| Author |
Message |
    
echase
Frequent Contributor
Username: echase
Post Number: 286
Registered: 07-2007
Rating: N/A
Votes: 0 (Vote!) | | Posted on Wednesday, 03 February, 2010 - 12:18 pm: | 
|
I have bought a Microchip1630DM-LED2 boost mode LED driver demo board. This is a PIC driving a boost mode regulatar and 5 350mA LEDs in series. Bit like the EPE/Jacar/ Silicon Chip high power LED unit.
I would like to use this for garden lighting where the 5 LEDS are spread out in a string along my garden path. Problem is what is the effect of adding the resistance, capacitance and inductance of 10-15metres of dual core wire into the output? As, like many LED drive circuits, it’s a high frequency pulsed output with current sensing resistor will this upset it?
Also would like to double up the LEDs to increase current to 700mA, being the max the max the driver delivers. Can these LEDs just be paralleled in 5 pairs in my 5 lights or will that cause current imbalance in each pair? Would a small resistor in series with each solve that? |
phonoplug
Regular Contributor
Username: phonoplug
Post Number: 32
Registered: 08-2009
Rating: N/A
Votes: 0 (Vote!) | | Posted on Saturday, 06 February, 2010 - 11:59 pm: |
|
Hi,
Best thing to do would be to try it. Measure its efficiency with a short set of wires on the bench, then do the same with the long wires you intend to use. A significant drop in efficiency would suggest its not happy. If necessary play around with the compensation pin to try and fix it.
As for paralleling LEDs, it is sometimes done but not very good practise. If you must do this, make sure the LEDs are from the same voltage bin, and as you say, adding a small resistance in series with each LED can help balance the currents a little. Something like 1R or 2.2R. Use one with each LED. |
echase
Frequent Contributor
Username: echase
Post Number: 288
Registered: 07-2007
Rating: N/A
Votes: 0 (Vote!) | | Posted on Sunday, 07 February, 2010 - 10:12 pm: |
|
I will try. Would adding a large smoothing capacitor across the load help? |
phonoplug
Regular Contributor
Username: phonoplug
Post Number: 33
Registered: 08-2009
Rating: N/A
Votes: 0 (Vote!) | | Posted on Sunday, 07 February, 2010 - 11:07 pm: |
|
It may. Again try it and see. You could also try adding a capacitor across each LED too if necessary. |
twintub
Frequent Contributor
Username: twintub
Post Number: 64
Registered: 02-2007
Rating: N/A
Votes: 0 (Vote!) | | Posted on Sunday, 07 February, 2010 - 11:29 pm: |
|
If you intend running this from the mains, why nor dispense with the fancy LED Driver PSU and just use an ordinary current limited PSU?
Using a special LED Driver PSU is obviously desirable when running from batteries in order to not waste too much energy, but mains power would be relatively cheap for your application.
An observation: A lengthy chain of LED's strung out in the garden may be an unintended RF transmitter unless the driver psu is adequately smoothed at its output. LED drivers ususally feed a pulsed current to the LED, so to smooth it would mean adding a diode (low loss schottky) in series with the PSU output and adding a smoothing capacitor at its cathode.
Obseration 2: If you were to use your proposed LED driver unsmoothed, and wire each LED with a 'balancing' resistor, the 'far-end' LED's could be less bright due to the extra lead inductance. The 'near-end' LED's would not have these extra losses. This is a good case for wiring LED's in series! |
echase
Frequent Contributor
Username: echase
Post Number: 290
Registered: 07-2007
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, 09 February, 2010 - 11:44 am: |
|
Linear supply is a thought. Say 5 LEDs in series, plus the long cable losses, comes to a supply of 20V. If I can make a constant current linear supply with less than 4V drop then the efficiency is 80%. That is not much worse than a typical LED switched mode driver especially as it may be less than optimal driving the long wires. I have to take off the mains transformer losses but that is needed whether is a linear or a switched supply.
Anyone know a good efficient constant current supply capable of 700mA? E.g. using a low drop constant voltage regulator and current sense resistor? The only dedicated single chip solution I know is a LM334 and similar but these only supply under 10ma. They do work though down to a ‘dropout’ of around 1V. 100 in parallel would work but these must be a more elegant solution!
There are very few single chip variable current controlled voltage regulators. LM200 is one that can work as a constant current source up to 2amp but it’s ‘dropout’ is a few volts I think. Anyone know of others? |
echase
Frequent Contributor
Username: echase
Post Number: 291
Registered: 07-2007
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, 09 February, 2010 - 02:42 pm: |
|
To answer my own question maybe I could use the lowest voltage low drop 1 amp regulator I could find, such as a 1.25V to something variable voltage one and set it to minimum voltage. Then add a 1.8 ohm load resistor. The current in that resistor is ‘stable’ at about 700mA and I can trim the resistor to make it exact. So can feed the LEDs from this. But low drop regs are difficult in stability terms, needing precise capacitance in the right places. If I drive all these metres of wire and LEDs will this create a stability problem, especially as the load is in series between the ‘earth’ of the reg and the bridge rectifier/smoothing capacitor 0V rail?
With 1.25V dropped in resistor plus the say 0.5V across the reg it’s about 91% efficient. In reality though it would be very difficult to find a transformer to output exactly 20 + 1.75 = 21.75V at 700mA under all temperature, aging, etc. conditions. So I would have to go a little higher to say 24V, doubling or tripling inefficiency. Maybe a switched mode 24V supply trimmed down to 23 or 22V might be more efficient. ‘Headroom’ above 21.75V needed depends how much the LED voltage changes over -5C to +25C, the night-time temperatures my home experiences. |
twintub
Frequent Contributor
Username: twintub
Post Number: 66
Registered: 02-2007
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, 09 February, 2010 - 07:27 pm: |
|
I don't understand why you're concerned about long cable loss - if you were to use a smoothed supply then the loss is purely due to the DC resistance of the cable & connectors - which will be relatively diddly in the great scheme of things. I would just use a linear current limited psu and accept whatever losses occur (mains power is cheap!).
I suppose you could use a switched-mode psu that has a variable voltage output, and feed it into your 'low loss' linear current regulator, then adjust the psu's output voltage to minimize the voltage 'wasted' across the current regulator. Though I would advise against encroaching unnecessarily near to the linear regulator's drop-out voltage, as even a 5v drop is only 3.5W wasted @700mA (again, mains is cheap!). |
twintub
Frequent Contributor
Username: twintub
Post Number: 67
Registered: 02-2007
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, 09 February, 2010 - 07:40 pm: |
|
...forgot to add:
The linear current regulator's resistor will also waste a bit of power - but not much.
e.g. An LM317 (a 1.25V regulator) will need a 3.1ohm resistor between its output & ref pins to get 700mA and will waste just 0.6W! |
terrym
Frequent Contributor
Username: terrym
Post Number: 709
Registered: 05-2005
Rating: N/A
Votes: 0 (Vote!) | | Posted on Wednesday, 10 February, 2010 - 06:30 am: |
|
If you go with the LM317 idea, put a .1uF cap from input to output pins to stabilise it.
(They make great LF oscillators with long leads on them)
TM |
echase
Frequent Contributor
Username: echase
Post Number: 294
Registered: 07-2007
Rating: N/A
Votes: 0 (Vote!) | | Posted on Wednesday, 10 February, 2010 - 11:35 am: |
|
Twintub I guess you are right, there is little voltage drop in the wire. I was confusing myself by thinking of the switched mode situation where the inductance, etc. may introduce a large effective drop.
Mains is ‘cheap’ but I was planning to take out the existing 10W tungsten spot bulbs out of the garden lights and replace with 1 W LEDs to improve efficiency (and life). Adding a wasteful mains supply circuit rather defeats the efficiency saving of going to LEDs.
Above should have said garden temperature, not home temperature! |
Closed: New threads not accepted on this page
