Welcome to the forum   
Welcome Guest! To enable all features please Login or Register.

Notification

Icon
Error

Share
Options
View
Go to last post in this topic Go to first unread post in this topic
Offline DaleSchultz  
#1 Posted : 16 October 2006 02:43:59(UTC)
DaleSchultz

United States   
Joined: 10/02/2006(UTC)
Posts: 3,997
I have had so little time to do train stuff that I have to relearn everything each time I get back to it... so I apologize if I repeat anything here... and any cross-post...

My original plan was to rectify 16V AC for powering accessories and building lights at each point of usage. i.e. run a bus of 16VAC and tap that as needed. My initial LED light circuits were based on that principle...
<http://layout.mixmox.com/1/LED_module>

Then when I did the lighting for my signal box,
<http://layout.mixmox.com/1/signal_box>
I found I burnt out some LEDs because I had done my calculations on the RMS voltage and not the peak voltage so I had to rethink my circuits. In doing that I again considered running a properly regulated clean DC bus wire under the layout for my lighting.

Then I was given a box of old (1960's) power supplies, transformers and magnetic voltage regulators, including a huge 15 Amp 24V power supply (360W) and a very fancy 15V 22W regulated power supply. Suggestions regarding the fancy power supply was to use it 'in the lab' for testing etc.

Now I want to run this logic past the electronic experts here to ensure that my thinking makes sense and not too many sparks...

The variable voltage power supply will only produce 1.5 Amps and each LED will use 20 to 30 mA so that would only drive about 60 LED lights, so I conclude that 1.5A will not be adequate.

After spending some hours looking at the circuit and huge capacitors on the monster power supply and figuring out what the magnetic voltage regulators are all about I retuned to my old idea of using an old PC power supply...

The label says that it produces the following outputs:

3.3V 24A
5V 35A
12V 12A

I believe that these voltages are smoothed and regulated suggesting an ideal layout DC power source. The unit is also in a nice compact case, is fused and has its own cooling fan. If it ever dies, I can find another one and plug it in.

Does anyone see a problem using a PC power supply?

Assuming that is a GO, my next decision is: "which voltage should I use... 3.3V, 5V or 12V..."

White LEDs have a forward voltage of about 4V - the ones I have, have a Max forward Voltage of 3.3V suggesting that the 3.3V supply would be ideal. If however later LEDs need more, I won't have any headroom... I am not sure on what the minimum forward voltage is for these LEDs - possibly 3V.
Also, I suspect that fluctuation and voltage drops may result in less than 3.3V resulting in some LEDs not lighting up.
24Amps would drive about 1000 LEDs. Plenty.

If I use the 5V supply I would not be able to drive two LEDs at a time but a series resistor of 56 Ohms to 85 Ohms would run a single LED at 30 to 20mA and each unit would use about 0.1W allowing about 1750 such lights, also plenty, but the disadvantage is I cannot place two LEDs in series as the forward voltage is not available.

If I use the 12V supply, a single LED would need a 290 to 435 Ohms resistor and at 30mA that starts exceeding a 1/4W resistor rating and would call for 1/2W rated resistors - probably not a big deal. One LED plus resistor would use about .36W and I could drive a total of about 400 such lights. Significantly fewer lights as more energy is going into resistors... and I have no idea how many lights I will need.
With 12V I could power up to three lights at a time which could be advantageous.

So can anyone suggest any further advantages or disadvantages for the 3 available voltages?

I am currently driving my k83 type devices using a separate 16VAC supply (not digital power) - I suspect I should not steal any of the 12V DC for such purposes though.


Should I consider any measures to prevent accidental 16V Ac or any other spikes into a low voltage line that is driving items that do no tolerate over-voltages? If so how? I guess I could place a diode on the DC line to prevent half of any AC from going the wrong way... is there a smart way to limit the voltage very fast so that a single accident does not destroy all installed LEDs in one flash of HO scale lighting?

Dale (sometimes in the dark)





Dale
Intellibox + own software, K-Track
My current layout: https://cabin-layout.mixmox.com
Arrival and Departure signs: https://remotesign.mixmox.com
Offline Tony  
#2 Posted : 16 October 2006 11:46:48(UTC)
Tony

South Africa   
Joined: 18/09/2004(UTC)
Posts: 726
Location: Cape Town
Hi Dale - I also use a PC power supply for all my lighting and some electronic bits and pieces. Really a good solution as it has built in short circuit protection etc.

Interesting you using white LEDs for model lighting - are they not a bit pricey??

I use the 5VDC to supply all my little glass "rice" lights which i use in all the models ( A previous thread discussed this: https://www.marklin-user...ult.aspx?g=posts&t=2777)

They provide an excellent warm shade in the models and as they are fine with 5V no resitors or anything are required. Can't recall what the current was at 5v per globe but was a lot more than an LED obviously.But these rice lights are soooo cheap; 100 would be about Euro 8. I think Standard White LEDs(3mm) cost about Euro3-4 each!!!! Not sure about in other countries

You mention that with 5V you use two white LEDs in series but have problem with too much Vf . Why don't you just put them in parallel and reduce the resistor accordingly?
Regards Tony
Offline rugauger  
#3 Posted : 16 October 2006 12:31:15(UTC)
rugauger

United Kingdom   
Joined: 19/12/2003(UTC)
Posts: 1,205
Location: Swindon, Wiltshire
Very interesting topic - thanks Dale.
Tony has a very good point regarding OTHER uses for the power supply - perhaps you may want to add some street lights or other non-LED consumers at a later stage? In which case why not go for a higher voltage and just put resistors in where needed?

Also, bear in mind that not all LEDs have the same Vf; I've used some white LEDs with a Vf of 2.2V, in which case you could run 2 LEDs of the 5V supply.
Richard
Offline jerdenberg  
#4 Posted : 16 October 2006 12:49:01(UTC)
jerdenberg

Netherlands   
Joined: 10/01/2005(UTC)
Posts: 1,011
Location: Den Helder, Noord-Holland
Hi All,

White LEDs have become very cheap by now: last May I bought 100 warm white 5mm LEds for €23 on Ebay (German seller, Shanghai despatch), but they can be had cheaper.

Dale, in the calculation for the 24V/1.5A PSU resulting in 60 LEDs you are apparently dividing the amperage by the individual LED amperage, but with the 24V you mention you can feed at least 5 or 6 in series, which raises the total to 300 or more. Another aspect you might check: when I was experimenting with my warm white LEDs they were way too bright at their standard current, and I found 10 mA more than sufficient. The warm-whites had a voltage drop of about 3V, so with a series resistor that takes off 3V you should be able to accommodate about 100 chains of 6 LEDS.

Jeroen
Figomima division, UP; mostly figment of my imagination yet.
Offline DaleSchultz  
#5 Posted : 16 October 2006 17:34:01(UTC)
DaleSchultz

United States   
Joined: 10/02/2006(UTC)
Posts: 3,997
Quote:
[size=1" face="Verdana" id="quote]quote:Originally posted by Tony

Interesting you using white LEDs for model lighting - are they not a bit pricey??

I use the 5VDC to supply all my little glass "rice" lights


They are getting cheaper all the time, earlier this year I bought some at 50c each (US) including shipping. I like the fact that they give off almost no heat and their life is very long - allowing me to embed them in buildings without having to ensure that I can get back to them to replace them...
Dale
Intellibox + own software, K-Track
My current layout: https://cabin-layout.mixmox.com
Arrival and Departure signs: https://remotesign.mixmox.com
Offline DaleSchultz  
#6 Posted : 16 October 2006 17:58:16(UTC)
DaleSchultz

United States   
Joined: 10/02/2006(UTC)
Posts: 3,997
Quote:
[size=1" face="Verdana" id="quote]quote:Originally posted by jerdenberg
<br />... with the 24V you mention you can feed at least 5 or 6 in series, which raises the total to 300 or more.


Yes, I just did the same calculation (one LED per usage) for each possible voltage just for comparison purposes. Some buildings may need 1 light others may need 10... and where I need more than one I will place them in series...
Dale
Intellibox + own software, K-Track
My current layout: https://cabin-layout.mixmox.com
Arrival and Departure signs: https://remotesign.mixmox.com
Offline hmsfix  
#7 Posted : 16 October 2006 22:37:14(UTC)
hmsfix


Joined: 06/02/2005(UTC)
Posts: 1,383
Location: Darmstadt,
Gi Dale,

Here my ideas to your lighting problem:

Using the 5 V output of a PC power supply is certainly a good idea. There are two points I would consider:

- You should protect your light circuit against an occasional short circuit. The power supply can provide 20 A or more. This might fry your electric wires in case of a short circuit. I would subdivide the light circuit into 4 or 5 separate branches, each protected with a 3 - 4 A fuse. No risk any more to damage something.

- M* transformers have a reliable insulation between the primary mains and the secondary 16 V circuit. I don't know how this is for PC power supplies. Probably there are good and bad ones. Perhaps it is wise to have an eye on the safety labels. Have a strange feeling in my stomach since I once had a glance on the wire jungle inside a PC power supply casing.[:0]

Hans Martin

-

Offline perz  
#8 Posted : 16 October 2006 23:09:20(UTC)
perz

Sweden   
Joined: 12/01/2002(UTC)
Posts: 2,578
Location: Sweden
Quote:
[size=1" face="Verdana" id="quote]quote:Does anyone see a problem using a PC power supply?

Yes, there is one problem. Whether it strikes you or not will depend.

PC power supplies are normally grounded through the wall outlet ground connection, if you use grounded outlets. The ground is connected to the "common" on the PC supply. If you use other equipment (e.g. computer control) that is grounded the same way, you may get "funny" effects. It happened to me. Everything worked well except that all my turnout lamps were always on even though they shouldn't. I found out that the power supply for the turnouts was connected through the wall outlet ground through my PC and through some PC-controlled accessories back out to the layout. If it had been with the wrong polarity, I would have got smoke instead of light... biggrin.

Quote:
[size=1" face="Verdana" id="quote]quote:Assuming that is a GO, my next decision is: "which voltage should I use... 3.3V, 5V or 12V..."

In this whole discussion, you should remember that it is the current that makes the diode light up, and that it is over-current that makes it break. Forward voltage on diodes is approximate, and may vary from diode to diode. If the voltage is below the forward voltage, the diode will not light up. If the voltage is above the forward voltage, the current may be too high. The relation between voltage and current is exponential, with a rather sharp "knee" at the forward voltage.

It is therefore in general not a good idea to drive LEDs without any form of current limitation. The simplest and most commonly used current limitation is a resistor. If, however, your supply voltage is close to the sum of the forward voltages (assuming LEDs connected in series) the variation in forward voltage will make the resistor solution unstable. The options are either to reduce the number of LEDs in series, or to use a more advanced electronic current regulation.

Because of the spread in forward voltage between different LEDs (even LEDs of the same type), it is not so good to have two LEDs in parallel with common resistor or common current regulation. If one of the LEDs has a slightly lower forward voltage it will get most of the current, and may burn, while the other one may not light up properly.

LEDs connected in series will all get the same current, and will therefore, if they are all of the same type, also give the same amount of light.

LEDs generally do not tolerate a lot of backward voltage. If reversed, they act as zener diodes (but with unspecified zener voltage). If you want to drive a LED with AC, you should always add a standard diode in series. If you look at e.g. Viessmann signals, they have such a diode.
Offline DaleSchultz  
#9 Posted : 16 October 2006 23:45:14(UTC)
DaleSchultz

United States   
Joined: 10/02/2006(UTC)
Posts: 3,997
Quote:
[size=1" face="Verdana" id="quote]quote:Originally posted by hmsfix
<br />- You should protect your light circuit against an occasional short circuit. The power supply can provide 20 A or more. This might fry your electric wires in case of a short circuit. I would subdivide the light circuit into 4 or 5 separate branches, each protected with a 3 - 4 A fuse. No risk any more to damage something.


Yes thanks..
I have drawn a diagram how I think this should be done. It can be seen at

&lt;http://layout.mixmox.com/1/zener_voltage_protection&gt;

Did i get it right ? Any suggestions?
Dale
Intellibox + own software, K-Track
My current layout: https://cabin-layout.mixmox.com
Arrival and Departure signs: https://remotesign.mixmox.com
Offline DaleSchultz  
#10 Posted : 17 October 2006 00:36:12(UTC)
DaleSchultz

United States   
Joined: 10/02/2006(UTC)
Posts: 3,997
Thanks Perz...

I had not considered that the ground of the supply might be different to the common ground I have on the layout and in the supply of the computer. I suspect that this is something I should be able to test though and it is it common then I should be OK right?

I do have grounded outlets and both outlets are on the same circuit in the wall. I should check that they are all wired the same way. I just checked my Eckert power supply and the main ground is tied to the chassis as is the common ground of the layout... so that looks good. Since the PC power supply puts out +12V and -12V I may have to use the right one and ensure its ground is the same... nice point thanks

As for wiring LEDs etc. yes, I won't connect them in parallel without their own resistor etc.
For over voltage protection I am proposing this:
http://layout.mixmox.com...zener_voltage_protection

The spread of forward voltages suggests that the 3.3V supply should not be used. Would you recommend the 12V or 5V ???



Dale
Intellibox + own software, K-Track
My current layout: https://cabin-layout.mixmox.com
Arrival and Departure signs: https://remotesign.mixmox.com
Offline perz  
#11 Posted : 17 October 2006 01:15:08(UTC)
perz

Sweden   
Joined: 12/01/2002(UTC)
Posts: 2,578
Location: Sweden
Quote:
[size=1" face="Verdana" id="quote]quote:is it common then I should be OK right?

Yes.

Quote:
[size=1" face="Verdana" id="quote]quote:Would you recommend the 12V or 5V ???

Doesn't matter really.

An advantage with using 12 V is that you can use the same power feed for other MRR accessories, which often also are 12 V DC.

An advantage with 5 V is that you can incorporate electronic circuits which usually use 5 V.

Users browsing this topic
Guest
Similar Topics
DC power supply (Digital)
by Bernd 08/12/2023 16:58:22(UTC)
Use AC HO Marklin turnouts with DC power supply? (H0-scale)
by Yigal 14/05/2023 07:41:52(UTC)
12 V DC power supply to Viessmann 5211 (Digital)
by DaleSchultz 10/11/2013 17:21:36(UTC)
Forum Jump  
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.

| Powered by YAF.NET | YAF.NET © 2003-2024, Yet Another Forum.NET
This page was generated in 0.757 seconds.