I'm largely staying out of this, but:
1) PWM for control of permanent magnet motors is done for voltage control, for example as in gregc's diagram immediately preceding.
2) in DCC, this is done entirely by the decoder, following digital commands it receives. The motor is isolated from 'track power' in that mode.
3) The DCC track power is also PWM, for digital logic purposes. (As the LocoFi material alludes, this is a modulation like that of audible Morse, where the duration of 'short' and 'long' pulses chosen to be much longer than usual noise sources are chosen to represent binary states)
4) If it is not obvious by now, the actual PWM in (3) bears no relationship to any PWM in (2).
5) It should therefore be no particular surprise that a power-level PWM signal modulated for motor control, imposed in place of DCC track voltage, will be wrong for anything expecting logic signals. How wrong, might be difficult to predict, but sure as hell into the world of monkeys on typewriters inadvertently sending commands computers might recognize -- and not balancing DC charge transfer across the logic connection.
We can get to a discussion of how 'DC compatibility modes' are arranged on decoders; I would be particularly interested in reading actual details. One logical but naive approach would be to bypass the motor leads to track power while arranging voltage-to-voltage conversion and at least keep-alive power that works with whatever other decoder-based functionality is provided in DC operation. If this is done expecting a potentiometer-controlled or minimal-ripple DC voltage, perhaps assuming some superposed signal or AC modulation for 'compatible-with-DC' device control ... there may be problems if interrupted PWM DC is encountered instead. (Not with respect to the motor, which would have to be of a type that would run on the bypassed PWM voltage... oh wait, wasn't there something about coreless motors not doing well on it sometimes...)
Now I'll grant you that PWM fine motor control has been a mainstream hobby technology for so long that it would be surprising if commercial manufacturers designed things ignorant of its existence. But it may be easier just to design what is essentially a crude compatibility mode for a 'least common denominator' kind of DC control, and just forbid the wrong kind of fancy DC that causes issues.
535
greg - Philadelphia & Reading / Reading
LastspikemikeModern DC powerpacks will run modern DCC equipped locomotives just fine.
From Rapido Trains:
Do NOT use MRC 1300-series DC controllers with any of Rapido's locomotives. The RailPower 1300 is notorious for voltage spikes and it WILL destroy your locomotive. There is no “if” about it. We will try to help you if we have the parts, but we are not responsible for locomotive damage due to voltage spikes in your power supply. As well, we will not repair any locomotive damaged by an MRC 1300-series controller (or any other "train set" DC controller) unless you have retired the controller. Otherwise the damage will soon reoccur.
MRC seems to believe that Rapido is at fault:
ATTENTION: MRC has proudly manufactured Power Packs since 1947 and has sold more than 1 Million products to satisfied customers in North America without any issue during this time. Evidently, Rapido Trains, a relatively new train manufacturer has allegedly not made their locomotives suitable for use with MRC's 1300 & 1370 Power Packs, which have used the same tried and true technology for the last 25 years with UL Certification.
2¢ Ed
LastspikemikeWow, these "conversations" go off the rails so fast.
That's a problem when you choose to use many words then make statements that are misleading.
https://tstage9.wixsite.com/nyc-modeling
Time...It marches on...without ever turning around to see if anyone is even keeping in step.
rrebellDCC has a constant higher voltage that gets released to the motor in varying amounts. Since it is always on a higher voltage, it is less likely to get stopped from turning if running real slow,
That is how I understand the benefit of the PWM signal instead of conventional DC.
-Kevin
Living the dream.
What I am trying to get at, correct me if i am wrong, but in DC you vary the voltage to the motor so a motor has to start running at low voltage. DCC has a constant higher voltage that gets released to the motor in varying amounts. Since it is always on a higher voltage, it is less likely to get stopped from turning if running real slow, lets face it, in some motors stop at certain points between the poles even when they try to latch on.
Lastspikemike Only sound equipped DCC has the higher starting voltage issue.
That is not true. My Bachmann Spectrum steamers with silent decoders still require a higher starting voltage when I run them on DC.
Back to the OP's question...
rrebellKinda wondered as I just hooked up my DCC system and it runs better on DCC at very slow speeds than on DC, ran engines as compatable as I could for comparison and no dual decoder for DC mode.
DCC proponents have explained it to me like this:
The DCC decoder provides a square wave Pulse Width Modulated (PWM) signal to the motor to control motor speed. This PWM signal provides smoother starts and better low speed control than if you just hooked the motor to the wheels and ran it on a filtered DC power pack.
I will not be offended by any of the DCC people that correct me.
To the motor, this is true. But DCC always has full power on the rails.
PWM has advantages in smoother starts and running, but it does have disadvantages, primarly being much reduced starting torque, because there is only power, and thus a magnetic field, applied for a very short time when the motor is running at a slow speed. Any decent decoder though will have what the various manufacturers call torque compensation or similar to help mitigate this.
--Randy
Modeling the Reading Railroad in the 1950's
Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.
LastspikemikeModern non sound decoders will start to drive a DCC locomotive at the same or very similar low voltage regardless of whether the power controller is DC or DCC.
Actually, no. Unlike a DC power pack, DCC is always supplied to the track at full-voltage. (With my NCE Power Cab that would be 13.8V.) The DCC command station tells the decoder how often to turn the motor on and off in "ones & zeros". At low speed you have more zeros than ones; at higher speeds you have more ones than zeros. This is VERY different than DC.
Tom
I should mention I had a DC throttle from DC days but found it a waste of time to run a dual mode decoder that way. Only tried it once as I recall.
I normally used my NCE Power cab. I did notice with a Scope the motor got 12 to 14 volt pulses, PWM, The wider the pulse, the faster the motor. No pulses, motor stopped. A meter would show voltage.
Older DC only locos before DCC could crawl at maybe a couple volts real nice.
rich
If you ever fall over in public, pick yourself up and say “sorry it’s been a while since I inhabited a body.” And just walk away.
The starrting voltage to the MOTOR - thus encompassing friction in the drive train and all that, is the same, regardless if the loco is running on DC, or DCC. What changes though is what's on the track to get you to that point.
When run on DCC, there is always power in the rails. The sound circuits can come alive with the loco completely motionless. Only when given a command to move does the decoder allow any voltage to reach the motor.
When running on DC, unless you want the loco to be rolling along at half speed before it even starts making sound, the decoder artificially stops power from reaching the motor until the track voltage reaches a certain value. This will be somewhat above the voltage needed to make the cound circuit work, so you can have the loco sitting still but not going silent. That leaves you with very little range of motion of the speed knob on the DC pack to go from stop to full speed, a slight twitch can be a large change in speed, and appear to mean that the loco does not move smoothly. There are possible workarounds, but no one, at least these days, builds a DC power pack with that fine of a control.
A non sound DC loco will almost always start moving before a sound unit, simply because of the voltage needed for the sound decoder to work.
Some sound locos do allow the start voltage when running ion DC to be adjusted lower. This can give you better control, with a larger range of motion on the throttle knob, but you can also then end up with the loco set such that it will satrt moving silently, and then as you speed up, the sound will kick in.
The decoder meters voltage to the motor. The motor does not receive rail voltage when the tracks are DCC-powered. You could have 220 volts to the rails, with a properly designed decoder receiving it, and it should still only meter out small volts to the motor to start turning the drive mechanism and then to speed up as you dial in yet more voltage.
As pointed out, the decoder can compensate some, or at least SOME decoders are capable of wrestling with sticky drives, and make them run a bit smoother. I can't say what it is, maybe Randy can fill you in, but things like BEMF can help the decoder to figure out how to compensate for 'stickiness'. So, if your engine really is smoother, and it's not something else, the decoder may be helping to improve the drive response.
Seems to be some confusion here. I was running non DCC engines on a DC only layout. Then I ran a DCC and sound engine on the layout converted to DCC only. Thought because of the higher start voltage for a sound engine, it might have an effect on the motor. All I really know is it seems to run better. The two engines are of comparible quality and only the DC one is broke in so to speak.
There is no speed step with a dual mode decoder in the DC mode from what I know. The loco will not start moving untill the track voltage is about seven volts or so. It takes at least five volts for the microprocesso on the decoder to wake up.
Many have reported simiar numbers.
For some years I ran locos very well a much lower voltage using a good motor with can wheel and good gears.
Rich
It's a little different with DCC. If the loco doesn't start moving until you get to speed step 10, for example, then you only have a range of control from 10-126 (for technical reasons explained in the NMRA DCC specifications if you are curious, the "128 step" mode is really 126 steps, the other 2 are for other purposes)
If you now change CV2 in the decoder to increase the start voltage, so that the loco starts moving at step 2 instead of 10, you now have a range of control from 2-126.
If it runs way too fast at step 126 and you asjust CV5 to cut the top speed in half, you STILL have from 2-126 as a control range, just each increment will be that much finer.
Example: Out of the box, the loco starts creeping at step 10, and at full throttle, step 126, it runs 100 smph. But this is a switcher, it really shouldn't go over 50 smph wide open.
Out of the box, steps 1-9 do nothing, it's dead area. Steps 10 to let's say around 70, it runs at appropriate scale speeds. From 71-126, it goes way fatser than it should, so there are another 55 steps on the throttle that are effectively useless.
Now you adjust CV2 so it starts moving on step 1. And you adjust CV5 so the top speed at step 126 is the same speed it ran on step 70. Now you can use all 126 steps on the throttle, from a crawl to moving as fast as the loco should in a realistic manner. You're actually expanding the range of control.
For those DCC systems that can run a DC loco, there is no such adjustment. And most DC locos run poorly in this mode.
For DCC Sound locos running on DC - frankly, I'm amazed anyone considers the operation to be acceptable. All increasing the start voltage will do is increase the dead band where the sound comes on until the loco actually moves, and reduce the control range. Bare minimum for running sound IMO is having something like the MRC Tech 6 which can run plain DC locos as plain DC locos and can run a DCC Sound loco as a DCC Sound loco. Just not at the same time.
'Full voltage' to the sound decoder is always there regardless of 'polarity', so it will be 'always on' when commanded.
Meanwhile the slow-speed control is synthesized in the motor-control decoder and is not affected by anything about the DCC track voltage. Expect it to be 'tunable' across the 128 nominal steps and to be reasonably high frequency and therefore 'kind' to motors turning very slowly.
Kinda wondered as I just hooked up my DCC system and it runs better on DCC at very slow speeds than on DC, ran engines as compatable as I could for comparison and no dual decoder for DC mode.