The driver screws have a biger impact on the Total Amps that the Q8 draws than i thought. I was checking some things and removed the driver to open up the screw holes a bit so that the screws fit nice. After that i put all back together and measured the Amps. I got ~15A. Not good. So unscrew the driver checked for dirt or something that was preventing a good contact. But nothing. This time i used panhead screws so that i can put a bit more preassure on the driver to be sure to make good contact and what said the meter? 21.8A (the cam was to slow picture shows only 21.6A)
I think i will remove the anodisation under the driver (where it makes contact with the outer ring from the driver) when my new thinheadscrews arrive. Maybe there is a little bit left for optimisation.
And :+1: for mod ledboards in XM and in Luxeon V size!
And it would be nice to be able to buy a clean switchboard. Maybe with pads for 3x0805 size leds and each with a (if space is to small) 0603 size resistors.
I imagine the driver has more contact with the tube when it is sits properly. Like the idea about removing the anno on the driver seating ledge.
I think my integrating tube is reading a bit high. With tail springs bypassed to the mounting holes (So basically bottom of cells connected to bottom of the battery tube via wire), I measured 7K lumens at startup using fairly new fully charged 30Q's flat tops, no solder blob. I would expect less output due to voltage sag. I only used 22 gauge wire because I had a scrap of it needing to be used and I figured the dual springs will already move current nicely.
Even if the measurement is high, this light clearly delivers the goods, both in the tube and outside. Next time I measure, will try to measure current at the same time.
I didn't measure out 30 seconds, but the output drop was slow and steady. So it appeared the drop was due to the cells discharging and not due to thermal sag.
Oh, swapped in one of those RBG slow flash emitters and a 56K resistor in the switch. I also swapped out the other green emitter for one of my green smd's for good measure. Has a nice effect.
I have built an M6 using a new battery tube made by MRsDNF for this purpose, it utilizes a TK61/TK75 battery carrier with an adapter to allow contact with the brass ring. This makes 2S2P really easy, and of course would allow a 4S carrier just as well. This particular M6 built with 3 70.2 emitters is making 13,558 lumens.
I don’t see why doing similar with the Q8 would be all that difficult, and with 4 emitters the total output should be better. Even easy to use 4 XHP-35’s if desired. I would use Richards driver for the 12V emitter, probably Richards driver for a 2S2P as well.
In my own M6, using a single 17mm FET with TK’s ramping firmware, I’m seeing 28.74A to 3 of the new SST-40s for 8866 lumens. This one has a large Omten clicky installed in the tail for cut-off, took 2 18ga loops to get the light to work with these SST-40’s due to the high amp draw.
The carrier modified M6 is manxbuggy1’s, pics here show the set-up
The way this was made by Steve, the adapter sat loosely inside and would fall out when removing the carrier. Manxbuggy1 had some issues with this so I made a retaining ring to cover it that is captured by the battery tube, it no longer can fall out and gets securely pressed onto the driver by simply screwing the battery tube down on it. I also modified the carrier with spring bypasses but did not remove the original brass supply lines (rods) as I have done previously. In this case I opted to not use 18ga Turnigy wires to replace these brass rods. Output is sufficient without radical overheating of the M6 head. Pretty confident the Q8 would fare well done similarly.
Am I missing something? I thought the negative contact is made by the battery tube pressing onto the backside (facing the batteries) of the driver. How can the screws influence the current level? Or are you talking about the rear PCB?
Tom, remember the 2 lights I got from you for my B-i-L and S-i-L? I simply soldered a copper disc onto the backside of the spring board up top and gained considerable output on each of those. Both those guys LOVE em! Your Narsil in each, they’re very happy. (didn’t wait for the Q8, they’ve been giggling over those for months now)
That’s probably the first thing I’ll do to my Q8, solder a thick copper plate onto the spring pcb for super current carrying ability, then test it to see the result. Then bypass springs. Maybe then replace the brass contact ring on the driver with a custom cut copper one. Maybe even work a copper heat sink into the overall design. And probably then go with SST-40’s. Possibly 2S2P 70.2’s and a carrier like above. The plan though is to make incremental steps and see what the output differences are… we’ll see how much of this I remember when it comes time to play…
FlashyMike, being able to press the driver into ground in the head facilitates current flow through the tube. Ground is of first importance, is also usually the cause for issues when it’s weak.
For those of you trying to remove the hold-down ring around switches, the shape of the semicircle cutouts is perfect to fit these in:
This will not work for large-diameter heads though, since they can only open so much.
What I have used to open large rings in the past is made a custom tool for the job, just a simple pipe with two screws through it at the correct distance.
Sorry, I don’t get this. The LEDs are connected with 2 wires, and the batteries through the backside of the driver. Positive at the inner brass ring and negative at the outer copper trace. Of course the outer ring needs a good contact to the battery tube, but how do the screws help here? They are just pressing the driver against the front part of the light, where no ground connection is needed.
But perhaps I just dont’t see an important point …
Ok, quickly, if the screws are threading into the PCB of the driver they are quite likely holding the driver up off the shelf of the head ever so slightly. This means the battery tube cannot press firmly against the entire ring, not as well as if the driver is firmly seated on that shelf. The shelf is non anodized for a reason, to allow ground contact through the head and to be nice and flat for full contact. So, with the hole enlarged slightly in the driver board so the screw can pass through without grabbing the pcb, it presses the driver firmly down, allowing the battery tube to make a good solid earth on the ground ring of the driver.
Oh boy, Yea, I need help understanding these 2 things - where exactly the copper rings were soldered - very interested in that. Think the copper disc simply connects all the traces together?
I'm also unclear with the driver mount issue. I would think all that matters is the connection of the battery tube to the driver - that's it. Contact of the driver to the head housing is not needed - no reason for that connection - you need a strong path of Batt- to the driver.
Dig in, Dale!
Just to get this clear, since I don’t have my lights yet and can’t check it by myself - does the outer edge of the battery tube rest on the drivers backside or is it’s diameter larger than driver and it rests on the shelf? If the first is true I still don’t see why any ground contact to the front casing is necessary. The light should even run isolated from the front case and without screws.
If you screw the battery tube in it makes contact with the driver outer ring but after screwing it in a few times, i can see that it did not make contact on the full surface. But the tube is also connected with the head through the threads so the current will flow partially from tube direct to the driver and through the threads to the head and then to the drivers other side groundring. I thought about this too and that is the only explanation i can think of. I think reducing the contact resistance in the light will be even more important when a LED change to SST-40 or Luxeon V leds will happen (because the Amp draw will get up to ~ double what it is now).
Ok, I see your point. If the tube doesn’t make good contact to the driver this will influence current draw. In the pictures of Tom E’s prototype tests the threads of tube and head look anodized - has that been changed?
Edit:
is it possible that the head of your screws is so big that it touches the batterie tube?
Threads on mine are anodized, and a bit goopy with lubricant.
The end of the battery tube is bare metal, and feels a bit rough to the fingertip (sawed off and not polished?)
Looking at the ground ring around the outside of the driver, it has some scuff marks around about a third of the circumference. The rest is pristine shiny.
One of the scuff marks is interrupted next to the screw head, suggesting the screw pulls the ground ring a bit out of plane so it’s not rubbing right there.
Are we wearing away at the ground ring trace on the driver every time we change batteries?
How thick is that trace?
Would adding a dab of conductive lubricant around the ground ring contact protect that trace?
Or polishing the end of the battery tube a bit?
I deeply hope we haven’t found another potential design flaw, I have seen this earlier in cheap SRK clones. Bad contact between battery tube and driver board. Last time I simple beefed up the trace with solder, since I didn’t have anything better at hand that time. Worked out better than expected but of course a brass or copper ring would be far better.
Hank is onto it…. and yes, a good finish on the end of the battery tube will allow for the most contact while protecting the trace. A rough end on the tube will be an abrasive to that trace, while also providing the peaks and valleys that will undermine a solid ground contact. And that 1/3 contact, all the current flows through there so if you can get the driver to seat firmly on the head’s shelf then the battery tube should also fit squarely onto the driver contact ring and provide the optimum ground path.
That’s the biggest trick to getting the most out of a light, making all contact points optimized to cut resistance everywhere possible, or at least in my experience. I’ve even gone so far as to solder thin copper discs on top of the springs so they make a large contact to the negative end of the cell. The little things, added up, yield the biggest gains.
Designers need to keep a thorough list of all the possible bad designs that have been used in the past on any similar product.
There has probably been a financial reason for every corner that’s ever been cut — watch out for those choices to happen again.
Does anyone build driver PCBs with a solid copper ring instead of a trace around the edge of the board?
When all that’s done, stick an UCLp in place of the glass lens and measure output again. I remember seeing a 17Kcd gain in an Maxtoch SN6x-2s by simply replacing the lens with an UCLp.
Seems the good drivers have thicker copper pours for the traces, like 3oz, or the same as the new Noctigons. If they’d wear out it’d take a lifetime of regular use I’m pretty sure.
