LED lamp input voltage on electric train is 18-137VDC

[Flyback]

Hello,
We have a spec for 20W LED lamps to be powered from a train power source which is 18V to 137.5VDC.

In fact, its wider than that, there is an emergency requirement of 4W LED power for times when this supply goes down to 8V. (-but that might be when the emergency battery pack switchs in?)

Do you know what it is on the trains which is providing this DC power source? Is it a big battery bank?

I believe they are electric trains. For example, it’s the trains that will be used in the HS2 project, soon to come in UK.

Do you know what kind of source impedance this 18-137.5VDC power source will present to our 20W LED lamps? We don’t have room for much input capacitance in the LED lamps, so we are hoping there’s not much wiring inductance going back to the power source. Do you know what the wiring inductance is likely to be?

Unfortunately, the client is not coming forward with much information on this for us.

 

[MaxHeadRoom78]

Many locomotives now are Diesel Electric and generate their own power.
Not sure what the actual system is in the UK now, but N.A. locomotives have a large capacity battery unit in place, it used to be 96vdc and up.
H.S. rail may have different systems now.
All the work I have done on Locomotives, the company supplies all schematics of the Loco.
Max.

 

[dknguyen]

Unfortunately, the client is not coming forward with much information on this for us.

You say they're not being forward with it. Did you try...requesting that information or anything that might give you a clue? If the customer doesn't know what you need they won't come forward with it. They're the customer because you know something they don't. They don't know what you need. Half the time they don't know what they need.

Doesn't really sound like obvious information they would have just have laying around in a file somewhere though.

 

[unclejed613]

sounds like a buck converter with a very wide input range is what's needed. output voltage and current for a 20W LED would be 3.6V@5.6A. the problem will be the range of the input voltage. i've seen buck converters that operate over a range of 5-40V. maybe cascading a couple of buck converters, the first one operates over the 40-140V range, but goes into pass mode below 40V, and the second one operates from 40V down. it will definitely have to be switching converters, because using analog regulators would dissipate a lot of heat (over 700W) at the 137V level

 

[ronsimpson]

I worked on train power before your were born. The power can be nasty. We saw it go negative. First thing should be a diode to remove the negative. Then an inductor to help with the spikes.

Do you know the LED is 3.5V @ 5A or 7V @ 2.5A or ? Choose wisely.
You can also use a boost/buck and pic a high voltage LED with current under 1A.

 

[eTech]

Hello,
We have a spec for 20W LED lamps to be powered from a train power source which is 18V to 137.5VDC.

In fact, its wider than that, there is an emergency requirement of 4W LED power for times when this supply goes down to 8V. (-but that might be when the emergency battery pack switchs in?)

Do you know what it is on the trains which is providing this DC power source? Is it a big battery bank?

I believe they are electric trains. For example, it’s the trains that will be used in the HS2 project, soon to come in UK.

Do you know what kind of source impedance this 18-137.5VDC power source will present to our 20W LED lamps? We don’t have room for much input capacitance in the LED lamps, so we are hoping there’s not much wiring inductance going back to the power source. Do you know what the wiring inductance is likely to be?

Unfortunately, the client is not coming forward with much information on this for us.

The electric trains i’ve seen back in the day (1980’s) were 900 Vdc provided via overhead power or third rail. Transit type trains.

eT

 

[Nigel Goodwin]

The electric trains i’ve seen back in the day (1980’s) were 900 Vdc provided via overhead power or third rail. Transit type trains.

I suspect they probably don't feed that directly to the lighting in the train?

 

[eTech]

I suspect they probably don't feed that directly to the lighting in the train?

Heh..No...but power isn’t constantly supplied to the train. The lights flicker or turn off when coasting thru “dead” sections of track.

 

[Nigel Goodwin]

Heh..No...but power isn’t constantly supplied to the train. The lights flicker or turn off when coasting thru “dead” sections of track.

What kind of crap trains do you have?

Obviously the OP needs full information from the client, if he can't provide it them drop him as a client, or make sure his contract includes paying you for HIS mistakes.

 

[kubeek]

AFAIK that range is most likely a range of different voltage systems on different trains, namely 24V, 48V and 110V, with -25/+20% margin. A typical wide range application, you don´t want to manufacture and keep stock of different units for diferent voltage system unless it is really necessary.

And as far as I remember, on-board rail equipment had pretty stringent requirements for power dips and interruptions, (100ms to 14.4V in 24V system, and 10ms to 0V IIRC), so good luck with your lack of space for capacitors.

 

[ronsimpson]

if he can't provide it them drop him as a client

You can not get specifications out of a rock. (Moses used a stick)
If he does not know the specifications then he can't give you much.
If he has a big check book then you make him happy. That is the specification. "make him happy"

 

[MaxHeadRoom78]

I would have thought British Rail would have an engineering dept that could supply maintenance data?
The Canadian rail companies do.
These old retro's we upgraded to dynamic braking for use in the Rockies, they had 8 battery's at 8vdc.
Max.

 

[ChrisP58]

Is your client the end customer? Or a middleman?

I've found that sometimes when a client is slow answering questions, it's because he's got to go through extra channels of communication to get the information.







Edit to fix typo.

 

[eTech]

What kind of crap trains do you have?

Obviously the OP needs full information from the client, if he can't provide it them drop him as a client, or make sure his contract includes paying you for HIS mistakes.

heh...actually that operation was common for transit type trains at the time. Don’t know if still that way now. The third rail (power) did not run the entire distance of track. There were dead sections where the train “coasted” (usually at a good speed) without power until it’s shoe made contact with the next section of track with third rail. The passenger lights would intermittently go dark...a little spooky if it was in the evening.

But I agree..the client should be more forthcoming...

eT

 

[schmitt trigger]

Beautiful vintage locomotives Max!
Very nice paint job.

 

[Diver300]

AFAIK that range is most likely a range of different voltage systems on different trains, namely 24V, 48V and 110V, with -25/+20% margin. A typical wide range application, you don´t want to manufacture and keep stock of different units for different voltage system unless it is really necessary.

I would say that is more likely than really terrible voltage regulation.

If you need to run lots of LEDs to get 20 W, you could put them all in series and go to 150 V or more with a boost regulator. Most boost regulators can't produce less voltage than the supply, so you will normally need to have more LED voltage than the maximum supply voltage.

The input current of a boost regulator doesn't have the fast current changes that a buck regulator has, so there is less need for an input capacitor. An output capacitor will reduce current variation in the LEDs, but if you are switching at 100 kHz or more, you won't need a big capacitor and no-one will be able to see flicker at those frequencies.

 

[ronsimpson]

Top half, buck, 3.5V 5A LED. FET turns on at 4.5A and off at 5.5A.
At high line, FET is on for a very short time. FET current is 5.5A pk. L1 is high inductance so LED current is continuous. C1 can be 0uF.
At low line, FET is on for most of the time. FET current is 5.5A pk.

Bottom half, boost, LED=150V 120mA,
At high line the FET is on for a short time. Peak FET current is 150mA.
At low line the FET is on most of the time. Peak FET current is 1A. D2 current is 1A pk to get 150mA average.
C2 must be there because D2 current drops to zero.

 

[ronsimpson]

These two are the same thing. I have used both.

Most boost regulators can't produce less voltage than the supply, so you will normally need to have more LED voltage than the maximum supply voltage.

This an interesting alternative. On the right side the LED voltage and the input voltage have less restrictions. An 50V LED will work.

 

[Diver300]

This an interesting alternative. On the right side the LED voltage and the input voltage have less restrictions. An 50V LED will work.
View attachment 112485

The one on the right is a buck/boost regulator, which will have discontinuous current in the supply, although there will be smoothing with C8. A boost regulator, as shown on the left, has the inductor in series with the supply so the supply current changes less. The downside is that the LED voltage has to be larger than the maximum supply voltage.

 

[ronsimpson]

We have a spec for 20W LED lamps to be powered from a train power source which is 18V to 137.5VDC.

I have been down this road before. After all the extra work to make 18 to 137.5 work, which is hard, then at the very end you learn that it is really 180 to 137 volts.
Why 137.5 why not 137.49999? or 137.5001? This number is so strange it should be questioned.