  If it can only withstand 23, it does not.

Yeah, my interpretation was a result of being one vodka over the line. I got that idea in my head and it wouldn't leave. It obviously means as you described.

Doug

[nkalanaga]

I don't know if you would get inductive spikes from momentary power loss, but you will get reverse-voltage spikes, if the loco has a flywheel, or the motor has enough rotating mass to act as one.  That's why, on DC, you sometimes see the rear headlight flashing when going forwards, or vice-versa.  It isn't noticeable with bulbs, but with LEDs it shows very nicely.  However, the voltage shouldn't be higher than the track voltage, and DCC decoders don't care about input polarity, so I wouldn't think it would be a problem.

[peteski]

I don't know if you would get inductive spikes from momentary power loss, but you will get reverse-voltage spikes, if the loco has a flywheel, or the motor has enough rotating mass to act as one.  That's why, on DC, you sometimes see the rear headlight flashing when going forwards, or vice-versa.  It isn't noticeable with bulbs, but with LEDs it shows very nicely.  However, the voltage shouldn't be higher than the track voltage, and DCC decoders don't care about input polarity, so I wouldn't think it would be a problem.

When the motor is controlled by a DCC decoder, it is isolated from the track by couple diodes and transistors.  So it does not put any voltage back to the track when the DCC loco is coasting.

[mmagliaro]

I don't know if you would get inductive spikes from momentary power loss, but you will get reverse-voltage spikes, if the loco has a flywheel, or the motor has enough rotating mass to act as one.  That's why, on DC, you sometimes see the rear headlight flashing when going forwards, or vice-versa.  It isn't noticeable with bulbs, but with LEDs it shows very nicely.  However, the voltage shouldn't be higher than the track voltage, and DCC decoders don't care about input polarity, so I wouldn't think it would be a problem.

The inductive spike back from a motor when the power is cut off definitely can be higher than the track voltage - much higher, in fact.  When the motor field collapses, if there is no load across the motor terminals, the voltage of that spike rises infinitely, in theory, until it finds a pathway to ground.  A 12v motor that gets cut off from power can definitely kick back spikes that are more like 50v or even more.   In fact, good transistor throttle designs often have a heavy diode reverse-wired across the outputs so that motor spikes can't kick back into the throttle and destroy the output transistor.

[nkalanaga]

Max:  Thank you!  Would that still be the case when the spike is shorted through an LED headlight?

Peteski:  That's good to know, but what about reverse voltage to the decoder itself?  Especially in light of Max's comment about the possible voltages involved?  Or is there enough capacitance in even basic decoders to prevent momentary contact losses from affecting the motor?

In any case, this wouldn't be unique to the Railpower 1300, as dirty track/wheels are independent of the power supply.

[peteski]

Max:  Thank you!  Would that still be the case when the spike is shorted through an LED headlight?

Peteski:  That's good to know, but what about reverse voltage to the decoder itself?  Especially in light of Max's comment about the possible voltages involved?  Or is there enough capacitance in even basic decoders to prevent momentary contact losses from affecting the motor?

In any case, this wouldn't be unique to the Railpower 1300, as dirty track/wheels are independent of the power supply.

Since in DCC the inductive loads are not directly connected to the track, there will not be any spikes generated when the model momentarily loses contact with the track.  And yes, the decoders have keep-alive type capacitors in them (a decoder is a complete miniature computer with its own power supply, memory, processor and input/output devices). Bu those caps only provide milliseconds worth of voltage and that is why many DCC users add extra larger capacitance keep-alive and stay-alive circuits.

As far as blown headlight LEDs go in DC locos, most are either wired in such a a way that each LED protects the other one, or they have a capacitor wired in parallel with the LEDs to shunt any spikes.  The older yellow LED are also quite robust, but the white LEDs are fairly susceptible to reverse voltage damage. But like I mentoned, most headlight circuits are designed to withstand the inductive kickback voltage.

[mmagliaro]

Max:  Thank you!  Would that still be the case when the spike is shorted through an LED headlight?

Peteski:  That's good to know, but what about reverse voltage to the decoder itself?  Especially in light of Max's comment about the possible voltages involved?  Or is there enough capacitance in even basic decoders to prevent momentary contact losses from affecting the motor?

In any case, this wouldn't be unique to the Railpower 1300, as dirty track/wheels are independent of the power supply.

I think Peteski covered it all, but yes, the inductive spike would go back through an LED in reverse.  Most LEDs have a reverse breakdown voltage high enough that they can handle it and will not conduct and will not burn out.  But white LEDs were notorious for being damaged this way because they have lower reverse breakdown limits than other types.   It is often recommended to wire a Schottky diode in reverse across the LED to conduct the spike around the LED.  Schottky types turn on quicker than a conventional silicon diode.   Sometimes people use a little ceramic capacitor across the LED.  You just need something that will conduct more instantly than the LED to short the spike out.
In my case, I used the little ceramic cap and it does seem to have worked.  I have not had the rectifier fail since I put the cap in.

[peteski]

I think Peteski covered it all, but yes, the inductive spike would go back through an LED in reverse.  Most LEDs have a reverse breakdown voltage high enough that they can handle it and will not conduct and will not burn out.  But white LEDs were notorious for being damaged this way because they have lower reverse breakdown limits than other types.   It is often recommended to wire a Schottky diode in reverse across the LED to conduct the spike around the LED.  Schottky types turn on quicker than a conventional silicon diode.   Sometimes people use a little ceramic capacitor across the LED.  You just need something that will conduct more instantly than the LED to short the spike out.
In my case, I used the little ceramic cap and it does seem to have worked.  I have not had the rectifier fail since I put the cap in.

The often used design with white LEDs is to wire the front and rear LED in parallel but with reverse polarity, then have a single current limiting resistor in series with that pair of  LEDs.  That way the forward conducting diode protects the other diode from excessive reverse voltage (which is what damages white LEDs).  This type of design uses minimal number of components (so it is cheapest).  But some manufacturers do use the design with a capacitor in parallel with each diode where each diode is then in series with its own current limiting resistor.