I did read the OP, of course, but I got side-tracked by the last line:
…
I still look for “off-the-shelf” solutions first, which means (assuming Modes are not the problem, only stupid blinky modes — I concur wholeheartedly with this philosophy, BTW)
Under $2 for this one
Nearly $3 for this one or
Just over $3 for the classic 105c.
It’s hard to compete with those prices, in my neck of the woods…
(not shilling for FT, just running late for work & needed a quick reference)
I can’t see the sim (no Java). Maybe H-Man can post a diagram / schematic?
Dimbo The Blinky, it sounds like the component count here is very low. Passive components like regular diodes (not Schottky or other special types), resistors, etc are very, very inexpensive. Much less expensive than the prices you are discussing in post #4. Doing a project like this might be 1/20th the cost, assuming you get volume discounts (100pc orders). Many electronics hobbyists may have the appropriate components on hand already from “stocking up” type orders.
I’m interested in this as an exercise / from an academic perspective. I’d rather build a replacement driver of course.
http://devjoe.com/?p=397 That is the mode IC I’m thinking of, but anything with a capacitor used like C1 should be able to use this circuit with some modification.
C1 is the normal mode memory cap on drivers.
D1 is the normal memory mode diode on drivers.
C1 needs to have a high enough capacitance that PWM doesn’t brown out the IC, but low enough capacitance that the off phase of strobe can cycle it. Easier to just pencil over C1 to make it discharge quick enough. (R1 represents this penciling in, adjust C1 or R1 to adjust how long of a brownout the driver needs to cycle through modes or reset to the first mode.)
D3 keeps the normal mode memory behavior. C3 is the mode memory cap, too low and it is hard to cycle through modes.
Q1 is a PNP transitor that cuts the mode IC off from C3 to simulate the mode cycling brownouts.
C2 controls how long the stobe/SOS triggered brownout is. Higher value means longer brownout.
R1 controls how long the strobe/SOS triggered brownout is. Higher value means longer brownout.
Updated sim reflecting these changes: Circuit Simulator Applet–1–0.5183103460024103–0.524256283640264+100.0%0Aw+272+256+272+288+0%0Aw+304+320+352+320+0%0Aw+400+320+352+320+0%0Aw+352+320+352+288+0%0Aw+352+288+352+256+0%0Ar+272+288+352+288+0+100000.0%0Ac+352+256+272+256+2+1.0E-8–2.582957651721599%0Ac+352+208+384+112+2+1.0E-6+0.6988929159787503%0Ar+352+208+400+208+0+50000.0%0Av+144+320+144+208+0+0+40.0+3.7+0.0+0.0+0.5%0As+304+320+224+320+0+0+false%0Ad+272+224+272+256+1+0.805904783%0Aw+400+208+400+320+0%0Ar+144+208+144+112+0+0.1%0Ac+208+224+208+272+0+4.9999999999999996E-6+3.1307086102016015%0Aw+208+272+208+304+0%0Aw+352+304+352+320+0%0Aw+224+112+176+112+0%0Aw+176+320+224+320+0%0Ad+144+208+208+208+1+0.805904783%0Aw+208+224+208+208+0%0Aw+208+304+352+304+0%0Aw+144+320+176+320+0%0Aw+176+112+144+112+0%0Aw+352+208+304+208+0%0Aw+208+208+208+192+0%0Aw+208+192+272+192+0%0Ao+8+64+0+35+5.0+9.765625E-5+0–1%0Ao+10+64+0+35+5.0+9.765625E-5+1–1%0Ao+8+64+0+35+5.0+9.765625E-5+2+–1%0A
Modding an Attiny13A and a mosfet into the circuit to replace the mode IC would be less complex and cheaper if you count soldering iron caused burns.
If the mode IC controls another component that in turn controls the LED, the pin that connects to the LED would be the equivalent to LX.
Err, either I’m misunderstanding or this sim is really complicating things. There’s no such thing as “the normal memory mode diode on drivers” for example, and certainly there is not normally more than 1x mode capacitor involved.
Can you post a schematic of the stock driver and then a schematic of the driver with your added components?
The stock part of the driver.
The one remaining capacitor is what I mean by a mode memory capacitor. The stock Ultrafire F13 driver (see post 22, capacitor is circled with green line and has a resistor stacked on top of it: Group Buy for the GearBest Ultrafire F13 - Closed? ) was what made me think up this circuit but I used this schematic for the stock circuit since it is stupid simple to start with:
All pretty normal looking. (D1 appears to be reverse polarity protection. I doubt that it’s normally included at all?)
How about your modified schematic?
Note I’m going to misuse terms here, I guarantee it.
The point of the modded circuit is that when the LX pin goes high, the mode IC (U1) gets power cycled to reset it to the first mode (U1 is assumed to be a mode IC that starts out in high mode).
When the base on the PNP transistor (Q1) goes high, Q1 cuts power to the mode IC U1. Q1 would cut the power to U1 and not resume it without C2.
The capacitor C2 will act like a short at first when LX pin goes high, thus pulling the base of Q1 up but as the capacitor charges it stops acting like a short so R1 can pull the Q1 base back down to resume power to U1 so that the flashlight will be in the initial mode without the user having to power cycle the light past the strobe/SOS modes.
D1 and C1 allow U1 to ride out short power cuts that would be caused by PWM so that modes other than high still work.
D3 and C3 preserve normal mode switching behavior by supplying power Q1 during the mode changing press of the switch so that U1 doesn’t get reset back to high when someone tries to put the light into the next mode. D3 keeps C3 from being discharged during a mode change. D3 and C3 needed because otherwise mode switching timeout would need to be shorter than one off period of the strobe for this circuit to work.
Shoulda bought a single mode clone for $2.99 from Banggood, unless you're just enjoying the exercise.
I understand your description better now. Sounds like a nice solution (assuming you are locked in a room with this modes chip and no other way to get modes and want to build a flashlight).
At first it seemed like you had way too many components, but your explanation helps show why you need them all. 
I’m actually surprised this circuit didn’t exist yet with the number of tact-i-cool flashlights being made and the number of times people have complained about the blink modes.
I’m trying to make a gateway drug for flashlight modders. The point of it is that all of these parts can be scavenged from consumer electronics, if the mod kills the driver nothing of importance is lost, and it is possible do it on impulse.
If I wanted a single mode, I could just remove the capacitors and it would reset to high every time. Option B is to just solder the LED - wire to the grounding ring and outright bypass the mode ic. The point of the circuit is to keep the useful modes while getting rid of the ones that aren’t useful. Single mode doesn’t work well when you need long runtimes/low output (say hatlamp duty while working on something close up) and periods of high output out of the same light.
Sorry, I thought you wanted single mode from the page you linked in post 6:
The Issue
This flashlight has one very poorly engineered feature and that is its modes.
It has three modes: Bright, Dim and Strobe, but to switch modes the light has to be turned off, and back on. This is a huge inconvenience when you switch it off in the Bright mode, as on the next power cycle it will be in Dim mode. You have to cycle modes to get it back to normal again.
This frustrated me, and I had no need for the other modes anyway, only full brightness. Time to fix it!
Are you saying that you expect this to work with an extremely wide range of component values? That does potentially make it more useful to the neophyte modder. I definitely like things that tempt new modders.
You’re aware of the common “modes IC” controlled drivers which allow the changing of mode groups through the soldering of jumpers, right? I’m not sure what percentage of the market is those chips and what percentage is the static (no options) 3-pin stuff. I suspect that the majority of the worst (no options) modes IC’s are placed in $3 SK68 & similar flashlights.
I know that drivers like the nanjg 105c exist, but I doubt that anyone starts by modding stuff they would actually miss if they killed it. I started by doing resistance mods to springs on a 3xAAA flashlight driving a K2 TFFC emitter without even a resistor.
I have a nanjg 105c sitting on my desk right now but the issue is that even if I order from IS, I still need a credit card and the willingness to wait a week. The under 18 demographic is significantly disadvantaged when it comes to having access to a credit card. I bought a two pack of flashlights from lowes with a H-L-Strobe driver in each for $12. The 2AA one has an XP-G2, the AA one has an XP-E2. Both had a decent tint for a CW LED. The driver and MPCB on the AA one are pressed into the hollow pill. The driver on the 2AA one is pressed into the pill but the MPCB has pill under it. Both drivers are held in by screw in retaining rings that screw into the pill, the pill screws into the body. Both pills and drivers are big enough to solder with a cheap soldering iron. The 2AA light has some semblance of regulation because it draws 1.7A on a single eneloop but only .8 amps off of 2 eneloops (and the same with a 3rd cell). The switches are pressed in, but the lens and reflector are held in by a screw in ring. The reflector has a clear plastic centering ring.
Since the lights are from a brick and mortar store, you can buy a bunch of them and only keep the ones that win the tint lottery, returning the others.
I’ll have to mod a cheap driver to test what values work for my circuit, but with resistance you can just adjust it with a pencil.
EDIT: Remembered the other reason for this: I can never seem to find a good boost driver for cheap.
We may eventually be able to help you out with that bit… [WIP] 15mm PAM2803 w/ ATtiny13A rough layout / possibility
I have disabilities that get in the way of soldering SMT. My hands will have 2-5mm of jitter when trying to solder, hence why I have a preference for through board components when possible.
That sucks! I assume that amount of jitter prevents you from doing most fine soldering? Including the type of intricate airwiring require to add a resistor, two caps, a transistor, and a diode to a small 15mm driver, regardless of whether they are through-hole or SMT? That is a lot of components to add to such a small thing.
I have workarounds, they just take a lot of time. I can make joints on small stuff, my issue is that I can’t avoid shorting pads on a PCB nearly as easy. My plan was to airwire the parts and then airwire them into the circuit. There is only one spot where I actually have to solder onto something other than a ground point, the battery positive spring, or the LED- so there is some space for forgiveness here. MK1 of the circuit will be on a breadboard. Through hole stuff gives me more space to take a razor saw to any shorted leads. I suspect that half my issue is that my soldering iron is bad. We are talking 10 year old tip, 25 watt iron that has faded to 15 watt, and mixed solders. Some of the solder I’ve had to use for the big stuff is so aggressive that it can eat a hole in the tip (I’m nearing the point where I just throw everything away and start over, it is that bad).
I have a newer variable power iron NIB, but I can’t seem to find any reliable guides and I refuse to ruin yet another tip by using it wrong. I know the iron is supposed to heat the items intended to be soldered together, but I at least have some decent solder.
One of the items on my todo list is to rig up an exhaust fan so that I can solder at a desk that is at the right height so that I can neutralize some of the jitter.
Standard 60/40 rosin core solder is what you want. Your local Radio Shack of Fry’s Electronics will carry it. There should be no reason to use acid-core. A wide range of temps will work. Is the new iron variable wattage or temp controlled? A temp controlled iron will tremendously increase your mileage on the tip.
Have you considered reflow soldering? I’m not sure it’s a great suggestion in this case (it doesn’t really sound like a great one), but you know your limitations and are a better judge. In that case you take a blank PCB and lay down solder paste on it, some members such as DBCstm normally use a stencil. Then you place components on the paste (this is the part I suspect might be trouble for you). If you make a mistake you can wipe the paste off and start over… but it just doesn’t sound like a practical solution here. I thought I’d better mention it just in case. You also don’t have any fumes to worry about so you can do it at any height or whatever, then carefully move the PCB to your reflow area.
Please keep us posted as you build MK1 on the breadboard. 
New iron is constant power. I’m considering selling it on craigslist and getting this: Radio Control Planes, Drones, Cars, FPV, Quadcopters and more - Hobbyking
Even though the large diameter solder eats away at the tip, it is labeled rosin core. I’m using almost nothing but 62/36/2 right now because that is the only solder I have that doesn’t eat at the tip (bought it for using with a cold heat iron, discovered that running NiMH cells in a cold heat causes it to melt from excessive power, discovered that it makes for extremely good joints with a regular iron.)
No reflow access. I used to have access to two hot air desoldering/rework stations, but both places went under.
I need a sharper tip I think.