Correct. Simply remove the driver’s screws and gently flip the board over.
Start here at Hoop’s how-to. This offers an overview of the flashing process and the hardware needed.
Hoop’s thread has more links at the top of the OP, including this page at the Flashlight Wiki.
The first Link in TK’s signature is also recommended reading material.
I just got started with reflashing lights. Expect some trial-and-error with getting your hardware working properly. The rest is pretty easy!
Yes, very easy to upgrade firmware - still need the clip and USBASP dongle. NarsilM is set up to make config changes with minor changes to settings in the code - don't have to write code, just change values, etc.
USBASP can be bought cheap, and the clip and wire - all depends.. Some cheap under $5, but you have to make up the wire yourself - don't know of anything bought off the shelf that just plugs in, unless a BLFer is offering that for sale.
Or there's other firmware options as well - the D4 version, or TK's new e-switch version,etc.
I ripped one wire from the switchboard out when I cleaned the screw holes. I resolderd it.
For my lamp I have to use longer wires, if I want to use a programmer clip. The LED wires are fine.
1st battery spring bypass. Sadly I didn’t have a measurement before bypassing and cleaning the shelf but it looks a little bit brighter.
The residue you see is flux that got cleaned off before assembly.
And “a little bit brighter” at around 5000lm it a lot 
Tailboard with be-cu outer springs installed. Original springs were retained and be-cu added outside. Tail bump significantly improved due to triple springs.
After adding the be-cu springs this light gets hot fast. One minute at highest output and the head is almost too hot to touch.
please tell me i can make simple change to ramp timing(the time from min to max)?
Yes please I’ve got all the kit except for the clip (clues to the correct one for this MCU please), being undecided when you were still bending MCU pins to fit big ones on smaller footprints.
Way back when I might have suggested header pins on the driver, negating need for a clip, just pop on the connecting cables directly. Still think its not a bad idea, plenty of space, but here we are. Just holes to put them in afterwards would have been nice.
A sub-4Hz flashy at maybe 500 lumens, quickly accessible in any mode, then back, would be very useful to me.
SOIC8 clip.
Search everyone’s favorite online auction site, or your own favorite online electronics retailer.
FWIW, I bought a cheap clip despite negative reviews about difficulties grasping an IC. Mine seems to work fine, but I’ve only connected it a dozen times or so.
Well... What exactly do you want to change about it? 150 levels, 16 msecs per level, 2.4 secs from min to max, plus pause time at moon if ramping up from OFF.
Say for example you want it to be done in 1.6 secs, then easiest thing to do is use a 100 level table, use TK's Python script to generate it (I have the command line params documented in the source code where the tables are), and just need to set the turbo stepdown settings accordingly. No code needs to be changed, just a table replaced and couple constants tweaked.
zeroflow - only 5,000 lumens from that mod? Hhmm, should be better I'd think, over 6,000 on good cells.
tekwyzrd - those springs look real nice! Yea, would be nice if they still could be found.
I had that question also lately is it possible to change the time between the levels?
@TomE: I didn’t measure it yet. 5000 was just a value to compare the term “a little bit brighter”. 100lm and “a little bit” is not much. But 5k lm needs a fairly big bump to look brighter to the average human.
It’s now about in line with a M43 Meteor.
It runs with Panasonic GA cells (VTC6 on the way) and the bump is clearly noticeable.
What Tom Tom said:
And, ah, you need a computer to do this flashing, right?
What else? Specifics most welcome.
Yeah, you’ll need a computer. 
I followed advice at Guide: how to flash ATtiny13a based drivers (NANJG, QLITE, etc.) with custom firmware and bought the Fasttech UsbAsp and ribbon cables, but held off on the clip, but have not yet tried to put them to use, realising that I could spend quite some time coming up to speed.
Edit: and that’s before trying to tinker with the code, being very much a hardware engineer, used to working with highly skilled software engineers to do the clever stuff 
I cut a piece of 12 awg copper residential wire, formed a circle, and fit it into the head. It’s pressed into the space after the threads and acts as a snap ring to help keep the driver secure. I used a round jewelers file to enlarge the driver screw holes. Good contact with the tube and driver, o-ring covered by the head, and as luck would have it the flats on the battery tube are perfectly aligned with the top and bottom of the head. With the triple springs and the copper ring tail impact is still much improved and vigorous shaking doesn’t cause a power interruption or battery rattling.
> computer
Virtual machine Windows 7.
OK, can do.
I had to replace the springs because melted two of them. so decided to check what I had in stock. Found a bunch of IOS "A" size springs. I see Hank doesn't sell them anymore, but they are a gold plated 11 mm height, stiff spring with 0.9 mm diameter wires, about a 8 mm wide base. The base diameter works out well, giving you enough space to add solder to the pad just outside the spring, so easier to work with than the stock springs, and I'd say this single spring is much stiffer than the double stock springs.
I was surprised I got a nice bump with them: 6,100 lumens at 30 secs,without any bypass wiring. Sure I could probably get 400-500 lumens more with bypasses, but previously this Q8 measured 5,400 lumens stock. Shame you can't buy them anymore, but these at KD seems about the same: http://kaidomain.com/p/S027055.DIY-Gold-Plated-Battery-Driver-Contact-Support-Springs-8mm-x-11mm-for-Flashlights-5pcs, but they are spec'd at 0.5 mm wire which is a lot thinner then they appear to be in the picture - dunno...
What happened? I thought you had bypassed all of your Q8s’ springs.
Not all. I was testing something, put 2 GA cells into a Q8, one upside down, and poof - no light then a small stream of smoke arose... The 2 springs melted down, no damage to the light or cells, just the springs. So thought I'd try replacing the springs and see how ti does without bypasses. Thinking of giving this to my daughter and son-n-law. They got a 4 month old, so the spring fuses may be a good idea for them - both not that into tech stuff.
With IOS effectively gone for DIY parts, MtnE limited offerings, CNQ gone, there's not much left for quality parts choices, but KD seems to have a good choice of springs, including silver coated phosphor bronze.
Whew, I’m glad to hear it was an experiment.
Does meltdown occur immediately after closing the light, or when the switch is pressed?
I’ll be providing protected GAs with my friends’ Q8s. What would happen if they insert one or more upside down?