So....I have been steadily installing DCC in a range of locomotives and, when looking online for how to open/install DCC on Trainorama 44cl, there are "stern warnings" to remove capacitor C1 - and other general warnings that any locomotive with a capacitor should have it removed.

Well, I didn't remove the capacitor and using a TCS DP2X, the 44cl seems to crawl at setting 1 (128 step) quite OK.  What dire consequences await me?

As many people may say, if it works to your satisfaction, leave them in  but they are easy to remove at the time of installing the  decoder.

I wasn't aware there was any such warning. A few years ago, when I got back into the hobby and bought about a dozen Trainorama models, I fitted DPX2 decoders to the factory boards without any other modifications. All those locos ran perfectly well.

I have since ripped-out all the original boards and wired in new Tsunami sound decoders (I'm hooked on sound these days and can't abide any loco that doesn't have sound).

Roachie

This has been raised many times before. If there is a capacitor across the motor terminals, you should be able to leave it there or disconnect it. Any decoder that does not handle the capacitor is not worth having.
The capacitor is there to reduce radio frequency interference from the motor caused by sparking between the commutator and brushes. There is doubt as to whether it is even necessary these days, as it was primarily a problem in the old days of analogue radio and television. With both television and radio moving to digital transmission systems, the likelihood of your locomotive interfering with these services is quite remote.

This has been raised many times before. If there is a capacitor across the motor terminals, you should be able to leave it there or disconnect it. Any decoder that does not handle the capacitor is not worth having.
The capacitor is there to reduce radio frequency interference from the motor caused by sparking between the commutator and brushes. There is doubt as to whether it is even necessary these days, as it was primarily a problem in the old days of analogue radio and television. With both television and radio moving to digital transmission systems, the likelihood of your locomotive interfering with these services is quite remote.

I understand the RF interference issue, but Googling suggests that it also interferes with smooth low speed running - I haven't noticed that (n=1) but I was looking for broader information from the experts

If my memory is still functioning, leaving the capacitor in MAY infer with BEMF to assist the decoder. This has been discussed a lot in UK forums and even on websites here in Aust. Bachmann recommend removal when installing decoders.

Wow, this is going to open a can of worms.

I make my own decoders, for me it's leave them in, I am inherently lazy and can seldom be bothered to go to the effort.

The only time I regard it as being prudent to remove the capacitors is when you find there are two capacitors soldered to one brush each and then soldered together on the motor can. This potentially violates the absolute DCC law that the brushes must be isolated from the chassis.

Older schools of thought will suggest removal, but the newer idea is that any decoder worth it's PCB is quite capable of functioning flawlessly anyway.

Bachmann don't know what they are doing, advocating the removal of suppression techniques, whilst quite elegantly supplying me with a DCC equipped (or 'DCC on board' as they call it) loco with suppression fitted to the motor. When I emailed them they responded that they recommend I operate their locomotives in an unaltered state for maximum pleasure and reliability. When I sought clarification as to whether this referenced just leaving their DCC equipped gear as is, or whether I ought not remove the suppression when converting their DC locos to DCC I did not get a response.

I understand the RF interference issue, but Googling suggests that it also interferes with smooth low speed running - I haven't noticed that (n=1) but I was looking for broader information from the experts

Well, that sure tells me. I would have though I WAS relatively expert in this area. The real problem is that what you find from Googling is probably NOT from an expert.

This is like a few DCC myths/concepts/recommendations/requirements/must do     ideas - some are good, some are suspect but it really boils down to each User - what works for one may not work for another.

IMO, leave them in if they do not cause problems.  Of course if you are the type of modeller who strips all the wiring & PCB's out and hard wires everything, then that is the ideal time to remove them.

For me, the simple answer in all things is if you are not experiencing any problems, then why change, because you read or hear somewhere from someone that you have to do this or that.

If you cannot get smooth control using a TCS decoder then there must be something wrong somewhere especially if its a DP2 type, I have had absolutely no problems with those decoders in several off the shelf RTR models from both TOR, & Austrains.

What Blacksmith says regarding googling is spot on, although you may find some good info there, its not always the case.

For me, the simple answer in all things is if you are not experiencing any problems, then why change, because you read or hear somewhere from someone that you have to do this or that.

If you cannot get smooth control using a TCS decoder then there must be something wrong somewhere especially if its a DP2 type, I have had absolutely no problems with those decoders in several off the shelf RTR models from both TOR, & Austrains.

What Blacksmith says regarding googling is spot on, although you may find some good info there, its not always the case.

Which is why I seek the counsel of experts in the field - people actually doing it.  I will report back once I get n>1

Which is why I seek the counsel of experts in the field - people actually doing it.  I will report back once I get n>1

Gremlin - take a look at the advice here on page 51 of the Zimo decoder manual:
http://www.zimo.at/web2010/documents/Small_Decoders.pdf

Basically, it says that noise suppression components can sometimes cause a problem, but it is more likely to be the chokes rather than the capacitors that are the issue.
In short, these noise suppression components are not needed with a DCC decoder fitted, so it is a safe bet to remove them.  

The manual referenced above actually gives advice on how to test to see if there is any issue.

Cheers,
John

UPDATE

I have installed DCC in a larger range (n=13) of Austrains, Trainorama, On-Track and Eureka without removing capacitors from any of them; they are all DCC-ready units, with either 8 or 21 pin plugs.  Mainly TCS decoders, some NCE.  The only difference I have seen is that all TCS equipped locos run with speed setting 1 whilst NCE need 3 or 4 to crawl.  Will get another TCS and try them in the NCE equipped locos to see if it is the loco on controller.  No other side-effects of not removing capacitors or chokes or....anything.  Perhaps I am just lucky?  I now have another 25 to wire in, over time.....

UPDATE

I have installed DCC in a larger range (n=13) of Austrains, Trainorama, On-Track and Eureka without removing capacitors from any of them; they are all DCC-ready units, with either 8 or 21 pin plugs.  Mainly TCS decoders, some NCE.  The only difference I have seen is that all TCS equipped locos run with speed setting 1 whilst NCE need 3 or 4 to crawl.  Will get another TCS and try them in the NCE equipped locos to see if it is the loco on controller.  No other side-effects of not removing capacitors or chokes or....anything.  Perhaps I am just lucky?  I now have another 25 to wire in, over time.....

What you are seeing is just the side effects of choice, see my other comments elsewhere when you originally asked for recommendations, I was not kidding.

What you are seeing is just the side effects of choice, see my other comments elsewhere when you originally asked for recommendations, I was not kidding.

Which is why (n=11) are TCS, the NCE were for comparison purposes and the sample size is large enough to be indicative and confirm your opinion

UPDATE

I have installed DCC in a larger range (n=13) of Austrains, Trainorama, On-Track and Eureka without removing capacitors from any of them; they are all DCC-ready units, with either 8 or 21 pin plugs.  Mainly TCS decoders, some NCE.  The only difference I have seen is that all TCS equipped locos run with speed setting 1 whilst NCE need 3 or 4 to crawl.  Will get another TCS and try them in the NCE equipped locos to see if it is the loco on controller.  No other side-effects of not removing capacitors or chokes or....anything.  Perhaps I am just lucky?  I now have another 25 to wire in, over time.....

So without any expertise on the subject you ignore the expert advice. Not a wise decision.

Leaving the capicators or chokes on the motor will not impact on low speed running, when there is minimal back EMF. However a non linear motor response is more likely to happen at higher speeds. The decoders ability to adjust for this sharp jump in voltage from the filtered voltage from BEMF  between pulses will depend on the decoder, motor and capacitors (or noise filter circuit).

Terry Flynn.

So without any expertise on the subject you ignore the expert advice. Not a wise decision.

Leaving the capicators or chokes on the motor will not impact on low speed running, when there is minimal back EMF. However a non linear motor response is more likely to happen at higher speeds. The decoders ability to adjust for this sharp jump in voltage from the filtered voltage from BEMF  between pulses will depend on the decoder, motor and capacitors (or noise filter circuit).

Terry Flynn.

"So without any expertise on the subject you ignore the expert advice. Not a wise decision."

Well, the expert advice so far ranges from members here saying do it and don't do it, the manufacturer's sites saying remove capacitors or remove chokes and Google saying both.  In all cases they refer to slow speed running and not higher speed.  This is why I am confused and have decided to do empirical testing, because the literature and conversations are contradictory....with all the best intent.  References for my comments can be supplied if required and I mean no offence, I am just trying to define a "best outcomes" position

Sorry, I should add that I am talking about locomotives that are marketed and sold as DCC compatible/ready - they all have NMRA plugs and that I can see, none have capacitors on the motors, only capacitors/chokes on the circuit board that provides the NMRA plug.

I would *definitely* remove capacitors from non-DCC ready/compatible locomotives!!

"So without any expertise on the subject you ignore the expert advice. Not a wise decision."

Well, the expert advice so far ranges from members here saying do it and don't do it, the manufacturer's sites saying remove capacitors or remove chokes and Google saying both.  In all cases they refer to slow speed running and not higher speed.  This is why I am confused and have decided to do empirical testing, because the literature and conversations are contradictory....with all the best intent.  References for my comments can be supplied if required and I mean no offence, I am just trying to define a "best outcomes" position

The best outcome is to remove the capicators and inductors, that is the 'expert' advice.  To 'see' the effect objectively the tool to use is a digital oscilloscope.

Terry Flynn.

The best outcome is to remove the capicators and inductors, that is the 'expert' advice.  To 'see' the effect objectively the tool to use is a digital oscilloscope.

Terry Flynn.

What? Like these images?

[img]http://noarail.com/members2/d/32455-2/Inductive+Motor+Load.jpg[/img]
This is what the output across a permanent magnet motor looks like 'au naturel' note that the RMS (pretty much what your multimeter will read is 12.72V. - That you would think was acceptable and perfectly normal, but then your multimeter does not seem the inductive spiking from the commutator as the armature rotates... 79.2Vpk that is possibly almost going to be enough over time to kill the output FETs on your decoder, few SMD FET devices (found on a decoder) are rated to withstand such a PIV.

[img]http://noarail.com/members2/d/2010-2/Diode+Clipped+and+Capacitor+Filered.jpg[/img]
This is what the output across a permanent magnet motor looks like when we leave (or in this case add) the capacitors in. The neat things to note here is that the peak volts has dropped to a very respectable 13V which is very safe for just about every active electronic device and that our RMS volts have DROPPED to 9.088V from the 12.72VRMS above.

What does that mean? The motor drive FETs are less likely to start smoking, AND the RMS volts (in lay terms power output) is LESS. What does this tell us? At a low throttle setting, where the motor is going to be given less power, adding/keeping the capacitors will see the motor get less power again, INCREASING the likelyhood of poor running at this range. At higher speeds, where the motor is provided with more power to run at increased speed, having a small reduction in power applied due to the capacitors is going to have a lesser impact on performance.

Filtering on the motor impacts most at low speed setting, but has little effect on BEMF because there is little BEMF being sourced anyway.

What? Like these images?

[img]http://noarail.com/members2/d/32455-2/Inductive+Motor+Load.jpg[/img]
This is what the output across a permanent magnet motor looks like 'au naturel' note that the RMS (pretty much what your multimeter will read is 12.72V. - That you would think was acceptable and perfectly normal, but then your multimeter does not seem the inductive spiking from the commutator as the armature rotates... 79.2Vpk that is possibly almost going to be enough over time to kill the output FETs on your decoder, few SMD FET devices (found on a decoder) are rated to withstand such a PIV.

[img]http://noarail.com/members2/d/2010-2/Diode+Clipped+and+Capacitor+Filered.jpg[/img]
This is what the output across a permanent magnet motor looks like when we leave (or in this case add) the capacitors in. The neat things to note here is that the peak volts has dropped to a very respectable 13V which is very safe for just about every active electronic device and that our RMS volts have DROPPED to 9.088V from the 12.72VRMS above.

What does that mean? The motor drive FETs are less likely to start smoking, AND the RMS volts (in lay terms power output) is LESS. What does this tell us? At a low throttle setting, where the motor is going to be given less power, adding/keeping the capacitors will see the motor get less power again, INCREASING the likelyhood of poor running at this range. At higher speeds, where the motor is provided with more power to run at increased speed, having a small reduction in power applied due to the capacitors is going to have a lesser impact on performance.

Filtering on the motor impacts most at low speed setting, but has little effect on BEMF because there is little BEMF being sourced anyway.

Your pictures are of a constant pulse width, not typical PWM used by DCC decoders.
Show some pictures of an actual model train motor driven by a DCC decoder, in particular show the BEMF  at different speeds and different pulse widths. The spike can be mostly removed by appropriate circuit design in the DCC decoder. I have been running DC  PWM control of model train motors for over 20 years, without motor capicators and without a motor failure due to this spike issue. Considering most DCC decoders are also fitted to motors with no capacitors across the motor, and DCC works, where are all the overheating motors?

Terry Flynn.

Your pictures are of a constant pulse width, not typical PWM used by DCC decoders.
Show some pictures of an actual model train motor driven by a DCC decoder, in particular show the BEMF  at different speeds and different pulse widths. The spike can be mostly removed by appropriate circuit design in the DCC decoder. I have been running DC  PWM control of model train motors for over 20 years, without motor capicators and without a motor failure due to this spike issue. Considering most DCC decoders are also fitted to motors with no capacitors across the motor, and DCC works, where are all the overheating motors?

Terry Flynn.

Those traces have come from the PWM output of a DCC decoder... They might look constant, but that's probably because there's only two of them.

When the motor is under an idle load at constant speed what you see is what it gets. PWM only alters in duty cycle according to speed and load.

I don't recall anyone suggesting that spiking was a danger from the motor, perhaps you might like to reread my post, the output FETs on the decoder are the suspects for failure under the PIV that comes FROM the motor switching. The similar FETs in your PWM DC throttles probably have an external diode to clip this spike, or it may be included within the discrete package as part of the device, or it may just be that the discrete device has an epically high voltage rating.

Why would the motor be expected to overheat? Where did that idea come from? Portescap motors might suffer heating under PWM DC and hence the slight reluctance to use them on DCC but unless I have misread this thread's title and OP has been mostly about TrainOrama locos, which unless I am very much mistaken use a fairly standard iron core armature.

Those traces have come from the PWM output of a DCC decoder... They might look constant, but that's probably because there's only two of them.

When the motor is under an idle load at constant speed what you see is what it gets. PWM only alters in duty cycle according to speed and load.

I don't recall anyone suggesting that spiking was a danger from the motor, perhaps you might like to reread my post, the output FETs on the decoder are the suspects for failure under the PIV that comes FROM the motor switching. The similar FETs in your PWM DC throttles probably have an external diode to clip this spike, or it may be included within the discrete package as part of the device, or it may just be that the discrete device has an epically high voltage rating.

Why would the motor be expected to overheat? Where did that idea come from? Portescap motors might suffer heating under PWM DC and hence the slight reluctance to use them on DCC but unless I have misread this thread's title and OP has been mostly about TrainOrama locos, which unless I am very much mistaken use a fairly standard iron core armature.

Motor heating of motors due to PWM has often been a claim by some on the net, I drifted off topic. Core less motors were a problem, however today some DCC decoders are designed to handle these types. Clearly any DCC decoder designer is going to use output transistors with a suitable maximum voltage rating to handle any resulting switching spike. Motor speed is determined mostly by average voltage, not RMS voltage. The switching spike has minimal impact on average voltage.  What does impact greatly on average voltage is the extra capacitance which in turn makes regulation more difficult to achieve, considering the decoders often use the back EMF as feed back to adjust the duty cycle.

If you had presented a 5% duty cycle pulse with and without capacitors and a 75% duty cycle pulse with and without capacitors we would see how the impact of capacitors is  greater at higher speeds.

It might be having the switching spike can be an advantage for extra low speed running, as it might give the motor a bit of a short kick, overcoming static friction earlier on.

Terry Flynn.

Off Topic I realise, but overheating of core-less motors is a real problem, as is damage by shock. Metal core motors are better able to withstand impact shocks and also to dissipate heat in the windings because of the heat sink effect of the metal core. Core-less motors have a birdcage of wire with no real structure to support it and if overloaded can heat rapidly. With no way inside a closed can to dissipate the heat, damage can be severe.

Motor heating of motors due to PWM has often been a claim by some on the net, I drifted off topic. Core less motors were a problem, however today some DCC decoders are designed to handle these types. Clearly any DCC decoder designer is going to use output transistors with a suitable maximum voltage rating to handle any resulting switching spike. Motor speed is determined mostly by average voltage, not RMS voltage. The switching spike has minimal impact on average voltage.  What does impact greatly on average voltage is the extra capacitance which in turn makes regulation more difficult to achieve, considering the decoders often use the back EMF as feed back to adjust the duty cycle.

If you had presented a 5% duty cycle pulse with and without capacitors and a 75% duty cycle pulse with and without capacitors we would see how the impact of capacitors is  greater at higher speeds.

It might be having the switching spike can be an advantage for extra low speed running, as it might give the motor a bit of a short kick, overcoming static friction earlier on.

Terry Flynn.

Have you been hanging out with David Peters too much?

The inductive spike is a result of a collapsing magnetic field (of a motor already moving) when the commutator cuts the current flow to the coil. There is no current of note produced, it's an open circuit. The ability to turn comes from torque, torque in a brushed DC motor comes from current, of which the spike produces ... none... Therefore at no time does (or can) this voltage spike provide assistance to overcome any amount of rotational or static friction.

Off Topic I realise, but overheating of core-less motors is a real problem, as is damage by shock. Metal core motors are better able to withstand impact shocks and also to dissipate heat in the windings because of the heat sink effect of the metal core. Core-less motors have a birdcage of wire with no real structure to support it and if overloaded can heat rapidly. With no way inside a closed can to dissipate the heat, damage can be severe.

Once the temperature in the windings reaches sufficient to soften the former made from plastic/resin/dog poo/whatever it is, the rotational forces displace the windings and the motor is good for the bin only. Old fashioned low frequency PWM controllers and early DCC decoders worked at something in the sub kHz range and would kill a coreless motor quickly. Newer, more modern DCC decoders and PWM controllers operate about 22kHz some even way up at 40kHz, these much higher frequencies of switching will allow the coreless motors to survive just fine. (But I suspect you already knew that anyway).