DX Driver Board Tests - Efficiency Results Included

[Aircraft800]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

Thanks, I didn't know the resistors numbering all started with a "R". How do determine the Ohm value? Is the numbering value R27 a .27 Ohm? What would a 1 Ohm say 1R? So now you have total of a .21 Ohm sense resistor?

I'll have to see if the local Radio Shack carries them for a little upgrade.

Thanks for the help! (sorry, been away from electronics since college)

 

[VegasF6]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

What would a 1 Ohm say 1R? So now you have total of a .21 Ohm sense resistor?

Wouldn't that just make too much sense (hah)
I think it depends on the manufacturer some. Here is a chart from vishay.
http://www.vishay.com/docs/20020/smdmark.pdf
See, it depends on the tolerance, at least in there case.
http://www.talkingelectronics.com/ChipDataEbook-1d/html/SM-Resistors.html

 

[eprom]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

Hi Friends,

I got theese boards from postal service and make some measurments. First batch of theese boards has PT4105 chip, now has PT4115 chip.

MR16 1*3W Driver, using PT4115 chip. (Don't forget this design have rectifier diodes to work on AC/DC applications. Theese diodes eat efficiency a little)

Results Drawn back until checked and confirmed again.

Thanks,
EpRoM

 

[TorchBoy]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

Thanks again for these results eprom, but can you please test to lower voltages? Enquiring minds want to know just when it drops out of regulation, and how. Is it sudden or gradual?

Does the efficiency improve running multiple LEDs? Have you seen any with the EQB8L chip?

 

[eprom]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

It starts regulation at 6.5V and 0,45A. Under 6.2V board does not work.

I Will measure the performance for 4 LED setup.

Thanks again for these results eprom, but can you please test to lower voltages? Enquiring minds want to know just when it drops out of regulation, and how. Is it sudden or gradual?

Does the efficiency improve running multiple LEDs? Have you seen any with the EQB8L chip?

 

[TorchBoy]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

:twothumbs

 

[Mitica]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

I thought you were talking about this one
SKU 26110 18V 5W Cree Circuit Board for Flashlights (16.8mm*5.5mm) ?

Which one is the R27 sense resistor? ^

I thought it was the new Kennan replacement?

I've burned 2 of those with 16v input :thumbsdow. They are rated for 18v. It seems that the yellow capacitor that sits in between chip and inductor was turned in ash. Seems that both capacitors are identical, so I hope I could still make a working one from 2.
Does anyone know the value of those? it saids on it S 160C
Thanks

 

[eprom]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

Hi, Like I promise I complete the tests for 2*, 3*, 4* serial connected LED's driven by PT4115

Results,

Results Drawn back until checked and confirmed again.

 

[TorchBoy]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

Here's a random thought. The spec sheet of the SS14 Schottky diode (which is used for rectifying in the EQB8L version of the DX SKU.13577 driver I have) has a Vf just over 0.4 V at 0.45 A. Two in the rectifier make 0.8 V, which at the peak efficiency an input current of 0.45 A will be 0.36 W, so 4.2% of the input power is being dissipated in the rectifier. 97.26% + 4.21 % = 101.47%. That's the sort of driver efficiency I like!

But seriously, I have no idea what's going on there. Thanks heaps for those test results, eprom.

 

[eprom]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

Hi TorchBoy,

I am using a Laboratory DC Power Supply,

Problem with this Supply is Amp. Meter has only two digits after dot and Volt Meter has only one digits after dot. So when I make a amp. measurment like 1.69A it could be 1.685A, 1.689A or 1.694A. And when I make a volt measurment like 4.0V it could be 3,95V~4,04V So it has a faulire rate on measurments. And it is effective on small measurments under 1A. But this is the best I can do. I dont have an amp. meter has an three digits after dot.

I wish this would be acceptable.
EpRoM

Here's a random thought. The spec sheet of the SS14 Schottky diode (which is used for rectifying in the EQB8L version of the DX SKU.13577 driver I have) has a Vf just over 0.4 V at 0.45 A. Two in the rectifier make 0.8 V, which at the peak efficiency an input current of 0.45 A will be 0.36 W, so 4.2% of the input power is being dissipated in the rectifier. 97.26% + 4.21 % = 101.47%. That's the sort of driver efficiency I like!

But seriously, I have no idea what's going on there. Thanks heaps for those test results, eprom.

 

[Justin Case]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

What you can do is turn the voltage dial ccl until you reach the min range (e.g. 3.95V, where the display fluctuates between 3.9V and 4.0V). Then turn cl to find the max range (e.g., 4.05V, where the display fluctuates between 4.0V and 4.1V). Finally, put the dial in the middle. Hopefully, the voltage value for the middle dial setting is close to 4.0V.

 

[eprom]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

If I make a new test, I will keep in mind? This is a brilliant idea.

What you can do is turn the voltage dial ccl until you reach the min range (e.g. 3.95V, where the display fluctuates between 3.9V and 4.0V). Then turn cl to find the max range (e.g., 4.05V, where the display fluctuates between 4.0V and 4.1V). Finally, put the dial in the middle. Hopefully, the voltage value for the middle dial setting is close to 4.0V.

 

[Justin Case]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

Built an LED Turbo Tower for use in SureFire TurboHeads.

I'm a little disappointed with the KD1640 driver that I used. Ordered six of them and this is the only one I've used so far. Driving an XP-G R4, which also underperforms compared to some other XP-G R4s I've gotten.

The KD1640 delivers about 1.05A, so that's good. Right in line with what you'd expect based on the sense resistors.

The first disappointment is that the XP-G R4 Vf ~ [email protected]. True that this is basically in agreement with what is in the datasheet. But I had previously measured about 3.1V for another XP-G at essentially the same drive current. Hot spot lux at 1 meter is also about 10% lower vs the same XP-G with Vf~3.1V. Measuring the XP-G die height is difficult, but the two towers look to be the same within 0.003", so I don't think LED focus is the issue. The reason might be due to LED lumens variation in the R4 bin.

The second disappointment is that the KD1640 seems to have lower efficiency than expected -- about 80%-83% vs the expected ~90%.

The third disappointment is that the KD1640 seems a little unstable for Vin of about 5.0V +0.3V/-0.2V. Power-in takes a repeatable dip starting at about 4.8V, hits the bottom of the dip at 5.0V (~15% drop in power-in), and finally returns to the steady state value at about 5.3V. At least for this specific driver board, using 2x123A might be less desirable than 2x16340 Li-ion. The Power-in vs Voltage-in curve is otherwise reasonably flat when running in regulation (I measured out to 12V).

The last disappointment is that the KD1640 voltage overhead is slightly higher than expected at about 0.7V. Still reasonably good.

 

[Ekke]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

Hi, Like I promise I complete the tests for 2*, 3*, 4* serial connected LED's driven by PT4115

Results,

http://dosyalar.ledcozum.com/CPF/Efficiency_3W_MR16_Serial.jpg

Thanks for those.. But have you measured the output current? Small change, but you have so many digits in your numbers.

From PT4115 datasheet:

I made diy-version today, I will measure the prototype tomorrow if everything goes well.. Currently running @ 1.13A with one led and 12V lead acid battery. About 76% efficiency.

 

[eprom]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

Waiting for your Results,

You are right I have not measured the amperage for every voltage change. For AX2002 it was very rigid so I think like AX2002 test when making PT4115 tests, I measured the LED current once and don't care about it again. You catch it well, thank you for your carefulness. I am drawing back my results for PT4115 chip. I will control them again when I have spare time.

Thank you,
EpRoM

Thanks for those.. But have you measured the output current? Small change, but you have so many digits in your numbers.

From PT4115 datasheet:

I made diy-version today, I will measure the prototype tomorrow if everything goes well.. Currently running @ 1.13A with one led and 12V lead acid battery. About 76% efficiency.

 

[Ekke]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

Waiting for your Results,

You are right I have not measured the amperage for every voltage change. For AX2002 it was very rigid so I think like AX2002 test when making PT4115 tests, I measured the LED current once and don't care about it again. You catch it well, thank you for your carefulness. I am drawing back my results for PT4115 chip. I will control them again when I have spare time.

Thank you,
EpRoM

I used cheap multimeters, so ±5% at least? Measured output voltage and amps, input voltage and amps I read straigh from the power supply. Measured in 1V steps, here's some graphs with a little filtering where you can get the idea: Efficiency and Current vs. voltage & leds.

Leds were XP-E, XP-G, 2x XR-E & 2x P4. 0.0825 Ohm theoretical R_sense so should be about 1.2A current. There were quite long and thin wire mess and not enough heatsink so it was pretty hard to measure because of heating so those isn't so accurate. Enough for me though, there seems to be consistent behavior. Datasheet says those should be like this @ 770mA. 3 LEDs are pretty close.

Your numbers were quite similar? Pretty good at higher voltage and >=4 leds, I think I will run those @ 27V with 4 leds.

 

[TorchBoy]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

Interesting that the 3 LED lines should be so close but the 1 LED lines aren't. Thanks for posting those.

 

[Paul Baldwin]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

Hi Eprom,
I'd just like to say thanks for posting the info on the ax2002 driver I ordered some, ignoring the title (which still hasn't been updated) and have tried one out on a 12 volt supply. I connected it to an xpg r4 and ran it for 5 hours straight with no additional heatsinking to see if it would cope. It certainly gets hot but didn't flicker or fail. All 4 are now glued to the back of their respective heatsinks before I potted them, hopefully they'll be ok. I'd been looking for a cheaper higher output than the MR16 12+volt capable driver for a while.

Paul.

 

[TorchBoy]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

BTW, it's quite easy to change the set resistor on AX2002 drivers, including the MR16 drivers. The EQB8L driver chip used on some of the MR16 drivers I think doesn't do very well at higher current levels though, but those that use the AX2002 do just fine.

 

[Paul Baldwin]

Re: Newly Designed DX Buck Driver (Kennan Like) 90% Efficiency

Ta for that, I might try it next time I buy some

Paul.