What kind of circuit does the KX2 have?

[Bullzeyebill]

I just purchased a used but excellent shape E2L Outdoorsman running the two stage KX2 head. The KX2 no doubt uses a buck type circuit, but it is exhibiting different characteristics that the typical SF buck circuit which can take up to 9 volts and maintain constant current regulation to the LED. The two stage KX2 does not run like that and added voltage above 6 volts is apparently running the driver in overdrive. I know this because using bounce with lightmeter I am getting higher output readings, the high reading an estimated 106 lumens and the low 46, both reading above the stock KX2 readings with 2 CR123's, of approximatey 85 lumens and 6.6 lumens. These figures are questimates and based on comparison with other lights of mine with more or less known lumen output.

I find this very interesting. The other part is that a 17670 runs the light in regulation. Not unexpected there. I have not yet done a full runtime of my KX2 with the 17670.

Has anyone delved into the two stage KX2 circuit board? Is the low mode using PWM? Why is the high mode not running in regulation at 8.4 volts or so? I seem to remember reading that the gen 1 KX2 single mode ran in regulation at 9 volts, providing longer runtime over two CR123's. Can the KX2 two stage be modded to for higher voltage with regulated runtime, not be overdriven at 9 volts? Other questions to answer.

Bill

 

[CM]

If I had a KX2 and was to characterize it, I would hook it up to a power supply and start with an input of about 3V. I would slowly raise the input voltage and record input voltage vs. current. As the voltage increases, the current should decrease. I would expect to see current decrease as I increase input voltage and it would continue this way. I would also expect to be able to go well above the nominal 6V input but I'd probably stop at 9V. Reason is that there are components exposed to the input voltage that have maximum voltage ratings. Capacitors for one thing. The switch mode controller is another. Going above this (especially the rating on the controller) may result in long term reliability issues. I say *may* because short term overvoltage generally have very little effect, only long term exposure to overvoltage can cause problems.

How high above 6V did you run? Most buck controllers rated 6V or above will generally go higher, about 12V or more. Without actually knowing the circuit, it's hard to give a definitive answer.

 

[Justin Case]

Assuming a buck driver, the voltage and current will both increase together initially since the driver will not be running in regulation until the input voltage exceeds the lower voltage limit (plus some voltage headroom) for regulation of the buck IC. After that, then the current should decrease proportionally as voltage increases, keeping the overall power constant.

Even without a bench power supply, you can test for the buck driver running in regulation by using two different battery configurations (assuming that the driver can accept those two configurations). For example, I have some LED Turbo Towers that use an SOB1000 buck driver. When using 2xLi-ion, I measure 0.50A current draw at the tail. When I use 3xLi-ion, I measure 0.33A draw. If you assume as a first cut that the Li-ion cell voltages under load are the same for both current draws, then the normalized voltage for 2xLi-ion is 1.0, and the normalized voltage for 3xLi-ion is 1.5. Multiplying normalized voltage by current, you get 0.50 for 2xLi-ion and 0.495 for 3xLi-ion. Looks like regulated buck behavior.

I think that this latter method is better than using a ceiling bounce to determine whether or not the driver is running in regulation.

 

[Bullzeyebill]

I know how a buck driver will drop current to a certain point as voltage is added. Running 2XRCR123's (8.4 volts starting) and using bounce with lightmeter I got the very high low mode of approximately 46 lumens (should be below 10) and 106 on high (should be about 85 with my numbers), so added voltage did not run the KX2 like the 2XCR123's, and like a typical buck circuit.

I run my SF P60L with 8.4 volts and I get the same light output that I get with 2X primary CR123's, and current at tailcap end is lower. Not so with the KX2.

Just measured current at tailcap end on high level with two RCR123 (8.22 volts) and got 260mA's, and then with two primary CR123's (6.22 volts) and got 260 mA's. ?????????

Yes, my DMM is ok.

Bill

See this runtime plot for 2 mode KX2 on high, CR123's and RCR123's. http://www.light-reviews.com/surefire_e2l/runtime_new.gif

Bill

 

[CM]

Assuming a buck driver, the voltage and current will both increase together initially since the driver will not be running in regulation until the input voltage exceeds the lower voltage limit (plus some voltage headroom) for regulation of the buck IC. After that, then the current should decrease proportionally as voltage increases, keeping the overall power constant...

I stand corrected, the current will initially increase if Vin < Vf + some small voltage and once this is overcome, the current should decrease. Efficiency does drop as the input voltage increases so that the drop is not exactly proportional to maintain constant input power. Input power increases as Vin increases even though power to the load is constant.

 

[Bullzeyebill]

CM, what do you think? You read my two posts. The KX2 does not respond like a typical constant current buck circuit. Why no constant current and regulated runtime at 8.4 volts with the KX2 circuit. It is responding like a boost circuit, but certainly is not a boost circuit. The LED is seeing more than vf to the LED at 8.4 volts or so and is running direct drive. Why? It acts more like a constant voltage circuit above 6 volts or so.

Bill

 

[Justin Case]

Can you gain access to the driver?

 

[Illum]

The 2007 gen KX2 operates on the 17670 quite nicely, see here:
http://www.candlepowerforums.com/vb/showthread.php?173122-E2L-CREE-great-runtime-w-17670

Expect the dual mode to function in similar terms.