Itās definitely a buck driver. Your earlier description left the possibility of a malfunctioning buck driver. We now know that itās a properly functioning buck driver.
A āproperly functioning, but incorrectly-specāed buck driverā (the 3 amp part) :)ā¦
Hereās a link to the resistor mod thread for the other driver. Whatās unknown is the resulting emitter current.
Hi,
Thisāll be a bit of cross-posting, but I just noticed that the driver in the Cofly KX-T21 is apparently this smaller driver:
Jim
Iām curious about something?
On the other thread about the Cofly light which appears to use the same driver, nofearek9 reported 1.4 amps at the tailcap with 2xLi-ion. So I guess thatād be like 1.4 x 8V = 11.2 watts, right?
On the emitter side, letās say the Vf was about 4V, so if the power-out was ~11.2 watts, then emitter current should be about 2.8 amps?
But, from the testing I did initially with the bench power supply, even though I didnāt go more than say 5.5V, current at the emitter didnāt increase to that level (seemed to go down instead)?
Shouldnāt I have seen, say, 3 amps at the emitter?
Iāll need to test with the single XM-L again, but this time, bump the Vin to the driver to maybe 8V (yes, Iām willing to lose that emitter)ā¦
EDIT: I just tested again, both with a single XM-L U2 (I think) and a new XM-L2, and am seeing the same behavior I saw before. With Vin up to 8V, the current both at Iin and Iemitter were dropping to about 0.7 - 0.8 amps. I also measured Vemitter (just in case) and was about 3V with the XM-L2 (I didnāt make a note of it for the XM-L, but I think that Vemitter was about the same, i.e., about 3V).
So Iām still puzzled!
I understand this is a buck driver, but where the heck is that power/watts disappearing to? Is the driver soaking up all the extra power/current?
I SHOULD be seeing about 2+ amps Iemitter, correct :(?
EDIT:
Although this is not the right chip, just to get an idea, I found this:
http://andrewpearson.org/?p=643
So, assuming that this driver is using a similar chip, I am guessing that the answer to my question about where the excess power is going is to the sense resistors in the R1/R2 voltage divider (fig. 11 in the above page), and probably the reason Iām seeing lower than 3 amps is that the sense resistorsā values are controlling that.
So thatās why you all are looking for the board without the toroid, to try to figure out which resistors could be/need to be changed to try to bump the driver output current?
Your description of whether emitter current (Iout) stays constant keeps changing! I asked for clarification earlier and you clarified by giving an example where it did not change.
Iout should stay the same once Vin is above a certain voltage. Your most recent description of Iout dropping shows a problem of some sort.
Letās really make sure we are on the same page now. Are you saying that Iout stays constant with 2*XM-L but changes with only 1*XM-L?
In the most recent scenario you described [With Vin up to 8V, the current both at Iin and Iemitter were dropping to about 0.7 - 0.8 amps. I also measured Vemitter (just in case) and was about 3V with the XM-L2] all your losses are in the driver.
After your good find in nofearek9ās COFLY KX-T21 review we are left with 2 or 3 possibilities (that I can think of):
- You got a bad batch of HX-1163c drivers, nofearek9 got a good one in the COFLY light.
- The version you got is on the same PCB but uses different components to support higher Vf than the one in the COFLY light. And itās just a crappy driver after those modifications.
- The version you got is on the same PCB but uses different components to support higher Vf than the one in the COFLY light. And the particular items you got are defective.
Without component values from a working board itās going to be difficult if #2 or #3 are the case. I could not find specifications for this driver on Lightmalls, they donāt even list that itās a buck driver, the input range, or anything.
I think your interpretation of where the extra power is going is probably wrong. If the sense resistor(s) was burning off 3.5W youād know it ;-). One or more things should be getting fairly hot though. Once you move the inductor out of the way we can check.
Basically this board is nothing but a (broken, non-functioning) mystery right now. Werner and I asked to see clear pictures of the board without the inductor because of that. We donāt know the parts in use, their configuration, or anything. I doubt that there is a magic button under there, so thereās only so much good to be had by guessing what we want to see under thereā¦ if taking the inductor off and showing pictures on the forum generates results youāll learn from that! A list of all legible component markings would also be useful - sometimes those chips are very hard to read even in good pictures.
Oops. As a follow up, I took a slightly closer look at the blog post you linked to. Figure 11 isā¦ not relevant. It shows a constant voltage configuration, not a constant current configuration. Iām no guru, but I strongly doubt that the modifications Andrew Pearson suggested to Fig11 work correctly. I would hesitate to use his blog post about a probably failed project as a learning resource. (no offense to AP, I encourage everyone to try and make high current drivers!)
Note that thereās no followup post with a success story. Iād say itās a safe bet that this project was abandoned because it did not actually function.
Hi,
Sorry, I was running to do some Sunday stuff earlier, and just got back.
Re. the AP page and Fig. 11, I was just pointing to that for some info.
I think that we may be working at different purposes, or maybe more accurately, for the same purposes but doing different things towards that.
What Iām (still) hoping is to get this driver pushing 3 amps to the emitter, if it can do that. I suspect you (all) might also want that.
Iām assuming that the driver should be able to do that (yes, I know what A-S-S-U-M-E stands for), so I keep thinking that I must have something āincorrectā, i.e., maybe a wrong emitter with a wrong Vf, or whatever, so Iāve been doing a lot of switching things around here, using different emitters and number of emitters, different wires, etc.
At the same time, I donāt have (easy) access to what Iād need to pull the coil off the driveā¦ or more precisely, Iām not very good with solder wick or with the solder suckers, so Iāve been putting off doing something that would potententially destroy the driver.
I think I understand the more methodical approach that you and Werner are taking, though, so Iāll give that a try at this point, since I have some time with whatās left of this weekend. Iām actually thinking of just cutting that sucker off of the board, because thatād definitely be the fastest way to get it off. Iāll post back if I can get some pics of the bare toroid-less board.
Later, and thanks,
Jim
No - Wait - Stop - Donāt Cut. ā¦just melt the solder at the joint and pull it loose from the board with needle nose. (just lurking here, waiting to see what this board is doingā¦)
LowLumen is right. Cutting is fine, but I was not suggesting cutting, sucking, or wicking. There is plenty of slack available to remove one leg of the inductor at a time. Just heat them up and pull them out, one at a time.
I see nothing to indicate that youāve got anything wrong unless itās wire placement, and that sounded fine. Youāve already tried several different LED setups and youāve tried a range of input voltages. Continuing to plug at it in that way is a waste of time IMO.
I donāt know why I missed this, but I found that I can re-position the toroid well enough to get some pics that might be good enough, so Iāll post those in a bit, then, if itās still not good enough to help, Iāll do what I need to get the darn thing off.
Some new, hopefully better info:
I used some new heavier leads from the driver to the emitter (an XM-L2 on Noctigon) and re-tested and I think that this is the behavior I am seeing (and I tried it several times):
- As I increased Vbat (Vin to the driver) when it gets to somewhere around 4.5 - 5V, the current to the driver (Iin/Ibat) starts dropping, and basically continued to drop as I increased Vbat/Vin.
- I have my clamp meter on the emitter lead, and watched Iemitter more carefully this time (sorry about earlier), and also measured Vemitter. I did these Iemitter and Vemitter measurements at Vbat/Vin at 5V - 9V at 1V increments. For all of these measurements, Vemitter stayed almost exactly 3.00V, and Iemitter stayed at 0.89 - 0.90 amps.
In other words, it looks like above a certain Vbat/Vin:
- Iemitter stays at 0.89 - 0.90 amps and Vemitter stays at 3.00V, but
- As Vbat/Vin increases, Ibat/Iin decreases.
Iāll post the pics that I can get in a bitā¦
Some pics from different angles. Some are duplicates, but just in case, and some have components labelled in RED:
EDIT: These pics are from the 2nd copy of the driver, which I havenāt actually used (notice no emitter leads on it yet). This board appears to have a solder bridge (you can see the big blob of solder in the 2nd pic), which Iāll have to remove and then add the emitter leads and then test, so Iāve been focused on using the 1st board.
If I just pulled one leg out, I think that the through hole would probably just fill when I pull the leg out. Iāll do that if needed, but let me know if the pics I posted are not good enoughā¦
Hi,
Can someone sanity check me?
I was just looking at the pic of the driver from the Cofly thread that I mentioned above:
And, I donāt know if Iām imagining it, but it looks like the hole where the toroid leg goes into and the hole where the positive emitter lead goes into on the above image is reversed compared to the boards that I have (well, the toroid leg hole anyway) - some new pics below where I tried to take the pics in similar orientation to the above pic, so you can compare:
In the pic from the Cofly thread, it kind of looks like the toroid leg is going into the left through hole/pad and the red emitter lead is going into the right through hole/pad, whereas in the 2 pics of my board, the toroid leg is going into the right through hole/pad and the red emitter lead (which I added because the toroid was already in the other hole/pad) is going to the left hole/pad?
Itās kind of hard to tell from the 1st pic, so Iām going to post and see if nofear can post a clearer picture.
Looks like that R200 resistor is your sense resistor: soā¦ I = .18 / R or therabouts if your LED current measures are correct.
Hi,
Can someone sanity check me?
I was just looking at the pic of the driver from the Cofly thread that I mentioned above:
And, I donāt know if Iām imagining it, but it looks like the hole where the toroid leg goes into and the hole where the positive emitter lead goes into on the above image is reversed compared to the boards that I have (well, the toroid leg hole anyway) - some new pics below where I tried to take the pics in similar orientation to the above pic, so you can compare:
In the pic from the Cofly thread, it kind of looks like the toroid leg is going into the left through hole/pad and the red emitter lead is going into the right through hole/pad, whereas in the 2 pics of my board, the toroid leg is going into the right through hole/pad and the red emitter lead (which I added because the toroid was already in the other hole/pad) is going to the left hole/pad?
Itās kind of hard to tell from the 1st pic, so Iām going to post and see if nofear can post a clearer picture.
NVM the above, I think that is a false alarm. It looks like the two pads where the toroid leg and the emitter lead go to are shorted together anyway, i.e., so it doesnāt matter which pad the toroid leg and the emitter lead are soldered onto :(ā¦
nofearek9 has posted info on news test he did with tailcap current on his thread:
i got from both 1amp on high 0.5 on med and 0.25 on low
have in mind now i have moded the flashlight with xpg2 on copper pcb
The above was with 2 different pairs of batteries. I donāt know what emitter voltage he was seeing, but letās just say it was also about 3V, so then his input power would have been:
1 amp x (2 x ~4V) = ~8 watts
If his Vemitter was ~3V, then his Iemitter would be (assuming no power loss between input vs. output):
8 / 3 = 2.67 amps. Good.
Whereas, in my latest measurements, I was seeing about .9 amps Iemitter at 3V Vemitter, or about 2.7 watts at the emitter end.
It definitely seems like the driver I have is underperforming compared to the one in nofearek9ās Cofly?
Just bump up that R200 sense resistor to the formula: I = .18 / R
Looks like that R200 resistor is your sense resistor: soā¦ I = .18 / R or therabouts if your LED current measures are correct.
Do you think I should try bridging that resistor with something in parallel? If so, what value would you suggest? I have some resistors, but I donāt remember what values, but maybe I can pull something from one of the drivers in my junk box?
Also, earlier it was not recommended to short across the current sense resistor, but what about jumpering just for a short test? Would that cause a problem (like āpoofā?)?