The PhD-M6 thread (programable hotwire driver for the SF-M6) ...

[wquiles]

************************************************** ************************************************** ***************************************
Edit: July 26, 2010

We are finishing up tweaking of various settings, so we are almost ready to start the formal sales thread. Given this, I am going to ask a moderator to go ahead and close this thread.

Thanks much,
Will


************************************************** ************************************************** ***************************************
Edit: June 28th, 2010

The production boards for the PhD-M6 V1 arrived today:


************************************************** ************************************************** ***************************************

I have been working with Alan and Jim for a couple of months on the PhD project (programmable hotwire driver) and I have always been wanting to create a variant for the SF-M6. The goal is to leverage the work done on the main PhD project thread but to tailor a version specifically for the SF-M6, since the M6 is a different beast, with unique battery choices, with two possible bulbs (MN20 and MN21), and no traditional switch like on a Mag. Here is the main PhD thread:
http://www.candlepowerforums.com/vb/...d.php?t=209098


Of course the precursor of this project/idea/variant is the outstanding M6-R project by js, which used rechargeable NiMH cells and a hotwire driver from Willie Hunt set for the MN21 bulb:
http://candlepowerforums.com/vb/showthread.php?t=79916

We had additional projects like the PIR1 and the HDM6, so this is not the first time this has been done. Still, in my book, we still owe Jim (js) credit for not only being the first one, but for giving us the bar by which such projects are judged.

So what do we have today? It is very early, but we have an outstanding development board done by Jim and great code created by Alan. I have been modifying this code somewhat, and will be making additional modifications to make this project M6-specific:



The goal is still to have a regulated circuit with soft-start, and I would like to offer the ability for the end user to select between the MN20 and the MN21 bulbs, so that the same pack can be used with either bulb (once selection has been made). Right now I am thinking to use a switch on the battery pack and a tiny LED to indicate what bulb has been selected, but I am wide open to suggestions/ideas. We can also consider and add other features, but we should perhaps keep the initial version simple and then consider adding other "stuff"

To get started, I would like to discuss:
- what batteries we should consider? NiMH, LiIon (protected!), or both?
- should we consider again a special tailcup insert like HDM6?
- should we require an extension tube for the body as sold by FM?
- should we initially just consider the MN20 and MN21 only?
- how would be the preferred way to select between bulbs?
- what do we want the driver to do when the battery pack is depleted?

Questions, suggestions?

Will

[zx7dave]

I think that you will find end users that are interested in just about all the options you mentioned...
But I think the most popular would be designs that keep the M6 looking factory externally...
Maybe bring the tailcap mod back?

[wquiles]

I agree - the same overall form factor would be best, as Jim did on his M6-R. Here are some photos of when I made my own M6-R pack with parts that I bought from Jim:
http://www.candlepowerforums.com/vb/...d.php?t=124058


At some point I also considered something like this one, but I don't think this could provide the amperage for the MN21:
3x 17670



And of course, the ill-fated HDM6 form factor that depended on the short tailcup. Short part in the middle, with a late model M6 cup on top and a very old M6 tailcup on the bottom:



I am also thinking that the 2x18650 holder from mdocod might be really good and maybe a version of this holder would allow a converter/driver on the sides of the cell:
http://www.candlepowerforums.com/vb/...d.php?t=204701


Will

[cnjl3]

Wow!
What a great project Will. I know that this will be a success only because there are so many CPF's interested in multiple bulb/battery solutions.

I wonder though about the 18650 battery carrier-because it is "only" 2200mah of juice.

You and I have AWR's HDM6 3300mah pack which uses six AW 17500's.
Are these the Li-ions you are talking about?

My vote is for:

- Li-ion cells which might or might not need the special tail cup.
- MN21 and MN20 initially with 'JS' recommended MN15 later?
- switch to select bulb selection [damn, I am smiling as I type this]
- driver needs to shut down before cells are damaged.

A big thanks to you and Alan and Jimmy for all your hard work.

[wquiles]

Wow!
What a great project Will. I know that this will be a success only because there are so many CPF's interested in multiple bulb/battery solutions.

I wonder though about the 18650 battery carrier-because it is "only" 2200mah of juice.

You and I have AWR's HDM6 3300mah pack which uses six AW 17500's.
Are these the Li-ions you are talking about?

My vote is for:

- Li-ion cells which might or might not need the special tail cup.
- MN21 and MN20 initially with 'JS' recommended MN15 later?
- switch to select bulb selection [damn, I am smiling as I type this]
- driver needs to shut down before cells are damaged.

A big thanks to you and Alan and Jimmy for all your hard work.

Like I was telling Jim (js), for this project my goal is not to make money, but to put my effort/time towards finding a way to try to fulfill on the desire/need for a rechargeable, regulated pack for the M6, for all of those who were left with nothing during the whole AWR HDM6 fiasco.

You and I were some of the very few that got "something" from AWR before he ran away with the money of many here in the forums. That being said, the technical aspect of the 6x 17500 Protected LiIon pack along with the new tailcup was a sound idea, and my two packs have been performing flawlesly, and it can drive both the MN20 and the MN21, so we should consider it again for this project as a possible way of implementing this battery pack.

Will

[LuxLuthor]

Will, great pursuit. I think having a dipswitch to choose between the two bulbs would be fine, as it is unlikely that people will be changing bulbs frequently. I can see having 3 choices available would be even nicer.

You will get more mileage with Li-Ion's, so some combination makes sense. The 6x17500 cells gives a lot of options, and makes more sense than 2 x 18650. Ability to replace cells would make this a long term & more useful idea. (i.e. once the JS & few AWR M6-R's reach end of NiMH life, most users will not easily replace those cells.)

For long term planning, you may want to consider a platform (in terms of batteries and form factor) that takes into account some of the flexibilities that DM51's thread outlined which would work with same battery solution. I don't think it's unreasonable to consider capability of setting other voltage choices with program update in with AVR.

I would rather see low voltage cutoff at 2.8 or 2.9V for Li-Ion, rather than most PTC's set at 2.4 to 2.5V. I like the idea of PWM Low/Med/High switch controlled settings...which would require a clickie. I also like the momentary on/off pushbutton. Wonder if there is a way to use both.

End of battery either shuts off light, or reduced output for 1-2 mins then shutoff...giving you enough emergency light to do something.

Just a few ideas off the top of my head.

[jayflash]

The ability to use stock MN21 or MN 20 is important to me as I don't want to use bi-pins. No extension tube; that would destroy the M6 for me. Low battery warning rather than abrupt cutoff. The soft start is a plus and availability of a lower level(s) would be nice, but not absolutely necessary.

How would six 17500 cells fit or did I misread?

Good luck and thank you.

[wquiles]

Will, great pursuit. I think having a dipswitch to choose between the two bulbs would be fine, as it is unlikely that people will be changing bulbs frequently. I can see having 3 choices available would be even nicer.

You will get more mileage with Li-Ion's, so some combination makes sense. The 6x17500 cells gives a lot of options, and makes more sense than 2 x 18650. Ability to replace cells would make this a long term & more useful idea. (i.e. once the JS & few AWR M6-R's reach end of NiMH life, most users will not easily replace those cells.)

For long term planning, you may want to consider a platform (in terms of batteries and form factor) that takes into account some of the flexibilities that DM51's thread outlined which would work with same battery solution. I don't think it's unreasonable to consider capability of setting other voltage choices with program update in with AVR.

I would rather see low voltage cutoff at 2.8 or 2.9V for Li-Ion, rather than most PTC's set at 2.4 to 2.5V. I like the idea of PWM Low/Med/High switch controlled settings...which would require a clickie. I also like the momentary on/off pushbutton. Wonder if there is a way to use both.

End of battery either shuts off light, or reduced output for 1-2 mins then shutoff...giving you enough emergency light to do something.

Just a few ideas off the top of my head.

Thanks - excellent ideas and suggestions. Since I will be using the Tiny85 controller, those settings would be easy to set.

What did you mean by this: "I like the idea of PWM Low/Med/High switch controlled settings.." what exactly you have in mind? Driving the specific bulb at a lower point than its rated voltage?

I also went ahead and posted an invitation in DM51's thread you linked to get additional feedback

Will

[wquiles]

The ability to use stock MN21 or MN 20 is important to me as I don't want to use bi-pins. No extension tube; that would destroy the M6 for me. Low battery warning rather than abrupt cutoff. The soft start is a plus and availability of a lower level(s) would be nice, but not absolutely necessary.

How would six 17500 cells fit or did I misread?

Good luck and thank you.

You did not misread. You can fit six 17500 LiIon cells, in three parallel banks of 2x serial cells each, which even when using protected cells (and with the soft start) allows enough drive current for the MN21. Of course, there is a small catch - namely that the original tailcup's insert has to be swapped out for a smaller (less tall one) which then allows the longer battery pack to fit inside the M6 without any body extenders. That is what I meant in my post above when I posted this picture showing the shallower tailcap insert:


Will

[LuxLuthor]

Thanks - excellent ideas and suggestions. Since I will be using the Tiny85 controller, those settings would be easy to set.

What did you mean by this: "I like the idea of PWM Low/Med/High switch controlled settings.." what exactly you have in mind? Driving the specific bulb at a lower point than its rated voltage?

I also went ahead and posted an invitation in DM51's thread you linked to get additional feedback

Will

Will, I meant along the lines of AW's multilevel driver. When you start using its 3 settings, like many LED driver options, you realize there are many changing scenarios that benefit from reduced/higher outputs. Admittedly, I don't use the 30% duty cycle setting all that often with incans because the color is too orange, but I flip between the 60/100% a lot. I suppose the better settings would be 40/70/100% duty cycle.

[LED61]

Will, believe it or not (I have already talked to Jim about it) I have been running my MN21 with the 7.5V regulated pack I bought from you (thanks so much). The runs have been limited to 1 minute and the bulb lasted about 7 hours total time in 1 minute bursts. The issue was that the heat created blackened my first MN21 and to avoid this I did not do extended runs.

But I would be willing to bet that with the soft start the MN21 could run at 7.0V with more decent bulb life (15-20 hours ?). We'd be driving it harder that the M6-R at 6.8V but I'd gladly trade the lower lamp life for a more white beam.

[tumblingdice]

Hi Will - An outstanding project. Having an M6 powered by 6X14500 cells is an excellent idea. My pennies worth: Avoid altering the external appearance/size of the fabulous M6 design. I'll be watching developments in this thread with great interest. Best of luck - Tumblingdice.

[tumblingdice]

Oops - that's 6x17500 cells.

Hi Will - An outstanding project. Having an M6 powered by 6X14500 cells is an excellent idea. My pennies worth: Avoid altering the external appearance/size of the fabulous M6 design. I'll be watching developments in this thread with great interest. Best of luck - Tumblingdice.

[DM51]

Great project! I've added a link to this in the shootout thread.

The safe high-current IMR18650 cells from AW would be a very good choice for the MN21.

[BSBG]

Here are my thoughts...

• Keep the form factor unchanged. I have a 2C extension and really don't like it, in spite of the admirable run time with the MN20 or a WA1111.
• The monetary option is important - one of my only objections to AW's 3 level soft starter is that it is constant on only.
• Less than 15 minutes on the MN21 is too little IMHO.
• Another option is 3x17670 with the MN60 and 61 - not too much current draw and a beautiful white light. Taking a couple of tenths of a volt off the pack and the soft start should extend the bulb life considerably.
• Teh 6x17500 is viable if the modified tail cup can be made readily available, or part of a kit.
• The bipin options on either 2 or 3 Li Ion should be considered - the WA1111 or 1331 are good alternatives to the SF bulbs.

I am sure more will occur to me...

[brunt_sp]

So far as running the MN21 is concerned, what is the advantage of the 6 X 17500 over the 2 X IMR18650 ? Is it just runtime ? Is softstart really necessary ?
I gather that the real aspect of this proposal is the regulation so there will be constant brightness level throughout the runtime of the cells - Interesting.

[wquiles]

So far as running the MN21 is concerned, what is the advantage of the 6 X 17500 over the 2 X IMR18650 ? Is it just runtime ? Is softstart really necessary ?
I gather that the real aspect of this proposal is the regulation so there will be constant brightness level throughout the runtime of the cells - Interesting.

Good questions. Nothing is perfect, but here are some advantages of the 6x17500 pack over the 2x18650:
- 6x 17500 gives you more capacity, so yes, longer runtimes.
- 6x 17500 sags puts less stress on the cells (each cell sees 1/3 of the rated MN21 current over 100% of the rated current for the 2x 18650)

Is soft-start really necessary? In my opinion, yes. I would not bother working in this project if we did not have it:
- Imagine that you "really" need the light to come "ON" - do you want to mess with double/multiple taps until you can fool the protection circuitry of the cells, or do you want a light with a soft-start so that it comes "ON" every time?
- A well implemented soft-start circuit extends the life of the bulb - pretty important in my book. It is during the initial ON that you have the highest risk of blowing the bulb.

On top of the soft start, we then of course use regulation to keep the bulb at the optimum voltage while the batteries drain, and with regulation using a micro-controller (Tiny85 in this case), we can also easily incorporate a low battery monitor which would protect the battery pack from over-discharge.

Lastly, another important factor in favor of regulation is that you can operate a bulb using a battery pack with a higher voltage than the bulb. That first photo on post #1 shows a bulb rated at about 5 volts happily being run from a 12 volt supply. Solutions like the 2x18650 (as shown by member mdocod) really over-drive the MN21, so you are already increasing the risk of killing the bulb, and of course this shortens the life of the MN21 bulb. With a regulated circuit, you keep the voltage to the bulb at a set point so you eliminate this over-driving condition..

Willie Hunt's "legendary" LVR did all of this, although it did not have the ability to be used with different set voltages/bulbs. We are just trying to duplicate it, and to provide the ability to have multiple voltages

Will

[leukos]

Will,

First of all, thanks so much for the hundreds of hours you, Alan, Jim and put into this project already!

I'll give my quick answers to your questions:
- what batteries we should consider? NiMH, LiIon (protected!), or both?
LiIon is the way to go, more energy density and less weight. 3x 17670 batteries is the most you can pack in there without modification, though I really do like the simplicity of 2x 18650 (6x 17500 is just too many batteries!). Besides, huge battery improvements seem to be around the corner, with all the talk of nano silicone wires we may soon have 18650's that can power the MN21 for 8 hours on a charge. Unlike js' M6-R packs, it would be nice if it wasn't a shrink wrapped kit, but something that we could pop the batteries in and out of for charging.

- should we consider again a special tailcup insert like HDM6?
Keep it simple and stock for now.

- should we require an extension tube for the body as sold by FM?
Keep it simple and stock for now.

- should we initially just consider the MN20 and MN21 only?
How much trouble would it be to make the Vout variable? With FM's excellent bi-pin adapters, there is a plethora of lamps I would like to experiment with at various voltages. I really would like a low powered lamp option with runtimes of up to four hours if the voltage could be user selected.

- how would be the preferred way to select between bulbs?
A switch makes sense if you only plan to offer two options.

- what do we want the driver to do when the battery pack is depleted?
I have always thought that blinking renders the flashlight useless. Makes more sense for there to be a moon mode for the last few minutes before cut off.

[Grox]

Thanks Will and everyone else who's been working on the PhD project. I have looked at that thread many times but have never been able to contribute constructively, lacking the knowledge.

Here's my 2 cents:

- what batteries we should consider? NiMH, LiIon (protected!), or both?
Li ion.

- should we consider again a special tailcup insert like HDM6?
I'd prefer not to have an insert. I appreciate the ability to swap out rechargeable packs and go back to primaries without having to fiddle with a tailcup. KISS is a good approach at least initially. Perhaps that could be an option further down the road.

- should we require an extension tube for the body as sold by FM?
Again, I'd prefer the M6 in stock form.

- should we initially just consider the MN20 and MN21 only?
For me, the most crucial feature would be MN20/21 support. However, it would be very handy if other bulbs could be powered. I'd love a variable Vout.

- how would be the preferred way to select between bulbs?
Switch or pot for variable Vout. Or... computer interface As simple as possible please!

- what do we want the driver to do when the battery pack is depleted?
It would be lovely to have a low batt indicator for example a dim pulse every few seconds.

[wquiles]

It is of course early, but so far:

- most want to use LiIon cells

- seems like most want to keep the overall original size of the M6 as is, with no extensions. Some are willing to replace the tailcap insert to gain more capacity (7x17500)

- most want to support the stock MN20/MN21, and some want even more bulb choices

- most like the idea of a simple way to select bulbs by the end user


Let me talk more about the voltage selection, since most of you are familiar with the AW soft-start/multi-level Direct Drive PWM circuit, the FluPIC, and LED drivers from TaskLED. If you go and read the PhD thread, you will note that the "normal" PhD platform (which from now on I will just call the PhD - I will continue using PhD-M6 for our unique variant) assumes the following:
- that the host is a Mag
- that the Mag has the switch converted to momentary (like we do for the drivers from TaskLED)
- that the battery is always connected to the micro-controller (which from now on I will call the Tiny), and it is constantly looking at the momentary button to determine when to turn ON, OFF, etc..
- that there are almost infinite battery possibilities (OK, I am exaggerating a “little”, but clearly having anywhere from a 1xC to something like the Elephant means far more battery alternatives than for the M6 host).


In the PhD-M6, since we want to keep the basic flashlight "as is", we have to consider some limitations that come from this unique form factor:
- We don't have a simple way to both having a momentary button AND having the battery always connected to the Tiny. If you release the button on the M6 the circuit is open and the bulb does not get to stay on. So the M6's switch can't be considered “momentary” - it is either ON or OFF, and in our case OFF means the circuit is open.

- For the M6, we have very few battery alternatives if we stick with the OEM body configuration, so this makes the design of the PhD-M6 both simpler, but also more limited compared to a Mag host.

True, for multiple levels we could come up with some solution like on the FluPIC where multiple taps get the driver in various modes, but unlike the FluPIC and the AW driver, the PhD-M6 is a buck regulator, so it “can” output a voltage higher than the bulb can take – a mistake in the menu and/or sequence of presses/taps could result in a very expensive blown bulb!

When Jim (js) came up with the superb M6-R, he wired the LVR from Willie Hunt so that it would be connected to the battery pack, and the main switch would just close the circuit and turn everything ON. The LVR was programmed for the MN21 (set to 6.8V rms) and that was it – nothing to select. It just worked. Simple. Reliably. No double taps :-)

At least to start, I am looking at a similar, very simple solution that works the same way for the PhD-M6. Here is my initial proposal (please feel free to punch holes in this, agree, disagree, offer alternatives, etc.):
- Each PhD-M6 will be wired like the M6-R, in a very simple manner. No assumption of multiple button presses. Just ON and OFF, just like the M6 works today.

- The PhD-M6 will provide a regulated output via a varying duty cycle PWM signal, and it will also offer a soft-start. For example, you can run a 7 volt bulb from a 12 volt supply.

- Each board will have two dip switches (or something similar, but changeable for the user), to select one out of 4 possible bulb combinations.

- Each time the Tiny in the PhD-M6 wakes up, it reads those values, and sets the output accordingly:
0 0 – bulb 1
0 1 – bulb 2
1 0 – bulb 3
1 1 – bulb 4

- Battery low level set depending on the battery solution used (9x NiMH, 7x 17500, 2x 18650, etc.).

Note: If in the future we have more battery choices, we could then use the next higher pin count Tiny processor (with more additional inputs) and have two or more battery alternatives. At least for this first iteration, I would like to suggest that we keep it to just one battery pack so that we can get the project moving along.

Why I am thinking this would be a good way to get started:
- This initial solution would therefore require no double taps, no alterations to how the M6 works today

- You would not have to set the bulb choice, nor the battery alternatives via button presses/multiple taps. You can't accidentally change those values by getting confused on the menu. Remember that here the driver “can” output a higher voltage than a bulb can take.

- If you forget what bulb you have programmed now, simply “look” at the battery pack and read the dip switches. This will prevent “ops” when assembling different bulbs in the M6.

- Keeping it to just a few single alternatives will be quicker and simpler to manufacture and implement.

OK, so lets hear it. Ideas/suggestions/changes/alternatives?

Will

[william lafferty]

Will, I would like to add my thanks to that expressed by others for your work and your willingness to share your skills with CPF members.

I think your basic proposal, keeping thing simple, is a good choice. My only question is what battery configuration would you utilize for your initial design?

I assume that what we are trying to do here is stuff as much power (that becomes usable through the buck driver) as we can into the M6 host. If that is the case, then it seems to me that the 6 x 17500 is the battery configuration that makes most sense. Since the 2 x 18650 or the NiMH configurations would have a lot less voltage, I'm not even sure why one would use them with a regulated driver.

[Lunal_Tic]

I'm really excited to see a regulated M6 solution come back around. I missed out on the M6-R and I'm hoping this comes to fruition. For the most basic parameters I'm definitely a +1 on the 6x17500s and keeping the stock appearance. I also like the KISS basic on/off of the original light (press momentary, twist constant on).

The only thing I'm a little fuzzy on is the soft start. I have a few of AW's Mag and SF C switches and I've noticed that when I have the light set to high the soft start takes a bit of time. One of the pluses for me of the M6 is a nearly instant huge blast of light when the button is hit. I don't know if the soft start can be made fast enough to provide this and still be a "soft start." I like my AW switches in the lights in which they are being used but the M6 is a different animal in my opinion and while I may end up getting one of his rumored M6 multi stage soft start switches for one of my M6s to try it out, I will definitely keep one as a "blast beast." If the blast beast happens to be a soft start and have more capacity then I'd be all over it.

Just my 2 yen.
-LT

PS Thanks to everyone involved in this, I can't wait to see how it turns out.

[wquiles]

I don't know how AW implemented his PWM soft start, so I don't know why sometimes is shorter/faster in some cases and not in others. I can however explain how the PhD-M6 implements the soft-start function (again, kudos to Alan for writing the PhD code!). First, to explain the soft start, let me explain a little the whole idea of PWM - pulse width modulation.

If I have a 12 volt battery but have a 6 volt rated bulb, if I turn ON and OFF the battery voltage to the bulb fast enough (about 254 Hz in our case), the bulb will only see the "average" value across its filament because the filament (unlike an LED) can't react quickly enough to the battery voltage going from zero to 12 and then to zero, back to 12, and so on. By varying the percentate of the time that the we connect the battery to the bulb (12 volts), vs. the time that we disconnect the battery to the bulb (zero volts), we can increase or decrease the "effective" voltage that the bulb sees.

This variation between the time ON and OFF cycles is called the duty cycle, and it is how we regulate the voltage in the PhD-M6. In the example above, we start at 50% duty cycle, so the bulb sees an average (or RMS - root mean squared) value of 50% of 12v, which is 6 volts. As the battery drains (which is monitored by the Tiny processor) the duty cycle is increased, which balances the lower capacity at the battery - the bulb still sees 6 volts.

This cycle of the battery draining and the duty cycle increasing continues until the battery is depleted enough, which is when the duty cycle will eventually reach 100% - at this point the battery is connected to the bulb all of the time (no OFF cycles), and we are basically running DD (direct drive) and regulation stops - if the battery drains more, the bulb will finally see less than 6 volts and will start to dim and turn yellow.

Now, back to soft start. If we apply a duty cycle of 50% right at the beginning, the cold filament in the bulb will have its lowest resistance, so it will draw a much larger inrush current through the filament - this high inrush current is what kills bulbs. The filament will eventually warm up, which will increase its resistance, which will then settle to the "normal" operating current for the bulb at that set voltage.

Since we want to limit that inrush current, what we do is that we simply apply a smaller duty cycle during the initial phase that is LESS than 50% (like say, 5-10%, which effectively means the bulb initially sees a voltage less than 6 volts), and then we slowly increase the duty cycle up to 50%, which is the operational point for the regulator in this particular example (12V and 6V bulb). My goal for the PhD-M6 is to have this soft start to last approximately 50-100mS, before going to steady state - based on my own experience with Willie Hunt's LVR on Jim's M6-R solution, this 50-100mS should be fast enough to be helpful to the bulb, while still being fast enough for users to almost "miss" the fact that the soft start is there (which is of course the ideal situation). The PhD-M6 would do this "soft" application of the battery voltage to the bulb, every single time it starts, which is why this is called a "soft start".

Will

[leukos]

The softstart is a perfect compliment to regulation. I wouldn't invest in an incandescent regulator without a softstart.

So there is the possibility of four voltage outputs? You have suggested the MN20 and the MN21, about 7.5 and 6.8 respectively. What other voltage settings might we be considering?

As far as a battery pack, maybe we could twist AW's arm to develop some 17670 IMR's for this project? 12V would be a nice battery supply for our regulator and no worries about needing a double tap from an IMR. I mention this because js recently was considering a new M6-R based on 3x protected 17670, but neither Pila's or AW's would power the MN21.

[Lunal_Tic]

Thanks for the explanation, not even noticing the soft start would be great.

I don't know how AW implemented his PWM soft start, so I don't know why sometimes is shorter/faster in some cases and not in others.
Will

On AW's switch the ramp up is still visible on lower outputs (same ramp speed I think) but it's the most noticeable going to full output since it has farther to go.

-LT

[Grox]

Will,

That sounds great. I think that starting simple is ideal. Softstart and regulation are musts of course

The dipswitch idea sounds good, but could be fiddly if people switch bulbs a lot. Are there multi-position click switches/dials that could serve? Is it possible to use a small rotary switch or slide?

Like LT, I'm not really enthusiastic about having multiple levels but I understand that others would desire it. Here's why: I like being able to use my light to signal and I like being able to give the light to someone who isn't used to multiclick setups and have them use it quickly.

What other bulbs apart from the MN20/21? For me the most useful would be WA1185 and HOM6R and maybe the MN15 and MN61.

Having the option to use IMR17670s would be great - we could run the upcoming LF IMR HO-M6 (whatever it's called), the MN61, and the 1185 with fewer concerns!

Cheers!

[mdocod]

As far as a battery pack, maybe we could twist AW's arm to develop some 17670 IMR's for this project? 12V would be a nice battery supply for our regulator and no worries about needing a double tap from an IMR. I mention this because js recently was considering a new M6-R based on 3x protected 17670, but neither Pila's or AW's would power the MN21.

Actually, An MN21, being ~30-35W, would run on 3x protected AW 17670s reasonably well if it were PWM regulated, and with the soft-start feature, there would never be a double-multi-tap required to start it up as there is basically no current spike (that the PCB can recognize) to speak of.

The PCBs on the 17670s are known for being right on the edge of being able to light up the 1185, which runs ~3.3-3.5A. But that's direct drive with the regular initial current spike associated with cold filaments...

If I did the math correctly, the MN21 regulated to 7V with a driver efficiency of 95%, would draw right around 3.3-3.4A from the 17670s... now, technically speaking, just like the 1185 configuration that is popular with this cell setup, this drain rate is right on the edge ore even slightly over the 2C maximum recommended discharge rate for most LiCo cells...

I brought this up more to point out that it would be possible.... let me throw my 2 lumens on the pile now:

It's an M6, with all of the effort to go into this project, aiming for reliability should be paramount, and this would be factored into all aspects. Bulb life, Cell health, regulator heat management, etc etc. With that in mind, any 3x17670 configuration should not stress the cells beyond 2C.... So LiMn would certainly be nice here, but we would be sacrificing some runtime, for sure. By my estimates, a 3x17670 LiMn setup regulated to an MN21 would run for close to 20 minutes, which would basically make this a direct replacement for the runtime normally associated with the MN21 on primaries, accept it would be regulated, soft started and brighter; not bad!

Now... having said that, a 2x18650 LiMn setup, with regulation, would probably also run for 15+ minutes on the MN21, so the simplicity of 2 cells, and higher driver efficiency (not bucking as much voltage should result in better efficiency correct me if I'm wrong here), so that type of setup would reduce thermal management issues in the driver.... If there are any issues (I haven't followed the whole project, forgive me if I'm off on a limb here that isn't connected to this tree...)..

Personally, what I would really love to see, is an after-market M6 replacement body, that shares pretty much the exact same design as far as the grip and everything is concerned, that would be bored to fit 3x18650s. This would open up a lot of options. LiCo cells in this configuration could drive an MN21 for 40 minutes and be totally safe for the cells... This type of offering could be done separately and would not necessarily be required as the 3x17670 option is on the table IMO...


----------

Will,
I'm very interested in this project and will be following it closely...

Let me know if you need some prototyping done for various battery adapter designs to work with your project. Just tell me where you want to fit the driver and what cells and the dimensions of the driver and I will see what I can come up with. I should be able to do 2x18650 or 3x17670 designs that leave you ~1+cm of the battery tube (lengthwise) freed up for the regulator (with the stock tailcap).

Eric

[wquiles]

The dipswitch idea sounds good, but could be fiddly if people switch bulbs a lot. Are there multi-position click switches/dials that could serve? Is it possible to use a small rotary switch or slide?

Nothing cast in stone yet - keep the ideas coming. In the end, it needs to be reliable and solidly built

Having the option to use IMR17670s would be great - we could run the upcoming LF IMR HO-M6 ...

Sorry about my ignorance - but what is the LF IMR HO-M6?

Actually, An MN21, being ~30-35W, would run on 3x protected AW 17670s reasonably well if it were PWM regulated, and with the soft-start feature, there would never be a double-multi-tap required to start it up as there is basically no current spike (that the PCB can recognize) to speak of.

The PCBs on the 17670s are known for being right on the edge of being able to light up the 1185, which runs ~3.3-3.5A. But that's direct drive with the regular initial current spike associated with cold filaments...

If I did the math correctly, the MN21 regulated to 7V with a driver efficiency of 95%, would draw right around 3.3-3.4A from the 17670s... now, technically speaking, just like the 1185 configuration that is popular with this cell setup, this drain rate is right on the edge ore even slightly over the 2C maximum recommended discharge rate for most LiCo cells...

I brought this up more to point out that it would be possible.... let me throw my 2 lumens on the pile now:

It's an M6, with all of the effort to go into this project, aiming for reliability should be paramount, and this would be factored into all aspects. Bulb life, Cell health, regulator heat management, etc etc. With that in mind, any 3x17670 configuration should not stress the cells beyond 2C.... So LiMn would certainly be nice here, but we would be sacrificing some runtime, for sure. By my estimates, a 3x17670 LiMn setup regulated to an MN21 would run for close to 20 minutes, which would basically make this a direct replacement for the runtime normally associated with the MN21 on primaries, accept it would be regulated, soft started and brighter; not bad!

Although we have more capacity in a 6x 17500 pack, I "really" like the 3x17670 battery pack since it would give us a higher working voltage, which with the regulation could mean working on bulbs as little as 4-5 volts to near the fully charge battery voltage - pretty nice range indeed.

Now... having said that, a 2x18650 LiMn setup, with regulation, would probably also run for 15+ minutes on the MN21, so the simplicity of 2 cells, and higher driver efficiency (not bucking as much voltage should result in better efficiency correct me if I'm wrong here), so that type of setup would reduce thermal management issues in the driver.... If there are any issues (I haven't followed the whole project, forgive me if I'm off on a limb here that isn't connected to this tree...)..

The efficiency in PWM is not related to the difference between the bulb set point and the battery voltage, but rather as to how little time the switching FET is keep in the linear region. Most PWM designs are like 99% efficient since by design the FET is switched fast through the linear region so the FET does not have much time to get "warm" and give up heat (which would lower efficiency). So for us, even 4x LiIOn in series would still yield nearly 99% on a low voltage bulb

Personally, what I would really love to see, is an after-market M6 replacement body, that shares pretty much the exact same design as far as the grip and everything is concerned, that would be bored to fit 3x18650s. This would open up a lot of options. LiCo cells in this configuration could drive an MN21 for 40 minutes and be totally safe for the cells... This type of offering could be done separately and would not necessarily be required as the 3x17670 option is on the table IMO...

I also love that idea (custom body for the M6), however, I find the width of the body pretty "fatty" as it is, and I never liked the size of the FM Elephant (which was designed to hold 3x18650 cells side-by-side), so we would have to consider some options to see what folks prefer, given that most of the feedback so far in this thread suggests that folks like to keep the same size/shape of the original M6 untouched. Definitely room for more talk on this later on

Will,
I'm very interested in this project and will be following it closely...

Let me know if you need some prototyping done for various battery adapter designs to work with your project. Just tell me where you want to fit the driver and what cells and the dimensions of the driver and I will see what I can come up with. I should be able to do 2x18650 or 3x17670 designs that leave you ~1+cm of the battery tube (lengthwise) freed up for the regulator (with the stock tailcap).

Eric

Eric - absolutely. I can lead us towards a solution, but I can't solve all of the technical/implementations details by myself. I am actually very happy of you joining as it would be great if you can own/drive the battery pack side of this project.

If is early, but here is an informal list of several areas where members can help or own part of this development:
- battery pack: We need to decide on what will be the "initial" configuration (right now deciding between 6x 17500, 3x 17670, and 2x 18650). Right now we can't proceed or do much until we really nail this one down.

- once we pick the initial battery pack configuration, which then sets the highest operating voltage, we need to pick/select those initial 4 bulb choices (I would personally like to keep the MN20 and MN21 as part of those 4, but this is a group project after all)

- once we do the above, we then need to design the actual PhD-M6 motherboard to fit within the constrains of the battery pack. If we pick something with a relatively long pack, we might be restricted to a long and not too wide circuit board like the LVR that Jim used on his M6-R which does not go on top or bottom of the battery pack, but rather on its side, like shown here:






If we don't have space on the "sides" of the battery pack, then we are restricted to something more like the HDM6 that used a circular board on top of the actual battery pack, like shown here in this mock up of a 3x17650 pack:


I can't start creating a working board for the Tiny until this is accomplished, but I can start implementing the code for the selector/dip switches in the meantime.


- Then we need to built 1-2, built-by-hand running prototypes to test the design with real cells, inside a real M6. Assuming this works, we probably need to make a few more and do a limited trial to make sure everything works reliably - not just on the bench.

- I will be working in parallel modifying Alan's PhD code and writing new code for the Tiny processors, and I am sure we will be making tweaks to the code/settings based on the real usage (which usually differs from paper design!).

- we then need to work on producticing (is that even a word?) the solution at least for the initial phase:
1) buying batteries

2) making the battery holders

3) making the Tiny boards, programming, calibrating them, and testing them (again, myself at this point)

4) assembling the completed packs, and testing them at the 4 set voltages - kind of tricky since most all RMS voltmeters can't read a PWM signal if it is not at 50% duty cycle.

5) perhaps work on some sort of charger for them, or assume the end-user will use an off-the-shelf charger capable of charging the cells/pack (Triton?)

6) selling the packaged solution: again my goal is not to make money or turn a "profit". At least initially, we should try to recover the cost of all of the raw materials and add a very small/nominal amount for the assembly/test time.


- After the initial phase of packs is done with, if we have enough demand, we "could" then consider offering individual pieces for end-users to assemble their own packs.


This above is my high level "project" view. Suggestions/comments/alternatives?

Will

[brandx]

I am quite pleased to see mdocod joining the fray... more of the 'pro' support for this project can only help. For input and commentary, I can only look at my own uses here.... we have basically 2 cell or 3 cell series options and I have stayed with the SF MN 20/21, and lets' not forget the MN61, and then the WA 1111 and 1185. I would like to have the ability to use any of them via a battery pack and bulb swap and setting the correct dips on wquiles magic board. Whatever you guys come up with.... I want one! As further down the road thinking, can the driver board be planned to be adaptable to the M3-M4 as well? The M6 does have it's following, but I had an M4 and M3 WAY before I jumped on the M6 platform.

[RobertM]

I think that 2xIMR18650 would work fine. At 1600mAh capacity, the MN21 should run for almost 20 minutes correct?

I agree with mdocod that since it is for the M6, it needs to be as reliable as possible. I like the idea of using a safe chemistry cell (i.e. IMR18650) that doesn't have any protection PCB's that could potentially fail. In additional, pushing 3x17670 to the limit doesn't seem like the best idea to me. At least not until AW has some IMR17670's?

I agree that support for the MN20 and MN21 would be at the top of the list too. Beyond those two, I'd like to see support for the MN15.


For me, a stock tailcap, stock body, regulated MN20/MN21 powered by AW's IMR18650's setup would be M6 nirvana!

Thanks for all your work on this project! I can't wait to see how it comes along. I've been wanting a safe/reliable rechargeable option for my MN21 for quite some time.

Thanks,
Robert