Mountain Bike Reviews Forum banner

Dual XPG, XML Hiplex build, thoughts please

8K views 55 replies 14 participants last post by  yetibetty 
#1 ·
OK, because of DX wonderful (joke) customer service I find myself with a a few spare of these drop ins. I have also ordered an xml drop in with what looks to be a similar reflector. I have on hand a hiplex driver that has been sitting waiting for a project for sometime now.

I'm thinking of building a three up helmet light and running two parallel driven xpg in series with a single xml @ 2.8A, the max output for the driver.

If I'm right the two xpg should see 1.4A each and the xml should see 2.8A, yes?

Is there anything I should be concerned about, apart from blinding the poor local wildlife?

How do I work out run time for this type of configuration?

Given that the hiplex needs +0.5v overhead 2 2.8A would I be able to run this off a 7.4v battery?

Thanks in advance
 
#2 ·
emu26 said:
If I'm right the two xpg should see 1.4A each and the xml should see 2.8A, yes?
Yes.
emu26 said:
Is there anything I should be concerned about, apart from blinding the poor local wildlife?
I think it should be fine. Except, of course, in case that one of the XPGs says good-bye (for whatever reason), the other one will follow it almost instantly......
emu26 said:
How do I work out run time for this type of configuration?
Think of it as all three leds are being driven independently and then sum up their power consumption.

Let's assume XML @ 2.8A has Vf around 3V, that is 8.4W.
Let's assume XPG @ 1.4A has Vf around 3.2V, that is 4.5W for one, 9W for two of them.
So, all together 17.4W.
Now, calculate how much Wh (Ah x V) your battery offers, account for some losses in driver and wires, divide that with leds power consumption and you have it...
emu26 said:
Given that the hiplex needs +0.5v overhead 2 2.8A would I be able to run this off a 7.4v battery?
Yes, it would be fine. Hipflex will see a Vf of one XML and one XPG as the output Vf, which should be around 6.2V. So, only at the very end of the battery discharge curve it should switch to direct drive, the main part of the curve it should be well in regulation.
 
#3 ·
Thanks Ortelius.

I was aware of the dangers of losing one xpg and the impact it would have on the other. One thing I hadn't considered though was if I do loose both xpg do I by default also loose the xml because no power will flow the xpg to the xml?

If that is the case is it possible to run a bypass cable around the xpg to the xml. Have it switched with a simple on off clicky switch. Keep the switch off all the time but if something goes wrong with the xpg, or I just want to run with the xml, turn the switch on?

Thoughts anyone?
 
#4 · (Edited)
emu26 said:
One thing I hadn't considered though was if I do loose both xpg do I by default also loose the xml because no power will flow the xpg to the xml?
Actually no, your XML will survive that situation just fine. I'm not sure how bunt-out LED behaves: does it act s an open circuit or as a short circuit. In first case, if both XPGs leave the output circuit open, then no current will go through XML, so no harm will be done (except that you'll be left in a total darkness ;) ). On the other hand, if dead XPGs behave as a short circuit, then driver will still provide the same 2.8 A of current on the output, so your XML will continue to work, as if nothing has happened. Of course, the driver's efficiency will be worse (because the output Vf will be halved), but as long as the driver can cope with the additional heat caused by that decrease efficiency, your XML will never know that its neighboring XPGs are dead.

EDIT: Actually, given it a second thought, it is obvious that burnt-out XPG will behave like an open circuit, otherwise first dead XPG would not cause the second one to die, too. So, in such a situation, you'll be left with non-functioning light, but no harm done to your XML.
 
#6 ·
emu26 said:
It was the being left in total darkness that I was referring to and why I was suggesting / asking about the bypass so that if the dead xpg stopped the juice flowing through to the xml I could flick a switch and provide an alternate route.
:aureola: Sorry, I've read your question too superficially. :prft:

Yes, your plan with a switch in a bypass should work just fine. But I doubt it is worth the trouble. I have couple of my lights that have leds in parallel configuration and never have had any problems with them.

Unless, of course, you want to use that second switch also for additional dimming purpose (turning both XPGs on/off at will). But with such an advanced driver, there is probably no need for this either.
 
#8 ·
Thanks guys.

Goldigger I already have the hipflex with no other project in mind for it and would like to keep this build cost to a minimum (as far as new outlay is concerned).

As for the battery I have a handful of old 12v NiMH which are almost on their last legs and a couple of 14.8v li-ions. I asked about the 7.4v simply because I am also building another twin DX b2flex helmet light which I plan on getting some small batteries for so I can put them on my helmet.

Problem is as I type this I suddenly remember that the new b2flex require much more over head than the old bflex and I don't think they'll run a twin xpg from a 7.4v battery :(
 
#9 ·
I read this a few days ago, and it might apply to your question.

"Electrical overstress (EOS) is often associated with catastrophic failures...

Beyond a certain threshold, a single-pulse event will lead to an immediate catastrophic failure of the LED. In this case, there are two general failure modes: either a short circuit or an open circuit."

Cutter. "Pulsed Over-Current Driving of XLamp LEDs.Information and Cautions." Cree: Application Notes.: (2010-2011) Online. http://www.cree.com/Products/pdf/XLamp-Pulsed-Current.pdf
 
#11 ·
Mr Postman arrived today. :)

Quick beam test against the ceiling with a couple of "jerry" rigged AAA batteries shows a nice even hotspot, a couple of rings around it, which should be lost with the xpg beam, and a fairly tight spill, even tighter than the DX xpg pill I used previously.

Sorry no pics, it drained down the batteries too quickly.
 
#12 ·
emu26 said:
If I'm right the two xpg should see 1.4A each and the xml should see 2.8A, yes?
My only concern would be that the forward voltage of the two XPGs might need to be pretty evenly matched. The one with the lower Vf will be sinking more of the current than the one with the higher Vf. The HipFlex wouldn't know - it's going to send 2.8A between the two, but if one is sinking 2A and one 0.8A then there might be problems. Doubt it would be that extreme though.
 
#14 ·
mfj197 said:
My only concern would be that the forward voltage of the two XPGs might need to be pretty evenly matched. The one with the lower Vf will be sinking more of the current than the one with the higher Vf. The HipFlex wouldn't know - it's going to send 2.8A between the two, but if one is sinking 2A and one 0.8A then there might be problems. Doubt it would be that extreme though.
Yes, that is one of the problems with paralleling LEDs. Still happy to roll thee dice with this one though. Everything except the xml pill was "stock on hand" :D

Itess, be patient my boy, be patient ;)
 
#16 ·
mfj197 said:
My only concern would be that the forward voltage of the two XPGs might need to be pretty evenly matched. The one with the lower Vf will be sinking more of the current than the one with the higher Vf. The HipFlex wouldn't know - it's going to send 2.8A between the two, but if one is sinking 2A and one 0.8A then there might be problems. Doubt it would be that extreme though.
The problem is even worse that that. Remember that Vf decreases with temperature. What happens is that one LED draws a bit more current and that heats it up. This becomes a cycle and one LED will hog more current. However, resistance in the circuit and the LED itself helps fight that effect. I would consider adding a tiny amount of resistance to help ballast the two XP-G's that are in parallel. You should only need a voltage drop of .05 to .1 volt to balance the LEDs. That's only 66 milliohms in series with each XP-G for that .1v drop at 1.5 amps.
 
#17 ·
Thanks Vroom, I have read about using resistors before to help balance parallel strings but never really understood.

Will one resistor before the cable splits to the two LEDs be enough or should I put a resistor between where the cable splits and each LED?

In simple terms if for a non electronics person, how does this help the leds stay balanced?
 
#18 ·
emu26 said:
Thanks Vroom, I have read about using resistors before to help balance parallel strings but never really understood.

Will one resistor before the cable splits to the two LEDs be enough or should I put a resistor between where the cable splits and each LED?

In simple terms if for a non electronics person, how does this help the leds stay balanced?
The resistors need to be in series with each LED so you need two resistors and they need to go between where the cable splits and the LED. With only a tenth of a volt of drop the LEDs will need to be as close to the same temp as possible. Try and design it so each LED has more or less the same geometry as far as heat sinking goes so they run at the same temperature. If you are going to build it three in a row that would mean putting the XM-L in the center.
 
#20 ·
Vancbiker I believe the theory behind it is sound but I'm not sure its enough to have me concerned, yet. It probably will be if I don't protect against it and I have problems.

Anyway, day one on my lump of Al.

A tad under 300gms on the scale at this point in time
Rectangle Natural material Collection

No work shop for me so about 3 hrs to this point including setting up, packing up and cleaning up
Product Rectangle Material property Silver

That time includes this housing which will replace my "cheap torch" that I have constant problems with
Rectangle Beige Parallel Paper product Square


Stay tuned, but don't hold your breath. Hopefully they'll both be done in time for The Mont.

Of topic, any other Aussies doing The Mont this year?
 
#21 ·
I'm runnng two XPGs in parallel up to 3.2 Amps for the pair and have not noticed any color shift or other indicators that there is a problem.

Granted these emitters were bought as a connected pair from cutter, so I would assume (I didn't test) that the VF are fairly close.

Another caveat, I have temp sensing set to 50°.
 
#24 ·
Day 3, + 2 hrs

Small light body now all but complete, still need to tidy up the top grooves which I have done by hand (last time I tried to groove a block this small it cost me a saw blade.)
Product Rectangle Metal Font Fashion accessory

Product Wood Tool Gear Metal

Circuit component Passive circuit component Electronic component Composite material Electronics accessory


Day 4 +1.5hrs

Big light, new face plate cut and drilled, I like to cut twice and measure once ;)
Holes for DX pills drilled but not yet milled
Wood Rectangle Circle Font Composite material


Lesson, never try to pick up a block of al that has been under the drill press for about 30 mins :madman:
 
#25 ·
Emu, I like the look of this Big light of yours. I assume you used hole saws.
I still have burn marks on the floor boards from using them when I picked the Ali up straight off the drill press and dropped it as quick as I could.

Nice work, looking forward to seeing it finished.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top