I think this info worth a new thread. After reading the very positive review from fellow user Budget I went ahead and bought one. Apparently the circuit design has changed. I only see three (3) resistors on the back:


img src: http://priuschat.com/forums/members/2009prius-albums-dealextreme-leds-picture2518-back-side-only-3-resistors.jpg


The highest power (which is temperature dependent) I measured with a small lead acid battery is 105.1mA at 12.33V or about 1.3W:


img src: http://priuschat.com/forums/members/2009prius-albums-dealextreme-leds-picture2520-105-1-ma-12-33-v.jpg


I have confirmed that there are three emitters per module by looking through a pin hole and also took a picture (difficult to get a good picture with my point and shoot camera!):


img src: http://priuschat.com/forums/members/2009prius-albums-dealextreme-leds-picture2519-pinhole-picture-showing-3-emitters-per-module.jpg


Finally I measured voltage drops across the three resistors: 328, 308, and 294 mV, voltage across the device was 12.12 V (the battery has gone down some) and total current 93.2 mA. So each resistor is only about 10 ohm. The good thing is that this seems very energy efficient - most of the voltage drop is across the LEDs. The behavior at charging voltage remains to be measured.

Some more pictures:
Packaging with the correct sku number:


img src: http://priuschat.com/forums/members/2009prius-albums-dealextreme-leds-picture2521-packaging-correct-sku-number.jpg


Reflection off the plastic packaging surface also shows the triple emitters clearly in the picture:


img src: http://priuschat.com/forums/members/2009prius-albums-dealextreme-leds-picture2523-reflection-off-plastic-packaging-surface-also-shows-triple-emitters-clearly-in-picture.jpg


Front side:


img src: http://priuschat.com/forums/members/2009prius-albums-dealextreme-leds-picture2522-front-side.jpg

Hi Prius
Good work here. I will take a while to do an analysis of your data and the pics. I will comment later.
It looks like DX have substituted a very good product with one which is likely to fail because of dumb circuit design and paralleling circuits on only three resistors. I would not be buying any more of these. [removed per terms and conditions]
DX have done the same thing with other good products. They really should change the SKU item when they change the product specs. It gets very difficult to know what you are going to get with the next order.
It is hopeless to talk to DX staff on these issues. They are not technicians.

They have three strands of LEDs in series with a resistor each, and those strands in parallel?

So actually no LEDs are in parallel without a resistor?
Thats an OK circuit, I fail to see the problem with it.
What did the old version have there?

Entropy
We have 36 emitters here.
Putting three emitters in series for 12 volts with a current limiting resistor, would require 12 resistors.
!2 resistors of 150 ohms is pretty good at protecting at automobile charging voltages.
Using three resistors to replace the original 12 is just false economy and likely to make the lamp fail. As soon as one emitter goes out, the current is increased to the others.
As I said. Dumb Circuit Design.
It looks like they are paralleling the three emitters in each SMD LED and then placing four in series. Normally you need more than 12 volts for four in series. It would be interesting to see the specs of these SMD LEDs in case they have also changed.
At the moment, I am glad I bought a few spares of the good original design with the 12 resistors. [removed per terms and conditions]
Bottom line is that there is no free lunch here. You cannot run 36 emitters, which should draw around a quarter of an amp, at around 100 mA and expect the same amount of light. Not unless DX has made some startling advances in LED design.
Good work Prius, with your excellent pics and data and the heads up, to avoid this one for now. Your comment about the high efficiency is correct, but you are seeing only about one watt from a LED lamp capable of more than three watts. It was that high power and very good light output which made the original 12 resistor board so good.
A real pity too. It was a first rate LED lamp before they messed with it.

Ah, Ok, there was my mistake. Didn't notice that there were three diodes in each LED case.

Its not that bad, though. The three diodes in one case are thightly thermally coupled, so their Vf can't differ that much, little chance for thermal runaway.
(Look at the SSC P7 for example, there you don't have an other choice but running the four diodes in parallel, while the CREE MC-E offers the option for running them in series. Still SSC P7 work and are sold in large volume)

That said, if the LEDs offer individual connections to their diodes, the design should use them. Saving a few cents for 6 resistors isn't worth it.

Hi Entropy
Here is the first product, to which you refer.
http://www.dealextreme.com/details.dx/sku.11809
There is a big difference between using the cheap current control of resistors and using an electronic circuit board designed to control current. The cheap method works well enough for low wattage LEDs provided that the resistors are chosen well and the voltage supplied is within reasonable limits. There has been a series of problems with poorly chosen resistors and I wrote a bunch of product description warnings for DX more than a year ago. They are still there as part of the product description. There were a whole bunch of LED boards using 120 ohms instead of 150 ohms and DX got a lot of returns from failures when used at charging voltages. There is another new 75 LED bar which is very good, but DX has shipped with 200 ohms first and then 180 ohms. We are back to adding a protective resistor in these as well, to prevent premature failure. It is tough keeping up with all these changes within a single product, but I do my best. [removed per terms and conditions]
I have had to remove products from the recommended list, several times, as product specs have been suddenly substituted with inferior stuff.
Despite these difficulties, DX remains as the best place to buy LED products at a good price.

What I wanted to say with my P7 Example:
When you put LEDs in parallel, one will get more current than the others. As a result it will get hotter, the hotter it gets the lower its Vf, the lower the Vf, the more current it gets...
So, switching LEDs in parallel == bad, one will die, the others will follow.
And it doesn't matter in that case if the pack of LEDs is driven with a simple resistor or a constant current source.

But you can get away with LEDs in parallel if you force them to have the same temperature, by mounting them close together on the same heatsink. That is what the SSC P7 does. There are four LEDs inside, all in parallel, but very closely thermally coupled.

Back on topic: Yes, if used in a Car, simple resistors are not enough.
I'd at least add a TVS, but car power supply requirements are a science in themselves, I usually try to avoid that part when designing circuits...

More data at various voltages:
11.22V 32.0mA
11.50V 46.6mA
11.73V 60.9mA
12.16V 92.6mA
12.48V 123mA
12.96V 163mA
13.22V 184mA
Have fun analyzing!

Hi Prius
You are doing a great job showing how they have sent you a dud substitute.
Let's just take the nominal 12 volts and about 100 mA from your measurements. The three resistors show division in to three parallel circuits, so each gets about 33 mA. Now each of those circuits has four triple emitter LEDs in series, BUT each of the three emitters are paralleled.
SO, each emitter gets about 11 mA
That means they are hugely underdriven and incapable of anything like the light output for the specs of 20 mA per emitter.
What a shame! [removed per terms and conditions]
Even at charging voltages, they are still barely driven to specs.
For this poor design circuit we should expect a current of:
3 times 20 mA = 60 mA
3 times for the three paralleled resistor circuits = 3 x 60 mA = 180 mA