LED lighting in caravans.

[ericmark]

DC to DC chips with 10.8 to 13.2 volt input and 12 volt output are reasonable cheap around £4 and there are many options but 9V to 36 volt and price jumps to £18 each.

Caravans if complying with regulations have a voltage 11 to 14 volt. (A721.55.4.1) and often they use stage chargers which means 14.8 volts.

Texas Instrument produce a chip which states:- “This design uses a 12Vac input which is rectified and fed to a boost DC to DC Converter with constant current feedback loop for driving 6 series LEDs at 0.32A. Also implemented is a simple circuit to decode a dimming signal from a TRIAC. This dimming decoder circuit controls the dimming level of the LEDs.” Number PMP7754 the details state Vin (Min) (V) 12, Vin (Max) (V) 12 input type AC. (I expect a chip like this is used in LED clusters) In other words it is designed for use with a regulated 12 vac power pack, normally used with 12 volt quartz bulbs. The output is 21 volt and divide that by 6 and 3.5 volt for each LED.

Car LED's often have a resistor built in so they have same load as tungsten bulb aim is long life not reduced power.

At the moment it would seem cheapest way to run LED lighting is to use an inverter and 230 volt rather than 12 volt fittings.

My thoughts are to simply fit a LM7812 chip in each lamp and accept when battery is not on charge they run a bit dim.

However there must be people who have come across this problem before, so question is how have others got around the problem?

 

[westie101]

Info here

http://www.caravantalk.co.uk/community/topic/44237-l-e-d-s-m-d-lights/

My understanding is that you do not need to be as precise as you are suggesting with voltage

 

[davelx]

These have built-in drivers/regulators and protection and will work from 10-30V, designed for use in boats/caravans/mobile homes. I've used a load of these in a narrow boat.


http://bedazzled.uk.com/12v_LED_Bulb_Replacement.htm

 

[ericmark]

That link is great 10 - 30 volt can't go wrong.

 

[flameport]

My thoughts are to simply fit a LM7812 chip in each lamp and accept when battery is not on charge they run a bit dim.

That won't work - with the input less than 14V, there will be no output at all.

 

[JohnW2]

My thoughts are to simply fit a LM7812 chip in each lamp and accept when battery is not on charge they run a bit dim.

That won't work - with the input less than 14V, there will be no output at all.

Indeed. I suppose that an appropriate resistor and a 12V zener would do what eric was intending - but that would be a pretty inefficient system.

Kind Regards, John

 

[ericmark]

I have found over the years not all 7812 chips are the same. It was a common trick to put a red LED between Com on chip and negative line so it would give 13.2 volts. However it did not work with all chips.

I know the bulbs linked to are expensive but they seem to be the answer. Can't see point in adding resistors, zenors or chips when one can simply buy a bulb with it all built in.

 

[JohnW2]

I have found over the years not all 7812 chips are the same. It was a common trick to put a red LED between Com on chip and negative line so it would give 13.2 volts. However it did not work with all chips.

I think that works with virtually all 78xx regulators. Indeed, it's conceptually the way in which one uses them to give variable voltage regulation. Let's face it, the chip merely regulates the voltage between it's output and COM. If you raise COM to above 'zero', that will raise the output voltage relative to your zero reference.

However, flameport's point was that all 78xx regulators have a 'dropout' voltage of about 2V. In other words, if the input voltage is not at least 2V greater than the regulated output voltage (i.e. about 14V in the case of a 7812), the the output voltage/current falls to zero - so quite useless for your application. Raising the potential of COM to above that of the negative supply line (with an LED or whatever) would merely make that even worse (i.e. you would need even more than 14V input to get any output).

I know the bulbs linked to are expensive but they seem to be the answer. Can't see point in adding resistors, zenors or chips when one can simply buy a bulb with it all built in.

Sure, if it's financially acceptable, then one doesn't want to fiddle about adding components. However, it was you who said that your thought was to add a 7812 to each lamp!

Kind Regards, John

 

[DetlefSchmitz]

However, flameport's point was that all 78xx regulators have a 'dropout' voltage of about 2V. In other words, if the input voltage is not at least 2V greater than the regulated output voltage (i.e. about 14V in the case of a 7812), the the output voltage/current falls to zero - so quite useless for your application.

This doesn't match my recollection of my experience with this device. I believe that there will just be a slow reduction in output voltage with a fairly fixed voltage drop (rather like Eric was suggesting).

I think I would have to see evidence to the contrary before I stopped believing that.

 

[ericmark]

Untitled

• ericmark
• 23 May 2015
• 1

The Input-output differential from the Fairchild data sheet does seem to show below 2 volt above regulated voltage the output voltage will be about 2 volt below input voltage which for 11.5 volt would mean around the 9.5 volt mark it would stop damage to the LED.

To regulate to 12 volt only real way is to first boost voltage to above 12 volt I found around 16 volt no load is required to get a regulated 12 volt.

Any inverter uses power so don't really want to run any inverter. I only have two lamps which use tungsten bulbs these

at £7 each will replace them nicely so for £14 job done.

Not worth playing for £14 just replace the bulbs. Just need to check actually BA15d bulbs are required.

 

[JohnW2]

However, flameport's point was that all 78xx regulators have a 'dropout' voltage of about 2V. In other words, if the input voltage is not at least 2V greater than the regulated output voltage (i.e. about 14V in the case of a 7812), the the output voltage/current falls to zero - so quite useless for your application.

This doesn't match my recollection of my experience with this device. I believe that there will just be a slow reduction in output voltage with a fairly fixed voltage drop (rather like Eric was suggesting). ... I think I would have to see evidence to the contrary before I stopped believing that.

Maybe my (and flameport's) recollections are wrong. I have to say that I've probably never had any reason to use a 78xx or 79xx in situations other than those in which the input voltage has not always been appreciably above the output voltage, so I can't pretend to have any 'experience' of eric's situation - although I certainly thought that the behaviour of these devices was as I described. When I have some spare minutes, I'll "do the experiment"!

Kind Regards, John

 

[JohnW2]

Untitled

• ericmark
• 23 May 2015
• 1

The Input-output differential from the Fairchild data sheet does seem to show below 2 volt above regulated voltage the output voltage will be about 2 volt below input voltage

I have to say that I'm somewhat struggling to understand how you draw that conclusion from the graph above (although what you say may still be true!) - but, as I just wrote to Detlef, I'll try to do the experiment later!

Kind Regards, John

 

[JohnW2]

... The Input-output differential from the Fairchild data sheet does seem to show below 2 volt above regulated voltage the output voltage will be about 2 volt below input voltage

I have to say that I'm somewhat struggling to understand how you draw that conclusion from the graph above (although what you say may still be true!) - but, as I just wrote to Detlef, I'll try to do the experiment later!

I don't actually need to do the experiment, since the Texas datasheet has already done it for me (for a 7805) ...

Untitled

• JohnW2
• 23 May 2015
• 1

... so, it seems that eric and Detlef are essentially right, and myself and flameport wrong! Sorry about that - but it goes to show that one learns something every day! All I can say in my 'defence' is the fact that, as I recently wrote, in 'normal usage' one would never usually contemplate a situation in which input voltage was not appreciably greater than the regulated output voltage - so it's probably not a situation I've ever actually seen!

Kind Regards, John

 

[DetlefSchmitz]

Ah - thanks! At least my memory isn't playing me up.

 

[JohnW2]

Ah - thanks! At least my memory isn't playing me up.

You're welcome, and again my apologies for perpetuating confusion!

Surprisingly, it took me quite a while to find a relevant graph, or even data. It seems as if most datasheets for these ubiquitous devices (like the Fairchild one eric quoted from) do not contain any information at all about what happens at input voltages below the regulated output voltage (perhaps because that it outside of the scope of 'intended use') - so I suspect your memory may possibly relate to actual 'experience', rather than what you'd seen in datasheets.

Kind Regards, John