Filtering out Noise

Dear All, has anyone put together a filter for cleaning up the output from DC controllers? I'm thinking of a simple LC circuit that I could knock up
in a few minutes to reduce the noise the ancient controllers at my club throw out and which present some difficulties for DCC fitted stock. Trouble is I've forgotten how to calculate the necessary values for the Inductance and the Capacitance (you tend to forget that sort of thing if you don't use it for 20 years.....). BTW I haven't forgotten I can't use an electrolytic ;->
Lose the spamtrap to reply
--
Ian Birchenough

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Ian Birchenough wrote:

Put your electrolytic _before_ the reversing switch.
Why not just use the old controllers as power supplies and adjust your voltage with a voltage regulator IC?
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

What sort of noise are you talking about?
There should normally be no electrical connection between a DC controller and DCC stock, or are you trying to run DCC fited stock in analogue mode? If it's a PWM controller then it will, indeed, upset a DCC decoder.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

All my stock is DCC fitted, but the clubs layout is still analog DC. The transformer rectifier and controllers are all.... somewhat ancient and I'm suspicious that they're not giving a clean output (haven't been able to borrow a scope to confirm it though). I was thinking of throwing an LC filter into the supply to track to clean any noise off the power so that I could use my stock in Analog mode.
--
Ian Birchenough

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Ian Birchenough wrote:

You've managed to attract my interest now! How susceptible to odd behaviour or damage are DCC decoders in analogue mode? I had assumed that they would be safe so long as the peak voltages remained below 25 volts and that any erratic behaviour would be for milliseconds and therefore not outwardly noticable.
Regards, Greg.P.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
(snipped)

Phil: A DCC decoder when used on DC requires a 'PURE' DC supply - NOT pulsed or half-waved rectified....
Why?
1/ First it needs to power its own internal circuitry - this can be with as little as 3.3V or as high as 8V or so, depending on the model of decoder. Therefore no 'movement' of a loco will occur until this point is reached, and the voltageis increased further.
Stage 2/ From this point upwards, the onboard circuitry interprests the incoming track voltage as a control signal - say (for discussion purposes) its from 5V to 12V: This will be used to act as a 'local controller' directly connnected to the motor.
Depending on the decoder, and some of its settings, it /may/ act as a pulse-width controller to the motor - with frequencies used from about 50Hz (NOT mains power supply related in ANY way) up to 'silent drive' decoders pulsing at 16kHz or above. (Equally, it could theoretically be a 'pure dc drive to the motor with the following relationship: (for example) Track voltage: 0-5V -- Motor Drive voltage 0v Track Voltage: 5.1V -- Motor Drive voltage 3V (programmed 'kick start') Track Voltage: 9V -- Motor Drive voltage 8V (programmable) Track Voltage: 12V -- Motor Drive voltage 11V ( reduced from track by diode drop(s))
and similarly in the reverse direction. (the decoder has an on-board full wave rectifier for its basic power supply, operating off the track voltage, and powering its own control circuitry) -- hence the need for BOTH motor terminals to isolsated from the track)
((the ZTC 3-wire decoders once available for some low power locos provided half-wave power at most (but NOT at 'mains frequency' -> track frequency is about 10kHz), and therefore, as with Zero-1 style decoders, were able to work on non-chassis-isolated motors))
THEREFORE if a pulsed / back emf / half wave rectifier controller is used as the 'dc' track supply, the decoder (stage 1) will keep turning on and off, and spend all its time resetting itself, getting nowhere, fast.
('Pure DC' electronic controllers are ideal dc sources) The benefits of back-emf are available WITHIN THE DECODER itself, on certain models of decoder.
Phil
--
Phil Spiegelhalter: snipped-for-privacy@fillin.co.uk
==== Technical Training for Broadcasters =====
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Phil wrote:

Given that some of us are never going to go DCC and that few DC models run particularly well on smooth constant voltage analogue DC, is there any other form of DC that decoders will cope with?
I'm thinking in terms of perhaps a constant base voltage with an imposed PWM signal at some frequency that would not conflict with the decoder. The PWM could be fast rise or slow rise or even a half sine wave.
I did ask basically this question by e-mail of Lenz and NCE (US) as I thought it would be nice to gain some of the DCC advantages such as feed-back control and speed tabling. I thought it might even be possible to have switchable lighting by using the 12 volts of PWM and sending a DCC command through the rails. Sadly, neither Lenz nor NCE bothered to respond.
Regards, Greg.P. NZ.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
(snippped - see separate original replay)

Phil: Gaugemaster controllers are, I believe, 'constant / steady DC' simply from an elctronic source (low impedance) instead of resistance mat.

Phil: The information to the DCC decoder is at about 10kHz - it is working very fast - and any humanly perceptible 'pulsed' waveform is simply switching the decoder on and off repeatedly, and will offer NO benefit, only problems.

Consider the hf signal to be the DCC control waveform, and stop the british model railway habit (I know you're in NZ - so no excuse/reason to follow british habit 8-) ) of trying to reinvent the wheel all the time - use the benefits of mass production, and leave full voltage on the track at all times, with the controller of your choice sensiobly located AFTER the dodgy wheel contact part, inside the locomotive.
There are NO advantages with 'conventional dc control' once you more past one loco in steam operation. Only unprototypical powered or unpowered sections and 'how should I switch the frog - which route should it be.
Just remember that the 'decoder' IS the controller - of your choice and tailored by you to meet the needs of that loco (eg matching max speed to others or for realism), and its located where the driver should be - in the loco.
For other (model) railway methods - be a signalman 8-)

--
Phil Spiegelhalter: snipped-for-privacy@fillin.co.uk
==== Technical Training for Broadcasters =====
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Phil wrote:

4.5/9kHz.
Is there a frequency above which the decoder(s) will not react?

That's the thing I'm trying to do - use existing DCC decoders on analogue control rather than building my own inbuilt feedback/speed controls.

Nice but ...

We're beginning to wander from the original question of using DCC decoders in analogue mode. Yes, there are a considerable number of advantages to conventional DC control over DCC in my opinion but that's not my purpose in asking.

I've been well aware of the advantages and disadvantages of digital control since the 1970s.

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Agreed, now we've summarised what the problem is, does anyone have an answer to the original question - values for Inductance and Capacitance for an LC filter to kill any noise induced by the rectification process and imposed on the DC supply?

--
Ian Birchenough

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Ian Birchenough wrote:

As far as I can make out from my book on the subject (not sure I understand it totally) voltage, current and frequency are all factors in the design of effective filters. It would seem to me that you would need an inductance with infinite resistance and an extremely large capacitor to achieve smooth DC from a rectified sine wave variable resistor controller!
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
.

Phil: No; all you require is a Voltage Regulator Circuit - available largely in a single IC with 3 legs - just a few external cpacitors. To be energy efficient (it is a green age?) - a switch-mode supply, running at high frequency is more effective, and considerably lighter, if one is starting from a mains supply: The reason being the amount of energy needing to be stored in each half-cycle of waveform: in a 50Hz transformer, the magnetically stored energy requires much more bulk than the smaller amount stored in a high frequency transformer. This is why modern 'plug top' low voltage supplies are so light (and cheap too) (see Maplins catalogue circa 10-20ukp for 3-12V 1A-2A a supplies)
But back to your 'original question' - is it really the correct question you are asking or answering?
What 'noise' are /you/ referring to? - hf switching noise (which may still be present in a switched mode power supply - but is easily filtered by relatively low values of capacitance, or the 100Hz ripple from full wave mains rectification? (and its harmonics)- this is inefficient to smooth due to the large amount of energy in each cycle (hence my comment about using a switched mode regulator above) Are you thinking simply that you 'need to ensure no content at dcc control frequency' - in which case this is a falicy - it is unlikely that random or cyclic variation on your power supply will be interpreted as valid command streams, and therefore (subject to CV29 setting) the decoder will endeavour to operate in dc mode, BUT edges present within the signal (even of not-full amplitude) might cause the processor to attempt to 'reframe' to the incoming 'data' and therefore hiccup.
Gregory wrote in an earlier message:

Phil: You miss the point: once a decoder is installed, even if it is used in 'dc mode', the dc track supply is NOT directly connected to the motor - it has simply been used as a power transmission line to the on board controller (the decoder) and the drive (usually PWM of preset frequency) is generated entirely on board the decoder. Therefore any superimposed signal will not be utilised intentionally. Harmonics of any other frequency added may cause random interference to the processor: you are communicating in a potentially noisy environment. THe decoders use edge detection (Lenz Gold series) or level change to detect the inversion of the incoming power supply, and respod to it as dcc commands (or Fleischmann or Marklin or Zero-1 in some modules) or in the absence of recognisable signal (and subject to CV29) respond as DC by monitoring the track voltage as a control signal as I described before. (THis start-up recognition delay can be modified by programming CV29 not to look for the unwanted/unexpected format eg dcc or dc)
If you were thinking of eg hf coach lighting being added to the dc, to be used directly for lighting, and not affecting the decoder, then think again. Maybe you should go for the type of installation used in some american models in which a switch is included to change between 'basic dc' and the dcc decoder: this takes up space.

Phil: no comment

Phil: I'm not surprised. The dcc system and modules already provide for that control, by a well designed and proven, reliable system. No manufacturer in their right mind will support/encoirage/respond to someone wanting to ignore all their work and effectively misuse the product - perhaps following it with a claim using a TV advertised no cost claim service?
--
Phil Spiegelhalter: snipped-for-privacy@fillin.co.uk
==== Technical Training for Broadcasters =====
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Phil wrote:

That's what I said umpteen posts back - try an LM317 for a 1 amp supply.

I was lead to believe that some DCC decoders do switch to direct connection in DC mode, but that's not particularly relevant.

Therein lies one of the potential advantages of using a DCC decoder in a loco that will only ever operate on analogue DC.

We're back on track! :-) Now I want to know if (say) a low smooth DC voltage and a PWM of perhaps a half sine wave or a saw tooth rise to 12 volts would be acceptable to a DCC decoder?

Why? The NCE DCC decoder copes with this. (it does not have feedback control however)

There has to be some purpose to my fitting a DCC decoder to my locos; the speed tabling and feedback control fall into that category, but I don't particularly want to provide specific analogue controllers for standard DC, DCC and Faulhaber type motors. The chances of error become too great.

Why?
I took the time to explain why I would want to utilize their products and pointed out that a considerable untapped market exists.
There are many modellers like me who have been in the hobby for a long time and have far too many models to ever convert. There is also the fact that many modellers operate by block just like the prototype does and the whole point of DCC is to overcome the block system. That forces one to add a further layer of technology to a DCC layout to get back to the starting point of block control.
Regards, Greg.P.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Phil wrote:

What rise times for the pulse will decoders recognise and what rise times will they ignore? Are fall times also likely to be recognised?
Regards, Greg.P.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.