Well for one thing, I'm pretty sure the switch PCB has been redesigned to have one 30K resistor per LED, and they are going back to the SMD green LED's they used in the prototype run. This should fix the unbalanced problem, and hopefully the less common flickering/failing problem.
I also know they addressed the driver screw issue. I think they went back to M2.5 panhead screws, but not totally certain of that. I'm hoping they also are addressing the quality issues in assembly - glue/loctite or whatever stuff they used on the tail PCB screws on some lights, plus the overtightening stripped screw head problems.
Basically we related all these issues to them and they seemed sincerely interested in correcting them.
Oh, yeah, the problem with not bypassing the LED first is that the voltage drops along the driver components (resitor, traces) should be very small while the LED is in the circuit. If the LED is bypassed, those get larger which just makes them easier to diagnose. The current will still be very safe with the indicator LED shorted, since there's still a 15K resistor in series. But the risk is not mine of course.
I did the measurements. So did Tom E IIRC. See about 10 pages back. It’s pretty clear that mine had current leaking around the LED. What’s not obvious is if that leakage was in the LED package/diode itself, through the PCB, through some dirt on the PCB, etc. No amount of poking with a DMM would have told me that. We can guess that it was heat sensitive as for both of us, poking around with the soldering iron fixed the problem temporarily. Doesn’t tell us much though.
Ok, so the voltage drop accros the resistor presumably told you there's current flowing and how much, and there are presumably no voltage drops anywhere else where there shouldn't be, just across the resistor and the LED pins. So the circuit is capable of delivering current and the problem is somewhere between (electrically between) the two LED pads. So you say it's leakage. That means the current is higher than you expect, or higher than for working lights? Of course the difference might be little, I understand, but there was actual evidence of excess current?
If so, ok, there could be shorting on the board across the LED I guess, but it seems the LED is still the primary suspect. Anyway if it's a trace problem it should be narrowed down to traces between the resistor and LED ground, not the trace from the mcu to the resistor for instance, and from what you say it sounds like that is narrowed down.
So whether it's the LED or the dirty state of the LED or its poor soldering, that's still fixable without altering the board. The other possibility is the board is leaky, but just there near the LED. It seems a little too contrived to be true, but only a little. Using two resistors will prevent one leaky led(or its surroundings) from shorting another of course. It won't make the one led(or its surroundings) not leaky, unless just by accidental reconfiguration of the board.
I think if they can't find more consistent LED's or clean up their pads, they should yes use two resistors (anyway, just because it's better in multiple ways) but they should also decrease the resistance, and if needed just use a darker cover to compensate. 200 or 300uA and moon mode that only goes for 6 months instead of 2 years is a fine price to pay to make production stable.
Funny all this homage talk. I never liked the SRK lights and never bought one. I DID like the M6 when I modified a friends and that led me to getting my own. It’s the one now making 8866 lumens with SST-40’s. That friends second M6 is on the way home to him now, making over 13,000 lumens.
I DID buy 3 of the Q8’s, one for me and two for family members. But I still don’t care much for the SRK style and I bought it largely for the BLF influence… I like it when several ideas merge from this “family” of flashaholics.
It may be worth mentioning that the issue was fixed last month, and very few people in this thread care about the details beyond that.
If it’s something anyone would like to investigate in detail, perhaps they could get someone with an affected Q8 to send it in exchange for a fixed Q8? It’d be good for satisfying curiosity, is a good dev host, and is also, um, pretty nice for its intended purposes as a flashlight.
I done some contact cleaning and opening screw holes in driver. And soldered the spring board contacts a little and now I gained 300 lumens. Now it has 5400. And it has 50800 Cd. Today i will do the spring bypass with wires.
I don’t bypassed springs yet.
Spring bypass can crack if not done carafully. It is time consuming precise thing and for this light you need to do 4 times. It will raise the time of manufacturing per light and that equals raising cost. It is much faster and safer and cheaper to make double springs. If some want the last 5-600 lumens they can do it themselves.
There have been a few flashlights with factory-made spring by-passes, so we could have asked it for the Q8 and we indeed debated about it.
Especially I was against it because the springs form a proven safety feature in case of shorting batteries when one is inserted the wrong way: lots of smoke, spring is toast but electrical contact is broken very fast by the retracting spring: the battery and flashlight owner lives! It will happen and happened before. With bypassed springs a battery short will continu longer, eventually a trace on the tailboard will burn through or the internal fuse inside the battery will blow (quality batteries all have that irreversable fuse), but the effects will be nastier because much more power is released.
As the light comes stock now, the output/efficiency/heat management of the Q8 is well-balanced, while the availability of 20% more light output (at cost of some efficiency and loss of some thermal stability) is coupled to the owner’s skill of performing spring by-passes and the owners awareness that a safety feature is lost.
Yes TK this may take 4-8 minutes depending on where one does it.
Do not set it too low, really don’t for you will think there is something wrong not being able to get to turbo or see it step down very fast.
Best setup for me is default 3 minute stepdown summer and now that it starts to get colder I plan on making that 5. When we get a cold winter with actual freezing I will make it longer.
Easy to do, easy to bump up the brightness when needed and not have it step down automatic.
This has been done to death, investigated, decisions made, and re-designed. Some time back.
Flintrock, yes we are told there will be two resistors.
DEL has shown them on his schematic. Value 33K each. If that value is being fitted by Thorfire that is equivalent to using one 16k5 resistor.
We think they will be using the same diodes as the prototypes. Not the ones used in the first batch, which were different.
So it will be a new PCB, new LEDS, slightly higher equivalent resistance (but not necessarily lower current, depends on the LEDs). All changed.
The new boards have been made, presumably samples carefully assessed, they are on the way to us in the second batch. Soon enough we will begin to hear reports from customers.
Frankly, this sort of speculation and second-guessing the designers is just noise at the moment. Your “should”s can be satisfied by your own subsequent personal modifications, should they seem to be necessary to you.
I bought a tube for $12.50 shipped from Ebay, will receive it on Thursday