CRX CuTi Piston Drive.

Hum, I guess youā€™ll have a hard time to decide that, wonā€™t you? :stuck_out_tongue:

Consistency is surely part of your building ethos :+1:

Very impressive! I really like the color!

mad scientist strikes again!
:beer:

you had me at
ā€œGITD LED centering ringā€

thanks for the gorgeous photos too

but, titanium should not be purple :smiling_imp:

Sweet build :heart_eyes: I like the cf ring, adds a little more character to what would otherwise be a pruple smoothy

Real noice CRX. Love your work. :beer:
Noice is an actual Aussie word. :wink:

Wow, CRX Iā€™m very happy to see you doing this. :slight_smile: I really like the concept of tail e-switch implemented this way and I know you like tail e-switches as well. Youā€™ve done a few before with a separate signal carrier, howā€™s the switch feel compared to them?
May I ask for more pics of the driver? Looks that unlike tterev3 you put the switch and spring side by side.

The tail construction is not clear to me either. I guess you donā€™t have pics of it during construction, do you?

Yeah just trying something different, maybe get more females interested in the hobby too, oooh purple :heart_eyes: :smiley:
Plus 19v was the highest I could go at the time :laughing:

This ano will scratch/ sand off fairly easily, considering a CF shell for it.

I was going to leave the section brass bare but liked the CF idea too :+1:

I believe it to be an English word too :smiley: :+1:

The switch feel is a little different to a standard setup, you can sort of feel that youā€™re pushing the cell onto the switch as opposed to just pressing a tactile switch, a little less crisp due to the involvement of springs.

The difficult considerations about this build are the spring tensions, switch activation point, cell construction and cell compartment length.
I built the light around the vapcell 16340 though the light will accept other cells it affects the button travel, an AWT 16340 for example is a half mil shorter so that travel is reflected in the switch operation.
Only buttontops will work, flat top cells wonā€™t hit the driver centre spring enough.

I took no pics of the build in progress. The switch is sitting across two 7135 chips and potted in place with JB weld.
A hole was drilled through the driver near the ground ring for the switch wire, the other side soldered to the pill.

I did really want to fit a switch in the centre under the spring as it would have shaved off a few mm from the length but the way the Lexel drivers are made with the LED positive wire hole going right through the centre contact pad and the pad itself being quite small there was just no way to do it with what I had on hand.

The tailcap is made from a M10 brass washer soldered into the formed copper inner lining then a piece of black industrial 1mm thick rubber was siliconed in place then the tail bezel pressed on top.
The button is from an old Nitecore Cobra I think, it sort of clamps around the rubber on both sides, siliconed underneath and the surround to that is an Emisar D4 switch bezel.
The rubber obviously gives the play needed to push the cell down, the inner button has a plunger on it which presses on the tail spring button contact.
It is important to have the driver spring more tensile than the tail spring, I used a standard 4mm QLite spring and a fairly thin tail spring possibly from a copper Maratac.
I did have it bypassed but it was throwing the spring out of alignment and made no real difference anyway due to the compression.
The locking system is fairly simple with an L shaped slot cut into the tailcap copper lining and a corresponding brass stud drilled & soldered into the brass body.

Thanks for the writeup, it explains a lot. :slight_smile:
Soundsā€¦.uneasy and finicky. But I still like it. :slight_smile:

Itā€™s still not clear to me whether the driver spring is mounted centrally or offset?

A thought:
What if you cut the driver spring so its top likes up with the switch top or extended the switch actuator to line up with the spring top, leaving cell-length adjustment fully to the back spring?
Would it remove the cell length effect on the switch action?

The driver spring is soldered in the centre, it is just the picture angle making it look offset.
ā€œSoundsā€¦.uneasy and finicky. But I still like it.ā€ Yes, it was a difficult build but also a good learning experience.

The problem is you want enough pressure on the driver spring to carry good current but not so much that it makes it difficult to press the switch button. The tail spring is just there to keep the cell negative contact to the body when the button is pressed and released.
It took some fine tuning of the setup to find an acceptable compromise between cell travel, battery tube length, cell spring contact and activation.
The important considerations are cell contact on that driver spring in resting state, battery tube length, switch plunger contact to the cell and the amount of piston travel. Remember you only want around 0.3mm to 1mm button travel at the most.
It is finicky to explain or actually make it.
It may well be possible to create a system with more play for different cell lengths but the main principle is having the cell making good contact with the driver spring which means a firm backstop, the tail spring only there to keep contact in movement and the total movement for switch activation to be no more than 1mm or so, so the tail switch must be very close to the cell in itā€™s resting state if that makes sense. I canā€™t see how to make it work with a more adjustable setup, might give it some thought thoughā€¦

Iā€™m getting another headache just thinking about it again :smiley:

Maybe a different designā€¦.?

The core idea is to use 3+ tiny springs supporting a battery contact plate which pushes a centrally mounted switch. The springs only carry mechanical loads and not electricity, so itā€™s fine to use any tiny springs as well as to epoxy them to any part of the PCB.

Sanity check 1, is it possible to order such tiny springs? The smallest springs I can quickly find in reasonable quantities have 0.4 mm diameter and 2 mm height, naturally larger are available as well. Definitely small enough and larger would likely be better. :slight_smile:
Sanity check 2, spring force calculator tells me that they are plenty stiff despite using a thin wire because the wire is twisted really tightly. Same wire with 2 mm spring diameter gives only marginal force.

Now the contact plate - take a piece of copper sheet and bend to U-shape. One side of the ā€œUā€ contacts battery, the other is soldered to the driver pad.

Switch is sandwiched in the middle of the ā€œUā€.

Itā€™s even more complexā€¦but should be quite independent from the battery used.

I look forward to your build :laughing:

Some good ideas I will think about for the next one but I would rather just say ā€œUse these cellsā€

I canā€™t visualize this. Where is the cell negative to the body? The switch plunger length travel still varied by cells length? How to incorporate? Do we want stiff springs at the tail end? How does the copper U shape soldered to the driver with a switch in it work, spring in there too?
Making a diagram may help.

:slight_smile:
As of now Iā€™m looking for a good host for EDC. Then I will consider modding options and tail e-swith is on the table so far. :slight_smile:

Sensible answer. :wink:


Hereā€™s a top view with the top copper layer transparent as well as a side view.
Both views omit switch wires. The grey things are springs (easier to draw this way).
Dimensions are probably wrong (f.e. with copper this thick the U arms wouldnā€™t move even close to parallel).

This sandwich should be assembled alone and then soldered to the driverā€™s positive contact. It should have much stronger springs than the tail ones, so the battery touches it but does not press it without user intervention. Even under g-forces.

Version 1: The tail is just as in your light. The switch has 2-parted travel. Initially with weak force (cell does not even move, only need to overcome rubber resistance) and the length of this movement indeed depends on the cell (I was wrong when I said it was not). But then thereā€™s firm movement which is largely consistent (depending on the ratio of spring force between the front and the back).

Version 2: The host must be DQG-like, with tail modeled such that flashlight length that changes with cell length. There is still a spring in the tail to carry current while the switch is being pressed but thereā€™s 0 initial switch travel. And the action is even more consistent as forces acting on springs are the same regardless of cell length.

ADDED:
Are the tiny springs actually needed? With a seriously stiff switch I think they are notā€¦

Sorry for a little hijack CRX. But as I started this side topic already Iā€™d like to add one more thought to it.

I think that the tiny spring may not be the only unneeded part in a DQG style host. What is the tail spring needed for? Only to deform a fraction of a mm without disconnecting power. It is always under pressure from both sides so it doesnā€™t need to add any pressure on its own. And a piece of copper foil, cut into a flat spiral should work just fine. Actually putting less pressure on the switch should be a good thing, so it should work better than a regular spring.

I purchased the light from CRX, it arrived a few minutes ago.
Build quality is awesome, the light looks great.
The switch works better than I was afraid of. Indeed not as good as a direct one but close, I like it a lot.
Unexpected drawback: power is connected significantly before the switch starts working which makes it impossible to do Andruil factory reset.

Well IMHO mad is not quite the right word, but then again: itā€™s not my native language.
I think it has more to to with an attitude of: who says it canā€™t be done? And then doing it!

Congratulations: CRX for being able to build such a light, and Agro for owning that light.

I think CRX didnā€™t like the purple colour either because the light is natural Ti now. And Iā€™m glad about that. :slight_smile:

I did some calculations as well as experimentation on that.
How strong must a switch be to act as the only spring?
For AA cell, gravitational force is 0.215N. The switch needs a some more to overcome friction. It also needs some more to not be pressed by shaking the light, this is probably more important than friction.

I tried with a 0.8N switch and no tube. As expected, the switch can push the cell against gravity easily. But with that weight it becomes very sensitive, the force needed to press it is surprisingly low. Seems too low.
I would be really interested in double-action switches in such use. A quick search shows that there are such switches with the first action of 1N. Might be OK for AAA but I doubt itā€™s enough for AA. Anyway, I intend to check.

Single action switches are available with forces up to 7N. This ought to be easily enough.

How is the cell positive connected?

A piece of copper foil.