Thats interesting, i didn’t know better than 1% existed.
But i still hope i can buy a kit with everything tuned so at least it is safe.
Personally i don’t need any LVP i just change the battery as soon as it starts to dim to much, but i want to be able to gift some really bright tiny lights to some less aware family members that need every safety feature in the book, like a short turbo timer, LVP & reverse current protection 
Yep, 0.5% is $0.05ea and 0.1% is $0.10- 0.20 at mouser.
The only one’s I see that Mouser has in stock are only 100mW. Is that enough? They show a .5% in the 1/4 watt range but it’s non-stocked.
Its just for sensing the voltage level, very little current. 100mW is certainly more than enough.
Are you looking at only 0805? I was in 0603.
It's kinda pointless to go for more precision than 1% on the resistors.
You’ve certainly built up more drivers. You haven’t seen enough variance in LVP to be worth it?
I have only seen LVP at work on a couple of occasions, when deep testing a light. I’ve never had one step down or shut down in actual use. So in the end, it’s not even really necessary FOR ME that it be there.
Rich is right. The internal reference less accurate than that anyway.
Not only the internal reference, but remember that down at 3V or so you are already down to virtually nothing left in the cell anyways. If you were to look at the discharge curve, you are already falling off the cliff, so to speak. Even if you are 0.05V off, what difference does it really make in terms of runtime? Not much.
I’m really happy that you built this DB. It looks great (even moreso after rebuilding the strap later) and it sounds like it performed well to boot.
Very, very nice. This definitely ups the ante a bit.
No doubt about it!
I usually have to calibrate each driver individually. If I go for a single calibration to apply to several identical drivers, I’m lucky if I can get within 0.1V. I basically have to measure a few and then aim for the average.
… and thermal calibration varies even more.
0.1V is probably close enough for LVP purposes, at least. And close enough for “is it time to recharge yet?” It might be a little coarse for an AA boost driver though, since the voltage range there is smaller.
Yeah, LVP is a pretty rough thing anyway since voltage isn’t a particularly accurate indicator of state-of-charge. How far out do you find the temp calibration to be? I’d think that overheat protection would also be an “in the ballpark” type of thing.
As far as LVP on an AA boost driver, there’s really no need… NiMH/primaries don’t care.
NiMH cells react pretty badly to boost circuits which draw low amounts of power for a long time. I’d want a low-voltage check of some sort to avoid damaging the cells or at least tell me when it’s low. One example is a 1xAA wireless mouse which drains cells to like 0.5V before it finally fails. If I let it go that far, I have to recondition the cell before it works correctly, and the whole process probably eats up a few dozen recharge cycles. I had similar behavior on a 1xAAA keychain light which only got used on low. After like 15 months between charges, I got curious… and found that the cell was as drained as it could get, using spontaneous recovery to recharge, and it had lost a great deal of capacity.
Anyway, low-voltage detection can be handy for stepping down even when the cell isn’t actually low. Like, if you’re direct-driving a triple on a relatively weak cell, it can hit LVP when the cell is still more than half full. But this merely makes it drop one level and keep going, which is arguably a desirable behavior.
The temp calibration seems pretty far off between units. I got the thermal ceiling calibrated just right on one X6, flashed another unit with the same value, and it stepped down at room temperature. At this point, I gave up on having a sane factory default.
In reverse order:
Oh, for high drain devices NiMH works fine.
What would you consider as a threshold between low and high drain for nimh’s? 100mA, 500mA, 1A?
It’s not a threshold. It’s more of a usage pattern.
If a device uses a small steady trickle for a long time, it may be hard to tell when the cell needs to be recharged. A wall clock, for example. It can drain too far, so that by the time it fails, the cell has already taken a hit.
If a device infrequently uses a small portion of the total charge, then rests a long time between, the cell might spontaneously recover between uses. For example, a keychain light used at 10 lm for a few minutes once every few days. This makes the charge rock back and forth across the “empty” line.
However, I also have a ZL SC52 running on moon mode (0.1 lm) all night. It lasts about two months per charge. I don’t have this problem on that device, because it lets me know when the charge is too low. It’ll run fine at first, then when it goes too low it starts blinking at me and I know to recharge it. So, a 3-4mA drain is fine, as long as it lasts long enough to push past the spontaneous recovery, and the device then warns about low voltage.
What I want to avoid is over-draining and repeated recovery without a charger. These both seem to reduce the long-term life span of Eneloops.
Li-ion cells are somewhat easier here, since they work better when the charge stays away from the top and bottom of their range. If they get drained to 2.8V and charged to 4.25V each time, their life span will be shorter than if they stay between 3.1V and 4.15V. They’re quite happy to stay in the middle.
In any case, if a light is on a low mode and voltage drops below 1.0V (Eneloop) or 3.0V (li-ion), it’s almost certainly time to recharge immediately.
/Eneloop%20AA%20HR-3UTGB%201900mAh%20(White)-Capacity.png)
The main place I run into li-ion issues is with phones. Apparently phone manufacturers think it’s imperative to always be at exactly 100% charge whenever possible, so they charge up to 4.37V or so (4.35V cell) whenever plugged in. And then drain a few percent, then charge back up a bit over full, then drain, then charge, etc, with a cycle time of anywhere from a few minutes to a couple hours. Leaving these phones plugged in all the time effectively over-charges them a dozen or more times per day. And then within a few months, people start to notice that the back of their phone is no longer flat. If they’re really unlucky, the screen will bend outward too. Then after perhaps a year, the battery capacity is a fraction of what it once was. I have phones only a year old which practically roll when I put them on a flat surface. And it’s an uphill battle convincing the relevant people that this is fixable and worthwhile.
Sorry for the long post and tangent. It’s almost time for the sun to come up, and I’m overdue for a nap.
My flash units hit on a set of Eneloops VERY hard! We sometimes have to shoot multiple multiple shots at a wedding or especially the reception and by the time the flash is starting to fail the cells are too hot to hold coming out of the flash unit.
BUT, I’ve still got working Eneloops from 2008.
That dovetails perfectly with TK’s point DB! Under very heavy loads NiMH will sag appropriately and make it very easy to tell when the cells are “fully” discharged (aka “empty”). Under very light loads it’s not clear that the cells are “empty” and they can be drained even more than we would normally expect.
Eneloops are tough and take repeated heavy loads like a champ. I’m sure TK is right when she says that overdischarging them causes problems - all NiMH chemistries I’ve heard of strongly dislike being drained all the way to 0v (or close to it).
This is such an advanced driver with 3D stacking of components…I’m really impressed by the design.
I see that Hoop has designed a simpler driver by using a smaller MCU and putting the 7135 vertically.
But I wonder whether nowadays it wouldn’t be easier to design complex 10mm drivers using rigid contact board and folded flex driver?
OSHPark now offers cheap flex boards…
Thermals could be an issue but likely to be solvable f.e. with potting…