Thanks to the links given here I pretty much understand DCC. But I have another question.
When the motor controller part of the decoded follows the directions of the cab controller for speed and makes up a variable pulse width pulse train to run the motor to control the speed, at what frequency does it do it at?
Does this variable pulse width speed control really give and allow better slow speed performance?
As a mental exercise I wondered if I could get better and more realistic slow speeds with a variable pulse width voltage to the track of my non-DCC point to point trolley. I might consider coming up with an auxiliary circuit to do that if I knew at what frequency the action of the pulses on the HO locomotion would look smooth.
I also wondered if anyone made such a power pak/throttle that used a variable pulse width for speed control rather the variable voltage. I do not need or want DCC.
Thanks:
LDBennett
PWM gives very good slow speed performance, BUT, a poor running loco via DC will not be improved by PWM.
I made my own PWM power packs many years ago using the 555 IC.
Today, one or two companies might still make a power pack with pulse power but I have not looked for those in many years. Much of the time, the pulse power was to get the loco moving and really not needed after running.
At one time, the power pack put a pulse with a constant pulse width. I think one company had the option to vary the pulse width a little.
Rich
Quote from: richg on May 30, 2012, 10:59:29 AM
PWM gives very good slow speed performance, BUT, a poor running loco via DC will not be improved by PWM.
I made my own PWM power packs many years ago using the 555 IC.
Today, one or two companies might still make a power pack with pulse power but I have not looked for those in many years. Much of the time, the pulse power was to get the loco moving and really not needed after running.
At one time, the power pack put a pulse with a constant pulse width. I think one company had the option to vary the pulse width a little.
Rich
I did a quick search. Look in the below link. Someone still sells what you would like.
http://tinyurl.com/7cduf36
Google should be your best friend. I find tons of answers on many, many different subjects and find a lot of good buys on the Internet with searches. I can even compare shipping charges.
Rich
most of the mrc power packs contain some form of pulse power. according to the block diagrams i've seen, the pulse is a constant width and amplitude. its effect fades out as the base dc voltage of the pack rises, as you turn the speed control knob. at soome point, the dc voltage equals the amplitude of the pulse, and completely cancels it out. pulse power as used in dc power packs is thus different than dc with an ac ripple riding the dc. it is more like square wave ac superimposed on the existing dc voltage.
as an example, it the amplitude of the pulse is 9 volts, and you set the dc control for 3 volts, you'd have 3 volts with the reamining 6 volts of amplitude as the pulse. at 6 volts, the pulse is 3 volts, at 9v or above the pulse is cancelled out completely.
OK. So the "pulse power" is not a variable pulse width at full voltage but a combination of DC and a pulse riding on the DC.
I would think just a variable pulse width at full DC voltage would do fine IF the frequency was high enough so the engine (or trolley in my case) did not surge with every pulses such that you could see the surging. What frequency would it take to achieve that? Anyone tried?
RICHG, is that what you made years ago with the 555 IC?
Some 50+ years ago I added a "slow switch" to my layout that gave a 50% duty cycle of DC (kind of) and it did indeed improve the slow running. I did it by half wave rectifying the 60 Hz AC then provided the resultant half wave rectified variable DC to the track. I was thinking that full voltage variable pulse width DC might do a better job. The question is what frequency works to keep the action of the trolley smooth and slow? The variable full voltage pulse width appears to be what DCC does(???). What frequency to they use?
In my situation there is no load to speak of as it is just a single trolley. I want it to go slowly at scale trolley speed. Slow speed operation of a DC motor is always hard to achieve without surging. Can the system I described make the motion of the trolley surge free or virtually?
Maybe I just have to experiment????
LDBennett
Pulse power started with low voltage AC connected to the DC via a resistor and capacitor to over come the cogging action of the motor. Not really a pulse though. It was a 120 Hz waveform.
Eventually electronics allowed a 12 volt pulse to be added to the filtered DC as many companies where making power packs with nearly pure DC. As the DC voltage was increased, the pulse would start to disappear. There where many variations on this.
I made the Pacematic power pack from a Model Railroader article and made quite a few modifications using the LM555. With a potentiometer , you can vary the pulse width a certain amount.
The link I showed before, the first one is a DC power pack with variable pulse width control.
http://www.awrr.com/throttle_Thorne.html
Rich
LD-
There's a 10-amp pulse pack here: http://www.hobbylinc.com/htm/crx/crx55401.htm
My MRC Power Command 9500 has switchable pulse power but its frequency doesn't change as far as I know.
-- D
Below is a little more DIY for DC throttles if you like to experiment.
http://home.cogeco.ca/~rpaisley4/CircuitIndex.html#Throttles
Rich
My thoughts were to emulate the DCC motor controller in the DCC trains.
According to my reading the DCC decoder drives the motor. The diagrams indicate a constant period of the wave form (frequency) with a pulse width controlled by the decoder to match the speed command from the train cab controller/throttle. At small throttle settings the period of the wave form appears to be the same as that at nearly full throttle. When at full throttle the pulses disappears and all that is left is pure DC. But the period of the pulse train appears to be constant at all other throttle settings. There appears to be NO pure DC mixed with the pulses. What is that period?
The stuff I read does not show the DCC decoder/controller mixing pure DC with the pulse as so many of the DC pulse power paks appear to do.
All I am after is smoother slow speed operation and perhaps slow speeds not attainable with a common variable voltage power pak/throttle. Does it take the mixed pure DC with a variable pulse to attain that???? or will just a variable pulse width, constant voltage, no pure DC mixed in version do the job better or as well???
Again this is just an mental exercise at this point as my layout is not yet far along enough to be able to run the trolleys for any length of time to even be able to break them in. I have run them to be assured the track has no shorts and the minimum speed was not to my liking but both trolleys are brand new and not broken in. This exercise is Plan B if I can not get the smooth low speed operation I desire after the trolleys motor are thoroughly broken in. I am days from finishing the Auto Reverse wiring and being able to run the trolleys point to point continuously to allow their motors to break-in. But I want to consider Plan B now, just in case.
Thanks to all those that offered help so far.
LDBennett
LD-
Smooth slow-speed operation is the product of both the motor and the decoder.
The decoder does vary the pulse width of the "power stroke." I believe you are correct that no pure, so-called pure DC accompanies the power pulses. It is not my understanding that the power wave is the same width at different throttle levels. On the contrary, the pulse width varies specifically to control speed/power. I do not believe there would be a way to control speed with equal length power pulses unless the waves' amplitudes vary, which, to the best of my understanding, they do not. That's the whole point of DCC: Power (speed) is controlled by varying the width of full-power pulses. Longer pulses, greater speed/power; shorter pulses, lower speed/power.
Modern electric motors, especially can motors, run at comparatively high RPMs so their operation smoothes way out when a gear box reduces the high revs to a reasonable rate. Combined with the consistently high power surges from the DCC decoder (albeit intermittently available) overall slow speed operation improves dramatically. I have even large locos which will smoothly creep as slowly as .3 smph.
I hope this clarifies things for you or, in the not unlikely event that I am talking through my hat, someone else will pick up the reins and make things right.
-- D
Some years ago way before DCC, I modified a Tyco 4-6-0 and 0-4-0 with a Sagami 5 pole can motor, flywheel and 45 to 1 gears for the 4-6-0 and 72 to 1 for the 0-4-0. Those two locos could crawl at an unbelievable slow speed using the Pulse power pack I made. I did this as an experiment.
Diesels with a good can motor and flywheels, good gearing can do the same.
As soon as I turned up the throttle, I had a 12 volt pulse and the DC would increase as I increased the throttle.
With DCC, there is a narrow 12 volt pulse which gets wider as you increase the throttle.
Below is from my NCE Power Cab.
FWD, minimum throttle.
(http://i98.photobucket.com/albums/l267/richg1998/DCC/FWDMinimumthrottle.jpg)
FWD, half throttle.
(http://i98.photobucket.com/albums/l267/richg1998/DCC/FWDHalfthrottle.jpg)
Rev. Minimum throttle.
(http://i98.photobucket.com/albums/l267/richg1998/DCC/REVMinimumthrottle.jpg)
Voltage stays the same.
Rich
Quote from: Hunt on May 31, 2012, 08:14:55 PM
Quote from: LDBennett on May 31, 2012, 11:17:05 AM
My thoughts were to emulate the DCC motor controller in the DCC trains.
. . .
Change the focus to a different technology.
Since you are leaning to using DC, not DCC, for model train control suggest MRC, Model Rectifier Corporation (http://www.modelrectifier.com/train-controls/dc-power.asp), TECH 4 DC model train control products as an option.
That is another option I have heard being used. I think one or two other companies sell a similar device.
Rich
LD-
I've been thinking about your project and I just have to ask, aren't you trying to reinvent the wheel? DCC is a mature control system which will do what you seem to want to do. Why not just use it? I appreciate the challenge of finding a new way to do something but I don't think you'll find a better way and I'm pretty sure you won't save any real money making a one-off prototype. So ... why are you doing this?
-- D
Doneldon:
There is a difference between pulse width and waveform period. Pulse width is what is varied in DCC and apparently the repeat or period for nearly full throttle and small throttle is the same. Look at the O-scope pictures by richg. Note that indeed the pulse width changes from 10us to 30us to 40us while the period of the waveform is every 60 us. That is, the pulse repeats every 60 us regardless of the pulse width. The pulse voltage is always about 12+ volts DC. This is as I expected. The frequency is a bit higher than I would have guessed.
I'm a retired electronics (digital design) engineer and figuring how things work is part of engineering. I decided to figure out how DCC worked. After studying it some I realize that the decoder probably makes equal period pulse trains of varing pulse width. Was that the key to smooth slow speed operation? That is why I visited here to see if anyone here knew. A key factor is the waveforms period or frequency. All of this is doable with a minimum of parts and I have the expertise.
My little layout is a straight section of single rail. It is a point to point city scene. I most certainly don't need DCC with all its computerized complexity. All I want is smooth slow speed operation of one trolley at a time. I have two trolleys, one with DCC and one without. The cost to use DCC would be pretty high to only get smooth slow speed operation. I know that using a pulse power will smooth out the operation from my own experience over 50 years ago. I am just trying to see if anyone else has experience using just pulse power without the mixing of DC with it as apparently several DC power paks do and determine what the period/ frequency of the waveform is.
This is not plan A (straight DC operation from a little Bachmann Power Pak) but plan B if the running of the trolleys is not smooth enough for me. I am not at the point where it is clear plan A will fail. But it does not hurt to research alternatives just in case.
LDBennett
richg:
Help me out here as I am not familiar with power paks by name.
Are you monitoring the DCC decoder controller for the motor as controlled by the NCE Power Cab? Is this the power form the engine motor sees?
If so then the motor sees variable pulse widths from virtually zero to 60us where the period is 60 us (??).
Does this result in very slow smooth engine operation?
The power pak you made years ago was straight pulse power with no DC mixed in and it gave smooth slow speed operation? Do you remember the period of the wave form? Was it about the same as the o-scope pictures? Do you happen to still have the schematic of the power pak or a link to where it might be found? I understand you added flywheels and low gearing but did you happen to test the setup without the mechanical changes? Was it smooth at low speeds?
Thanks,
LDBennett
The Scope test probe was right on the motor leads for the DCC test. It should be very clear, the width of the pulse determines the motor speed. That is typical for DCC operation. I do not remember what the frequency is. It was not important to me. I do know original decoders had a lower freq. And could cause motor noise but the newer generation decoders changed all that.
My locos start moving at around 0.7 to 0.9 mph. Some can do better than that but the decoders need a lot of fine tuning, sometimes.
The DC power pack had a pulse as soon as I moved the pot a little but the DC level was very low. It is a combination of the pulse to overcome the cogging of the motor. All that I have read back in the 1970's pulse power was to overcome the cogging as was not needed after the loco started to me. The design I used just kept the pulse in operation.
I am not at hope until Sunday night to check my power pack.
I will put the schematic together that I built but I already posted a link to this which is quite similar as it is by the same author, Peter Thorne. Did you look at the links? You did not give any indication if you did.
I do not use this power pack anymore. I only refer to it when someone has a question about DC with pulse power. I and many others have found that DCC is much better.
The locos I modified where old worm to worm gear, about 33 to 1 ratio Tyco's.. I used NWSL Sagami parts, A worm, worm gear and two spur gears My original intent was to experiment with modifying a loco for better DC operation in the 1970's. I don't remember what the start speed of the locos was. I do remember being very satisfied with the results. I do remember having to play with the 555 pulse width/period to achieve no motor noise. Also, minimum gear lash while not being too tight for no gear noise.
I also modified the 10k pot with a resistor to be able to use more of the pot for the 0-4-0 that had 72 to 1 ratio.
This is old school stuff. I am not interested in arguing with anyone. It worked for me at the time. I like to experiment anyway.
Sometime, check your locos for gear reduction as an experiment. I do know the Spectrum 0-6-0T has excellent gear reduction.
Rich
richg(Rich):
I hope you don't think I am arguing with you. I am just seeking information. So far you have offered very good info indeed and the O-Scope pictures were excellent. Thank you!
OK. The DCC motor controller is getting 12 volt pulses from virtually zero pulse width to 60 us with a period of the wave form of 60us. Got it. Thanks for the info.
That means the frequency is about 16KHz if the O-Scope pictures are correct. The Thorne link schematic seems to be operating at about 150 Hz. It also mixes the pulse with DC. Unless that makes the engines run smoother I don't understand why they do it that way. It appears that is a common theme in these older pulse DC power paks (??). I'd be interested in seeing your schematic if it was just pure pulses.
I got my layout wired last night but have to set up the train detectors for the Auto Reverse before I can run the trolleys to see if they will run smoothly at slow speeds after they break in. I'll be busy all weekend so it will have to wait until next week.
Again, thank you. I really appreciate your efforts.
LDBennett (Lynn Bennett)
Caution - long, boring technical comment follows. READ AT YOUR OWN RISK of falling asleep.
Some of the better decoders implement CV 9 which is used to set the total pulse width modulation period, that is, the on time plus the off time. This is the inverse of the pulse repetition rate, a.k.a. frequency. With these decoders, you can set the frequency for the best compromise between low speed performance and noise.
The reason that decoders use pulse width modulation rather than pulse rate modulation is that pulse width modulation will work with all motors while pulse rate modulation, using fixed width pulses, must to tuned to the motor for best operation. This is particularly true of older or cheaper motors with a lot of cogging. It the pulses are too narrow, a slow repetition rate will not be able to kick the motor through a cog and it will just vibrate until the pulses are frequent enough to cause jack rabbit starts and poor or nonexistent low speed performance. With pulses wider than necessary, the motor will advance the locomotive in large steps, or at least, steps larger than really necessary. With pulse width modulation, you can turn up the speed control until the pulses are wide enough to kick the motor through a cog. An ideal solution that was, to my knowledge, never implement commercially but was used by some home builders had both a repetition rate control. which served as a speed control, and a pulse width control. The idea was to set the repetition rate to some low value, then advance the pulse width control until each pulse was just wide enough to advance the locomotive a tiny amount. Thereafter, the pulse repetition rate control was used as the throttle.
With cheap motors, the sort used in the seventies to power train set locomotives, pulses in the 5 to 10 millisecond range were required. For better motors with the armatures more tightly coupled to the fields and five instead of three poles, pulses of around a millisecond were long enough. Really good can motors with five poles and skewed windings that appeared in the eighties needed as little as 100 microseconds pulses. Having to tune the controller to each locomotive separately was a bit of a pain but it meant that cheap, train set locomotives that normally started at about 10 smph could be reliably run at one tie per second, or less than 1 smph. This meant that just about any locomotive, including the brass clunkers of the day, could be used for realistic switching.
Nowadays, just about any decent locomotive can do the same, even on pure dc. The use of pulse width modulation in DCC decoders is simply a way of reducing decoder heating. With ideal transistors driving a motor, the heating in those transistors is E (voltage across the transistor) times I (current through the transistor. When the pulse is turned on, there is no voltage across the transistors so E * I is zero. When the pulse is turned off, there is no current through the transistors so E * I is again zero. If E * I is always, zero, then there is no heating in the transistors. Of course, the transistors are NOT ideal and so there is some voltage across the transistors when they are on and there is both voltage across the transistors and current through them during the finite time it takes the transistors to switch from on to off and vise versa. This means the transistors do have to dissipate some power, but only a very small fraction of what they would have to dissipate if used to control pure dc.
Jim
Hunt:
Review of the richg O-Scope photos reveals a frequency of about 16KHz and that is close enough to 22 Khz to be the same. The 150 Hz pulse power pak in the Thorne link is not even close and implementation of it might be problematical. If I attempt this pulse power scheme then it will be at the 16 to 22 KHz frequency, not 150 Hz.
Jim Banner:
Thank you for the good explanation. It kind of makes me believe that after I break-in my two trolleys I may be able to operate them smoothly without pulse power.
This exercise has been very educational for me and I appreciate all the good info provided by everyone. Knowing how DCC works takes away all the mystery of DCC. If I ever expand to a bigger layout (probably not in the future as I am space limited and age limited) I most certainly would implement DCC at the highest level as it is surely a neat system. For now DC control ((either pure DC or maybe (??) pulse control)) is all I need or want.
LDBennett
I finally got the little layout all wired with the Circuitron auto reverse controller and the intermediate train detectors that allow stops along the way with a delay. The Bachmann PCC Street Car runs smoothly but only moderately slowly while the Spectrum trolley will run slower than a tie per second (plenty slow enough). My concerns about slow smooth running are for naught as both satisfy me.
My initial test was with the Spectrum Birney trolley with the DCC controller still in place and the little trolley was balky for starts. After I removed the DCC board and installed the supplied dummy plug the trolley seemed to perform better at slow speeds. But I did spend a couple of hours on each trolley letting it run continuously to break them in. I think that might have had something to do with the increased smoothness too.
Anyway there is NO pulse power in my future. All is well as is. What is remarkable to me (apparently a product of current motors having more poles than my trains of 50+ years ago) is the smoothness at slow speeds of both these trolleys but the Spectrum Birney is superior to the Bachmann PCC Street Car in that respect.
My idea was to kind of create my remembered environment of the 1950's with the Los Angeles Yellow and Red cars (it does not have to be exact and the Birney is wrong for LA but it is red). I got so stoked after running the two trolleys that I ordered the Spectrum LA Lines Peter Witt street car.
Thanks again all for the help.
LDBennett