CMI SuperBlue 156 vs. MRC ControlMaster 20
j...@redrose.net
CMI SuperBlue 156 vs. MRC ControlMaster 20
Does anyone have and use these walk-arounds? I'm considering buying
one but haven't heard much about them. The CMI is cheaper but the MRC
has been around longer. Any comments/suggestions appreciated.
Jeff, PA
j...@redrose.net
Mark A. Holmstrup
Jeff-
I have an older model of the CMI Hogger (2 of them, actually). I had to
send one unit back to the factory three times before it was properly
repaired. Talk about frustrating! Since I bought the units, I have heard
that others have experienced problems as well.
I don't have the MRC unit but have also heard that some people have had
problems with these units--I believe it had to do with the unit
malfunctioning when the blocks were accidentally bridged. Hopefully,
you'll get something better than hearsay information.
If I were to do it again, I would spring for a DCC unit. MRC has one that
would be very nice for a one or two-person layout. It's not too much more
than the MRC ControlMaster for the basic unit. Of course, buying decoders
will add to your cost but the changeover can be done in stages. You'll
have much more operational flexibility and simplify your layout wiring at
the same time.
Regards,
Mark Holmstrup
Andy Harman
j...@redrose.net wrote:
>CMI SuperBlue 156 vs. MRC ControlMaster 20
I think the experience of many is conclusive that the CMI stuff is
unreliable. I could tell you some stories, but I'll let the info flow
for a while. I highly recommend staying away from CMI (or their *3*
predecessors, Troller, Tri-Tec, and Starr-Tec).
Andy
Visit the Prototype Modelers Group Web Page at:
http://w3.one.net/~aharman/index.html
TOM HEFNER
I have one of the Hoggers that I got last year. It went dead during a show
shortly after that, but to CMI's credit I must say that I had the unit
repaired and back in my hands within a week. Since then there have been no
problems. I was told by the guy at CMI that I talked to that there was
evidently a problem with the power transistor in the early models. That
supposedly was fixed in the later units, but I imagine that there are still
some of the older ones sitting on shelves out there.
The Modular club that I belonged to purchased two of them, even after my
experience, because they work so smoothly. Those too went sour, but were
promply repaired. For those that have problems, I would recommend dealing
directly with CMI rather than going back through a dealer. Of all of the
outfits that I have dealt with, I found their customer service to be
outstanding. I also have the earlier incarnations of the Troller throttles
that have been chugging along for 15 years now. I got the Hogger because
of the memory feature of the walk-around control.
Having said all of that, I would seriously look into DCC if I were starting
out again. The cost is high for those of us with a fleet of locomotives,
but I have used the Digitrax throttles with my trains at a couple of joint
module meets and they work nice!
--
Greenbrier Valley & Potomac Rwy, (Piedmont Division)
Mark A. Holmstrup <Mark.Ho...@ost.dot.gov> wrote in article
<Mark.Holmstrup-...@152.120.98.103>...
> Jeff-
>
> I have an older model of the CMI Hogger (2 of them, actually). I had to
> send one unit back to the factory three times before it was properly
> repaired. Talk about frustrating! Since I bought the units, I have
heard
> that others have experienced problems as well.
snip-snip>
> In article <3300127e...@news.redrose.net>, j...@redrose.net wrote:
>
> > CMI SuperBlue 156 vs. MRC ControlMaster 20
> >
bv...@aol.com
>Does anyone have and use these walk-arounds?
I own a CMI hogger. It had to be sent to the manufacturer for a full voltage problem twice. The second time I asked for a new unit, the 156, they sent it out and called me to make sure it was working as expected. They treated me like a valuable customer even though they already had my money. The problem was handled profesionally and without question. What more could a person want? I also love having a walkaround throttle. If I were to do it any different I too would (will) go with dcc.
Brian O.
dh...@aol.com
>I don't have the MRC unit but have also heard that some people have had
>problems with these units--I believe it had to do with the unit
>malfunctioning when the blocks were accidentally bridged. Hopefully,
>you'll get something better than hearsay information.
>
>
Our club has been using two MRC 20's for the last few years without any problems. No more problems associated with bridging blocks (this has been "tested" more than a few times) with the 20 than any standard power pack. The MRC is just a very large DC Power Pack with a walk around (and memory for that). We use it at the corners of our layout so that folks can keep an eye on there trains and stop them if needed without the old "holler back" of a fixed throttle.
Regards,
Dave Henk
UNRL...@ix.netcom.com
Isn't the CMI SuperBlue the old Troller units reborn?
Andy Harman
UNRL...@ix.netcom.com wrote:
>Isn't the CMI SuperBlue the old Troller units reborn?
Yes, same stuff. Nice design when it works, which is rarely. Every
Troller I ever owned is in the garbage. The hand cabs had about 40-50
actual hours of operation before croaking. The TwinMomentum 5 died
(went to full voltage, a common problem reported) and was sent back
when it was one year old and repaired. A few years later after
sporatic use it went south again and Troller was OOB. Into the
dumpster it went. My 1972 MRC Throttlepack is a pile of rust but
still works fine, in comparison. Module club has two 8-amp Tri-Tec
hoggers, their electronics nut keeps them going after numerous
failures. A friend had a new CMI pack catch on fire on him; for his
efforts he received a bill for $40 in return shipping from a warranty
repair. I think he let it rot.
Notice how many used Troller, Tri-Tec, or Starr-Tec products you see
for sale at swap meets? My point exactly, they don't live long enough
to be sold "used". Unless this company has done something drastic to
improve quality control, I'd stay away from them completely.
Ken Willmott
I am researching throttle designs in preparation for (of course) my
own design. This is what makes the difference between re-inventing
the wheel and re-engineering it.
I recently completed and tested Rich Weyand's "Cooler Crawler".
I did not purchase the kit, instead I used locally available parts.
Please understand that I speak without malice, when I obligingly
report that, although this throttle may be a nice kit with plenty
of "failsafe" built in, it contains no "magic". In operation it seemed
almost identical to the MRC 1400 I used to own. I tried every
loco I could find, from balky to silky smooth, in HO and N.
In one case (N 0-6-0 steamer) the loco siezed on the track
and hummed. The same unit runs okay on regulated DC.
The ties-per-minute test was the same as with every pulse
throttle that makes this claim - it works fine as long as you
are not running between about 0.5 and 2.0 mph. Then it lurches
to a stop with a conviction dependent upon brand quality.
The locos I have that respond best to this throttle also happen
to (surprise!) run very well on pure DC. The exception was with
a Model Power FA-2. Low speed performance was amazing, but
I could not help wondering if this would not be the case when it
is run with any pulse or fullwave DC throttle. The less flywheel
mass, the better a loco seemed to work (no good for my roster!).
It has some quirks, too - big deadzone at the bottom end, cannot
supply the full voltage which the transformer is capable of, and the
current regulating circuit limits the output to about 2A, even though
the rest of the circuit seems to be designed for the advertised 6A.
Having designed and published some circuits myself, I should conclude
by saying that I recognize the enormous effort that is required to put
a kit like this together, with all the documentation, and website to
boot. I only wish to point out that the circuit design yields no
magical powers, and in fact leaves some room for improvement. I
believe that better results can be achieved with fewer components.
The key to this is *understanding* the way the output waveform affects
the engine load... if Rich Weyand's published waveforms are accurate,
there are many equivalent (but competing) methods for a circuit
implementation.
I am currently thinking about two approaches to this:
1) a variable voltage shunt regulator in series with a full wave
rectified AC supply would do the same thing better, and also
eliminating the "witchcraft" surrounding the bootstrap capacitor in
the output drive circuit.
2) a series DC regulator with pulse injected prior to the output
stage. This would smooth the medium speed performance.
I expect to put details online sometime before spring.
Ken Willmott _ _ _____ _ _ _____ _____ _ _
####################[<_(_|=@=@@>]-[<_(_|=@=@@>]-[<@@=@=|_)_>]#######
Rich Weyand
>It has some quirks, too - big deadzone at the bottom end, cannot
>supply the full voltage which the transformer is capable of, and the
>current regulating circuit limits the output to about 2A, even though
>the rest of the circuit seems to be designed for the advertised 6A.
The deadzone at the bottom of the throttle setting depends on how you
build your hand unit; a dead band resistor below the throttle pot
eliminates this dead band. The current limiting of the unit is set
by the value of the zener voltage used; my assumption is that you used
the zener value for a 2 amp limit.
>I am currently thinking about two approaches to this:
>1) a variable voltage shunt regulator in series with a full wave
>rectified AC supply would do the same thing better, and also
>eliminating the "witchcraft" surrounding the bootstrap capacitor in
>the output drive circuit.
Been there, done that. Not as good.
>2) a series DC regulator with pulse injected prior to the output
>stage. This would smooth the medium speed performance.
Been there, done that, too. Not as good.
Rich Weyand | _______ ___,---. ---+_______:_ |Rich Weyand
Weyand Associates| |_N_&_W_| |_N_&_W_| |__|________|_ |TracTronics
Comm Consultants | ooo ooo ~ ooo ooo ~ oOOOO- OOOO=o\ |Model RR Electronics
wey...@mcs.com | http://www.mcs.net/~weyand/ |wey...@mcs.com
Ken Willmott
wey...@mcs.com (Rich Weyand) wrote:
>In article <5ea1db$6u3$1...@goblin.uunet.ca> k...@sos.on.ca (Ken Willmott) writes:
>>It has some quirks, too - big deadzone at the bottom end, cannot
>>supply the full voltage which the transformer is capable of, and the
>>current regulating circuit limits the output to about 2A, even though
>>the rest of the circuit seems to be designed for the advertised 6A.
>The deadzone at the bottom of the throttle setting depends on how you
>build your hand unit; a dead band resistor below the throttle pot
>eliminates this dead band. The current limiting of the unit is set
>by the value of the zener voltage used; my assumption is that you used
>the zener value for a 2 amp limit.
Oh. I built it as specified, just to keep things honest. If I wanted
more current, though, I'd lower the value of the current sense
resistor instead. I don't recall seeing anything about alternate
zener values. Is it possible I missed part of the documentation?
>>I am currently thinking about two approaches to this:
>>1) a variable voltage shunt regulator in series with a full wave
>>rectified AC supply would do the same thing better, and also
>>eliminating the "witchcraft" surrounding the bootstrap capacitor in
>>the output drive circuit.
>Been there, done that. Not as good.
>>2) a series DC regulator with pulse injected prior to the output
>>stage. This would smooth the medium speed performance.
>Been there, done that, too. Not as good.
>Rich Weyand | _______ ___,---. ---+_______:_ |Rich Weyand
>wey...@mcs.com | http://www.mcs.net/~weyand/ |wey...@mcs.com
Rich, the failure of an experiment never proves anything - the
possibility always remains that the theory is sound, but something
got overlooked. I now have your throttle on the bench to A/B test
against anything that I come up with. Once I have a design that
I like, I'm interested in developing fair and scientific test
procedures to determine its effectiveness objectively. This
should be fairly easy if it is limited to testing the low speed
linearity.
Rich Weyand
>>The deadzone at the bottom of the throttle setting depends on how you
>>build your hand unit; a dead band resistor below the throttle pot
>>eliminates this dead band. The current limiting of the unit is set
>>by the value of the zener voltage used; my assumption is that you used
>>the zener value for a 2 amp limit.
>Oh. I built it as specified, just to keep things honest. If I wanted
>more current, though, I'd lower the value of the current sense
>resistor instead. I don't recall seeing anything about alternate
>zener values. Is it possible I missed part of the documentation?
Yup. Bear in mind that the unit you built, from the magazine article,
is not identical to the production unit. The production unit includes
a second diode bridge for a second transformer (getting to 6 amps), a
lower value for the current sense resistor (1/2 ohm instead of 1 ohm),
and several values of zener for different current values. The
production unit also includes the RC interface and the reverse-blocker,
which prevents reversing a moving train.
More info on different hand units and other enhancements were sent to
owners of record about mid last year as well. All of these things
were still in the works when the article went to press.
We are working on some other things as well, the most notable at the
moment being a permanent memory for use at open houses and on display
layouts, and direction indicator lamp connections. These are able to
be added on to current units, and owners of record will get a mailing
on this.
So there _is_ a difference between building a unit from the mag article
and building our latest production unit.
>>>I am currently thinking about two approaches to this:
>>>1) a variable voltage shunt regulator in series with a full wave
>>>rectified AC supply would do the same thing better, and also
>>>eliminating the "witchcraft" surrounding the bootstrap capacitor in
>>>the output drive circuit.
>>Been there, done that. Not as good.
>>>2) a series DC regulator with pulse injected prior to the output
>>>stage. This would smooth the medium speed performance.
>>Been there, done that, too. Not as good.
>Rich, the failure of an experiment never proves anything - the
>possibility always remains that the theory is sound, but something
>got overlooked.
Understood.
>I now have your throttle on the bench to A/B test
>against anything that I come up with. Once I have a design that
>I like, I'm interested in developing fair and scientific test
>procedures to determine its effectiveness objectively. This
>should be fairly easy if it is limited to testing the low speed
>linearity.
One thing to make sure of: you want to make sure that the locomotives
do not have other problems. How well a throttle controls a loco with
poor lubrication or some other problem is not interesting, if for no
other reason than all throttles will have poor performance with such
a beast.
Rich Weyand | _______ ___,---. ---+_______:_ |Rich Weyand
David Forsyth
>Subject: Re: Cooler Crawler - review
>Date: Mon, 17 Feb 1997 20:35:03 GMT
>wey...@mcs.com (Rich Weyand) wrote:
>>In article <5ea1db$6u3$1...@goblin.uunet.ca> k...@sos.on.ca (Ken Willmott)
>writes:
>>>It has some quirks, too - big deadzone at the bottom end, cannot
[]
>>Been there, done that, too. Not as good.
>>Rich Weyand | _______ ___,---. ---+_______:_ |Rich Weyand
>>wey...@mcs.com | http://www.mcs.net/~weyand/ |wey...@mcs.com
>Rich, the failure of an experiment never proves anything - the
>possibility always remains that the theory is sound, but something
>got overlooked. I now have your throttle on the bench to A/B test
>against anything that I come up with. Once I have a design that
>I like, I'm interested in developing fair and scientific test
>procedures to determine its effectiveness objectively. This
>should be fairly easy if it is limited to testing the low speed
>linearity.
> Ken Willmott _ _ _____ _ _ _____ _____ _ _
>####################[<_(_|=@=@@>]-[<_(_|=@=@@>]-[<@@=@=|_)_>]#######
Ken, Send me a set of circuits and etch patterns and I'll build a test
model. If it can run my Lima locos like a CC can, it is good, if it is
better, then *I'll* use it for sure. But I doubt it can be done....
David Forsyth da...@iwr.ru.ac.za
Keeper of the listserver for South African Railways railfans
TRAINS http://www.ru.ac.za/departments/iwr/staff/daf/sartrain.html
GUITAR http://www.ru.ac.za/departments/iwr/staff/daf/guitar/guitar.html
Ken Willmott
da...@SPAMLESSiwr.ru.ac.za (David Forsyth) wrote:
>In article <5eafbm$8qg$1...@nntp1.uunet.ca> k...@sos.on.ca (Ken Willmott) writes:
>> Once I have a design that
>>I like, I'm interested in developing fair and scientific test
>>procedures to determine its effectiveness objectively. This
>>should be fairly easy if it is limited to testing the low speed
>>linearity.
>> Ken Willmott _ _ _____ _ _ _____ _____ _ _
>Ken, Send me a set of circuits and etch patterns and I'll build a test
>model. If it can run my Lima locos like a CC can, it is good, if it is
>better, then *I'll* use it for sure. But I doubt it can be done....
> David Forsyth da...@iwr.ru.ac.za
Okay. Actually, when I get something done, I will be looking for beta
testers so I will bookmark you. However, I think the etch pattern
stage comes only after second-party testing. After all, my CC is built
on a piece of perfboard.
It would be good to
make one with better low speed performance than a CC, but that is
not what I'm trying to do (I'm only trying to match that). Instead,
I'm mainly interested in overcoming some deficiencies that are
inherent in the design. Lowering the output impedance and allowing a
smooth transition to DC at higher speeds are worthwhile aims that can
only be achieved by radical changes.
Less obvious reasons make this project worthwhile. For example, AC
power is not always available to a controller, as when the source is a
battery or rectified DCC signal applied to the motor through a remote
control reciever.
The feature that is most important to me, is to have some control over
the pulse amplitude. I hate that humming sound! To get an idea of how
much AC gets through to the loco with pulse, put your hand on a loco
sometime. Aaargh! I only need that to get started, then I want it to
go away. I don't know about your Limas, but my Atlas locos pull
smoother on pure DC. I can't prove it, but I'll just bet that their
drive trains will last longer, too. That heavy 120Hz energy has to be
disturbing the "wear pattern". So like many design projects, the idea
here is to get the best of both (without compromising, of course).
The "cooler" part of the CC throttle deserves the attention of all
throttle designers. This is where is sets itself above designs that
simply switch DC on and off rapidly to vary the speed (PWM). These
sharp voltage transitions produce high frequency energy in the output,
which cause some motors to overheat. The CC produces "rounded"
pulses, which have much less harmonically related HF energy.
Want to see another CC-like design? Here, a battery charger is used as
a power source, to keep electro-neophytes away from 120/240VAC.
It's simpler, and boasts an overload indicator. It's the work of Pekka
Siiskonen of the Tapiola Parish railroad club in Finland.
A detailed sketch of its design history is included. See:
http://www.mikrolog.fi/~psi/taprk/eng/pakki.html
I think where I "got off the bus" was when the claim was made, that a
circuit that would impress the same waveform on the track as a CC,
would still not work as well. This is bad science. There is nothing
that can be communicated on the track except voltage (and current, but
in this case, it's a function of load and applied voltage... hence
cannot be influenced independently by the throttle). Any difference in
drive characteristics that *could* affect motor performance *must*
show up as a different track waveform. If not, Shannon and other
information theorists failed to discover a hidden information channel.
Perhaps this could be developed further into a new kind of modem...
extra bits could be transmitted over plain telephone wires by
alternately connecting the Cooler Crawler, and the one I'm trying to
design. A different signal would appear at the reciever, even though
the line signal is invariant. But I digress...
I have recieved some really useful mail on this subject already. Let
me make it official - I am collecting all the info I can get on the
subject while experiments proceed. I want this to be a public project,
with myself as an improptu director. Anyone who participates will
recieve full credit, and the design will be strictly public domain.
The idea is not to produce a mature product, but online design
resources which would fully address the needs of DIY'ers.
The exception would be printed circuit boards, as they are much easier
to have made in quantity (and much freindlier to the environment, if
you choose your PCB manufacturer carefully). I, or someone else, might
invest in a batch and therefore justify a small markup to cover costs.
Perhaps it is too soon to worry about this.
I'm setting up a new website, and this project will be homed there
indefinitely if it shows sign of potential. If I recall correctly, the
CC design was a team effort (but offline!). They took an existing
design, and improved it. That's all I want to do, but I want to use
this amazing net technology to solicit a wider input first.
Might as well... it's too cold outside now to do any model or layout
construction around here. :-)
john purbrick
In article <5ecmcj$kfd$1...@nntp1.uunet.ca>, Ken Willmott <k...@sos.on.ca> wrote:
>da...@SPAMLESSiwr.ru.ac.za (David Forsyth) wrote:
>
>>In article <5eafbm$8qg$1...@nntp1.uunet.ca> k...@sos.on.ca (Ken Willmott) writes:
>I think where I "got off the bus" was when the claim was made, that a
>circuit that would impress the same waveform on the track as a CC,
>would still not work as well. This is bad science. There is nothing
>that can be communicated on the track except voltage (and current, but
>in this case, it's a function of load and applied voltage... hence
>cannot be influenced independently by the throttle).
The bad science is in neglecting all the possible events that can occur.
To say "nothing can be communicated except.." is a fatal error because it
assumes only one direction of communication. In the case of the CC,
the throttle drives the locomotive, and the loco also drives the
throttle. The throttle works by sending out pulses of voltage/
current, as it must because the power stage is fed by a rectified
but unfiltered AC waveform. During the intervals between pulses, the motor
sends back a voltage to the throttle, and this voltage ("back EMF") will
be roughly proportional to the speed of the motor. Somehow the Cooler
Crawler responds to this by decreasing its output as the back EMF gets
higher--exactly how it works is something I'd like to see explained, but
I think it would challenge Bob Pease himself. I don't think Rich Weyand
knows, and in fact until recently he was denying that it happened. Rich
still has my congratulations, for recognizing a good thing when he got it.
Anyone can prove the voltage-feedback effect by putting a voltmeter across
the rails, then starting a locomotive out slowly and stalling it. Even
the best op amp circuit will show a slight dip in the voltage on the track,
but the CC shows an increase.
If you take the waveform of a Cooler Crawler and feed it through a power
op-amp, whose output would be a pure voltage source coming through a low
impedance, I assure you that the throttle's performance will be far
poorer than the original design. It has been tried and described, as good
science should be. You can roughly duplicate this, as I have done, simply
by making a Cooler Crawler drive a load through a diode. Sure enough, the
performance gets much worse.
> For example, AC
>power is not always available to a controller, as when the source is a
>battery or rectified DCC signal applied to the motor through a remote
>control reciever.
Note that Lenz has a version of his DCC design which does use voltage
feedback right in the decoder (obviously not using the Cooler Crawler).
I saw this demonstrated at the NMRA DCC booth at the Springfield show
last year and it was extremely impressive; the utterly steady motion
of the loco made it seem very heavy and powerful. And if you tried to
shove it down on the rails it would really fight you!
There are also several voltage-feedback throttles available commercially,
especially from British manufacturers.
John Purbrick
Ken Willmott
jpur...@nyx.cs.du.edu (john purbrick) wrote:
>In article <5ecmcj$kfd$1...@nntp1.uunet.ca>, Ken Willmott <k...@sos.on.ca> wrote:
>>da...@SPAMLESSiwr.ru.ac.za (David Forsyth) wrote:
>>
>>>In article <5eafbm$8qg$1...@nntp1.uunet.ca> k...@sos.on.ca (Ken Willmott) writes:
>>I think where I "got off the bus" was when the claim was made, that a
>>circuit that would impress the same waveform on the track as a CC,
>>would still not work as well. This is bad science. There is nothing
>>that can be communicated on the track except voltage (and current, but
>>in this case, it's a function of load and applied voltage... hence
>>cannot be influenced independently by the throttle).
>The bad science is in neglecting all the possible events that can occur.
>To say "nothing can be communicated except.." is a fatal error because it
>assumes only one direction of communication. In the case of the CC,
>the throttle drives the locomotive, and the loco also drives the
>throttle. The throttle works by sending out pulses of voltage/
>current, as it must because the power stage is fed by a rectified
>but unfiltered AC waveform. During the intervals between pulses, the motor
>sends back a voltage to the throttle, and this voltage ("back EMF") will
>be roughly proportional to the speed of the motor. Somehow the Cooler
>Crawler responds to this by decreasing its output as the back EMF gets
>higher--exactly how it works is something I'd like to see explained, but
>I think it would challenge Bob Pease himself. I don't think Rich Weyand
>knows, and in fact until recently he was denying that it happened. Rich
>still has my congratulations, for recognizing a good thing when he got it.
B.S.! I just ran the CC with an Atlas C424 (no lighting) and closely
examined both the peak and average voltage compared to the loco's
speed. I did not detect the faintest trace of the above mentioned
behaviour. Want to check again?
>Anyone can prove the voltage-feedback effect by putting a voltmeter across
>the rails, then starting a locomotive out slowly and stalling it. Even
>the best op amp circuit will show a slight dip in the voltage on the track,
>but the CC shows an increase.
Not. See above. If you don't believe me, try it yourself.
When it stalled, the voltage even drooped a little tiny bit.
>John Purbrick
Ken Willmott
wey...@mcs.com (Rich Weyand) wrote:
>In article <5ecmcj$kfd$1...@nntp1.uunet.ca> k...@sos.on.ca (Ken Willmott) writes:
>>I think where I "got off the bus" was when the claim was made, that a
>>circuit that would impress the same waveform on the track as a CC,
>>would still not work as well. This is bad science. There is nothing
>>that can be communicated on the track except voltage (and current, but
>>in this case, it's a function of load and applied voltage... hence
>>cannot be influenced independently by the throttle). Any difference in
>>drive characteristics that *could* affect motor performance *must*
>>show up as a different track waveform.
>Nope. Impedance plays an important role as well.
>Two different statements here. One is about impressing the same waveform
>on the track, presumably at a lower output impedance, as you mention
>elsewhere in your note. This does not work as well as the same waveform
>at a higher output impedance such as CC. I have heard a lot of theories
>why, and some arguments that it just can't be, but it's so.
Well, if you just follow the path that current would take going from
the motor to the controller, you will see that the only place it can
go is the 1.0K load resistor (and it has a short memory ;-)
>The second point is different. The same voltage through a lower impedance
>will look identical on open circuit track, but will not be the same
>voltage with a loco present, and will not react the same to the periodic
>differences in loco current draw. So the waveform on the track will now
>look different. The output impedance plays an important role in the CC
>performance.
Actually, you've convinced me on this point.
>Just add a voltage follower to the CC you have and note the performance
>difference.
Right. However, there is a big "gotcha" waiting here. A voltage
follower with what kind of input impedance? To perform a fair test, it
would have to be a motor. So you take the track voltage as it drives a
loco, buffer and amplify and feed to an identical loco. This is not
the same as running it into a load resistor, then amplifying.
>Rich Weyand | _______ ___,---. ---+_______:_ |Rich Weyand
>Weyand Associates| |_N_&_W_| |_N_&_W_| |__|________|_ |TracTronics
>Comm Consultants | ooo ooo ~ ooo ooo ~ oOOOO- OOOO=o\ |Model RR Electronics
>wey...@mcs.com | http://www.mcs.net/~weyand/ |wey...@mcs.com
My experiments are coming along at an incredible rate. I now know why
the CC works so well... because I have a ridiculously simple circuit
that does the same thing. I have to go out today, so I don't have time
to get into it right now, but as soon as I do, I promise to share it
with you and everyone else here.
Suffice it to say... I was partly right and partly wrong. Stay tuned.
:-)
Rich Weyand
In article <5emumr$9lq$1...@nntp2.uunet.ca> k...@sos.on.ca (Ken Willmott) writes:
>wey...@mcs.com (Rich Weyand) wrote:
>>Nope. Impedance plays an important role as well.
>>Two different statements here. One is about impressing the same waveform
>>on the track, presumably at a lower output impedance, as you mention
>>elsewhere in your note. This does not work as well as the same waveform
>>at a higher output impedance such as CC. I have heard a lot of theories
>>why, and some arguments that it just can't be, but it's so.
>Well, if you just follow the path that current would take going from
>the motor to the controller, you will see that the only place it can
>go is the 1.0K load resistor (and it has a short memory ;-)
The capacitor around the output transistor plays a role as well. There
is only a 1 ohm resistor in the path to this cap.
>>Just add a voltage follower to the CC you have and note the performance
>>difference.
>Right. However, there is a big "gotcha" waiting here. A voltage
>follower with what kind of input impedance? To perform a fair test, it
>would have to be a motor. So you take the track voltage as it drives a
>loco, buffer and amplify and feed to an identical loco. This is not
>the same as running it into a load resistor, then amplifying.
Ah, but then you are not using the waveform the CC generates, but the
waveform it generates as modified by the interaction with the loco
motor. I am not sure how this would change the results.
>My experiments are coming along at an incredible rate. I now know why
>the CC works so well... because I have a ridiculously simple circuit
>that does the same thing. I have to go out today, so I don't have time
>to get into it right now, but as soon as I do, I promise to share it
>with you and everyone else here.
Interesting.
Rich Weyand
>I think where I "got off the bus" was when the claim was made, that a
>circuit that would impress the same waveform on the track as a CC,
>would still not work as well. This is bad science. There is nothing
>that can be communicated on the track except voltage (and current, but
>in this case, it's a function of load and applied voltage... hence
>cannot be influenced independently by the throttle). Any difference in
>drive characteristics that *could* affect motor performance *must*
>show up as a different track waveform.
Nope. Impedance plays an important role as well.
Two different statements here. One is about impressing the same waveform
on the track, presumably at a lower output impedance, as you mention
elsewhere in your note. This does not work as well as the same waveform
at a higher output impedance such as CC. I have heard a lot of theories
why, and some arguments that it just can't be, but it's so.
The second point is different. The same voltage through a lower impedance
will look identical on open circuit track, but will not be the same
voltage with a loco present, and will not react the same to the periodic
differences in loco current draw. So the waveform on the track will now
look different. The output impedance plays an important role in the CC
performance.
Just add a voltage follower to the CC you have and note the performance
difference.
Rich Weyand | _______ ___,---. ---+_______:_ |Rich Weyand
Brouwer family
This is a message not strictly in reply to any particular posting on this
subject but in response to the way the entire discussion is going. The
poet said "a little knowing is a dangerous thing". It is possible that
this discussion is spreading that knowing and that is a real problem for
the hobby.
Rich Weyand has been attributed by John Purbrick as admiting that he does
not know how the CC actually works and Ken Wilmott and Rich Weyand have
responded with arguments that seem to support that theory. Despite Ken's
denials, John Purbricks discussion actually has some theoretical basis.
(Please do not accept arguments that say "in theory this should work but in
practice...." In that case the theory is wrong!).
I write this reply as a small hobby manufacturer of a controller which has
been available for over 5 years and which has been favourably reviewed
internationally. That is not a commercial but just to present the context
in which I write.
The class of controller which the CC represents is the "Voltage Feedback"
type, or perhaps more practically, a "constant voltage" type. The
controller attempts to keep the output voltage at a constant level
regardless of load (high impedance, low impedance, whatever). This it does
via the diode from the output back to the base of the output Darlington
pair. This type was described in various forms by Roger Amos in his 2
volumes about model railway electronics (UK). The Britsh EM gauge society
also publishes a circuit. So John Purbrick is on the right track but for
this type of controller, to take note of the "back EMF" which stays in the
system in the spaces between the pulses, would need to incorporate a
capacitor to remember the signal level when the next pulse starts.
This capacitor would also however remember the output pulse so is not
really much use. A more sophisticated circuit is needed to allow it to
remember just the EMF signal.
And that is what the next class of controller does. The "back EMF
feedback" type or, "constant speed " controller. These have been used
for years and are especially marketed by the Brits as John Purbrick hints
at also (he may have combined the types although their effects are quite
different). For some reason I have never seen any published North
American circuit and no manufacturer notes it as a feature in commercial
units . Am I right????
Most of the UK units and some from the Continent are phase angle or PWM
units however with all the evils Rich Weyand notes. No good for coreless
or even modern iron core motors. This type of control is frequently used
in industry and variable speed handyman power drillunits are often of this
type.
As far as I know (but would love to be told otherwise) my own units and the
UK Pentroller, are the only ones which incorporate the reading of motor EMF
WITHOUT PWM or phase angle. The whole voltage line moves up and down
according to demand. These controllers are unfortunately however in
another league of complexity and need to use IC's for the specialist tasks.
The EMF is read by a "sample and Hold" circuit which is clocked by a window
signal which interupts the output purely to sample. These units indeed do
increase the track voltage as the locos are slowed by grades, curves or
whatever and the Lenz unit John Purbrick describes will incorporate such a
characteristic although I suspect it is PWM as all DCC is??? Locos
therefore are NOT slowed by grades etc and so try to keep up a "constant
speed". As well, they are superb for slow speed as the outgoing signal is
entirely dependent on the speed feedback signal from the loco, on an
osciloscope the trace can be seen gently moving up and down.
A simple way to keep the more simple voltage feedback type of circuit more
constant is to take good note of the speed input signal from the handset.
If this can be affected by the load requirements on the power transformer,
ie drop as the load comes on, then the track regulation will just follow
and lower the output! The CC circuit on the Website attempts to remedy
this somewhat by feeding from separate diodes but a better system would
involve a separate (low power) transformer and a regulator to absolutely
lock the control supply in place. (a good use for all those small
transformers that are useless for loco power).
Another thing I learned over 10 years ago when I first started on these
things is that the inertia and brake systems need to be a bit more
sophisticated than a simple resistor/capacitor unit. The characteristic
curves of these is that on "fill up" they start rapidly and then fill
slower and slower as the final voltage is reached. A curve starting
almost vertically which levels off. On discharge, they go the other
way,drop off quickly then tail off level. In practice I never knew where
exactly the train was going to stop! The answer is to use a "constant
Current" circuit to fill the capacitor and another to discharge it. You
then get a straight line characteristic which leads to measurably better
realism and control. Especially in stopping you DO know where the train
will stop. There are other factors but this posting is long enough and I
do not want to encourage mutterings of "smartass". I only want to help!
John Brouwer
bro...@es.co.nz
Ken Willmott
bro...@es.co.nz (Brouwer family) wrote:
>This is a message not strictly in reply to any particular posting on this
>subject but in response to the way the entire discussion is going. The
>poet said "a little knowing is a dangerous thing". It is possible that
>this discussion is spreading that knowing and that is a real problem for
>the hobby.
If you mean that unproven theories might be taken as fact, I agree
that it's possible, but as any serious design will undergo field
testing, I see no risk to the hobby at large.
>Rich Weyand has been attributed by John Purbrick as admiting that he does
>not know how the CC actually works and Ken Wilmott and Rich Weyand have
>responded with arguments that seem to support that theory. Despite Ken's
>denials, John Purbricks discussion actually has some theoretical basis.
>(Please do not accept arguments that say "in theory this should work but in
>practice...." In that case the theory is wrong!).
Thus demanding a new theory... I read J. Purbricks letter with
interest but accidentally lost it. I seem to recall that he described
the CC throttle as implementing back EMF feedback. This is simply
not the case. Unfortunately, I cannot delve into a proper refutation
right now, and still leave myself time to do the experimental design
and bench testing I'm doing. The feedback action described by J.P. and
others exists, but not for the reasons I have heard given so far.
Every bit of news, email, and mail I have read has been a great help
in trying to figure this puzzle, and I believe (by George) I've got
it. A proper article will appear on my website soon.
>I write this reply as a small hobby manufacturer of a controller which has
>been available for over 5 years and which has been favourably reviewed
>internationally. That is not a commercial but just to present the context
>in which I write.
>The class of controller which the CC represents is the "Voltage Feedback"
>type, or perhaps more practically, a "constant voltage" type. The
>controller attempts to keep the output voltage at a constant level
>regardless of load (high impedance, low impedance, whatever). This it does
>via the diode from the output back to the base of the output Darlington
>pair.
Nope. That's there to limit the output current. The CC falls somewhere
between a constant voltage and constant *charge* source (because of
the bootstrap capacitor).
>This type was described in various forms by Roger Amos in his 2
>volumes about model railway electronics (UK). The Britsh EM gauge society
>also publishes a circuit. So John Purbrick is on the right track but for
>this type of controller, to take note of the "back EMF" which stays in the
>system in the spaces between the pulses, would need to incorporate a
>capacitor to remember the signal level when the next pulse starts.
>This capacitor would also however remember the output pulse so is not
>really much use. A more sophisticated circuit is needed to allow it to
>remember just the EMF signal.
Right. Exactly. The output capacitor gets its "charge memory" during
the SUPPLY phase of the cycle, while for a true back EMF circuit, the
IDLE phase is when this measurement is captured.
>And that is what the next class of controller does. The "back EMF
>feedback" type or, "constant speed " controller. These have been used
>for years and are especially marketed by the Brits as John Purbrick hints
>at also (he may have combined the types although their effects are quite
>different). For some reason I have never seen any published North
>American circuit and no manufacturer notes it as a feature in commercial
>units . Am I right????
For the most part. There might be a few out there.
>Most of the UK units and some from the Continent are phase angle or PWM
>units however with all the evils Rich Weyand notes. No good for coreless
>or even modern iron core motors. This type of control is frequently used
>in industry and variable speed handyman power drillunits are often of this
>type.
Although, work-arounds have been published to fix this problem (the
attendant motor heating).
>As far as I know (but would love to be told otherwise) my own units and the
>UK Pentroller, are the only ones which incorporate the reading of motor EMF
>WITHOUT PWM or phase angle. The whole voltage line moves up and down
>according to demand. These controllers are unfortunately however in
>another league of complexity and need to use IC's for the specialist tasks.
>The EMF is read by a "sample and Hold" circuit which is clocked by a window
>signal which interupts the output purely to sample. These units indeed do
>increase the track voltage as the locos are slowed by grades, curves or
>whatever and the Lenz unit John Purbrick describes will incorporate such a
>characteristic although I suspect it is PWM as all DCC is??? Locos
>therefore are NOT slowed by grades etc and so try to keep up a "constant
>speed". As well, they are superb for slow speed as the outgoing signal is
>entirely dependent on the speed feedback signal from the loco, on an
>osciloscope the trace can be seen gently moving up and down.
I'll bet... but one attraction of the CC is it's simplicity from the
standpoint of construction. Once you get into IC's you practically
have to have a PCB. I have really two throttles in mind... one simple
and one advanced, like the above. In fact, the way my experiments are
going, it may be possible to have a CC-like output stage with true
back EMF.
>A simple way to keep the more simple voltage feedback type of circuit more
>constant is to take good note of the speed input signal from the handset.
>If this can be affected by the load requirements on the power transformer,
>ie drop as the load comes on, then the track regulation will just follow
>and lower the output! The CC circuit on the Website attempts to remedy
>this somewhat by feeding from separate diodes but a better system would
>involve a separate (low power) transformer and a regulator to absolutely
>lock the control supply in place. (a good use for all those small
>transformers that are useless for loco power).
Worse, if the momentum is set to zero, the memory cap ends up charging
directly from those secondary diodes... influencing the average
voltage applied to the control pot. I changed the CC I built so that
the walkaround gets its power from an additional cap/diode DC supply.
This voltage is sufficiently stable, unless the transformer is
drastically under-rated (which it isn't). I was hoping to keep this my
little secret, as if I say much more, there may be a reward out for my
head. :-)
>Another thing I learned over 10 years ago when I first started on these
>things is that the inertia and brake systems need to be a bit more
>sophisticated than a simple resistor/capacitor unit. The characteristic
>curves of these is that on "fill up" they start rapidly and then fill
>slower and slower as the final voltage is reached. A curve starting
>almost vertically which levels off. On discharge, they go the other
>way,drop off quickly then tail off level. In practice I never knew where
>exactly the train was going to stop! The answer is to use a "constant
>Current" circuit to fill the capacitor and another to discharge it. You
>then get a straight line characteristic which leads to measurably better
>realism and control. Especially in stopping you DO know where the train
>will stop. There are other factors but this posting is long enough and I
>do not want to encourage mutterings of "smartass". I only want to help!
I have to question whether this is prototypical. At least in the case
of acceleration, the dV/dT where V is velocity, the tractive effort is
not constant with speed. Hence, non-linear. I have been trumped before
on this sort of question, though, so someone else should comment.
It's an interesting idea.
>John Brouwer
>bro...@es.co.nz
John Purbrick
In article <brouwer-ya0232800...@newsch.es.co.nz>,
Rich Weyand thinks of the principle of the Cooler Crawler as having something
to do with impedance, and I don't agree with this. I don't see how any kind
of impedance would cause the output voltage to rise if a train stalls.
I might well have confused the two types. Actually, based on limited
observation, I think the CC uses "partial voltage feedback". It increases the
output voltage in response to decreasing motor speed, but not to the full
extent possible (i.e. the full power supply voltage) and I think the feature
only functions significantly at slow speeds. But I think this is perfectly
acceptable--the reason I want voltage feedback is to give smooth motion
at low speed, not to make a train run at a constant speed up hill and down
dale, unrealistic as that would be.
>For some reason I have never seen any published North
>American circuit and no manufacturer notes it as a feature in commercial
>units . Am I right????
As far as I know, yes.
>As far as I know (but would love to be told otherwise) my own units and the
>UK Pentroller, are the only ones which incorporate the reading of motor EMF
>WITHOUT PWM or phase angle. The whole voltage line moves up and down
>according to demand. These controllers are unfortunately however in
>another league of complexity and need to use IC's for the specialist tasks.
>The EMF is read by a "sample and Hold" circuit which is clocked by a window
>signal which interupts the output purely to sample. These units indeed do
>increase the track voltage as the locos are slowed by grades, curves or
>whatever and the Lenz unit John Purbrick describes will incorporate such a
>characteristic although I suspect it is PWM as all DCC is??? Locos
>therefore are NOT slowed by grades etc and so try to keep up a "constant
>speed". As well, they are superb for slow speed as the outgoing signal is
>entirely dependent on the speed feedback signal from the loco, on an
>osciloscope the trace can be seen gently moving up and down.
I have a design for a voltage feedback throttle (using a sample-and-hold
technique as described above) which I designed to be part of a "one throttle
per block" control system for the MIT Model Railroad Club. It does use IC's--
one LM324 quad op amp--but the cost is about $3 in components (not PC board,
connectors and so forth). However, as I said before I don't think full
feedback is necessary; so it's only when starting a train that it's active on
my design.
JP
John Purbrick
In article <5epour$ru2$1...@nntp1.uunet.ca>, Ken Willmott <k...@sos.on.ca> wrote:
>bro...@es.co.nz (Brouwer family) wrote:
>
>>Rich Weyand has been attributed by John Purbrick as admiting that he does
>>not know how the CC actually works and Ken Wilmott and Rich Weyand have
>>responded with arguments that seem to support that theory. Despite Ken's
>>denials, John Purbricks discussion actually has some theoretical basis.
>>(Please do not accept arguments that say "in theory this should work but in
>>practice...." In that case the theory is wrong!).
>
>Thus demanding a new theory... I read J. Purbricks letter with
>interest but accidentally lost it. I seem to recall that he described
>the CC throttle as implementing back EMF feedback. This is simply
>not the case.
If by "implement" you mean it was deliberately designed that way, then no.
But if you mean that it's there in demonstrable form, then yes, it's
implemented, and if it weren't the CC wouldn't be much of a throttle.
Now, HOW does it work? That's a question I'd like to know the answer to.
Some of the discussion which I didn't quote here had the right sound to it.
JP
Brouwer family
> bro...@es.co.nz (Brouwer family) wrote:
>
>
> >Another thing I learned over 10 years ago when I first started on these
> >things is that the inertia and brake systems need to be a bit more
> >sophisticated than a simple resistor/capacitor unit. The characteristic
> >curves of these is that on "fill up" they start rapidly and then fill
> >slower and slower as the final voltage is reached. A curve starting
> >almost vertically which levels off. On discharge, they go the other
> >way,drop off quickly then tail off level. In practice I never knew where
> >exactly the train was going to stop! The answer is to use a "constant
> >Current" circuit to fill the capacitor and another to discharge it. You
> >then get a straight line characteristic which leads to measurably better
> >realism and control. Especially in stopping you DO know where the train
> >will stop. There are other factors but this posting is long enough and I
> >do not want to encourage mutterings of "smartass". I only want to help!
>
> I have to question whether this is prototypical. At least in the case
> of acceleration, the dV/dT where V is velocity, the tractive effort is
> not constant with speed. Hence, non-linear. I have been trumped before
> on this sort of question, though, so someone else should comment.
> It's an interesting idea.
>
I understand what you are saying and thought about it too but linear is
better than the plain cap/resitor arrangement. In the case of
acceleration,in that trains do not race away because of the steep charge
curve and in the case of braking, because the braking effort does not fade
away due to the discharge reducing with decreasing charge. On real
trains, for any given brake application the deceleration increases with
reducing speed, the opposite to a plain cap/resistor arrangement. Try it
next time you stop your car at traffic lights,you involuntarily back off on
the pedal as you draw to a stop to prvent lock-up and tyre squeal (well
most of us do! :-)) In accelerating, train crews have to control the
throttle to limit tractive effort to the prevailing adhesion conditions.
Once moving then there is rapid acceration until the available horsepower
curve limits it up to final speed. So for models we need a compound
curve but that increases the complexity of the circuits further and you
want to keep it simple Ken! Unfortunately we do not get most things for
free :-).
> >John Brouwer
> >bro...@es.co.nz
>
>
> Ken Willmott _ _ _____ _ _ _____ _____ _ _
> ####################[<_(_|=@=@@>]-[<_(_|=@=@@>]-[<@@=@=|_)_>]#######
John Brouwer
bro...@es.co.nz
Ken Willmott
jpur...@nyx.cs.du.edu (John Purbrick) wrote:
>In article <5epour$ru2$1...@nntp1.uunet.ca>, Ken Willmott <k...@sos.on.ca> wrote:
>>bro...@es.co.nz (Brouwer family) wrote:
>>
>>>Rich Weyand has been attributed by John Purbrick as admiting that he does
>>>not know how the CC actually works and Ken Wilmott and Rich Weyand have
>>>responded with arguments that seem to support that theory. Despite Ken's
>>>denials, John Purbricks discussion actually has some theoretical basis.
>>>(Please do not accept arguments that say "in theory this should work but in
>>>practice...." In that case the theory is wrong!).
>>
>>Thus demanding a new theory... I read J. Purbricks letter with
>>interest but accidentally lost it. I seem to recall that he described
>>the CC throttle as implementing back EMF feedback. This is simply
>>not the case.
>If by "implement" you mean it was deliberately designed that way, then no.
>But if you mean that it's there in demonstrable form, then yes, it's
>implemented, and if it weren't the CC wouldn't be much of a throttle.
>Now, HOW does it work? That's a question I'd like to know the answer to.
>Some of the discussion which I didn't quote here had the right sound to it.
>JP
Okay, here goes. The simple way to understand it is to look at the
Armstrong throttle. This uses a capacitor in parallel with a rheostat,
placed in series with the track. Think of the dynamics. At the lowest
setting, when power is applied the cap charges through the load until
it reaches almost the peak value of the applied half wave rectified
power supply. Subsequently, the cap spends most of the cycle slowly
discharging through the rheostat, and a small portion charging rapidly
through the load. It is the latter portion that propels the loco.
During the discharge portion, the power supply diodes are reversed
biased, and non-conducting. Hence, a diode may be added to the output
of this circuit without affecting its performance.
Now suppose that during some charge cycle, the motor/track impedance
is momentarily reduced, as when the commutator hits a "not so
favourite" spot. Charge current is reduced and the voltage on the cap
decreases (relative to the norm). On the next cycle, the reduced
voltage on the cap (corresponding to an increased voltage to the load)
results in a heavier charge current pulse - corrective feedback!
This circuit without the cap behaves as a current source - a high
impedance. With the cap, and half wave supply, you get the long term
behaviour of a current source, and the short term behaviour of a
voltage source (hence, low impedance) during the charge pulse
interval.
The CC works similarly, except that the discharge cycle is controlled
by the amplifier block - restricting the range of feedback action, and
therefore providing a lower drive impedance.
Yesterday I built a MOSFET version of the Armstrong throttle. The
MOSFET is perfect for this because it is a voltage controlled current
source, with a radically high input impedance. I was able to set it up
for momentum and memory with no other transistors for this reason. The
other nice thing about it is that if the drive voltage is properly
limited, the output is inherently current-limited.
In my admittedly biased opinion, it performs identically to the Cooler
Crawler. The next step is to get this out to the many people that have
expressed an interest, so that the first round of R&D and testing can
begin. This will require some tinkering with my website, and it
involves a change of subject for this thread, so I'll announce it when
it's ready.
Right now, my objective is to combine the current from the MOSFET
circuit with a regulated DC supply, so as to smooth the power at
higher speeds, and improve loco "speed lock" due to the low output
impedance of the DC supply.
flyt...@aol.com
I am not fully knowledgable about this kind of stuff but wouldn't a square wave generator help with the curve?
richardt...@gmail.com
It back for repair. When got it back it b burn up on me.