Odd shorting behaviour in Lenz equipped Heljan 33

[A.B. Window Cleaning]

Evening all.  So I have a Heljan 33 with Lenz silver decoder which will frequently short out the dcc system. Usually over most points, sometimes on straight track and this behaviour can always be induced by physically stopping the loco with one's finger, whereupon the shorting will instantly occur.   HOWEVER, … if I disconnect the power to my points decoders board, ( a DCC concepts ads-8fx), the problem vanishes, as completely as the pound coin in your pocket on the approach to a motorway toll booth.

The loco also runs perfectly on DC.

All my other locos are unaffected. (They also use Lenz decoders).

My system is a sprog / jmri which, this particular issue aside, I find amazing in terms of what it can do, (and for under £50 quid too).

 

Is anyone else experiencing anything similar?

[Enterprisingwestern]

 

Is it an original run 33? If so, and this may not be a relevant issue, but they were prone to shorting out the decoder pins due to insufficient clearance between the underside of the socket and the chassis block.

 

Mike.

[Nigelcliffe]

1 hour ago, A.B. Window Cleaning said:

Evening all.  So I have a Heljan 33 with Lenz silver decoder which will frequently short out the dcc system. Usually over most points, sometimes on straight track and this behaviour can always be induced by physically stopping the loco with one's finger, whereupon the shorting will instantly occur.   HOWEVER, … if I disconnect the power to my points decoders board, ( a DCC concepts ads-8fx), the problem vanishes, as completely as the pound coin in your pocket on the approach to a motorway toll booth.

The loco also runs perfectly on DC.

All my other locos are unaffected. (They also use Lenz decoders).

My system is a sprog / jmri which, this particular issue aside, I find amazing in terms of what it can do, (and for under £50 quid too).

 

Is anyone else experiencing anything similar?

 

No experience, but "short" and "overload" are not the same thing, though have identical symptoms - the command station shuts down when the current exceeds its limit.

 

Heljan locos have a reputation of drawing quite a lot of current. The DCC Concepts board is drawing some power.   A standard Sprog setup is fairly low power (1A for some of them).    Stalling a loco will generally push its power demand to "maximum".    So, all that lot suggests to me "overload", ie. trying to draw more power than the system can deliver.

 

 

- Nigel

 

[Flood]

I would always add insulation tape to the underneath of any Heljan circuit board. There are exposed stubs under the DCC socket and a number of other items which I have known to cause a short when pressed onto the metal chassis (I was using a Lenz direct at the time in a 47 and the body pressed down on the chip which caused the circuit board to short).

 

Since adding tape to every Heljan loco I have had no more problems - that's about 10 years now.

[NIK]

Hi,

 

Adding to what Nigel has said it appears Sprogs are not currently supplied with a power supply.

 

All of my Heljan 33's draw  a bit less than one Amp (on 12V DC, on a rolling road) but some out there might do (such as if the gear towers are not properly lubricated).

 

The power supply input requirement on the Sprog II can be up to 15 volts DC - so if the voltage is towards 15V there may be more current draw than on 12V.

 

The power supply itself may be shutting down rather than the Sprog if the power supply has not enough to power the Sprog and the DCC gear attached to the Sprog.

 

Does your Sprog have an indicator that the current drawn is over the Sprogs limit?.

 

Regards

 

Nick

 

 

[Phil S]

The current draw to run at the same speed will increase if the supply voltage is reduced ... ohms law and P=IV

It is therefore more efficient to use a 16Vdcc track voltage than 12V ... as this is compensated for, at 16V, by needing only 75% of the pulse width at 12V ... less heat is dissipated in the decoder.***  HOWEVER - when a short occurs, the higher originating voltage would result in a higher track current - up to the limit of the power supply/protection level. ...which in turn makes short detection easier.

 

I fully agree with earlier comments about the risk/probability of the plug shorting by being pushed down 'too far' - and the need for insulation beneath the socket .... Roco use a piece of plastic in many cases . A thin piece of plasticard may be more durable than the insultating tape alone. Unfortunately, WHICH pins get shorted is not so predictable - it could be the pickup/track pins [shorting the controller only] or the motor pins [ possibly damaging the decoder], or even a function, or combination of the above.. which may bypass any protection in the decoder itself.

 

The [ Heljan ] locos current draw will be most when stationary ['stall current'] - just starting (and a Lenz decoder is usually  rated for a higher current for 3 seconds, BUT any load hauling, gradient or curved track will increase the load from that of a simple test on level track with a light engine. More margin/headroom can be provided by replacing the bulbs with LEDs and resistors.

 

I have recently 'successfully' tested TTS decoders in a Heljan diesel: by adding a 10 ohm resistor in SERIES with the motor - I have ensured that the motor current CANNOT exceed the current rating of the (TTS) decoder (again ohms law and a multimeter check).

This will reduce the maximum speed and power of the loco in use - slightly.  But the unlimited top speed (no CV5) of the TTS decoder is more than required [with my 16Vdcc track voltage - Roco 18Vdc smps ] ... but I have yet to test the result on my 2%-4% gradient with a 'full load' behind. ...but that would be no more than 5-6 coaches and 600mm radius curves in my case.

 

***Back in Zero-1 days, when the track power was NOT full wave rectified (so may not apply if ZTC Zero-1 decoders are used)

the 4A current limit could be reached 'easily' if all the locos were travelling in the same direction, but possibly twice as many locos could be running if equal numbers were running in each direction ... because each direction had its OWN power cycle time-slot.... Locos  in 1 direction would ALL be taking power at the END of that direction's power cycle (because higher speeds turned the thyrister or Triac on early in the time-slot).  

One of the benefits / assumptions of DCC decoders is that they full-wave rectify the track signal, and then pulse-drive the motors at their own (decoder controlled) rate ... so hopefully different decoders even out their current demand on the controller ... maximising the number of locos able to run within the limit.

Early decoder designs (including Zero-1's) also used a low frequency (often mains-related) pulse rate for motor drive .. resulting in large heavy pulses, whereas modern decoders tend to use high frequency pulse rates for motor drives giving smoother, more even control and current demand.  .... I recommend monitoring the system current = often displayed by advanced controllers/handsets/smartphones ... to see what quiescent current is taken by other decoders, and then the loco motor  as it traverses the layout.

 

[34theletterbetweenB&D]

Nigel Cliffe's diagnosis fits all the evidence presented, overloading the system supply to cause shutdown rather than shorting.

 

I would suggest removing the decoder temporarily from the class 33, putting in the blanking plug and giving it a DC test. Firstly it is then easy to measure current draw, both  running and stalled at 12V, and compare to system current supply rating. Also to assess how free running the mechanism is, by getting it moving at full speed on a 12V supply, and measuring how far it coasts on straight and level track when the power supply is disconnected (don't wind down the power, break the connection to one rail). I don't have a 33, but the identical design drive line in the 26, and very similar design in the class 23 are capable of coasting at least 4 feet on flywheel action. If they do not, then it is usually the grease in the gear towers preventing free movement. The operating current draw comes down very significantly once this is corrected, by removing most of the grease.

[A.B. Window Cleaning]

Thanks and apologies to all of you who contributed such detailed replies.   I had assumed that any replies would pop up in my email as per the jmri forum. I did think it a little odd that no one had chipped in!

      Haven't had any real diagnostic time due to (life) but I'll certainly insert some plasticard as a first port of call. I also believe the loco is more prone to this behaviour when another loco is also running on the layout which seems to point a finger at the small (iirc) power supply for the sprog. Although, I find it weird that upon removal of the points decoder board from the system that the problem never occurs, even with three locos on the go.  Now that I know that you guys are here I'll keep you posted, and thanks again.

[RAF96]

If the points decoder has a built in CDU then this will cause an in-rush of current to recharge the caps. Can you replicate the problem by firing a point or two as the loco is running normally.

 

I would agree that upping the Sprog power supply unit if available will likely cure the ‘fault’.

 

As an example the Hornby 1-amp wall-wart PSU controls an overload by dropping its output voltage which causes the attached controller to reboot until the cause of the voltage drop (usually a track short) is removed. Could this be happening with the Sprog (where is Crossland when you want him).

[newbryford]

3 hours ago, A.B. Window Cleaning said:

Thanks and apologies to all of you who contributed such detailed replies.   I had assumed that any replies would pop up in my email as per the jmri forum. I did think it a little odd that no one had chipped in!

      Haven't had any real diagnostic time due to (life) but I'll certainly insert some plasticard as a first port of call. I also believe the loco is more prone to this behaviour when another loco is also running on the layout which seems to point a finger at the small (iirc) power supply for the sprog. Although, I find it weird that upon removal of the points decoder board from the system that the problem never occurs, even with three locos on the go.  Now that I know that you guys are here I'll keep you posted, and thanks again.

 

An ADS8FX decoder takes somewhere between 120 and 240mA depending on how many frog relays are energised (all off = 120mA, all on = 240mA)

 

Yes - there will be some surge as mentioned above but when firing a motor, this will only be for one channel/capacitor rather than the larger surge on startup of the system.