The Advantages of DCC Keep Alives

 
  NSWGR1855 Deputy Commissioner

OK if the above is true, can somebody explain why all my NON KA loco's wheels stay just as clean as the KA loco's wheels.

Only 10 0f my 50 locos have KA. 8 are DCCconcepts decoders with KA. 2 are Lenz gold with USP KA's

Since going DCC , 18+ yrs ago) my track, loco wheels & rolling stock wheels stay cleaner longer than yrs op DC.

BTW, my layout is in an unlined tin shed and as I said now that I use Track Magic my track is cleaned every 4-5 months whether it needs it or not.

Before TM I used Iso alcohol & cleaned track 6-8 weeks.

Rarely clean loco wheels. Can't remember when I last cleaned R-stock wheels.

Oh yes, all my locos steam & diesel do 2-3 MPH on speed step 1 using 28st (personal choice) thru electronic speed trap.

OH dear, after all this my head hurts I need a Bex and a lie down.

Cheers

Ian
sunnysa

Ian,

Your results are consistent with my results as it is hard to see a see an improvement when you already have decreased arcing significantly by having models with good pickup, use an oil that makes the dirt conductive, and stops the wheels and track from oxidizing.

The speed step 1 was only an experimental method to compare the improvement in smooth low speed running between keep alive and non keep alive models.

I use 128 step mode, and only use the push buttons on my NCE controller with lots of inertia for prototypical acceleration and braking. Typically I use the fast buttons which increase the speed by 10 steps, resulting in  13 pushes to get from 0 to full speed. When shunting, I can use the 1 step push button to slacken the coupler to take up slack when stopped, or position the train very accurately without changing the inertia setting. Most people I know do not use the push buttons for speed control, but that is one advantage of DCC, you have flexibility to set things up the way you like.

Terry Flynn.

Sponsored advertisement

  NSWGR1855 Deputy Commissioner

Speed step 1 is set so that the loco just moves, and this may vary for different locos, for some it may be 1kph others it might be 15kph that is why it is a speed step, not an actual speed. Just as 128 is maximum, on some locos that might be 40kph others might go to 200kph or more.

If the loco will start and run smoothly and consistently at 5kph but not at 4 then step 1 is set to run it at 5kph where it will happily do so all day.

Here is an experiment for you, power your loco via an on board DCC supply (wires from the track), I bet it will run just as slow and even without the KAs which would prove that the KAs are only helping with constant power not smooth running.

Tony
miktrain

Tony,

For smooth low speed running constant power to the motor is required. Without a keep alive capacitor power can easily be momentarily lost to the motor due to dirt between the wheels and rail. This will  cause the model to jerk or stall. The keep alive capacitor solves this problem and is easily observed. For high speed running, the inertia of rotating parts will smooth out the momentary no power to the motor. Your experiment above does not take dirty track into account.

Terry Flynn.
  Aaron The Ghost of George Stephenson

Location: University of Adelaide SA
I can not be certain but judging by most of user names on this thread is everyone apart from Terry, Geoff and myself modelling or from SA? Just seems weird ha ha
linton78

Kevin is not from SA and neither is SAR523 - he might be modelling Mitcham IIRC, but he's not even in AU.
  Aaron The Ghost of George Stephenson

Location: University of Adelaide SA
As I think I have read somewhere on here that Aaron has built his own decoders, he may be able to elaborate a little more on how the decoders motor control circuit reacts to a motor starting and stopping.
linton78

Black arts and trade secrets Wink

In a permanent magnet DC motor, speed is proportional to applied EMF alone (minus the losses due to winding resistance).

No decoder I have ever bothered to scope (ESU, TCS, Lenz, QSI) has seemingly bothered to implement a 'soft start', the motor drive pulse train seems to start up with fixed peak amplitude and in general the fixed pulse widths as defined for the speed step.

The exception to this is decoders that have bits set for momentum delay on acceleration/deceleration. In the case of decoders with values set for these variables the width of the drive pulses is moderated according to the CV setting.

BEMF is sensed at the H bridge driving the motor, and they are not returned anywhere beyond there. That is, the track will never see it, and neither would a KA if fitted. - The KA being fitted across the output of the input bridge.
  linton78 Train Controller

Location: South Coast NSW
Kevin is not from SA and neither is SAR523 - he might be modelling Mitcham IIRC, but he's not even in AU.
"Aaron"


Aah, I thought there was some SA modellers secret hand shake thing going on ha ha. Still have to get down there one day. I would like to stop in at the Orient Express and have a look at the Zimo stuff they have there. Re the conversation we had about the Zimo DCC command station, I have been looking at the new ESU radio throttle also. It's sort of like a smart phone throttle but has a rotary knob, which is something I like.

I know this is off topic but I like the ESU command station because it has a function that allows any DCC systems output to be connected to the ESU controller. This would allow my Powercab to still be utilised as a throttle which in turn controls the ESU command station. Not sure if any other manufacturers have that feature?

Linton
  linton78 Train Controller

Location: South Coast NSW
Black arts and trade secrets Wink

In a permanent magnet DC motor, speed is proportional to applied EMF alone (minus the losses due to winding resistance).

No decoder I have ever bothered to scope (ESU, TCS, Lenz, QSI) has seemingly bothered to implement a 'soft start', the motor drive pulse train seems to start up with fixed peak amplitude and in general the fixed pulse widths as defined for the speed step.

The exception to this is decoders that have bits set for momentum delay on acceleration/deceleration. In the case of decoders with values set for these variables the width of the drive pulses is moderated according to the CV setting.

BEMF is sensed at the H bridge driving the motor, and they are not returned anywhere beyond there. That is, the track will never see it, and neither would a KA if fitted. - The KA being fitted across the output of the input bridge.
"Aaron"


Thought I had read it somewhere.

For decoders you have put on a scope, do they show a higher power consumption briefly during motor start?

Out of interest, how were the decoders built? Did you design multilayered boards? What was the physical size of the end product?

Linton
  NSWGR1855 Deputy Commissioner

Black arts and trade secrets Wink

In a permanent magnet DC motor, speed is proportional to applied EMF alone (minus the losses due to winding resistance).

No decoder I have ever bothered to scope (ESU, TCS, Lenz, QSI) has seemingly bothered to implement a 'soft start', the motor drive pulse train seems to start up with fixed peak amplitude and in general the fixed pulse widths as defined for the speed step.

The exception to this is decoders that have bits set for momentum delay on acceleration/deceleration. In the case of decoders with values set for these variables the width of the drive pulses is moderated according to the CV setting.

BEMF is sensed at the H bridge driving the motor, and they are not returned anywhere beyond there. That is, the track will never see it, and neither would a KA if fitted. - The KA being fitted across the output of the input bridge.
Aaron

The track or keep alive capacitor will see the result of the motor BEMF as a change in current drawn by the motor. No keep alive, the track sees all the current drawn from the motor.

One objective way to see if there is a noticeable difference in arcing is to compare  keep alive and non keep alive models by looking at the track waveforms with a CRO.

Terry Flynn.
  miktrain Deputy Commissioner

Location: Adelaide SA
Yes that is all true....

...Where I saw that there may be a benefit is when the track interruption results in the motor to stop rotating, I.e. The model stops for a brief moment, which happens. In this case you would have a microsecond of very high current as the motor restarts.

What I didn't think about enough, was that the decoder may change the amount of motor start current felt at rail and wheel. The affect of the motors collapsing field would be different to a simple DC run model. I guess a higher current is still required to restart the motor but how is it managed.

The KA theory was all about keeping the motor running to avoid motor start currents. As an electric motor draws maximum current when its armature is not rotating, that initial start current can be very high (model train high that is), higher than both constant current draw + KA recharge current.

What bugs me the most about RP is that no one can ever be wrong. Tony you were only half correct with you KA capacitor knowledge , yet were quick to correct me 'on my maths'. Once you were corrected you said nothing about it. Rather you just keep banging on.
linton78

Glad you agree

There would only be a very small increase in current as the decoder uses high frequency pulse width modulation to drive the motor, that is basically full voltage for a very short time, zero volts, full voltage for a very short time, zero volts, which is why DCC can run the motor very slow without burning it up. There is quite a bit of electronics testing and constantly fine tuning the power.

If this were not so, running a few locos very slow would shut down the boosters due to the over current draw.

The BEMF will never get to the rails as the motor and drive circuit are isolated from them. That may actually be where the difference you see in arcing between DC and DCC, DC has lots of BEMF on the rails DCC does not

I was only unsure as to the arrangement inside the KA pack not how it works/what it is supposed to do. I was not, in my opinion corrected, but told which of my two options was the one used. Re-read my post I said "IF in parallel" and "IF in series"
As I have said I do not have any KAs so could not look inside.

Tony
  miktrain Deputy Commissioner

Location: Adelaide SA
OK if the above is true, can somebody explain why all my NON KA loco's wheels stay just as clean as the KA loco's wheels.

Only 10 0f my 50 locos have KA. 8 are DCCconcepts decoders with KA. 2 are Lenz gold with USP KA's
sunnysa

Because the KAs are not doing anything in regards to keeping the wheels clean.

Tony
  miktrain Deputy Commissioner

Location: Adelaide SA
Tony,

For smooth low speed running constant power to the motor is required. Without a keep alive capacitor power can easily be momentarily lost to the motor due to dirt between the wheels and rail. This will cause the model to jerk or stall. The keep alive capacitor solves this problem and is easily observed. For high speed running, the inertia of rotating parts will smooth out the momentary no power to the motor. Your experiment above does not take dirty track into account.

Terry Flynn.
NSWGR1855

Yes, but the discussion is not "Does the KA help to keep the loco running?" it is "Do they keep the wheels clean?"

My experiment for you had nothing to do with dirty wheels/track, only to show you that the KAs help to keep the power constant just like having a direct wired connection, they have nothing directly to do with slow, smooth running, ie. the constant power lets the decoder do the job that is asked of it.

Tony
  linton78 Train Controller

Location: South Coast NSW
Glad you agree

There would only be a very small increase in current as the decoder uses high frequency pulse width modulation to drive the motor, that is basically full voltage for a very short time, zero volts, full voltage for a very short time, zero volts, which is why DCC can run the motor very slow without burning it up. There is quite a bit of electronics testing and constantly fine tuning the power.

If this were not so, running a few locos very slow would shut down the boosters due to the over current draw.

The BEMF will never get to the rails as the motor and drive circuit are isolated from them. That may actually be where the difference you see in arcing between DC and DCC, DC has lots of BEMF on the rails DCC does not

I was only unsure as to the arrangement inside the KA pack not how it works/what it is supposed to do. I was not, in my opinion corrected, but told which of my two options was the one used. Re-read my post I said "IF in parallel" and "IF in series"
As I have said I do not have any KAs so could not look inside.

Tony
"miktrain"



Hi Tony,

I am familiar with PWM. Large electric motors controlled by PWM circuits still have to deal with motor starting currents. The average power for a time period would increase. There are ways of starting motors with PWM controllers which allows the pulse voltage and frequency to be adjusted to ramp the motor up so as to reduce inrush current. It is a real problem which trips over current circuits/devices if not dealt with.

One of the aircraft I have worked on has a 28 volt DC motor which drives an oil cooler fan. The CB is rated at 30 amps. Although some times ambient temperature related, this CB can trip during motor start. This has been the cause of quite a few engine oil over temps. You should see the starting inrush current when starting some aircraft. We are talking in excess of 600 amps and then settling to around 200 amps during engine light. I would hate to see the instantaneous current spike! I can't because the amp meter is not sensitive enough. It doesn't matter if the power cart was a PWM power supply or one plugged into the snowy mountain scheme there is still a huge increase in power required. I have seen this many many times.

Simply speaking, while the motor is stationary the windings are a set resistance say 100 ohms ( just made that up). When a pulse of  6 volts is felt, instantaneously 60 mA is felt. This decreases as the motor begins to rotate due to BEMF etc. What your saying is, because the pulse is a higher voltage, switched for a short time there would only be a minor current increase. Let's say the voltage level of the pulse is raised to 12 volts. Instantaneously with the same motor winding resistance the current has doubled. I may be wrong here but this is what I remember learning 15 years ago.

Maybe you should buy a KA before correcting peoples posts then. You could have asked how the capacitors were arranged then told us all how configuring in series and parallel worked. The ESU power pack has only one capacitor and it is also rated around 2.7 volts. That little device is a whole new ball game.

This is all becoming a little tiring. DCC has helped make my engines run better and appears to have helped keep wheels cleaner. The stay alive installs have improved the reliability, in that I have never had a locomotive stall and the sound has never reset un-commanded. By the sounds of it you have never had a model stall either. Good on you, you have saved some money.

For me this is over until someone, including myself can show proper evidence.

Linton
  NSWGR1855 Deputy Commissioner

Yes, but the discussion is not "Does the KA help to keep the loco running?" it is "Do they keep the wheels clean?"

My experiment for you had nothing to do with dirty wheels/track, only to show you that the KAs help to keep the power constant just like having a direct wired connection, they have nothing directly to do with slow, smooth running, ie. the constant power lets the decoder do the job that is asked of it.

Tony
miktrain


Tony,

If the discussion is about "Do they keep the wheels clean?" then why did you have an experiment that has 'nothing to do with dirty wheels/track?' (your words)

Slow speed running has everything to do with electrical pickup and dirty wheels and track, amongst other variables.

Compared to non KA DCC models, he keep alive capacitors improve low speed running, because the motor and decoder are less effected by poor pickup and dirt, having a stable on train power supply, allowing the model to move at the lowest possible speed the motor / gearing / decoder combination will allow without jerking over dirty track. This is easily observed and if you had a KA fitted model, you would see the lower practical minimum speed for yourself.



Terry Flynn.
  miktrain Deputy Commissioner

Location: Adelaide SA
Tony,

If the discussion is about "Do they keep the wheels clean?" then why did you have an experiment that has 'nothing to do with dirty wheels/track?' (your words)

Slow speed running has everything to do with electrical pickup and dirty wheels and track, amongst other variables.

Compared to non KA DCC models, he keep alive capacitors improve low speed running, because the motor and decoder are less effected by poor pickup and dirt, having a stable on train power supply, allowing the model to move at the lowest possible speed the motor / gearing / decoder combination will allow without jerking over dirty track. This is easily observed and if you had a KA fitted model, you would see the lower practical minimum speed for yourself.



Terry Flynn.
NSWGR1855

Because the thread had drifted that way

Tony
  miktrain Deputy Commissioner

Location: Adelaide SA
...By the sounds of it you have never had a model stall either.
linton78

Yes, just like most over the years I have had my share of stall problems but I still rarely have to clean wheels. Must be everyone else's KAs working Smile

Tony
  Aaron The Ghost of George Stephenson

Location: University of Adelaide SA
Thought I had read it somewhere.

For decoders you have put on a scope, do they show a higher power consumption briefly during motor start?

Out of interest, how were the decoders built? Did you design multilayered boards? What was the physical size of the end product?

Linton
"linton78"
Not to put too fine a point on it, but they were built with much difficultly, or at least designing one is not trivial.

I mostly build accessory decoders in truth, because they are cheaper than what I can buy and I can have them do EXACTLY what I want them to do. Our club has a large layout and space is not at a premium, so that are built as cheaply as possible - which means entirely by hand using through hole components. A single decoder to drive a servo motor for points is maybe 60mm square, don't have one handy to measure. It could be made much smaller, but I keep these boards spacious and single sided. 8 solenoid points are controlled on a board not much bigger, maybe 60x80mm, again single sided board, but that's about the minimum size that board could be anyway, multiple laying would not help.

Locomotive decoders are built entirely from SMT componentry, my 'draft' design is a two layer board, approx size of a postage stamp. I have a production run design which is about a quarter of that size, four layer board, I will need to get the boards made because this is beyond the limits of the technology I have access to. Might be able to make it smaller, but as it is, the design in pushing the limits in terms of what I can achieve in SMT, the smaller design and possibly even this current design may mean I have to get someone else to produce them. To be honest that probably won't happen, the price of TCS gear is so good that it'll be more economical to use them than my own made professionally, and TCS frankly is a superior product. Little wonder, they have a dedicated team of engineers and firmware designers, I am just me extending my hobby.
  miktrain Deputy Commissioner

Location: Adelaide SA
Simply speaking, while the motor is stationary the windings are a set resistance say 100 ohms ( just made that up). When a pulse of 6 volts is felt, instantaneously 60 mA is felt. This decreases as the motor begins to rotate due to BEMF etc. What your saying is, because the pulse is a higher voltage, switched for a short time there would only be a minor current increase. Let's say the voltage level of the pulse is raised to 12 volts. Instantaneously with the same motor winding resistance the current has doubled.
linton78

In theory yes, but as I said earlier, theory and practice are often not the same. Doubling the voltage with the same resistance does create double current draw but with PWM the voltage does not change, it is the width of the pulse, and hence the average voltage that changes and that fits your claim, but only while the motor stays stationary, and because the decoder is reading and acting on the BEMF which would also be higher the pulse would be shorter for the same amount of motor rotation.

If on the other hand a higher voltage pulse was used (as you are saying) then the motor would get moving faster and as you know a moving motor has much higher resistance so the current would be limited by that. Yes it would be higher for that bees knee of time that the motor is not turning but the decoders internal power supply would cover that. Once moving the average voltage would still be the same and hence the current because the pulses would be half as wide (or the motor much faster).

If you are seeing an lessening of pitting of the wheels with DCC it is most probably the lack of BEMF getting to the track because the decoder blocks it, with DC that would be directly connected to the track via the wheels.

Tony
  Aaron The Ghost of George Stephenson

Location: University of Adelaide SA
Doubling the voltage with the same resistance does create double current draw but with PWM the voltage does not change, it is the width of the pulse
"miktrain"
Correct, there is not a decoder I have seen that is capable of modulating the amplitude of the motor drive pulses, the MOSFETs are always operated in saturation mode, never triode, hence the energy to the motor is strictly adjusted by time. If the applied amplitude varied you have to ask 'where the volts went?' and hence 'where did the heat go?'

but only while the motor stays stationary, and because the decoder is reading and acting on the BEMF which would also be higher the pulse would be shorter for the same amount of motor rotation.
"miktrain"
The is no BEMF in a motor that is not rotating.
  miktrain Deputy Commissioner

Location: Adelaide SA
Correct, there is not a decoder I have seen that is capable of modulating the amplitude of the motor drive pulses, the MOSFETs are always operated in saturation mode, never triode, hence the energy to the motor is strictly adjusted by time. If the applied amplitude varied you have to ask 'where the volts went?' and hence 'where did the heat go?'

The is no BEMF in a motor that is not rotating.
Aaron

Yeah, I put that a bit wrong I was meaning reading the BEMF the instant it started to turn.

Tony
  linton78 Train Controller

Location: South Coast NSW
Correct, there is not a decoder I have seen that is capable of modulating the amplitude of the motor drive pulses, the MOSFETs are always operated in saturation mode, never triode, hence the energy to the motor is strictly adjusted by time. If the applied amplitude varied you have to ask 'where the volts went?' and hence 'where did the heat go?'

The is no BEMF in a motor that is not rotating.
"Aaron"



My point about doubling the voltage was not aimed at thinking the decoder was adjusting amplitude. It was more based on just because it's PWM driven doesn't mean that motor starting current would be less. It's probably actually higher due to the voltage level compared to DC. This is one of the reasons motors can be so well controlled at low speed settings as the current is higher during the pulses which in turn equals higher torque values. More current more magnetic force.

Your post regarding your built decoders was interesting, thanks for that. Yeah surface mount stuff is getting so small these days. It would be a nightmare to assemble something like an n scale decoder. When we were learning SMD techniques we used a microscope to solder under. I have not seen how a mass produced SMD soldering process works. I would imagine there would be a lot of solder paste and hot air involved. Would be interesting. Might you tube it ha ha.

Linton
  linton78 Train Controller

Location: South Coast NSW
In theory yes, but as I said earlier, theory and practice are often not the same. Doubling the voltage with the same resistance does create double current draw but with PWM the voltage does not change, it is the width of the pulse, and hence the average voltage that changes and that fits your claim, but only while the motor stays stationary, and because the decoder is reading and acting on the BEMF which would also be higher the pulse would be shorter for the same amount of motor rotation.

If on the other hand a higher voltage pulse was used (as you are saying) then the motor would get moving faster and as you know a moving motor has much higher resistance so the current would be limited by that. Yes it would be higher for that bees knee of time that the motor is not turning but the decoders internal power supply would cover that. Once moving the average voltage would still be the same and hence the current because the pulses would be half as wide (or the motor much faster).

If you are seeing an lessening of pitting of the wheels with DCC it is most probably the lack of BEMF getting to the track because the decoder blocks it, with DC that would be directly connected to the track via the wheels.

Tony
"miktrain"



Completely agree on the BEMF effect when using DC. Something I had not considered until reading a few of the threads here on Railpage. I too think this is where DCC really shines and most likely has the biggest impact on improving wheel cleanliness.

Linton
  Aaron The Ghost of George Stephenson

Location: University of Adelaide SA
Your post regarding your built decoders was interesting, thanks for that. Yeah surface mount stuff is getting so small these days. It would be a nightmare to assemble something like an n scale decoder. When we were learning SMD techniques we used a microscope to solder under. I have not seen how a mass produced SMD soldering process works. I would imagine there would be a lot of solder paste and hot air involved. Would be interesting. Might you tube it ha ha.
"linton78"
There'll be piles of vids on youtube of the process. There is actually very little solder paste, it's applied via a screen (stainless steel or Mylar usually) and a squeegee. We bake the boards in a very specific oven, although I solder SMT in a 'standard' toaster oven, which I have heavily unstandardised to achieve very accurate temperature and time profiles.

See:

http://m.youtube.com/watch?v=82W0OJylh_Q

I had something like this but have evolved mine somewhat to cope with my different 'baking' needs and I also use multiple temperature measuring points to keep a better average of temperature rather than a single point measurement. That, and I keep all my control gear within the oven case so it looks a little less 'bomb like'.
  MRHD66 Station Staff

Location: Wingham, New South Wales, Australia
after adding a tcs kat22 (inbuilt keep alive) into my austrains c36 the difference is amazing   next to get one is my c35  then will also fit a keep alive to my powerline 48 (2012) and my austrains 442  so glad i only have 3 to fit them into atm BUT any new locks i get will get them straight away!  the best thing  is that even if the track isnt perfectly clean  it will run smooth (perfect for the club rooms) i recently   saw a club members  austrains 442  running really erratically then my 36 runs over the same section  and it just runs so smooth  and also i can juust run it on speed steep 1 of 128 all day and it just runs no more getting frustrated when trying to shunt it just dose it so  after seeing thhis with my own eyes the keep alives are worth every $ and all my locos will get them

Cheers Oscar

Sponsored advertisement

Display from:   

Quick Reply

We've disabled Quick Reply for this thread as it was last updated more than six months ago.