The layout of the BQ chip and related components is very restrictive, but I could maybe swing it if I flip that part of the circuit 180 degrees. Was even considering having only a 4-6 pads for the brass ring in that case, spaced equally for better stability. Will give it another go.
As you mentioned 4-layer is the other way out. That does take complexity and cost up a notch, especially for the blind/buried vias we need. We would need to talk to the manufacturer before we go this way.
Given that a separate small PCB is going to be needed for the USB connectors, would there be any merit in including some or all of the powerbank/charger function onto this PCB, if it would make the main PCB and brass ring easier ?
It might be more cost effective than e.g. a 4 layer main PCB, or extra machining of the brass ring and special masking and assembly processes.
Please could consideration be given to mechanical reverse polarity/short circuit protection, I note that we have now had a report of this occurring “in the field” on a Q8.
A remote control using RF “plipper” would also be a great feature if possible.
Narsil was the sword of Isildur, which once defeated Sauron. Andúril was the sword of Aragorn, the blade Narsil reforged. If it looks similar, it’s because it is.
Basic operation hasn’t changed much, but it has changed:
“Hold” always goes up… unless it was already at maximum or it has been less than a second since the last ramp-up. So, no need to remember which way the ramp is going.
“Click, release, hold” was added, and always ramps down. No need to ramp up first.
The mode group UI is exactly the same as the ramping UI, except with stair-steps instead of a smooth ramp. No need to remember two different UIs.
Accepts slower clicks (~0.4s instead of ~0.3s), since people were having trouble with that.
So, usage is very very similar. However, under the hood it’s completely different, a from-scratch rewrite using totally different abstractions.
One of the bigger UI changes is how things are configured. Instead of one large config mode, config settings are attached to the mode they affect. To change a config option, go to the mode you want to configure and then either hold the button (for things like strobe speed)… or click three times (for numeric config settings) then click N times (to set the value to N). For example:
Want to change the bike flasher brightness? Go to bike mode and hold the button.
Want to change the beacon timing? Go to beacon mode, triple-click to enter beacon config, then click N times to set the cycle to N seconds.
Want to change the button brightness for soft lockout mode? Go to lockout then click three times to cycle it between low, high, and off.
Want to change between smooth ramp and mode group? Go to a normal illumination mode then click three times.
There are also a variety of other little details changed, like the soft lockout mode also acts as a momentary moon mode so you don’t have to unlock the light to see while digging through a bag or something. And moon should get about 3X to 4X as much runtime, with more stable brightness, due to under-clocking the MCU at the lowest levels.
Er, it uses the same BAT+ contact ring design as the Q8… which means no truly flat-top cells should work. It could perhaps have a plastic ring for some sort of mechanical reverse polarity protection, but it would restrict cell choice even more.
To eliminate the contact ring, it’d need an almost complete redesign.
I was only referencing the fact that mechanical polarity protection does require the use of button top cells and hoping to be able to use flat top cells.
I was not familiar with the similarities between this driver design and the Q8 design and given the success and design work already put in, and the issues with burnt springs on the Q8 like you linked to Tom, I think it’s a good decision to incorporate something to eliminate the 14 potentially hazardous situations.
I’ve heard that people occasionally put batteries in backward. I don’t really understand it, but I guess it happens sometimes. Fortunately, the Q8 already has a way to deal with potentially-hazardous situations. When a cell was inserted backward, the springs worked exactly as designed and acted as a fuse to prevent any serious issues.
Maybe a thin plastic ring could be added for additional protection. Maybe not. Not sure if it’ll be a feasible option. One was part of the Q8 design for a while, but then it didn’t happen. It’d be nice, but not having it isn’t a huge loss.
Not a huge loss on a lantern, unless you actually were relying on it to be a lantern, for light, in the outdoors, and to “just work”, with maybe little or no backup. Rather than a plaything. Better carry a few candles too.
Could be quite a big loss, if the springs are bypassed on a tuned-up Q8. We haven’t seen that, yet.
Surely this is a fundamental for any lantern to be taken seriously, i.e. ruggedness, reliability and fool-proofness, in all hands, young to old, not necessarily indoctrinated in arcane LiIOn procedures. I’m not sure I am hearing this loud and clear.
Just a worrier, who sometimes doesn’t get cells in the right way around, an imperfect human.
A worrier paints a vivid picture which sticks easily in the minds of people who see it, often overriding a more accurate map of reality in which the meaningful risks lie elsewhere. People spend inordinate amounts of time and effort worrying about one-in-a-million risks while ignoring things which are actually likely to happen. This has a cost, and can be contagious. On an individual basis it’s a relatively small opportunity cost, but when repeated on a larger societal scale it has bigger and more meaningful consequences. So I try not to encourage people to focus on unlikely risks.
Several companies have been using the contact-ring design for several years with a near-zero rate of incidents, so I expect the risks involved are very small. This is probably because it’s common knowledge that batteries have a right way and a wrong way; it is not an arcane or chemistry-specific procedure.
If you know you are personally at a high risk though, it’s fairly easy to prevent issues by applying a bit of non-conductive tape or shrink wrap. For example, a bit of tape over the outer edge of the contact ring, or an extra layer of tape or wrap on the batteries to recess the negative pole. I’ve done something like this with good results — my bench power supply is on a wire shelf so I wrapped nearby surfaces with kapton tape to avoid accidental shorts where I do electrical testing. And, after burning myself one day during a test, I make sure my clip leads are shrouded.
Anyway, if it’s feasible to do so, I’d like to add a plastic layer on the edges of the contact ring to improve fool-proofness. Even a small risk is worth avoiding if there is no significant cost in doing so. But given how rarely it has been an issue in past products, I don’t think it’d be a big problem if it didn’t happen. Other design details have a higher priority.
Then we are agreed. By the way, the Q8 brass ring is different, it is small in diameter, and only just meets the centre of the cell buttons, so they have a natural tendency to tilt when put in the wrong way around, so are more likely to make contact when reversed.
Traditional SRKs with no brass ring, just a flat driver surface, or other designs with brass rings that have better contact with the buttons, are less susceptible. For these, decent wrappers on the negative of the cells are probably enough to prevent contact when reversed.
Not necessarily so on the Q8, and it cannot be denied that it has happened, at least twice, and I am sure will again.
I will be performing some tests eventually soon (if the weather don’t get to bad when i get it ready to test) (outdoors under controlled conditions of exactly those scenarios. ( 1 cell in reverse against the other three and full connection made. ) and other tests like a depleted cell thrown in the 4-paralell config with three fully charged. ( if any cells survive the first tests) I will build a jig to hold the cells with 10-gauge solid copper wire 8 feet to a 50 amp switch)
Because the lantern will be of a lower amps draw/load, maybe even four fusible links between the four negative traces on the bottom cap PCB is possible? ( 5 or 10 amp flat SMD fuses are cheap, and if one cell manages to contact then the fuse for that leg can blow to open the short.
I do like this plastic riser to prevent reversed cells from contacting though, especially if one cell gets inverted accidentally at night. >> Q8 modding - #1202 by 5ar