Q8 modding

[quote=joechina]

[quote=Tom Tom]

Excellent.

I’ll be ordering a set myself anyway, for other applications.

I'm sure someone has answered this, but I couldn't find anything in my scanning of this thread. Is there any problem with boring the driver mounting holes enough to clear the threads and then adding a counter sink for the screw head to fit flat?

I haven't look at the driver pics to determine if the screw holes are the only thru-holes for Batt- connection, but that seems unlikely.

^

Yep, that is what I'm talking about. Does that cause any problems? It all looks good from here. Plenty of thru-holes and no shorted or broken traces appear evident.

All fine

Indicator resistor change to 22K. It’s too bright for me still. :person_facepalming:

^

Great. I should get mine today and would like to do that to make sure the driver is nicely seated. Thank you steel_1024 :)

Unless I have missed something, the Batt- connection is entirely through contact between the end of the battery tube and the outer ring on the driver PCB (and all of those vias in it). The need for any contact through the screws or the reverse of the driver and the head is I think a red-herring.

So a DIY fix to countersink the PCB (and maybe a little way into the metal of the head) is a great zero-cost way to improve the mechanical integrity.

But this would not be practical for Thorfire in production. Putting a countersink into the driver PCB as a production step is just not practical.

I am hoping that we can steer Thorfire to do the right thing in future production by using a correct thin pan-head screw, in a correct sized PCB hole.

Good that I have 90 degree countersink drillbits
They are cheap

Like this in different sizes
https://de.aliexpress.com/item/High-Quality-3-Flute-90-Degree-HSS-Countersink-Drill-Bit-Wood-Steel-Chamfer-Cutter-Tool-6/32767774041.html

Which is to say — that’s not the BLF way.

LEDs and drivers keep getting better, and we keep using the new stuff in the old bodies.

Can someone with a Q8 in hand check those screws please? The pictures that I have seen do not look like standard countersink angles. this one in particular:

For anyone who is planning on changing the board to make it fit the screws, be aware that there are different angle countersink bits that are not exactly the same. 82 degree and 90 degree are a couple of standards, but the screws look different from either of those - perhaps due to the small size.

Don’t just assume that they are all the same.

Could we ask Thorfire what defines the screws that they bought for this use?
I notice they have a “service” email address on the packaging for the light.

please ask questions to us, we collect them :slight_smile:

If in doubt, swap it out! I'm not wasting any more time on the stock screws - they be going in the circular file. Way too easy to replace.

The Q8 arrived a few hours ago :partying_face: .
The Spring PCB is already modded :smiley: . Pictures will follow when my cam is charged. Now i have to go modding the other end. New switch LED´s and a new resistor (the LED´s are way to bright for my taste) and then change the XP-L wires with 16AWG.
Only with spring bypass and thicken the traces i got 20,8A out of my VTC6 @4,05V. With one VTC6 i got 15,8A. With two i got the full 20A.
I am curious what i will get after changing the XP-L LED wires?

Are you sure you want more than 20.8 A? Let me calculate.

With 5.2A per led, there’s just 14% possible output gain leftover (the maximum is at about 7.5A). Say you want that maximum output, the ~220 extra lumen costs you 8W extra power. So that last bit has an efficiency of 27lm/W

(my Nitecore P12 at second highest mode at 200 lumen has an efficiency of 112 lumen/W)

90% of the time the light will run in the lower levels. But to be able to get the max possible on turbo is my goal. Would a FET change bring some additional lumens? I have a few SIR800DPleft over from my last order? I think i will get the biggest lumen bump if i change the LED´s with the SST-40. But i will wait with that until the new ledboard is ready. I hope someone will take over for me because i will have to go into rehab in a few days.

I’m really surprised Mcmaster Carr doesn’t have wafer head machine screws. I’ll do some looking around to see if I can find anything good.

You don’t really wanna push them THAT much harder though because they are on the precipice of being overdriven as they are. Look into the Luxeon Vs that djozz did some testing on. They seem like a really good stock replacement candidate apart from being a slightly different die size and finnicky to reflow.

Do consider seriously whether you really want to push the max. LED current way past the reasonable efficiency level (lumens/watt).

What some may not appreciate is that, in all modes where the FET is in use, the instantaneous peak LED current is always at that max. level (simplistic analysis ignoring various second-order interactions but in the ball-park).

So the LED efficiency is always at the compromised over-current level, way past the peak of the curve where efficiency is decent.

Not just at turbo/ “maxx powwer”, but at all lower levels, until only the 7135 is engaged. The dimming for the lower levels is done by (high frequency invisible) PWM, but the peak current remains at the same inefficient level.

That’s not entirely true. A good part of the inefficiency is caused by the increased temperature, which you won’t get at a lower output level. I doubt you can directly use one of the constant current test curves to figure out efficiency once PWM is applied.

When I saw steel_1024 ’s photo in post #121 it did look to me like like a bog standard 45 degree (90 included) head. That is to say I just assumed it was.

Of course, as you said, modifying the PCB to fit the screw is not really the correct approach (i.e., a bodge), even if sure of the match. Another reason to wait, until the way forward is clearer. Meanwhile leaving things as-is is safe, and for many, good enough.

We can be a fussy bunch, and maybe missing the bigger picture, which seems to me to be that this Q8 is turning out to be pretty good. If the worst problem that emerges is a few dicky switch LED boards, I will be very happy.

Here is a link to that photo: