I've got an HO steam loco in which I want to install the subject lighting. I'm using d.c. on my railroad. The loco begins to move with a track voltage of 2. For realism I want the loco/tender lamps at full brightness before movement starts. Although I'd prefer to use very small, closer-to-scale size white/clear LEDs they don't come to full brightness under 3 volts so I'll use 1.5 volt grain-of-rice-size incandescent lamps.
I have another loco in which I installed this type of lighting and it works very well. In that loco I used a commercially-built lighting circuit that's encased in a potting compound so I don't know what components were used. There was no schematic diagram with it either. From review of constant lighting circuits in general I find that I'll need at least a couple of diodes to light one lamp and shut off the other depending upon the direction of travel of the loco. It would seem I would also need to somehow provide a voltage drop with resistors in a voltage divider circuit to limit the voltage to the lamps to 1.5. Since the track voltage varies, though, depending on the loco speed, a conventional voltage divider circuit wouldn't work since it uses a constant input voltage level to produce a given voltage drop.
Does someone have a schematic of a circuit that would work, including diode specs.?
Not all 'constant-lighting' solutions apply to this situation.
Personally, I strongly recommend that you stick with LEDs, and design your voltage networks and switching around them.
What I think you're saying you want is to have the lights come on at an 'earlier' stage of powerpack advance than when the trains physically move -- say at around the nominal 1.5V that the bulbs were rated for. The problem is that to maintain a consistent clamp of 1.5V above that point is going to involve more than simple diodes and a resistance network... think some designs of ESD protection, where any voltage 'higher' than 1.5V-to-the-load is shunted around the light to some sort of dissipating resistance and separate ground return, so the effective load voltage is 'clamped' at your 1.5V or whatever. A proper clamping diode and some relatively simple circuitry will provide this.
https://resources.pcb.cadence.com/blog/2020-clamp-diodes-and-transient-voltage-protection-in-circuit-boards
A diode in a usual light board may be more than just the 'one-way switch' that enables default directional lighting.
A potential issue is that it may be hard to source clamping diodes (e.g. Zener or Schottky) with a voltage as low as 1.5 but adequate ampacity for the bulb starting loads. Most uses for these are as voltage-controlled references at 3V or above, which for your purposes are worthless if the train moves by 2V...
Resistors only reduce what goes through them. To get them to approximate "1.5V" over the range of speed increase, presumably up to 12V applied to the track, would require some sort of voltage-controlled switching to multiple resistors or taps in a resistor ladder -- not simple, not cheap, and still showing stepped brightness. Theoretically you might use a varistor to do what the clamping diode does, but I think one that would be sensitive enough to work to 1.5V over the range up to 12V accurately might be expensive -- this is a cheap approach to surge and spike safety, where the voltage may run into the thousands of statvolts or whatever and the effective response time has to be in nanoseconds, but I'm not sure if practical marketed materials science has caught up to the requirements of low-voltage high-current devices...
Now, if you're going to do this, you have to start by considering whether you have a PCM power supply, and if so, what sort of waveform this puts out as you increase the speed control ... you may have visible effects of 'ripple' in your light output even at higher 'track voltage' out of some pulse=power or switched supplies unless you compensate with some appropriate filtering. At this point it may become cost-effective to consider actually using a capacitor as a full keep-alive and voltage source for the lighting, charging it quickly from whatever voltage becomes available at the track, and using its smooth DC output to run the actual lights. Here you'd implement a physical switch that turns the keepalive lighting output off when the track voltage falls below a certain value for a certain time, or just 'size' the capacitor to run out xxx seconds after the track voltage drops. That isn't difficult either, but requires parts and implementation.
Pity Randy isn't with us any longer. Who else has ideas for him?
Back in my DC days I installed several of these in locomotives. Circuitron made kits back then. I had a diagram of one written on the back of an envelope but I lost that long ago.
Try this:
http://www.mrollins.com/constant.html
Or if you want to buy something pre-made there's this:
https://www.richmondcontrols.com/constan.html
I tried to look at Paisley's site for a circuit but couldn't find anything. Maybe you can? http://www.circuitous.ca/CircuitIndex.html
Good Luck, Ed
This gets a bit confusing but I’ll give it a shot.I did this with all my locomotives before DCC. Back then I used 4 diodes in series parallel for each locomotive. Two diodes in series have a 1.4 volt drop. You put a pair of diodes in series with the motor and the 1½ volt GOW across the diodes. This drops the motor voltage by 1.4 volts.The single diodes above have .7 volts drop across them, put two sets of diodes in series with the motor and you will have constant 1.4 volts continuously across the diodes in ether directions after the locomotive starts moving. If you want the bulbs to be directional You will have to use three sets of diodes in series for a 2.1 volt drop to the motor and diodes in series with the bulbs.Three sets of diodes in series with the motor (.7 volts drop per set of diodes) and .7 volts drop to the bulb.Mel My Model Railroad http://melvineperry.blogspot.com/ Bakersfield, California I'm beginning to realize that aging is not for wimps.
Thanks, guys, for your replies.
Overmod- you mentioned that "Randy" is no longer with us. Sorry to hear that. I'm pretty sure it was he, a couple of years ago, who suggested some circuits for speed control ("throttles") using d.c., one of which I put together and have been using ever since.
Ed- you said you used commercial kits by Circuitron. The "black box" one in potting compound I have was made by Hobbytronics a couple of decades or more ago. Think they've been out of business for a while now. Could be similar circuit.
You and Mel point to circuits using only diodes to drop the track voltage to ~1.5 for the lamps. Guess I need more insight into solid-state electronics to understand how that works. I "came up" in the vacuum tube days of the '50s and '60s before solid-state had taken hold in the "general" population!
Anyway, maybe I'll try the circuit at http://www.mrollins.com/constant.html which looks like will give me what I want.
stayboltFrom review of constant lighting circuits in general I find that I'll need at least a couple of diodes to light one lamp and shut off the other depending upon the direction of travel of the loco. It would seem I would also need to somehow provide a voltage drop with resistors in a voltage divider circuit to limit the voltage to the lamps to 1.5. Since the track voltage varies, though, depending on the loco speed, a conventional voltage divider circuit wouldn't work since it uses a constant input voltage level to produce a given voltage drop.
i believe you're looking for a current limiting IC that can be put in series with the LED replacing the resistor. like the resistor, it limits the current. but unlike the resistor, it maintains the current even as the voltage increases
... as long as the voltage is above some minumum
greg - Philadelphia & Reading / Reading
staybolt Thanks, guys, for your replies. Overmod- you mentioned that "Randy" is no longer with us. Sorry to hear that. I'm pretty sure it was he, a couple of years ago, who suggested some circuits for speed control ("throttles") using d.c., one of which I put together and have been using ever since. Ed- you said you used commercial kits by Circuitron. The "black box" one in potting compound I have was made by Hobbytronics a couple of decades or more ago. Think they've been out of business for a while now. Could be similar circuit. You and Mel point to circuits using only diodes to drop the track voltage to ~1.5 for the lamps. Guess I need more insight into solid-state electronics to understand how that works. I "came up" in the vacuum tube days of the '50s and '60s before solid-state had taken hold in the "general" population! Anyway, maybe I'll try the circuit at http://www.mrollins.com/constant.html which looks like will give me what I want.
That circuit works good and is simple. It does lower total and available motor voltage a bit.
Not sure what throttles you are using, but, the best DC motor control is full voltage high frequency pulse width modulation, like the Aristo Train Engjneer or several other throttles on market from people like Ken Stapleton, search Kens Model Railroad Electronics.
The side benefit of PWM throttles is that lighting circuits see the power as full voltage as soon as any voltage is on the rails, lighting the circuit before the loco moves, almost every time with almost every circuit design.
Sheldon
Using this throttle takes care of having to wire each individual loco. Its the basic version of Sheldon's Train Engineer he mentioned. Incandescent light bulbs illuminate to full brightness before the loco starts to move.
Wire the small box in between the track and the power pack, and then use the tetherless, wireless, throttle. Its always nice to have a wireless throttle.
Hard to find these days but they do appear on the market once in a while.
- Douglas
ATLANTIC CENTRALThe side benefit of PWM throttles is that lighting circuits see the power as full voltage as soon as any voltage is on the rails, lighting the circuit before the loco moves, almost every time with almost every circuit design.
don't understand. the average voltage is still low (that's why the loco moves slowly)
gregc ATLANTIC CENTRAL The side benefit of PWM throttles is that lighting circuits see the power as full voltage as soon as any voltage is on the rails, lighting the circuit before the loco moves, almost every time with almost every circuit design. don't understand. the average voltage is still low (that's why the loco moves slowly)
ATLANTIC CENTRAL The side benefit of PWM throttles is that lighting circuits see the power as full voltage as soon as any voltage is on the rails, lighting the circuit before the loco moves, almost every time with almost every circuit design.
The lamps see the full votage pulse and light full brighness, any flicker from the pulses is so fast your eyes can't see it.
That is the simplest explaination.
The lamps are a resistive load, it still takes time for the load to colapse when the voltage is removed, by the time it starts to colapse, the next pulse is there.
The motors on the other hand are inductive loads, so they see the pulses as "average voltage" maintaining the the field current. So the pulse width determines motor speed, but the pulse makes the motor resist stalling and have a lower possible slowest speed.
Here is a picture, not a great on, of the Train Engineer throttle I use. It too is a base station throttle that requires no recievers in locos, but this one has 10 channels for multiple throttles on the same layout.
Here's the basic circuit I used for years when I was still running DC. Should be self explanitory. Constant / directional lighting for 1.5 volt bulbs ....
Mark.
¡ uʍop ǝpısdn sı ǝɹnʇɐuƃıs ʎɯ 'dlǝɥ
When you install the TE, the DC power pack is set to full throtttle. Because the TE is installed between the PP and the track, the TE calibrates the voltage that gets delivered to the track based upon how you throttle it up or down. I don't understand PWM modulation, but the "spectrum" of the voltage that lights need gets delivered full power because your power pack is set at full throttle, while the motor needs the other part of the spectrum to start moving, which it does as you continue to advance the TE throttle. Putting it into terms I can understand.
As background: The old Life Like Protos used to have a bunch of "resistors" (some other do-dad that escapes me) on their directional light boards, to hold back power to the motor as the light began to illuminate at slow speeds.
You could sort of get the bright light effect before the loco moved with just a DC power pack (and moreso if you switched to LEDs), but many complained about that slow speed lighting feature because the loco would then run considerably slower than other brands. And LL seemed to change the lighboards every six month run so locos were not real consistent with each other if you had some from various runs.
I assume this phenomenon would also happen if you built your own directional light circuit and installed the same circuit into locos with slightly different motors, but I'm no electronics expert for sure.
For a DC guy with a PWM throttle, those LL P2K lightboards were great, but it did take a few extra notches on the throttle to get them to move. For a switching layout, those LL protos were really nice.
ATLANTIC CENTRALThe lamps are a resistive load, it still takes time for the load to colapse when the voltage is removed, by the time it starts to colapse, the next pulse is there.
because the lamps are resistive, no voltage means no current (V = IR). it doesn't take any time for the "load to collapse" as with inductive loads. (inductive loads resist changes in current).
with PWM, both resistive and inductive loads see an average voltage and power (the power supply certainly does)
however, the resistance of incandescent bulbs increases as the bulb gets hot and i assume temperature and resistance will be lower (i.e. more current) even with partial voltage
using pulses, like pulse width modulation (PWM) is a common way to control the brightness of LEDs. AC dimmers work on the same principle, delaying the start of each AC cycle and power to the bulbs even though the bulb my see peak voltage
you may be right ... but i don't agree with the explanation.
DoughlessFor a DC guy with a PWM throttle, those LL P2K lightboards were great, but it did take a few extra notches on the throttle to get them to move.
There are a couple of types of 'PWM' as the term might be broadly applied to DC voltage control. We might start by trivially noting that typical model-train motors run more quickly when fed a higher voltage, but make shaft torque/'power' in proportion to amperage. Likewise that increasing voltage to an incandescent lamp of fixed wattage rating will make it burn more brightly.
If we take DC current and interrupt its flow periodically during a particular time, the effect will be as if we decreased the overall voltage. We can do this as the kind of "PWM" in DCC modulation, where nearly the full DC supply voltage is available most of the 'on' time (this is what is meant by a 'square wave modulation') or we can do it with half- or full-wave rectification of sinusoidal AC potential followed by SCR/GTO gate-controlled interruption of part of each of the rectified half-cycles, before or after the voltage has reached 'full'. The result is effectively a lower average voltage, but possibly relatively high voltage at part of each rectified half cycle. We also note that the (brief) peak voltage in the rectified sine wave is greater than the 'average' voltage, a point that becomes significant later.
For practical electronic reasons the easiest way to build a device to do this kind of 'PWM' is to have it reset itself to conduction when the driving voltage goes near or to zero, which occurs 120 times a second for full-wave rectified house power.
Now consider what a capacitor does when looking at a 'chopped' pulse from this sort of device. It charges very quickly depending on instantaneous voltage, so if the applied pulse goes up to 'peak' before being gated off, the cap will charge to that higher voltage even at 50% duty cycle (where the motor is seeing half the RMS average voltage over time)
There are some interesting (to us tech nerds at least) discussions on the Web about how certain manufacturers (MRC being prominent among them) use cheap components and 'dirty' power output to get a pulse-power effect out of a voltage-controlled motor. At least some of this serves to give a higher instantaneous voltage somewhere in the 60Hz (half-wave rectification) to 120Hz cycle, which a capacitor will gleefully use to charge and which a proper circuit can then meter into, say, a low-current lighting load like a low-voltage LED string...
Meanwhile, what modern PWM systems do is completely different from anything dependent on rectified AC power. A system like the Train Engineer takes what is basically like the DC-Link supplying traction inverters in an AC-drive locomotive and switches the available voltage potential many times a second. In the original bad old days of DCC, the primitive cost-effective decoder electronics might give an output as low as 200Hz, which will make many types of motor buzz and heat up rather than turn happily. But those days are long gone. Current state of the art is to put the PWM frequency not only out of the 'audible' range, but above the highest frequency from a good CD=quality sound system (20kHz) and at that pulse rate, the overall voltage 'seen' by the motor can be very smooth, approximating a gentle sawtooth profile around the 'desired' voltage, but the peaks may still allow high-voltage capacitor charging over time.
Overmod Doughless For a DC guy with a PWM throttle, those LL P2K lightboards were great, but it did take a few extra notches on the throttle to get them to move. That, of course, is a kludge. And not a very skillful one, either. There are a couple of types of 'PWM' as the term might be broadly applied to DC voltage control. We might start by trivially noting that typical model-train motors run more quickly when fed a higher voltage, but make shaft torque/'power' in proportion to amperage. Likewise that increasing voltage to an incandescent lamp of fixed wattage rating will make it burn more brightly. If we take DC current and interrupt its flow periodically during a particular time, the effect will be as if we decreased the overall voltage. We can do this as the kind of "PWM" in DCC modulation, where nearly the full DC supply voltage is available most of the 'on' time (this is what is meant by a 'square wave modulation') or we can do it with half- or full-wave rectification of sinusoidal AC potential followed by SCR/GTO gate-controlled interruption of part of each of the rectified half-cycles, before or after the voltage has reached 'full'. The result is effectively a lower average voltage, but possibly relatively high voltage at part of each rectified half cycle. We also note that the (brief) peak voltage in the rectified sine wave is greater than the 'average' voltage, a point that becomes significant later. For practical electronic reasons the easiest way to build a device to do this kind of 'PWM' is to have it reset itself to conduction when the driving voltage goes near or to zero, which occurs 120 times a second for full-wave rectified house power. Now consider what a capacitor does when looking at a 'chopped' pulse from this sort of device. It charges very quickly depending on instantaneous voltage, so if the applied pulse goes up to 'peak' before being gated off, the cap will charge to that higher voltage even at 50% duty cycle (where the motor is seeing half the RMS average voltage over time) There are some interesting (to us tech nerds at least) discussions on the Web about how certain manufacturers (MRC being prominent among them) use cheap components and 'dirty' power output to get a pulse-power effect out of a voltage-controlled motor. At least some of this serves to give a higher instantaneous voltage somewhere in the 60Hz (half-wave rectification) to 120Hz cycle, which a capacitor will gleefully use to charge and which a proper circuit can then meter into, say, a low-current lighting load like a low-voltage LED string...
Doughless For a DC guy with a PWM throttle, those LL P2K lightboards were great, but it did take a few extra notches on the throttle to get them to move.
That, of course, is a kludge. And not a very skillful one, either.
I'm not sure what a skillful kludge would look like since I don't even know what a kludge is.
Those deliberately slow LL P2K locos were great with the TE throttle because the locos were even slower (or could be run slower by choice) than when using the unfettered power pack. They were still slower than an Atlas. Great for a switching layout. But terrible if I wanted to consist with Atlas locos.
Is that unkludgy?
I entered the thread to simply offer a solution to buy one used throttle that seemingly would accomplish OP's goal instead of possibly building 52 circuit boards, in case he was unaware of that possible solution.
A 'kludge' is an inelegant, but operationally functional, 'solution' to an engineering problem -- though ugly to technician eyes, it works after a fashion. (There are less polite terms for 'solutions' that don't work right or fail purpose... )
What we'd need to see is how the Train Engineer produces its modulated output, and what the electrical characteristics of that output are. That will have a direct bearing both on the kinds of 'light board' the OP might use and on the various alternatives to get better lighting appearance that he might use if he adopts the TE solution.
If the electrical explanation of PWM is too arcane, I can do the 'garden hose' version I use teaching STEM students...
I don't have the time or the energy to debate why or how this works.
But it does.
A few base line facts.
The Aristo throttle uses a pulse rate well above 60 cycles, I'm not home where I have that data, and per Mr Einstein, I don't memorize stuff I can look up.
It also has a very clean square waveform, decades ago several of the large scale guys scoped the output.
I know this, 140 different locos, steam, diesel,, most made in the last 30 years, different brands, different lighting circuits, some homemade, many factory lighting boards, none with DCC decoders, all work pretty much the same on the TE throttle.
The TE throttle uses push buttons and has "steps". One or two quick taps steps you up enough to light the lights. The locos typically require about 5 steps to start to move at a speed between one and three scale mph, depending on the loco.
Some lights eventually get just a little brighter by half throttle, but I would say all are above 85% brightness when they come on and with many there is no change you can notice.
1.5 incandescent or LED seems to make no difference.
And I have allowed locos to idle with the lights on for periods as long as 20 minutes with no motor damage.
Have a nice day guys,
Overmod A 'kludge' is an inelegant, but operationally functional, 'solution' to an engineering problem -- though ugly to technician eyes, it works after a fashion. (There are less polite terms for 'solutions' that don't work right or fail purpose... ) What we'd need to see is how the Train Engineer produces its modulated output, and what the electrical characteristics of that output are. That will have a direct bearing both on the kinds of 'light board' the OP might use and on the various alternatives to get better lighting appearance that he might use if he adopts the TE solution. If the electrical explanation of PWM is too arcane, I can do the 'garden hose' version I use teaching STEM students...
ATLANTIC CENTRALI don't have the time or the energy to debate why or how this works. But it does.
And that is all that matters.
I don't know how my Blu-Ray player works, but I don't question it either.
-Kevin
Living the dream.
DoughlessOf course, never confuse inherent archane-ness with "simply too uninteresting to pay attention"...
Some people find the nuances of why stuff works interesting. Others don't care... until something doesn't 'just work'. Then it is fun to see what they do if they don't understand it... and in no few instances, when they do understand it but it doesn't respond as they understood it would.
There was a famous quote in one of the Doc Smith Lensman novels about 'women driving automobiles' -- if you know how to turn the key and pull the levers and work the pedals, you don't need to know about spark and combustion and gear surface finish.
Until it makes a funny noise and stops... and you have to decide whether to keep going or call AAA...
ATLANTIC CENTRALThe Aristo throttle uses a pulse rate well above 60 cycles,
i'll assume this works for the Aristo Throttle, but will incandescent lamps work the way you describe with any pulsed system?
gregc ATLANTIC CENTRAL The Aristo throttle uses a pulse rate well above 60 cycles, i'll assume this works for the Aristo Throttle, but will incandescent lamps work the way you describe with any pulsed system?
ATLANTIC CENTRAL The Aristo throttle uses a pulse rate well above 60 cycles,
Maybe not "any", but I would expect that any clean square waveform full voltage system would perform in a similar manner.
Sone cheapo dirty half wave pulse washed out by a capacitor at higher voltages, likely not.
But then again those cheapo throttles don't perform as well driving the motors either.
ATLANTIC CENTRALMaybe not "any", but I would expect that any clean square waveform full voltage system would perform in a similar manner...
It needs to be a high frequency. I don' tknow how high that frequency needs to be, but I do know that with at least some DCC decoders (which have a clean, full voltage, square wave output), if you leave the light wired to the motor it will vary significantly in brightness with the speed of the motor.
ATLANTIC CENTRAL would expect that any clean square waveform full voltage system would perform in a similar manner.
by square waveform full voltage, do you mean like DCC that constantly puts full voltage on the tracks alternating polarity to provide signalling (which is great for providing constant lighting)
or do you mean any approach that puts full voltage for a fraction of each cycle, a pulse. if the pulse is 10% of the cycle, the average voltage is 10% of full voltage. such a pulse is not likely to look like a square wave except at 50%.
does the Aristo Throttle just put a pulse on the track?
ATLANTIC CENTRALThe Aristo throttle uses a pulse rate well above 60 cycles, I'm not home where I have that data, and per Mr Einstein, I don't memorize stuff I can look up.
I presume we're talking about the units that go with the 2.4GHz Revolution TEs -- apparently, the in-model wireless units used 7.9kHz for the "PWC" even when sound-equipped into the end of the 2010s. The Trackside TE had a frequency of 22.96kHz, above 'CD-quality' digital sound output.
Either of these is safely above the frequency where traditional kinds of motor would suffer significant 'damage' regardless of waveform.
The older units I knew about (27MHz CB carrier range) had an interesting switch between "PWC" (their name for proprietary PWM modulation) and what they called 'linear'. This was interesting in part because ALL the output was PWC; the 'linear' setting switched in some filtration to smooth between the pulses (which I believe was just a 2mf capacitor across the DC input). There was no 'bypass' to get regulated DC from the base powerpack through the device.
We have discussed here in the past what PWM frequency the earliest systems used (ISTR it was in the 2400 to 4800Hz range but it has been a long time) and the references I now find on the older systems are coy about actually coming out and stating it, other than that the peaks (as Sheldon noted) were reasonably square and at high relative peak voltage -- important if you want light boards that will work with the construction he describes.
2400 is probably just fine for iron-core motors. It sure won't work well with coreless; even the 16.6kHz of the early Crest "HO" systems isn't quite high enough. My motor references all indicate 20kHz or better (which is, of course, what the current Tracksides provide).
gregc ATLANTIC CENTRAL would expect that any clean square waveform full voltage system would perform in a similar manner. by square waveform full voltage, do you mean like DCC that constantly puts full voltage on the tracks alternating polarity to provide signalling (which is great for providing constant lighting) or do you mean any approach that puts full voltage for a fraction of each cycle, a pulse. if the pulse is 10% of the cycle, the average voltage is 10% of full voltage. such a pulse is not likely to look like a square wave except at 50%. does the Aristo Throttle just put a pulse on the track?
The "square wave" part refers the close to 90° corners of the wave form, so a "rectangular" wave form would still be a square wave.
My understanding is that the Aristo system, including the recent high-frequency version, has reasonably good square-wave characteristics on the actual PWC pulses, which implies relatively short risetime and clean decay on each pulse.
Between the pulses the output voltage still would drop briskly to zero in the absence of filtering or 'connected equipment' effects, so the overall waveform (if you measured it with a scope) might be a bit dirty, and the 'average' voltage (that the motor speed would correspond to) would be lower if you integrated. As I noted, the "linear" setting on the earlier PWCs switched a fair-size (2000microfarad) capacitor across the DC input from the 'base' powerpack, which had the effect of slowing the decay of the sharp pulses without compromising the ability of the high pulse voltage peak to keep 'persistence of vision' lighting bright (as 'advertised' here)
CSX RobertThe "square wave" part refers the close to 90° corners of the wave form, so a "rectangular" wave form would still be a square wave.
thanks, that helps my understanding (not a technical term i've heard during my career)
so Robert, do you think low duty-cycle PWM would light an incandescent lamp very bright? you mentioned that "the light wired to the motor it will vary significantly in brightness with the speed of the motor", but i think sheldon would suggest that's because of the motor.
gregc CSX Robert The "square wave" part refers the close to 90° corners of the wave form, so a "rectangular" wave form would still be a square wave. thanks, that helps my understanding (not a technical term i've heard during my career) so Robert, do you think low duty-cycle PWM would light an incandescent lamp very bright? you mentioned that "the light wired to the motor it will vary significantly in brightness with the speed of the motor", but i think sheldon would suggest that's because of the motor.
CSX Robert The "square wave" part refers the close to 90° corners of the wave form, so a "rectangular" wave form would still be a square wave.
Do you mena a lower frequency, low duty cycle PWM. No, at least not in all circumstances. There are a lot of different factory lighting ciruits out there, and some would probably do fine, especially directional circuits with capacitors. The example I described is the light wired straight to the power pick-ups in parallel with the motor (this is what I meant by wired to the motor - normally in DCC you wire the light to the decoder seperate from the motor, which is how all of my permanent installs are done, but sometimes when testing stuff out I just leave them wired together).