Now This Is Weird - A Major MRR Supplier Dissin' Another on Their Website

[thomasjmdavis]

Peteski, thanks, I think I get it now.  After reading your most recent post, I took a couple minutes to read up a bit. 

[mmagliaro]

Guys, read the NMRA standard here:
https://www.nmra.org/sites/default/files/standards/sandrp/pdf/s-9.1_electrical_standards_2006.pdf

EDIT - I PASTED IN THE WRONG SENTENCE EARLIER  THIS IS NOW CORRECT
It reads in part:
"..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. ...
(I pasted this back in because the quote of the quote didn't keep it in the forum --- Max)

So...if Walthers is now saying that the decoders can't withstand more than 18V...then...

ESU has released a decoder that doesn't meet NMRA standard. 

Now that isn't a crime, but ESU should have said so, or at least stated a max DC voltage clearly somewhere in the documentation.  And to be fair to ESU, every other DCC manufacturer so far as I've noticed is equally guilty of not specifying a max voltage.  I find that to be a travesty by the way... In what other industry do manufacturers release electronic parts where the user is expected to apply a raw voltage to the device from any source they please, but they don't specify or remind the user of the max voltage in the product documentation?    Note that the NMRA standards do not even specify a max voltage for DC track power!  So all DCC decoder manufacturers should be stating max DC voltages for every decoder that supports analog mode and so far as I've noticed none of them do, so that's problem #1.

Now let's go back to the MRC powerpack...
Max tested it and found peak waveform voltages of 25V but an RMS voltage of below the nameplate 15V.  So unless someone can show that their quality control isn't keeping the voltage there, none of this is really their fault.  After all, the decoder per NMRA standard is supposed to withstand more than 25V RMS.  I don't know why the Railpower 1300 should be getting the call out but in my opinion it's moot if no one can show that the RMS voltage is more than nameplate.

It's not just the RMS voltage on the track.  There is ANOTHER line in that same NMRA spec that says:
"In no case should the peak amplitude of the command control signal exceed +/- 22 volts"

Wouldn't 25 volt peaks from a DC power pack like the MRC 1300 violate this spec?  Don't peaks of *any* kind that appear on the rails and are above 22v violate the rules?  It doesn't matter that the MRC's peaks aren't intended to be a DCC control signal.

[C855B]

Max, I don't recall - did you look at the 'scope traces under load? I would put enough load on that power pack to be just below the protection threshold - 1-1.5A - and see what that waveform looked like. I have a working hunch that at this level slightly higher than the 7VA spec that the post-rectifier filtration will be overwhelmed and you would be seeing major transients, well above the 25V peaks.

Again, just a hunch.

[mmagliaro]

Max, I don't recall - did you look at the 'scope traces under load? I would put enough load on that power pack to be just below the protection threshold - 1-1.5A - and see what that waveform looked like. I have a working hunch that at this level slightly higher than the 7VA spec that the post-rectifier filtration will be overwhelmed and you would be seeing major transients, well above the 25V peaks.

Again, just a hunch.

I only tested the pack at full loads after I made my modifications to bring down the peaks, in order to assure myself that I wasn't pulling down the maximum RMS output too much.
Initially, I was testing just with a 60 ma  grain-of-wheat lamp load.

[eric220]

Surely you mean VA is the rating for DC

No, DC was carved out of MD, not VA.

[jagged ben]

It's not just the RMS voltage on the track.  There is ANOTHER line in that same NMRA spec that says:
"In no case should the peak amplitude of the command control signal exceed +/- 22 volts"

Wouldn't 25 volt peaks from a DC power pack like the MRC 1300 violate this spec?  Don't peaks of *any* kind that appear on the rails and are above 22v violate the rules?  It doesn't matter that the MRC's peaks aren't intended to be a DCC control signal.

Max, the peak amplitude spec you quoted is for DCC power supplies.  There is no such max voltage for DC power supplies.     Meanwhile there is a spec for the max DC voltage that a decoder should be expected to withstand.  So no, the Railpower is not out of standard, and the ESU is.

Now, if it were shown (which it has not been) that the ESU would not die at 27V linear DC but it dies on 'pulsed' DC with peaks below 27V, then we have a deeper more mysterious problem and an inadequate standard.  But I'm very skeptical, due to physics, that this will be shown.

[jagged ben]

No, read the link I posted earlier.
Here is an excerpt:
A volt-ampere (VA) is the unit used for the apparent power in an electrical circuit. The apparent power equals the product of root-mean-square (RMS) voltage and RMS current. In direct current (DC) circuits, this product is equal to the real power (active power) in watts. Volt-amperes are useful only in the context of alternating current (AC) circuits.

If a manufacturer states power rating for some AC device in Watts, then they already did the conversion from VA to Watts.

Notwithstanding Wikipedia, I believe power factor of the harmonic distortion variety does have relevance when DC  has a waveform like in these PWM power packs. It would be related to the BEMF stuff.

[peteski]

Notwithstanding Wikipedia, I believe power factor of the harmonic distortion variety does have relevance when DC  has a waveform like in these PWM power packs. It would be related to the BEMF stuff.

In my dealings wit electronics I have never encountered anybody specifying power rating (in either VA or Watts) of a square wave signal.  But I'm also not saying that it can be done. VA is most commonly used with sine wave AC.

I'm a bit rusty on alternating-current theory, but I'm pretty sure that power rating of any shape alternating current wave (sine, triangular, sawtooth, square, etc.) can be described in both VA and Watts (but the equation for calculating Watts to VA ratio will depend on the shape of the AC waveform).  I think with a square wave, Watts to VA ratio will be almost 1:1.

[mmagliaro]

Max, the peak amplitude spec you quoted is for DCC power supplies.  There is no such max voltage for DC power supplies.     Meanwhile there is a spec for the max DC voltage that a decoder should be expected to withstand.  So no, the Railpower is not out of standard, and the ESU is.

Now, if it were shown (which it has not been) that the ESU would not die at 27V linear DC but it dies on 'pulsed' DC with peaks below 27V, then we have a deeper more mysterious problem and an inadequate standard.  But I'm very skeptical, due to physics, that this will be shown.

Ah, I see that now.  Your point is fair.   I was reading it more with the mindset of what is "reasonable",  but what I find reasonable  isn't in there.  My point is this.  Even though a DC power pack is not obligated to keep its pulsed output peaks below the DCC signal level of 22 volts, if I were selling a DC power pack, I think that in good conscience I should keep my rectified DC pulses below 22 because who can ever say if a decoder could be confused or damaged by them?  Sure, they aren't the right shape, or frequency.  I get that.  But it certainly isn't prudent.

My common sense tells me MRC should know better than to be using 25v pulses.

By what's written in the spec, however, the obligation falls on ESU, and I do see your point there.

[learmoia]

How are we coming up with 25v pulses on a throttle that should be producing max 12v (is it 14v?)... either way..  almost double seems more of a voltage surge than a pulse.?

~Ian

[jagged ben]

In my dealings wit electronics I have never encountered anybody specifying power rating (in either VA or Watts) of a square wave signal.  But I'm also not saying that it can be done. VA is most commonly used with sine wave AC.

I'm a bit rusty on alternating-current theory, but I'm pretty sure that power rating of any shape alternating current wave (sine, triangular, sawtooth, square, etc.) can be described in both VA and Watts (but the equation for calculating Watts to VA ratio will depend on the shape of the AC waveform).  I think with a square wave, Watts to VA ratio will be almost 1:1.

I don't know why we're talking about square waves when the equipment at issue is a DC power pack.  VA came up because that's how the Railpower 1300 is marked. 

VA is apparent power, a.k.a. the product of average (RMS).  This is often not the same as real power (watts) when there are waveforms involved because if the voltage and amperage waveforms do not line up then the RMS of the product that is the product of the waveforms is a different value than the product of the RMS values of each waveform measured separately.  This phenomenon is (I believe) universal with AC motors.  I know of no theoretical reason why a DC power source with a waveform like the Railpower 1300 has wouldn't cause a similar phenomenon in a DC motor, although I also don't know to what extent that does happen in real life.  I believe I know enough to say that when decoders use pulse-width-modulation to control motor speed, and then measure the circuit voltage to find BEMF, they are measuring a very similar effect which may also involve a difference between apparent power and real power, however tiny.

So what does it mean that the Railpower 1300 lists its power in VA instead of watts?  Well, very possibly nothing, since it's a big assumption that there was any kind of thought-out decision involved.   Or maybe it could mean that the overload protection is based on devices that look at amps and volts in such a way that in effect looks at apparent power instead of real power.  Note: apparent power can be more than real power, but not less.

[peteski]

How are we coming up with 25v pulses on a throttle that should be producing max 12v (is it 14v?)... either way..  almost double seems more of a voltage surge than a pulse.?

~Ian

Ian,
this throttle's output is not filtered DC - it is unfiltered DC, which is basically like a folded over sine-wave (so, it will have a peak-to-peak voltage component).  And since the voltage is also unregulated, it will vary somewhat with the amount of electrical load connected to the output.

From the throttle analysis thread Max did, here is that revealing scope capture.


It came from this post: https://www.therailwire.net/forum/index.php?topic=44170.msg569520#msg569520

I recommend reading Max's entire lengthy thread. It is quite educational.    See https://www.therailwire.net/forum/index.php?topic=44170

Also worth remembering is that the model 1300 throttle was likely designed before DCC, and none of the electrical components used in DC models (light bulbs, motors) are very sensitive to peak voltages.

[peteski]

I don't know why we're talking about square waves when the equipment at issue is a DC power pack.  VA came up because that's how the Railpower 1300 is marked. 

Because I thought you brought that into this conversation with your statement: Notwithstanding Wikipedia, I believe power factor of the harmonic distortion variety does have relevance when DC  has a waveform like in these PWM power packs. It would be related to the BEMF stuff.

And I agree that with the 1300 we are dealing with simple sine-wave-type signal.  As for why certain manufacturers use the VA rating, while others use Watts for their AC-powered devices, I'm not sure why they do that.  I suspect it is because that device supplies unfiltered DC signal.

[jagged ben]

How are we coming up with 25v pulses on a throttle that should be producing max 12v (is it 14v?)... either way..  almost double seems more of a voltage surge than a pulse.?

~Ian

Nominal voltages typically mean RMS (average), and Max measured the Railpower at 14.5.  Your 120V outlets in your home peak at around 170V 120 times a second.  (They also cross zero 120 times a second.)  There's no problem plugging  120V appliances into that.

 It seems to me that when the NMRA says that a decoder should withstand 27V DC at the track, it is reasonable to assume they mean RMS voltage and not waveform peak, since they don't get more specific    I think a decoder built to NMRA standard should withstand more than 27V for waveform peaks.   I think that MRC is being treated unfairly here.  I also think it's the sort of thing where the NMRA would have a real purpose stepping in and defining a new standard for DC power packs that would be compatible with decoders that offer an analog mode.  Maybe the 27V requirement is also not realistic for producing economical decoders that fit in small spaces.

Thanks for posting the link Pete, I was about to go looking for it.

[jagged ben]

Btw Max...

The testing you did...

Was any of it under load?