Now that there is a reason, there will be new drivers made to take advantage of the lower Vf, just a matter of time.
The easiest fix right now is to simply turn the max PWM down on the TA driver to get the current you want, it will still drop off as the battery drains but at least you will limit current to what you want.
I do this with 219C’s or small lights that I want to limit the heat in. I simply set the max PWM to keep total power where I want.
A firmware upgrade could allow for very crude “regulation” using this setup. While not perfect, it is better then we once assumed. In my testing you only loose about 10% efficiency @ 50% duty PWM. This is a worthwhile tradeoff for the low cost and compactness of a PWM driver over a buck driver.
Although I do have slightly better improvements in mind PWM regulation is the way of the future for cheap and compact drivers. For lights that can use larger drivers buck drivers will become much more popular.
If you PWM a XHP50.2 at 12A, you still get the inefficiency that the die has at 12A, just not the heat sag. The output drop at high currents is not solely caused by heat. Do you know what the individual contributions of those two effects are?
I agree with Dale’s post above. I think these low-Vf emitters can be used advantageously in multi-emitter lights (triples and quads). Direct drive with a single emitter (XPL2 or XHP50.2) usually results in a lot more current but not a huge output increase because the output curve tends to level out at high current. But in a triple or quad the low Vf allows to drive each emitter more optimally. For example a triple XPL with a single cell might draw 12A, or 4A per emitter, but a triple XPL2 might draw 5-6A per emitter which is near the optimal current before the output curve levels off too much.
Of course these low-Vf emitters also tend to not have as good throw, but that’s another story.
I don’t have a function generator to test this as completely as I would like but in my testing I have consistently seen results in the same ballpark as the XHP35 PWM I posted.
In that test it lost about 11% efficiency @ 50% duty tapering to 0% loss at 100% duty. In other tests I have seen similar numbers.
Thus as the battery voltage drops the efficiency loss will taper down to 0%. So it could actually be brightest on a slightly drained cell.
Yeah, it does depend on the pulsed current although the xhp35 is a pretty good comparison from my un-publised testing.
I have tested a few other low Vf emitters and they had very similar numbers, some slightly better like the XP-L2 I played with and some slightly worse (like the 144A I played with).
At 50% duty it appears to range between 10-15% loss. Which like you say, is really not that bad, the higher the duty the lower the loss. Using thin wires and low drain batteries improves the numbers as well. As the battery voltage drops the efficiency loss drops to 0 as well.
The lower ranges can be taken care of with current regulation in most cases without much trouble.
PWM is not ideal but it is cheap, effective, simple and compact. It does have some advantages as well such as little to no tint shift in different modes, seemingly better tint / CRI in my experience ect.
Future drivers will most likely use a combo of constant current and PWM if I had to guess.
One of my build plans was to take a single 26650 light (ultrafire F13) and run a pair of 26350 batteries from Banggood along with a direct drive 22mm driver from Mountain Electronics then put in the xhp50.2 that I tried out in my Convoy L6.
Now I know that will pull way too many amps. I can’t switch to amp limited batteries as they don’t exist. I don’t know of any non direct drive drivers that will work. Do you guys know of any?
I guess I’m gonna have to shelve the 50.2 and buy a new xhp70 to run in it. Does that sound right?
One thing to keep in mind with the 26350 cells is they can’t handle much more than 7A. See discharge curves at the bottom of the banggood listing. Banggood says they are out of stock, by the way.
These 26350 cells with an XHP50 V1 resulted in about 7A on a full charge. Using an emitter with a lower Vf would, in my opinion, probably be asking a bit too much of these cells.
Update, I was looking around at MTN and saw they had a linear driver for 6 volt use. This is a 7135 based driver.
Can someone give a summary of how the new low voltage emitters will respond to different driver types?
I think there is going to be a lot of confusion over this.
Direct drive = We know that they will draw way more amps than the regular xhp50 and xhp70. Not good at all unless you can limit the amps by using certain batteries or some other means.
Boost and Buck = It should draw the same currents. We may get improvements in efficiency, right?
Linear = It should draw the same currents. Anything else I’m not sure.
BTW, I tried my L6 with the 50.2 in it with some fully charged protected KeepPower 5200 26650. They are supposed to have a 10 amp limit. I expected them to trip the circuitry on turbo, but they didnt. Maybe they limit it to 10 amps?
Compared to fully charged Liitokala it’s not quite as bright and it’s not heating up as quick. In fact, the KP are getting a lot less current in all modes. Even slightly less current than the xhp70 with the KP batteries. Pretty interesting.
I can only compare amp draws as I don’t have a clamp meter to get exact numbers. I know where I can borrow one if anyone is interested. Idk, maybe I’m boring you guys. Lol
That 22mm driver will work, you can simply not use the highest mode until the battery voltage drops or have RMM set the max pwm in turbo to say ~200 to limit current. It depends on how much self control you have.
So I was thinking about these new generation LEDs with FET drivers. If the MCU is monitoring cell voltage with an ADC for LVP, would it be possible to reprogram the MCU to use the value it reads to limit the PWM until cell voltage falls?
Ex. If it sees the cell voltage is at 8V, only command a maximum of 75% PWM or something, then when the cells are down at 7V let it get to 90%, etc.
If that does work it should let us take advantage of the efficiency increase of the new emitters as well, which linear does not.
Thanks, I’ll make a note of it. Richard at MTN has me covered, though. This one.
6 amps is no slouch and I can remove 7135 chips if I want to bring it down some for better battery life and less heat build up. The best thing is I can have Guppydrv with off time memory and it’s user setable turbo timer. :+1:
Some speculation... They are binned the same, so same amps = same lumens. For the lumens output issue, depends on the driver, whether it's more amps limited or total power (watts) limited. The lower voltage of the 50.2 could be a good thing (more lumens output) or potentially bad thing - might have to burn off more voltage, so more heat. For the beam/tint, of course depends on what you are going from/to. The 50.2's don't have the same tendency to black hole the hot spot, but they do tend to yellow around the hot spot. Generally these new design LED's don't throw as well as the old ones - I know the XPL2's don't throw as well as the XPL's running at the same amps.
I'm not familiar with the EA81 or it's driver, so again, can't get specific.Those that know details on buck and boost drivers and worked with the 50.2 could probably answer better than me. Few trade-offs.
Yes, this is possible and something I have been wanting to get worked into Bistro from awhile now.
Although at this point I think we might end up skipping that rather crude method and going to something better unless someone gets it implemented before we make the move to the newer MCU’s.