Adding TCS stay alive to Zimo decoder

[Tickplan]

No its not a question of whether this can be done it is a question of how. The Zimo MX 645 comes with two wires to connect to a capacitor. The two wires are coloured blue (to the capacitor positive) and black (to the negative or decoder ground). The TCS KA1 also has two wires - a blue marked as 12v positive and a black with white stripe marked as ground. So far so good and all looks very simple - blue to blue and black to black and white stripe - or is it?

The suggestion is that the blue of the KA1 should go to the blue positive of the decoder. However, the Zimo has a second blue wire close to the black one that goes to the capacitor. Is this the same blue wire or a different one? In other words do the two blue wires connect internally on the decoder together or are they doing something different? If it is a different blue wire which one should I connect to? I am sorry if all this sounds very simple but I do want to be very sure since there is over £100 worth of electronics at stake and I am sure warranties will not cover this sort of thing.

On that subject the Zimo instructions clearly state, "Total capacity of all connected capacitors must not exceed 5000uF. NO gold caps!" Does this apply here and therefore connecting the KA1 will invalid the warranty or worse or am I worrying unnecessarily?. I have Zimo decoders in most of my locos and love them. I also have TCS complete decoders with stay alive functions built in and they too are brilliant so was hoping to combine the best of these two worlds. I know the likes of Nigel Cliffe have done this and report that it works but I just wanted some reassurance before committing the soldering iron to the job.

 

Thanks in advance of your comments as always.

 

Tickplan 

[Nigelcliffe]

Its not "dead simple".  And, this reply gets a bit technical.

 

 

Some newer Zimo decoders have internal charging circuits for capacitors (resistor/diode/inductor), and then present those as two solder points or two wires.  The two contacts are decoder positive (blue) and a capacitor negative.  These contacts are designed to have ONLY capacitors connected to them.  The Zimo manual gives a voltage rating of the capacitor depending on decoder (the PLuX decoders seem to need 16v rated capacitors, other types depend on track voltage).

 

 

All Zimo decoders (old and new) have a decoder ground and a decoder positive.  The positive is the normal decoder Blue wire. The Ground is a solder pad.  Consult the Zimo manuals for the location of the solder pad(s).  On some designs (eg. MTC 21Pin), the ground is presented on one of the connection pins. 

Excluding really small capacitors, any capacitor connected to the ground and positive requires a charging circuit (resistor/diode) and ideally an isolation inductor.  A large capacitor needs a discharge resistor.  The relevant circuits are in the Zimo manuals for some decoder types (the diagrams on page 61 for the MX623 and MX630 is the clearest), and can be generalised to any decoder. The arrangement of the charging 68ohm resistor and discharge diode can be moved to the negative wire if that is more convenient (assuming one understands what each component does, and which way current is flowing !).    The voltage rating for the capacitors in this case will be based on the track voltage, and the 16v track voltage rating of the TCS KA units should not be exceeded.

 

 

I think the Zimo rating of 5000uF (0.005F) refers to their charging circuit (first paragraph above).  The manual describes it in terms of current handling (which usually translates to "heat").  I'm not totally convinced by the manual's explanation, but for now assume it applies.   

The TCS KA units are in the region of 0.1F.  

 

 

 

The TCS KA1 and KA2 are a circuit containing both capacitors AND a charging (resistor/diode) circuit.   It doesn't seem wise to duplicate this charging circuit (because it drops voltages, which means things don't work as well as they should).  So, the choice is

• Either : Chop up the TCS unit interally to by-pass the diodes/resistors, and then use the bare capacitors (in series) using the Zimo capacitor connections.  It may be over-volting the TCS capacitors as I don't think the KA2 series quite add up to 16v rating (in the original TCS package there is a little bit of voltage lost in the diodes).  And, it's approx. 20 times the maximum capacitance quoted by Zimo in their manuals.
   
• Or :  Connect the TCS to the Zimo Ground and Positive.  Ideally add a discharge resistor and decide if an additional isolation inductor is also required.  

 

 

I've used the second method - connect to the Zimo Ground and Positive, and then use the TCS KA1/KA2 without modification. I fit a discharge resistor across the terminals to force the KA unit to fully discharge when off the track - I measured 25 minutes discharge time with the resistor fitted.  I don't bother with the inductor, instead arrange things so they can be disconnected if I need to blow new sounds or new firmware into a decoder. 

 

The time duration that the stay-alive actually runs the Zimo decoder can be set in a CV.  So, you can limit the stay-alive run time to, say, 2 seconds. This is plenty of time to deal with any pickup issues, but avoids a loco ploughing along deliberately dead track (or a bare baseboard) for a yard or more without any control. 

 

 

I think you need to be sure you know how these circuits work, and what you're doing before starting.  There is a huge amount of energy stored in a TCS KA unit, and the potential for explosive discharge is present.  I'm not accepting any responsibility for damage you might cause !

 

 

- Nigel

[Tickplan]

Thank you Nigel - I was hoping you would respond. I accept and understand your warnings.

Just to be clear however for an electronics numpty like me is the Zimo black wire that leads to the capacitor on their drawing (page 9 of the instructions) the same as your "Zimo Ground" -and therefore to adopt your second choice involves connecting that to the TCS black with white stripe (and the blue of the TCS unit connected to the decoder blue (positive)?

 

Tickplan

[Nigelcliffe]

Thank you Nigel - I was hoping you would respond. I accept and understand your warnings.

Just to be clear however for an electronics numpty like me is the Zimo black wire that leads to the capacitor on their drawing (page 9 of the instructions) the same as your "Zimo Ground" -and therefore to adopt your second choice involves connecting that to the TCS black with white stripe (and the blue of the TCS unit connected to the decoder blue (positive)?

 

My interpretation of Zimo's drawings is that the black wire to the capacitor on page 9 showing the MX645 is "decoder ground". The "capacitor positive" may be something specific to the MX645's charge/discharge circuit and is not the same as the decoder "common positive".  The "common positive" appears to be next to the index-pin on the PluX connector.

 

But, this is interpreting drawings.  I've not had an MX645 to install - the big batch of 4mm locos I did for someone were mostly MX640's.  My own stuff uses smaller decoders. 

So, to be absolutely certain, ask Zimo. Or, I guess, look up the Plux socket on the Morop website (NEM 658, probably in German only), where it should be specified.

 

- Nigel

[pauliebanger]

The MX645 has a pair of wires, one blue (positive) and one grey (negative) connected at the opposite edge of the decoder to the main wiring harness.

 

These wires are the one's which should be connected to external capacitors. They provide a regulated 16v and are indeed connected to ZIMO's on-board management circuitry. No additional components are required, just connect blue to capacitor positive and grey to capacitor negative. 

 

It is safe to add supercapacitors to the MX645 (it is a 'PluX-like decoder' per the decoder manual and I have it directly from ZIMO proprietor, Dr Peter Zeigler), but there are less expensive ways to acquire them than cutting up TCS KA modules. Buy 2.7v 1F supercaps from suppliers like 'RS online' for about £1 each or much less if you can buy in bulk. Then wire six of them in series to give 16.2v and a massive 166,666 uF.

 

You can restrict the duration that the motor will be operated after it looses DCC signals with CV153.

 

Kind regards,

 

Paul

[pauliebanger]

Thank you Nigel - I was hoping you would respond. I accept and understand your warnings.

Just to be clear however for an electronics numpty like me is the Zimo black wire that leads to the capacitor on their drawing (page 9 of the instructions) the same as your "Zimo Ground" -and therefore to adopt your second choice involves connecting that to the TCS black with white stripe (and the blue of the TCS unit connected to the decoder blue (positive)?

 

Tickplan

 

No.

 

The ground/negative is as you say, but the blue wire shown in the diagram does not exist on the actual decoder. It is not the blue common positive, but a solder pad to which you can connect capacitors' positive. 

 

These are not common to all ZIMO decoders, my response is for MX645 only. eg MX644 specifically states NOT to use normal ground for capacitor connection, so as Nigel suggested, be very sure before connecting caps.

 

The blue and grey at the 'capacitor end' of the decoder (I've never seen a black wire here) is the safe way to connect external caps of any description. (Minimum 16v) 

 

Paul

[Nigelcliffe]

It is safe to add supercapacitors to the MX645 (it is a 'PluX-like decoder' per the decoder manual and I have it directly from ZIMO proprietor, Dr Peter Zeigler), but there are less expensive ways to acquire them than cutting up TCS KA modules. Buy 2.7v 1F supercaps from suppliers like 'RS online' for about £1 each or much less if you can buy in bulk. Then wire six of them in series to give 16.2v and a massive 166,666 uF.

Thanks for that Paul.

I couldn't understand why there is a 5000uF limit in the manuals, and its good to know from Dr Zeigler that it doesn't really apply.

- Nigel

[pauliebanger]

Thanks for that Paul.

I couldn't understand why there is a 5000uF limit in the manuals, and its good to know from Dr Zeigler that it doesn't really apply.

- Nigel

Nigel,

 

To complete what I was told regarding the MX645 and gold or super caps is that they are OK provided the track voltage is maintained below 20v. This is due to potential heat bulid up (which you anticipated) and damage to certain components on the decoder at high track voltage.

 

The 5000uF 'safe' limit still applies to normal electrolytic caps at all track voltages due to the duration of charging causing heat build up on the decoder.

 

BTW, the 16v regulated cap charging output does not apply to all ZIMO decoders with cap management on-board. So,e.g. MX644D which is an MTC design, requires cap voltage to match or exceed the maximum track voltage.

 

Currently 'all MX633, MX645, MX695, MX696 and MX697 have internal circuit that limit the requirement to 16 Volts'

 

 

Kind regards,

 

Paul

[Simond]

Paul

 

I was talking with one of my colleagues at work about stay-alive capacitors, and remarked how effective I find them to be. He is a very knowledgeable & experienced electronics engineer, with a hobby interest in model engineering generally. He asked what I did to ensure the capacitors were not subject to excessive voltage - I replied that I connected 6 in series, giving me a nominal voltage of 16.2V for a bus voltage of about 15, thus giving a hopefully reasonable margin.

He was concerned as the voltage distribution across series capacitors will be in proportion to their capacitance, (only roughly the same as the fraction of the overall applied voltage) - and the tolerance on this kind of device is in the range of -10 to +30% - which could lead to significant differences in voltage across individual capacitors, and this will only worsen as they age. I have attached the data sheet for your reference.

http://www.digikey.c...df?redirected=1

When subject to excessive voltage, the failure mode of the capacitors is to burst, and dribble electrolyte - this is undesirable at best, but would be a real shame inside a model, which might be significantly damaged by it. I understand that it is unlikely to be dangerous.

His recommendation is to fit a quite high value resistor across each capacitor - say 100k, as this will ensure the voltages across the capacitors are related to the resistors' values, and these are more stable than the capacitances.  This will marginally reduce the KA duration..

 

Do you have any background info on this, or thoughts about it?

 

I'm planning to fit resistors to the locos to which I have already fitted caps, and all future ones.  Would appreciate your views before I start stripping stuff down.

 

kind regards

Simon

[Nigelcliffe]

Simon,

 

your engineer is correct on electronics theory.     But, practise seems to be that manufacturers don't bother.   The TCS KA units lack a resistor ladder (I've not dismantled a decoder with integral KA module, but would be surprised if they built them differently)  and the guidelines from Zimo omit resistor ladders.   I've not done the maths, but wonder if even with every component at extremes of tolerance ranges, the capacitor's tolerance of over-voltage may still save the day.  

 

Some manufacturers use different approaches; a single capacitor with a voltage multiplier to ring the lower voltage up to that needed by the decoder will avoid the issue.    

 

If fitting a resistor ladder, and tight for space, surface mount devices will be a lot smaller than conventional devices with metal legs; there will be hardly any current flowing in the resistors.  Or, if sticking the legs, metal film resistors are usually a lot smaller than carbon film resistors.    And, if you do fit a resistor ladder, then,  if you choose the total resistance of the ladder wisely,  the discharge resistor described in the Zimo manuals is redundant. 

 

 

 

I covered resistor ladders in an article on Stay Alive devices in ScaleFour News some time ago.   Its only an issue when using lower voltage rated capacitors in series to cope with a higher applied voltage.   Its not relevant to parallel capacitor banks to increase total energy storage.   I've never fitted a resistor ladder.  

 

 

- Nigel

[Crosland]

The theory is correct but given the way some railway modellers ignore best practice in other areas (e.g. the comment "it works for ME!") ...

 

Andrew

[Simond]

Nigel, Andrew,

 

Thanks for the comments - after discussion with Peter (my colleague) I had no doubt of the theory (I think they call it "lifetime learning". ) but the question becomes one of "can I also get away with it or should I take the three locos I have with KA fitted apart and fit the resistor ladder?"

 

One of these will be refitted with a sound decoder just as soon as someone does one for a 13xx, so it's no hardship.

 

I can take the other two to bits, one of them is a 28xx and thus has bags of room. The other isn't, and doesn't!

 

I also wrote a bit of a "how to" on KA at the back of last year - I will amend this as I would hate someone to have an "oops" based on my musings.

http://www.rmweb.co.uk/community/index.php?/topic/79052-porth-dinllaen-in-0/?p=1590807

 

Best

Simon

Edited January 19, 2015 by Simond

[Crosland]

If space is at a premium you could stick tiny surface mount resistors to the caps and wire them with fine copper wire. The charge balancing current will be negligible.

 

Andrew

[Art Dent]

Sorry for necro-posting in a 4-year old thread ...

 

This is a question more for Nigel (and/or Simond's friend) ...

 

Would this circuit be appropriate for wiring six 2.7V super-capacitors in series?

 

 

The value of the 1k-ohm resistances are purely arbitrary (but hopefully high enough that negligible current flows across the capacitors whilst connected to track power and also when the stay-alive is powering the decoder/motor) but provides an equal voltage (of 2.5V) across each capacitor, so ensuring that none are over-voltage.

 

I propose using SMD resistors to minimise the 'footprint' of the stay-alive and to encapsulate the whole array (assembly / unit) in heat-shrink to avoid shorts. 

 

Comments and/or improvements welcome.

 

Art

 

BTW, Happy New Year!

[Nigelcliffe]

Charging resistor of 500ohm is too large, 100ohm or even less is appropriate.

 

Balancing resistors of 1kohm, probably too small - at least 10k if not more.   They also make your discharge-the-caps 100kohm resistor redundant, the capacitors will discharge via the balancing resistors (5kohm resistance to discharge, as opposed to 100kohm). 

 

 

- Nigel

[Art Dent]

Thanks Nigel,

 

This more in line with what you are thinking?

 

 

With six 2.7V caps in series they can handle 16.2V and the Zener ensures that they only ever see 15V (plus tolerance of 5%) so 15.75V max in theory.

 

With six 1F capacitors, the effective capacitance would be 0.167F (or 166.7mF / 166,66.7uF).

 

Ideally, I'm not looking for such a huge energy store, but I can't seem to find any 2.7V super capacitors much below 1F (like these: https://www.mouser.co.uk/ProductDetail/Illinois-Capacitor-CDE/DGH105Q2R7?qs=sGAEpiMZZMuDCPMZUZ%252bYlzGTqubQTce6AVq05nm1JIg%3d ).

 

There are some button-type 3.6V 470 000uF  (= 470mF / 0.47F), 330 000uF (= 330mF / 0.33F) and 220 000uF (= 220mF / 0.2F) but they are around £1.50-£2.00 each and have an 8 week lead time.

 

I have also found some 5.5V 0.5F capacitors but again nearly £2.00/unit.

 

This means (IMHO) that it seems to be uneconomic at the moment to fabricate your own super-capacitor stay-alive unit.

 

Regards

 

Art

[Izzy]

 

Thanks Nigel,

 

This more in line with what you are thinking?

 

 

With six 2.7V caps in series they can handle 16.2V and the Zener ensures that they only ever see 15V (plus tolerance of 5%) so 15.75V max in theory.

 

With six 1F capacitors, the effective capacitance would be 0.167F (or 166.7mF / 166,66.7uF).

 

Ideally, I'm not looking for such a huge energy store, but I can't seem to find any 2.7V super capacitors much below 1F (like these: https://www.mouser.co.uk/ProductDetail/Illinois-Capacitor-CDE/DGH105Q2R7?qs=sGAEpiMZZMuDCPMZUZ%252bYlzGTqubQTce6AVq05nm1JIg%3d ).

 

There are some button-type 3.6V 470 000uF  (= 470mF / 0.47F), 330 000uF (= 330mF / 0.33F) and 220 000uF (= 220mF / 0.2F) but they are around £1.50-£2.00 each and have an 8 week lead time.

 

I have also found some 5.5V 0.5F capacitors but again nearly £2.00/unit.

 

This means (IMHO) that it seems to be uneconomic at the moment to fabricate your own super-capacitor stay-alive unit.

 

Regards

 

Art

 

 

I have found using packs of 220uf tantalums to be good enough to overcome basic poor running. The packs are cheap to generate if you get the bits off ebay. I recenty just got another 30 case D's (7x3x3) for £11.99 posted to use with the zeners/diodes/resistors I already have.

 

Here is a basic pack and a 4x one recently fitted into a Bachmann 03.

 

 

 

 

This uses a MX648 and small Zimo sugar cube speaker. The shot was a test one before final fitting/conversion to P4. At speed step 10 of 28 you get around a quarter of a wheel rev when power is lost. So less at lower speeds, but even at creep speed there is no stalling at all. More than enough for me. Especially at the cost effective nature of the arrangement.

 

Izzy

[Art Dent]

Hi Izzy,

 

Thanks for posting.  That first pic - is that three tantalums or six - I can't make it out.

 

What voltage are the tantalums?  Are those two identical 1N4007 diodes??

 

Do you have a curcuit diagram or any more pictures?

 

Looks encouraging.

 

Cheers,

 

Art

[Izzy]

 It's just 3, so 660uf. More than enough for small, efficient motors, especially in 2mm.

 

Full details are here:  http://www.rmweb.co.uk/community/index.php?/topic/107235-2mm-coal-tank-test-build/page-6      starts at post 143

 

The ebay links are old of course, but new searches using them will bring up alternatives. With the mini-melfs/SMD bits I reckon about £2-3 per pack all told. You can of course arrange the tantalums in a variety of ways to fit the space available.

 

All this comes because of the ground work of Nigel Cliffe. Details and more links here: http://www.2mm.org.uk/articles/DCC%20Stay%20Alive/index.html

 

cheers,

 

Izzy

[Nigelcliffe]

Art's last diagram is good enough to start experiments.  Don't be surprised to change some of the resistor values.  A voltmeter over the capacitors will show how fast they discharge via the resistor ladder. 

 

The Tantalum's Izzy is referring to will be 16v rated, and a Zener to stop over-voltage is highly desirable.  The voltage rating for the Zener is a trade-off for; go lower and the capacitors will max out at that lower voltage, which may lead to a speed drop as the loco looses power.  But if the track voltage is, say, 14v, then a 13v Zener is probably fine as the average decoder voltage is around 13v.   Also, if chasing every last ounce, using a Schottky diode rather than a standard rectifier buys another 0.7v output from the stay-alives. 

 

A block of about 10 Tantalums is enough to actually see Zimo's "never stall when stationary" feature working (4mm scale small 0-6-0 with Mashima motor).  With that, if the decoder is coming to a halt (following deceleration) and as it comes to a stop it detects no track power (its on some dirt), the decoder will nudge the motor to drive the loco to some pickup.  Very useful additional trick in the quest for locos which never stall.   

 

 

 

 

- Nigel

[Art Dent]

Thanks Nigel,

 

The Tantalum's Izzy is referring to will be 16v rated, and a Zener to stop over-voltage is highly desirable.  The voltage rating for the Zener is a trade-off for; go lower and the capacitors will max out at that lower voltage, which may lead to a speed drop as the loco looses power.  But  if the track voltage is, say, 14v,  then a 13v Zener is probably fine as the average decoder voltage is around 13v.

 

The difficulty, of course, is actually measuring the DCC track voltage accurately - the DCC signal being not DC and not a sinusoidal AC waveform - which most cheap multimeters expect to measure.  I don't propose buying an £80-odd RRAmpmeter to check - hence playing safe with a 15V Zener and six 2.7V super-caps that should handle 16.2V safely.

 

Looking again at my v2 diagram, would a 30V, 1A schottky be OK - something like a 1N5818?

 

 

As drawn, this circuit would provide 15V (max 15.75V) across the capacitor array, the 10k-ohm resistor ladder provides balancing of the capacitors for charging and draining of the capacitors when the loco is removed from the track.

 

The purpose of the schottky diode is to more-fully drain the stay-alive unit when track power is interrupted?

 

Like I said, my electronics is a bit rusty so any clarification helps my understanding.

 

Just need to source some suitable, cheap 2.7V super-capacitors in the 200mF range

 

Cheers,

 

Art

Edited January 3, 2019 by Art Dent

[Crosland]

The difficulty, of course, is actually measuring the DCC track voltage accurately - 

 

Just connect the decoder to the track and measure between the blue and black wires shown in your diagram, before fitting the stay alive. This is the rectified track voltage. You can ignore the voltage drop in the rectifiers since the measured voltage is what you will actually be applying to the stay alive.

[Art Dent]

Just connect the decoder to the track and measure between the blue and black wires shown in your diagram, before fitting the stay alive. This is the rectified track voltage. You can ignore the voltage drop in the rectifiers since the measured voltage is what you will actually be applying to the stay alive.

 

Hi, yes, I am aware that you can do this to measure the decoder-rectified DC voltage (as applied to the motor) although others may not.

 

What I meant was the AC voltage as presented to the decoder (i.e the voltage between the rails - or the output of the DCC Command Station).  For some reason I was thinking that the stay-alive was presented with the unrectified track voltage (call it a brain-fart).

 

Only a True-RMS digital voltmeter, the RRAmpmeter ('RRAmpmeter' - arrghh! -  that so annoys me!) or indeed an oscilloscope will do this accurately as the DCC voltage is a pulse-width modulated AC square-wave (as subject of much discussion on a  Model Railroader thread in the US - see this link: http://cs.trains.com/mrr/f/88/t/69828.aspx).

 

It is AC as the voltage is symmetrical about 0V - the voltage varies between +/- 22V (the maximum allowed for in the NMRA standard (so that a DC loco placed on the tracks essentially sees the mean voltage - which is zero - and doesn't take off at high speed in any particular direction) but the pulses (representing the digital signals to the decoders) alter in size (see NMRA Standard S-9.1) and the frequency is around 100 kHz.

 

Most 'cheapo' digital mltimeters can't read AC at this sort of frequency and 'expect' to see a smooth sinusoidal AC waveform (which the DCC track voltage decidedly isn't).  You need the likes of a True RMS DMM (such as the cheapest Fluke - the 115) costing upwards of £125.

 

You are right, of course, in saying that the stay-alive will see the rectified track-voltage and that is, of course, all we really need to be worried about here!

 

Art

Edited January 4, 2019 by Art Dent

[Simond]

Art

 

Don’t believe everything you read on the RRampmeter website

 

There are lots of true RMS meters (can’t tell you how true they are, they might be lying) available online

 

Google “true rms multimeter”

 

Best

Simon

Edited January 4, 2019 by Simond

[Graham Radish]

I highly recommend you don't mess with these capacitors, i tried fitting a dcc concepts one to my very expensive zimo class37 sound decoder and it overheated and smoked, lesson learned never again, i went over the decoders current limit and fried it. Just not worth the hassle tbh, theres no need if you maintain your system

Edited January 11, 2019 by Graham Radish