Triang DC loco's blowing LED lighting (Possibly Back EMF issue)

[Gazbo]

Hi all.

 Hope I found the right section for my question.

I planned to re-lamp my R155 and R57 loco's (Purely DC), with Bi-colour / directional LEDS. 

I bought some 3 leg Red/white 3mm LEDS for the job.

I built the circuit, and connected  it to my section of test track, which also has my rolling road fitted to it. (Minus any Loco at this point.)

The circuit is capable of running to 16VDC, in Both directions, 10k resisted, to protect and keep the brightness down to a reasonable level, up to that Voltage. 

I presently run my single road track from a standard  Benchtop PSU, so that I can monitor power draw, and control the loco speed, and normally only run locos at up to 12V. I will be using an original controller once repaired, which may be able to peak at around 16VDC, hence the protection level.

I am also presently using this benchtop PSU for the circuit testing.

The leds performed amicably, right to 16v, maintaining around 50% of full brightness, from around 6v, up to 16v, without issue. they were successfully red/white directional, when the polarity was switched, so I was delighted ! (I'm a mechanical engineer, not a sparky !) 

So I crock-clipped the flying leads to the power bogie solder points, and set that up on the rolling road, with the circuit positioned down one side, away from metalwork ! Double checked for any short circuit areas. All clear.

So again I slowly wound on the volts. loco usually kicks in around 7.4V anyway. LEDs came on steady at around 6v, but as soon as the motor began to spool, the red led, suddenly went bright white, and popped, before I could back off ! My mate was convinced I must have shorted something, so we changed out the bust led, and tried again.

Same result. Pop ! 

 I reckon this must be some kind of back EMF issue. 

Question. Have others had this issue?

if so - what's the workaround?

IE. does anyone have info on a circuit I could build from components, to fit between the motor, and the led circuit, to eliminate the issue

( remember - it needs to run with the loco In both forward and reverse ) ?

I know there are some circuits out there, for example using a back to front P FET in the circuit to block reverse polarity,

but all I can find, will only suppress back emf in one direction/polarity.

I've attached my LED circuit schematic (The colour one !) and the P-fet example - for ideas.

I don't want to blow any more LEDS !

any ideas please?

Thanks in advance

Gaz.

 

[RJS1977]

I don't think it's a BEMF issue (though I may be wrong).

 

Your bicolour LED is effectively containing two LEDs connected to different legs.

 

The circuit as shown as well as feeding one LED at each end the 'right' way, is also feeding another LED the 'wrong' way. That isn't something LEDs like, particularly at higher voltages.

 

Try including a diode in series on each leg, facing the same way as the LED.

[Dagworth]

Put the resistors in the outer legs of each LED.

 

Andi

[DCB]

How do you control the voltage?   Does it vary with load?  Because the symptoms you describe are those of trying to run LEDs off a Rheostat, and that won't work.  Try connecting the LEDs to the 16 volt supply then adding the motor, the LEDs shouldn't be affected, I have had red and yellow LEDs track powered in Triang coaches and Transcon diesels for over 20 years with no problems, but reds need different resistors to yellows for equal brightness.

I reckon your 16 volt supply is delivering a lot more than 16 volts off load and the voltage is climbing as the amps drop off as th tired old motor spools up.

 

  I use  diode controller, 11 diodes in series selected by a Rotary switch in its original form it now has various combinations of diodes to give 4-16 volts in unequal stages to give useful speed steps. When I select 16 volts I get 16 volts until the load exceeds 3/4 amp after which it drops away and I arrange my resistors so they can survive full power and no other load.  I actually use them for testing track on full power on occasion.    

 

The circuit looks like you are over thinking this.

 

 

[AndyID]

If the resistors really are orange-orange-brown as shown they are not 10k ohms. That would make them 330 ohms which could well make the LEDs pop (something like 36mA at 14 volts).

[Gazbo]

Thanks guys. The schematic is inaccurate/misleading - sorry. it was meant more as a wiring diagram, and the values were not meant to be relevant, and which would also be corrected once the final solution was settled on. It was simply a drawing, to figure out what "went where". 

Originally I intended to use different resistors to the 10kohm - presently installed- which would have been 330s as per the instructions from a "popular online video channel".

the problem with that resistance was the uncomfortable brightness that ensued, even at around 6V.(Note : that was still minus any loco being present at that stage.) I have actually got 10kohm resistors installed on the centre leg, per LED, and one diode per LED leg. The circuit test ran admirably, for an hour, without issue, and in both directions.

I am using a proper bought Lab style Benchtop psu, where the voltage is limited to what I dial in. The 16v power  rating was only selected, due to the fact that old DC controllers can exceed the stated 12v output, (I believe them being capable of a possible16v output max) although I never intended to run anything at that voltage. Regardless, the damage occurred at only around 7.4 v input, and only once the loco motor was connected in parallel with the LED circuit. We went for 10kohm resistors, after testing various Kohms, and found this resistor rating to bring the brightness down to an acceptably realistic  level, (both red and white) and believing that would also allow the LEDS to run more gently, hopefully giving them a longer life. Yes the Red LED does run slightly dimmer than the white, but not really too far off to look badly out of balance. I'm not sure/don't understand about the rheostat argument. Because each LED circuit is wired as it's own circuit, and linked  in parallel, to a central tapping point, each with their own resistor and pair of diodes, I didn't think they would interfere with each other. If memory serves the circuit current draw was only around 0.12mA.. Others have used this very circuit in their locos without issue. this is what I couldn't understand. I was told it was a known BEMF problem. I don't know personally, I just need to make it work, without destroying itself at the first instance. Data sheet for the red/white LEDs included here. Thanks again guys. PS. don't forget, I'm a newby to electronics !

[AndyID]

OK, I see what's causing it. The reverse voltage across the LEDs makes them "break down" and pass a large current through the other diode in the pair. That current completely bypasses the resistors.

 

One way to prevent that is to put a diode in parallel with each LED pointing in the opposite direction from the LED. There's probably a more efficient way to do it but you could try testing it that way first.

[AndyID]

Just short out the two resistors and put one 4k7 resistor in either one of the feed lines from the motor (not both). That will protect the LEDs from reverse voltage.

[Gazbo]

Aha ! Thankyou AndyID. I get the 4k7 resistor idea, but what did you mean by shorting out the other two resistors? IE - where am I shorting them to ? They are  on common anode, and are already linked through to each other, and the feed out from them is via the existing diodes, so are you suggesting extra diodes in each led's other leg too, so there would be a diode before, and a diode after - in each case ? That would kind of make sense as guarding both anode and cathode. let me modify my  sketch, and upload it for you to check, to see if I have understood you correctly. 

Fingers crossed I have understood.

[AndyID]

Remove those extra diodes and make the 10k resistors zero ohms i.e. a piece of wire. The 4k7 is correct.

 

The original placement of the resistors prevents the two diodes from protecting the Leds from excessive reverse voltage. When you eliminate the resistors the diodes will protect the Leds by limiting their reverse voltage to less than one volt.

 

The 4k7 limits the forward current through the Leds.

Edited May 29, 2021 by AndyID
More info

[melmerby]

Looks over complex.

 

2 Leds back to back with one series resistor will work perfectly well, the forward voltage of one protects the reverse voltage of the other

Depending which way round the supply is, one or the other lights.

 

Is there no way the red and white Leds cannot be separate?

[AndyID]

15 hours ago, Gazbo said:

Aha ! Thankyou AndyID. I get the 4k7 resistor idea, but what did you mean by shorting out the other two resistors? IE - where am I shorting them to ? They are  on common anode, and are already linked through to each other, and the feed out from them is via the existing diodes, so are you suggesting extra diodes in each led's other leg too, so there would be a diode before, and a diode after - in each case ? That would kind of make sense as guarding both anode and cathode. let me modify my  sketch, and upload it for you to check, to see if I have understood you correctly. 

Fingers crossed I have understood.

 

This is probably the simplest configuration for bi-color LEDs

 

 

I left out the second LED to simplify things (obviously it just gets tacked on in parallel with the first one.)

 

The diodes conduct to protect the LEDs from excessive reverse voltage and the resistor limits the LED current in either direction.

 

The problem with the original circuit is the resistors are actually defeating the protection diodes

Edited May 30, 2021 by AndyID
typo

[melmerby]

 

2 hours ago, AndyID said:

 

This is probably the simplest configuration for bi-color LEDs

 

 

No.

This is the simplest, a two wire bi-colour LED:

 

These are readily available, just add a resistor, reverse the LED at the other end

2 Leds + 2 Resistors. Total.

[AndyID]

16 minutes ago, melmerby said:

 

No.

This is the simplest, a two wire bi-colour LED:

 

These are readily available, just add a resistor, reverse the LED at the other end

2 Leds + 2 Resistors. Total.

 

Yes, but he's already bought the other type.

[Gazbo]

Thankyou melmerby, but as AndyIO says - I already bought the other three leg type. (I found loads of 2 leg bi colours, but couldn't see any that were red / white, most were red/green, or red/blue,  which is why I opted for these 3 leggers.)

The 3mm "single" unit fits perfectly into the existing Triang Rivet -lamp holes. using two seperates would mean having them inboard of the hole, and rigging up some kind of director /reflector to send the light out. I did consider that originally, before I found these leds, which solved that particular issue.

 

Andy, thanks for your time and input too, as I thank all of you equally. I didn't know the existing resistors were causing that particular "resistors defeating the diodes"  issue. After scratching my bald patch for a few minutes, The penny dropped ! Yes- you are right.

(Funny how the #tube video example worked for him, but for how long ?)

I will now try the simplified idea with just the 4k7 in one of the feed lines (And the existing diodes- obviously) and see how that goes.

Just before I retry  -Is this how you meant ? (Pic) -

Edited May 30, 2021 by Gazbo

[melmerby]

Red/White bi-colour are available at several places, including this model railway shop:

3mm

https://www.railwayscenics.com/3mm-21v-colour-red-white-led-resistor-required-p-993.html

 

or 2mm:

https://www.railwayscenics.com/2mm-colour-red-white-12v-water-clear-tower-led-with-resistor-p-2430.html

 

Bookmark for next time you need some, saves a lot of hassle.

 

 

 

[AndyID]

4 hours ago, Gazbo said:

Thankyou melmerby, but as AndyIO says - I already bought the other three leg type. (I found loads of 2 leg bi colours, but couldn't see any that were red / white, most were red/green, or red/blue,  which is why I opted for these 3 leggers.)

The 3mm "single" unit fits perfectly into the existing Triang Rivet -lamp holes. using two seperates would mean having them inboard of the hole, and rigging up some kind of director /reflector to send the light out. I did consider that originally, before I found these leds, which solved that particular issue.

 

Andy, thanks for your time and input too, as I thank all of you equally. I didn't know the existing resistors were causing that particular "resistors defeating the diodes"  issue. After scratching my bald patch for a few minutes, The penny dropped ! Yes- you are right.

(Funny how the #tube video example worked for him, but for how long ?)

I will now try the simplified idea with just the 4k7 in one of the feed lines (And the existing diodes- obviously) and see how that goes.

Just before I retry  -Is this how you meant ? (Pic) -

 

Yes, that's it.

 

If you need to adjust the relative intensity between red and white you can always add resistors at the cathodes.

 

The maximum reverse voltage for your LEDs is around 5V but in practice you can usually get away with quite a lot more than that before they actually fail. That's why it worked without the motor in circuit.

 

When you added the motor the inductive commutation voltage spikes on top of the supply voltage were the straws that broke the camel's back 

[Gazbo]

AndyID -   Fascinating, and fully digested !

 

I will be breadboarding first. But that's all great info - and explained at a level us non-sparky numpties can follow !

(I can rebuild a chemical factory, or overhaul a motorcycle engine, mechanically, but LEDS are -well - a newer, more alien and delicate subject for me. I'm Learning on the fly ! 

many thanks. Gaz.

 

melmerby - thanks for the links. Interesting for future use. I'll save these to my notepad on desktop. Will persevere with these for now though, seeing as how they are here ! haha. If I seriously get the bug, I do have some ideas for later...... Cheers. Gaz.

[Gazbo]

We since jury rigged my setup to a digital Oscilloscope, to see what is going on. At 12v feed from a lab bench psu, ( loco running on a rolling road) and the scope connected to the motor pickup wires, we were getting 12v input (with Minimum ripple) but a high reverse pulsing spike. When measuring the peak spikes, this was 478V in the wrong direction!  

I did see a lot of brush sparking, so we have replaced the old bridging Cap, with a new 470pf Cap. That has calmed 90% of motor arcing, so maybe the old cap was duff anyway?

A new revelation.

when testing again, back at home (Minus any  scope) anything past 8.5v input, began to light both red and white elements in a single LED again, although not quite as fierce as before.

BUT, if I apply light finger pressure to the top of the bogie, ie: just less pressure than to slow the motor, I found the Red Led would remain RED only, right up to 12v input !

I am now suspecting the serrated wheels are having a lot to do with it, as in forcing them to maintain contact with the rolling road, seems to eliminate the reverse spiking almost completely.

The tyres and flanges have been cleaned and polished, checked for run-out, and any surface lumps or bumps. A re-try proved an improvement, but not enough to completely eradicate the issue. (NOTE: I'm also now rigged up with a pair of UF4007 schottkys, which are good for 1000v reverse voltage protection, and 75 Nanoseconds recovery. but the issue remains. Perhaps the old serrated wheels are the culprits ? They do spark a fair bit on the rolling road.

(This was supposed to be a simple project ! wow !)

Edited July 15, 2021 by Gazbo

[SamThomas]

2 hours ago, Gazbo said:

 Perhaps the old serrated wheels are the culprits ?

 

Do you get the same effect if you feed the chassis direct without the power going through the wheels ?

 

If not, then the serrated wheels are the cause.

[Gazbo]

Hi Sam. Yes - or rather - No. if I piggyback the power straight onto the motor wires, I get zero emf issues. (Only discovered this aspect an hour ago). and the LEDS are happy with this setup.

Not sure that it is the actual serrations though. it's as if it happens when one area of the wheels pass over the rollers, ie - not throughout the whole wheel- revolution. (Imagine you stuck a minute strip of insulation tape across the wheel, so one area quickly disconnects  as the wheel rotates). don't know what the issue is. I had thought to buy some smooth re-profiled sets from Chambs123. but at 26 quid a set, requiring 2 sets for this switcher, my 55, and my dock shunter the they will be more than the loco's are worth. My Mini Lathe is in bits for overhaul at the minute, so a re-machining of my existing rims is out of the question just now. So annoying, but at least I'm getting to the bottom of the problem now.

[AndyID]

Now you know why Tri-ang said a suppressor track was required for every six* pieces of track

 

 

 

*Maybe it was five?

[SamThomas]

4 hours ago, AndyID said:

Now you know why Tri-ang said a suppressor track was required for every six* pieces of track

 

 

 

*Maybe it was five?

That was in the days of VHF TV reception which was very suseptable to interference from model railways.

 

In the days of "Triang" I don't think there were even red LED's.

[AndyID]

14 minutes ago, SamThomas said:

That was in the days of VHF TV reception which was very suseptable to interference from model railways.

 

In the days of "Triang" I don't think there were even red LED's.

 

Tri-ang even predated the first silicon transistors and affordable LEDs only became available in the late 60's.

[Gazbo]

Hi again Guys.

After much head scratching - a light came on ! (pardon the pun).

I have now removed the plastic wheels from the rear (Dead) bogie, and replaced them with Peco metal (Axles fully insulated) items and brass pin bearings.

They are ever so slightly smaller than original, but it's not noticeable.

I then fashioned twin pickup whiskers, from 0.6mm brass sheet, and soldered a wire to each, which then fed up through the central rivet.

Being the caveman I am, I have glued a 3mm plastic plate to the long flat chassis top, and stuck Mini Chocolate blocks to it, for ease of connect / disconnect.

feed in from pickups, with inline 1k resistor - to 1st twin chock block. out of there to the twin, +ve linked  diodes, and forward to 3 seperate chock blocks, mounted vertically (Held steady by the diode legs). the leds are a good push fit into the original lamp hollow rivets in the body, so it was just a case of bending the legs to follow the inside contour of the walls. another 1k soldered to each red negative leg, has balanced the brightness nicely.                                                             (If you recall, attempting to light a red along with a white, one of them always saps all the juice and the other one dies out ! the inline 1k on the red leg has balanced things, forcing them to share, and has cured that issue completely !)

All that remained was to run a 3 core (RC aircraft Servo wire), with centre core to +ve chock-block from linked +ve of diodes, and one each negative, to the other two neg chock blocks. If it worked backwards, simply switch these two around. WELL !!!! It seems if I pick power up directly from the track this way, as opposed to piggybacking off the motor pickups at the other end - as was originally suggested with the circuit , the motor interference issues are now completely negated. She works perfectly, and bidirectionally now, totally separated from all the motor shenanigans !

Who would have thought it could cause so much grief !!!

Cheered My old fella up no end though. He says they haven't lit, since 1966 ! But that's another story....

So - just wanted to say thanks for all your help guys.

Got there in the end.

Attaching a rather garbage schematic/ wiring / hybrid sketch  here, so you might follow the thinking.

Coz I ain't taking that top off again !