b-bassett
Newly Enlightened
superb, let us know how they turn out. the piezo latching part will be especially useful to me, and anyone else who has been having to make our own latches, then try to fit them in with a driver and a fet.
15A tailcap controller, multimode pwm, piezo/magnetic/clicky, feeler thread
- Thread starter DIWdiver
- Start date
Help Support Candle Power Flashlight Forum
superb, let us know how they turn out. the piezo latching part will be especially useful to me, and anyone else who has been having to make our own latches, then try to fit them in with a driver and a fet.
DIWdiver
Flashlight Enthusiast
I've been working on the UI which results in looking at the schematic again, and have a few updates to the specs:
Typical quiescent current should be 20-35 uA, max 50-80. This is higher than previously mentioned, but should still be irrelevant unless you expect to put the batteries in and leave it for years.
"No wiring" option is only available if you meet certain conditions. This option means you can connect only ground and the lamp to the board. This would be ideal for tailcap operations where the battery+ doesn't easily connect to the board. Incans would meet the conditions if the battery voltage is above 3.0V and the bulb draws at least 3W full on (it gets complicated if these requirements are not met, but it may be okay). LED lights are more complicated. The minimum battery voltage minus the voltage dropped by the LED/driver at low current must be at least 3.0V. It gets pretty hairy trying to predict this, but if the battery voltage minus the LED voltage is greater than 3.0V, you're probably okay. Also, if you use the "no wiring" option the maximum duty cycle of the lamp is about 95%. The exact number depends on many variables, but shouldn't be any less than this.
For LED lights and low-voltage batteries, it's best if you make three connections to the board: B+, B-, and L- (the low side of the lamp, driver, or LED). If you do this, things should go pretty well until the battery voltage drops below 2.5V. Adding the third connection also allows 100% duty cycle
Updates on the UI will be available soon...
D
Typical quiescent current should be 20-35 uA, max 50-80. This is higher than previously mentioned, but should still be irrelevant unless you expect to put the batteries in and leave it for years.
"No wiring" option is only available if you meet certain conditions. This option means you can connect only ground and the lamp to the board. This would be ideal for tailcap operations where the battery+ doesn't easily connect to the board. Incans would meet the conditions if the battery voltage is above 3.0V and the bulb draws at least 3W full on (it gets complicated if these requirements are not met, but it may be okay). LED lights are more complicated. The minimum battery voltage minus the voltage dropped by the LED/driver at low current must be at least 3.0V. It gets pretty hairy trying to predict this, but if the battery voltage minus the LED voltage is greater than 3.0V, you're probably okay. Also, if you use the "no wiring" option the maximum duty cycle of the lamp is about 95%. The exact number depends on many variables, but shouldn't be any less than this.
For LED lights and low-voltage batteries, it's best if you make three connections to the board: B+, B-, and L- (the low side of the lamp, driver, or LED). If you do this, things should go pretty well until the battery voltage drops below 2.5V. Adding the third connection also allows 100% duty cycle
Updates on the UI will be available soon...
D
Last edited:
DIWdiver
Flashlight Enthusiast
Comments are VERY welcome on the following:
The hardware offers thermal protection for both the on-board FET and the LED. The LED protection is only available if you wire a thermistor from the board to the LED. Actually you can stick the external thermistor wherever you want, but the LED heatsink would be the most obvious place.
My current intention is to set up the on-board thermistor to derate the current linearly from 100% to 0% as the temperature rises from 80C to 115C. The max junction temp of the FET is 175C, so this should provide ample margin. I wouldn't expect this to be a limitation in most circumstances. Keep in mind that the thermistor temperature is somewhat lower than the junction temperature.
The off-board thermistor will derate the current linearly as the temp rises from 70C to 90C. Since the LED max junction temp is probably 150C (true for both the SST-90 and the XM-L), this would appear to be only slightly less conservative than for the FET. However, the thermal resistance of the FET is much lower than that of either the SST-90 or the XM-L. Also, the placement of the thermistor next to the FET is the same in every case, and the results easily measured by me. Not so for the external thermistor. This makes the settings for the external thermistor much less conservative. But I hesitate to lower them, lest I prematurely shut down a good design.
Comments?
D
The hardware offers thermal protection for both the on-board FET and the LED. The LED protection is only available if you wire a thermistor from the board to the LED. Actually you can stick the external thermistor wherever you want, but the LED heatsink would be the most obvious place.
My current intention is to set up the on-board thermistor to derate the current linearly from 100% to 0% as the temperature rises from 80C to 115C. The max junction temp of the FET is 175C, so this should provide ample margin. I wouldn't expect this to be a limitation in most circumstances. Keep in mind that the thermistor temperature is somewhat lower than the junction temperature.
The off-board thermistor will derate the current linearly as the temp rises from 70C to 90C. Since the LED max junction temp is probably 150C (true for both the SST-90 and the XM-L), this would appear to be only slightly less conservative than for the FET. However, the thermal resistance of the FET is much lower than that of either the SST-90 or the XM-L. Also, the placement of the thermistor next to the FET is the same in every case, and the results easily measured by me. Not so for the external thermistor. This makes the settings for the external thermistor much less conservative. But I hesitate to lower them, lest I prematurely shut down a good design.
Comments?
D
b-bassett
Newly Enlightened
as long as everyone knows to put the thermistor as close as possible to the LED, with the best possible thermal interface, i cant see any real danger in a max temp of 90C at the heatsink. no-one would be able to hold a torch that hot.
90C will allow a margin of error, in thermister placement/thermal interface, whilst still providing protection to the LED.
Im sure a properly heatsunk LED would heat up the inside of a head enough, even if the thermistor was floating with little contact.
im more of a 'hands on' person though, thermal specs and such arn't my strongest subject.
90C will allow a margin of error, in thermister placement/thermal interface, whilst still providing protection to the LED.
Im sure a properly heatsunk LED would heat up the inside of a head enough, even if the thermistor was floating with little contact.
im more of a 'hands on' person though, thermal specs and such arn't my strongest subject.
Epsilon
Enlightened
Make the switch battery powered like the FETtie, then you are never dealing with voltages of 3V to power this thing 
. And what I found: Chose your FET, so the part that needs to be cooled, is the BAT-, not the LED-. This way you can use any conducting compound instead of thermal adhesive tape, since there normally is nothing that has the LED- and a heatsink capability.
. And what I found: Chose your FET, so the part that needs to be cooled, is the BAT-, not the LED-. This way you can use any conducting compound instead of thermal adhesive tape, since there normally is nothing that has the LED- and a heatsink capability.
DIWdiver
Flashlight Enthusiast
Make the switch battery powered like the FETtie, then you are never dealing with voltages of 3V to power this thing
Unfortunately, the current draw isn't low enough to make a small battery practical. The largest battery that would be practical in a tailcap is a 1632, at 120-140 mA-H. With a load of 35uA, it would only last only around 6 months, which I wouldn't consider acceptable. If you wanted to, you could wire a battery between B+ and B-, if you removed a diode from the board. It ought to work fine.
All the FETs I've looked at have thermal conduction primarily to the drain connection. And no matter how you wire the circuit, the B- never connects to the drain. Do you have an example? If so I'd love to see it.
But if you mount the board in the tailcap, it works great with the L- needing the heatsink. Once you bypass the original switch, the body of the light becomes L-, and the FET connects the body to the B- terminal. Since the board was designed to be mounted in the tailcap, it ends up using the tailcap as a heatsink. The L- terminal on the board includes a ring of bare copper at the edge of the board rests on the ledge of the tailcap, making both electrical and thermal contact. If you use compound, it must be conductive or it will interfere with proper operation, but in most cases I wouldn't bother.
Except for a moderate sized circle in the middle, the flat side of the board (opposite the L- ring) is B-, and attaching that to a heatsink would help a little. I wouldn't want to guess at how much.
P.S. I got notice that the boards shipped today!:twothumbs
DIWdiver
Flashlight Enthusiast
Make the switch battery powered like the FETtie, then you are never dealing with voltages of 3V to power this thing
Okay, you got me thinking. If I gave up the Hall sensors, making the board useful for either piezo or contact type switches like a reed or a clicky, and intended it to be run off a separate 3V battery, the quiescent current (lamp off) could be gotten down to a few microamps. The battery wouldn't exactly last forever, but 28,000 hours (3 years) is good enough for me.
I could probably get the operating current as low as 0.5 mA, giving you around 250 hours runtime on the 1632 battery. Of course you'd get 3 years or 250 hours, not both. You'd get 1yr/160 hrs, or 2yr/80 hrs, or 2.7yrs/25 hrs, etc.
This would take modifying the firmware somewhat, and ideally re-doing the board to allow a standard battery configuration to be soldered in, so there would need to be some serious interest for me to want to do it. Otherwise, you could just use the regular board and accept the less-than-optimized battery life.
Comments welcome.
D
Epsilon
Enlightened
About the L-, you are correct. I forgot this has to be mounted in the tail .
Personally, I don't mind changing the battery every 6 months if it is not that hard to do i.e. it does not require soldering parts. But thats very personal.
.Personally, I don't mind changing the battery every 6 months if it is not that hard to do i.e. it does not require soldering parts. But thats very personal.
DIWdiver
Flashlight Enthusiast
About the L-, you are correct. I forgot this has to be mounted in the tail
Personally, I don't mind changing the battery every 6 months if it is not that hard to do i.e. it does not require soldering parts. But thats very personal.
It doesn't HAVE to be mounted in the tail. It was designed to work well there, but it can be mounted wherever it's convenient. Most applications (up to 10A) will not require heatsinking.
A quick-change battery solution is not obvious to me, unless I solder some wires from the board to a connector and provide you some batteries with mating connectors. I could do that, but it's not a very elegant solution. Even if I re-spin the board, I don't see a good solution, without giving up some other features.
D
DIWdiver
Flashlight Enthusiast
I got the boards today!
Actually the postman tried to deliver them yesterday, but I wasn't home, so I had to go to the post office today to get them.
This was quite ahead of when I expected them. I ordered 10 boards, got 12, less then 3 weeks, for $13.50. That's hard to beat! Even if I did have to go to the post office, show my ID, sign both electronically and on paper. That price is AFAIK is impossible to beat or even come close to. Nobody else seems to want to make boards for less than $100.00.
The boards look quite good, especially considering the price, but they made one little error. On one side the copper is supposed to go right to the edge of the board, but they pulled it back from the edge a little. Of course I had to try it out, and soldered a wire across the FET connections, then dropped it into a 2D MAG, and it works.
Now I just have to finish the firmware, build a board, debug both the board and the code, then I can offer some boards for test. I'll keep you posted on the progress.
D
Actually the postman tried to deliver them yesterday, but I wasn't home, so I had to go to the post office today to get them.
This was quite ahead of when I expected them. I ordered 10 boards, got 12, less then 3 weeks, for $13.50. That's hard to beat! Even if I did have to go to the post office, show my ID, sign both electronically and on paper. That price is AFAIK is impossible to beat or even come close to. Nobody else seems to want to make boards for less than $100.00.
The boards look quite good, especially considering the price, but they made one little error. On one side the copper is supposed to go right to the edge of the board, but they pulled it back from the edge a little. Of course I had to try it out, and soldered a wire across the FET connections, then dropped it into a 2D MAG, and it works.
Now I just have to finish the firmware, build a board, debug both the board and the code, then I can offer some boards for test. I'll keep you posted on the progress.
D
b-bassett
Newly Enlightened
superb, let us know as soon as they becom avaliable, iv been putting off re-doing a few builds awaiting a decent driver for the piezo switch.
Epsilon
Enlightened
10 PCB's for 13.50? Thats a bargain indeed
.Interested in the results
. You are doing a very good job!
DIWdiver
Flashlight Enthusiast
The piezo circuit works!
I mounted one piezo in my MagD tailcap and a smaller one in the MagC tailcap. The smaller one works fine, but the bigger one takes a LOT of force to activate. This seems strange, since it appears almost identical to the one I did my initial experiments with, which takes very little force to activate. They have the same size, capacitance, and voltage rating, but vastly different performance in this application. I guess I'm off to Radio Shack to get another to play with.
Tomorrow I try programming the PIC.
I mounted one piezo in my MagD tailcap and a smaller one in the MagC tailcap. The smaller one works fine, but the bigger one takes a LOT of force to activate. This seems strange, since it appears almost identical to the one I did my initial experiments with, which takes very little force to activate. They have the same size, capacitance, and voltage rating, but vastly different performance in this application. I guess I'm off to Radio Shack to get another to play with.
Tomorrow I try programming the PIC.
b-bassett
Newly Enlightened
b-bassett
Newly Enlightened
any luck getting these programmed?
DIWdiver
Flashlight Enthusiast
Had a bit of trouble with the build. Those parts are way too small for easy assembly. My second board actually works, and programming is easy once you get the board built right.
I've found that only one of the four piezos I've tried works very well. After my initial success with the proof-of-concept, I've been gaining respect for the folks who make reliable switches with much smaller devices.
I've gotten it to work, but before I try to make a big deal of it, I want to understand what makes one piezo work splendidly and another seemingly similar one not work (expletives deleted) well at all.
Mixed success plus a fairly low level of interest from the community, along with being rather busy with other things, has me not working on it much, so progress will be slow.
b-bassett
Newly Enlightened
fair enough.
if i knew what i was doing id make and program my own, maby ill look onto it again.
DIWdiver
Flashlight Enthusiast
Obviously there hasn't been much progress on this lately. I ran into technological hurdles I didn't expect, there hasn't been the kind of interest that I hoped for, and I've been busy on other things. As some of you know, that's a triumverate that's difficult to oppose.
However, recent events have led me to discover a new circuit that dramatically improves the performance of piezo sensors, compared to the '555 based circuits I had been using. It's so simple it's actually embarrassing that I haven't thought of it before, and it allows ALL of the piezo elements I've tested to achieve a reasonable performance.
For anyonw interested, there will be some activity on this thread soon.
However, recent events have led me to discover a new circuit that dramatically improves the performance of piezo sensors, compared to the '555 based circuits I had been using. It's so simple it's actually embarrassing that I haven't thought of it before, and it allows ALL of the piezo elements I've tested to achieve a reasonable performance.
For anyonw interested, there will be some activity on this thread soon.
Hi DIWdiver,
Thanks for bringing this to the table.
I noticed in a couple of posts that you said interest from the community was not what you hopped.
This is just MHO but if you could post either a schematic or a couple of pictures of what this circuit would look like installed in a light and how it hooks up etc I think you would see at least a bit of a gain in the attention given to it.
Thanks for bringing this to the table.
I noticed in a couple of posts that you said interest from the community was not what you hopped.
This is just MHO but if you could post either a schematic or a couple of pictures of what this circuit would look like installed in a light and how it hooks up etc I think you would see at least a bit of a gain in the attention given to it.
b-bassett
Newly Enlightened
thanks for the update. im sure there will be plenty of response once the right ppl see the potential
