Just a quick question in hopes of understanding this driver. I’m a new owner of the FW3A. I’ve always had kind of a bad perception of pwm, but I don’t really see it with my data logging light meter setup on the FW3A. In theory, you just need one FET to pwm from 0% duty cycle all the way up to 100, yet I know this driver has (I think) seven 7135 current regulator chips. Each 7135 can source 350mA. And they are in parallel, so that means, you get 350mA, 700mA, 1050mA, 1400mA, 1750mA, 2100mA, 2450mA, and then direct drive with the FET (100). So for lower levels than 350mA, you need to pwm the FET? I’m assuming the 3 LED’s are in parallel as in series would require a boost circuit. Did I get that right.
I saw the pinout of the Atmel microcontroller, but didn’t really pay much attention. So I know there are 8 pins, 2 of which are power and ground and other pins used for stuff, that I recognized, but in no way are there enough pins to turn the 7135’s on one at a time. Anyhow, just curious about the hardware workings. I think this driver is pretty typical and not proprietary. What makes it unique in my opinion is the software control and getting the E-switch signal from tail cap to driver.
Some people shorten the term “PWM flicker” to just PWM so people tend to think of it in a negative way. The flicker is caused by low pulse frequencies like 4k to 8k or so Hertz. These lower frequencies can be noticed as flicker in fast moving objects like a fan or running water, etc… Some people are more likely to see it than others. The drivers designed by blf members running Narsil or Anduril use a higher frequency of around 14k or more Hertz which looks very smooth.
Using PWM can be a good thing because it’s more consistent as far as tint shift is concerned. Linear drivers, which don’t use PWM, can have noticable tint shift as the brightness level drops.
The FW3A driver uses 3 channels.
Channel 1 is a single 7135 chip (0.350A)
Channel 2 has 7 x 7135 chips in parallel (2.45A)
Channel 3 has the big FET.
All 3 channels use PWM.
Let’s assume we set the FW3A ramp ceiling to “1” which means it ramps up from moonlight to Turbo. (The default FW3A ramp setting stops when channel 2 is at 100%)
When you start ramping up from moonlight you are using channel 1 and the PWM signal climbs from 1% to 100. At this point the led is drawing .350A and is about 120 lumen or so. (IDK if 1 is enough to make the led glow so it might actually start a bit higher than 1% on this channel)
As you ramp up more, channel 1 stays at 100% and channel 2 starts at 1% and goes up to 100. At this point there are 8 total chips at 100 which is 2.8A and maybe 900 lumen or so.
As you ramp up more, channels 1 and 2 stay at 100% and the FET starts at 1% and ramps up to 99. When the FET goes to 100, channels 1 and 2 turn off. It sounds odd, but you get better performance from only the FET running when at 100%.
It’s a bit more complicated as far as the software coding, but this is the basic outline. I think this driver is based on Texas_Ace’s Texas Avenger line of drivers. You see those a lot in BLF lights. Some designs use 2 channels, some use 3. You can use a single channel as well, but it’s not very common.
this was sitting on a surface head down, I was plugging in a charging cable for my phone, I bumped the light, the tail fell down 2-3” and it stopped working, I unscrewed everything and reassembled, at first nothing, then it started to work again. I wish this light was more robust and not so finicky, it might really get damaged with a short fall on a hard surface, whereas I have dropped many others and just get some chips in the annodizing
So not sure what happened or why it suddenly works again
so far mine’s been very reliable and i have killed many other lights by either dropping or just using
i;ve dropped it many times
i think once it quit but started right back with the switch
(battery may have lost connection )
My snap ring pliers are too chunky too reach down in the head. I got creative and used them with one straight pin and one 90° pin but ended up chewing up the threads a little in the process.
Then I remembered seeing this post about binder clips. I have various sizes and none matched so I squeezed a slightly-too-big one in a vise until it was the right size.
I don’t know if it would’ve worked from the get-go, though; my ring was on tight.
I just used a screwdriver to tighten, but you could make a custom took with a grinder, probably, some T shaped thing with a metal bar that would hit both the notches and could be turned from the top./
In my case it’s hard to say, I was getting the flash when reinstalling the battery after it tipped over, but no response from switch, I can’t say for sure why it suddenly started to work again, If it flakes out again I will try and tighten the retaining rings, I do have a pair of snap-on pliers.
It would be nice if the design was tweaked a bit in coming generations, I really like the beam and power and UI, but think the QC could be improved.
You can get the flash without the E-switch making contact. Once the head starts to be twisted on, you get power to the driver. What you are experiencing is what has happened to me off and on. The problem is that the switch has a hard time making electrical contact all the way from the back to the front where the driver is because everything has to be aligned just right. It happens occasionally just by changing out the battery.
If you think about inserting a battery, the tail and the front are always connected with the battery because there are springs providing tension and large surface areas to make contact. But in the case of this E-switch and driver which are at opposite ends there isn’t anything like a spring providing constant pressure at the contact points other than the tightness of the head and tail cap along with proper alignment of the electrical edge of the inner tube to the rear and front pcbs.
I think it’s pretty clever what they have done to get that switch signal to the front via the inner tube, but everything has to line up correct. So I kind of doubt its a QC issue, but more the nature of the beast.
Jason is right about this: “It is all about the inner tube design and it’s tight tolerances.”
I ended up getting the snap ring pliers from harbor freight but they did fit the fw3t. I still cant get the retaining off :/. I did get the flashlight to turn on by tightening everything down with a bit a force. I have a few question if anybody can help. I noticed when the light ramps up or down its not as smooth as I seen in some videos about the flashlight, it blinks a few throughout the ramping is this normal? The other question is, when turning off the light there is a delay before it turns off is this normal?
From the Illumin site, “RAW ALUMINUM MODELS DELAY, EXPECTED ~9/12”.
From Nealsgadgets, “NOTICE: TI+CU VERSION OUT OF STOCK, ESTIMATE RESTOCK AT SEPTEMBER 10TH. OTHER VERSION IN STOCK. ”.
As I ordered from Neal a month ago. Seems like more waiting is in order. I know both my and Neal’s emails work because Neal sent me coupon codes when I ordered. Just wish Neal would have responded to my request for an update.
My wife says I don’t have enough patience. Watch, I’ll have a box full FW3Xs, on my doorstep, in the morning.
Admittedly I may be worrying over nothing.
[/quote]
I got an update from Neil. My order will be available this week.
I just checked. It seems the retaining ring is too thin and too deep for the plier tips to grab. You would need to grind or file some material off the plier sides so that it can reach.
They work fine on more normal size retaining rings.
It blinks at the transition from channel 1 to 2, from channel 2 to 3 and at the ramp ceiling. These were all features that were voted on.
The turn off delay is also normal. This allows you to do a double click without the light turning off then back on.
So, I think I have a bad driver in one of my FW3A’s. Let me start by saying… I am not a noob when it comes to building and mods.
The issue is that when I ramp up to above the 7135 ranges, sometimes the light just goes out. Not dead, just out because if I ramp down to the 7135 levels it comes back on. It does not do this all of the time, but when it does I can unscrew and screw back on the head and it will work for a bit. This happens when the light is cold or hot. Everything is tight and everything is clean.
As soon as I get to my shop and get some time I am going to pull the driver out and resolder the FET, and maybe just heat the entire driver with hot air soldering station and see if a component is not soldered well.
Has anyone seen this happen and if so, how did you fix it?
Not that I think it’s related but anytime I have to pull a driver (especially FW3A’s) I take a kid sized old toothbrush with alcohol and scrub the whole board before and after I do whatever i’m doing. So many lights that are covered in flux goop, solder balls, alu shavings, grease, you name it.
Yup, been there done that. But this one has never had a button top or a long protected flat top. It really does seem that when the current needs to go up, the light will fail. My concern is that it is not 100% of the time. But I can get it working, put it down and go back later and pick it up and it will not work! Driving me nuts. I have thought about a think layer of solder on the driver ring that makes contact with the center tube, but this light is so sensitive to dimensional tolerance that I really do not want to start down that path.
I had a similar problem (although not exactly the same behavior) that I thought had to be the driver but it ended up being the switch. Trade the switch for one in another FW3A and see if the problem moves with the switch
