Circuit design inquiry

[ThumperIII]

I'm pretty new to flashlight modding. Previous experience is only with upgrading a 2AA XM-3 with a new heavier pill and DX 15880 driver to a P4. Works great.
Now working on 3D M*g mod and have Q's. I want several brightness levels. I liked the 15880 as I will be running from 3*D NiMH. If I understand correctly, I can parallel in some 7135 boards for higher controlled current levels. Here's my Q's. Should I go for one 1400mA 7135 board and the 15880 to drive a P7 at abt 2.4a or could I use two (or 2 1/2 ) 1400mA boards to drive a SST-50 at at 3.8a (or 4.5a)?
1. Are 3 NiMH batteries capable of providing enough output? (rated 12000mAh each)
2. Will the 15880 hold up to 3.8 or 4.5a throughput? (hoping the voltage from the 15880 can keep the 7135s in regulation for awhile)
3. Considering if both would work, which option would you use?

I made a heavy pill to heatsink all parts and have DX Alum OP reflector.

Thanks for any comments/suggestions or advise.

 

[TorchBoy]

1. Normal NiMH cells will fall flat on their face at 12 amps.

2. Do you mean "3.8 or 4.5 A"? How are you hoping that board will work on the other board? Like a saturated transistor? Can you explain again why you are considering it?

 

[ThumperIII]

OK, maybe it was too late/early to present my idea clearly. I could just direct drive the P7, but I want brightness control. I have read numerous threads like http://www.candlepowerforums.com/vb/showthread.php?t=195358 where the poster used a multimode (buck/boost) driver of abt 1.0 mA output in a parallel combination with a 3*7135 or 4*7135 circuit board to boost the output to the LED to 2.4 to 2.8a at around 3.6v, using 3*D NiMH batteries as the power source. The multimode driver was tapped at the control leg (? not sure of the correct term) of the PWM to connect at the + of the 7135 board. That gave brightness control modes with higher output to the LED than just the multi-driver alone for driving the P7. I was interested to know if it would be possible to sandwich additional parallel 7135 boards to provide sufficient current to drive a SST-50 LED at a higher current -- in the realm of 3.8 to 4.5a. I know the 7135s don't 'add' any to the battery output but they should allow a higher total current (if available) to pass to the LED while still being controlled by the PWM of the mode board. Sorry, I don't remember the conventions for when to capitalize a vs A or v vs V (it's been over 50 yrs since high school electronics) but you should get the gist of my intent. With 3D's at abt 1.2v each for 3.6v total (near the LED Vf) I didnt think it would require a 12a draw. The 3*D NiMH batteries are rated at 12000mAh each for 36000mAh total, so I figured I should have some run time -- IF the batteries can handle the required level of draw (not sure of how they rate for internal resistance). They show 4.09v in series at present charge. This will drop (or get bucked briefly) and then the boost driver should pickup and hold abt 3.6v for some period of time. The 15880 works from 0.9 to 4.5 v input to output abt 3.6v. So, if that explains what I want to do; then back to my original questions from 1st post. Perhaps this won't work - don't know - that's why I asked. Educate me. Use small words

 

[Justin Case]

The multimode driver used in StefanFS's Maglite P7 build thread is not a buck/boost driver. It is merely another AMC7135-based driver that adds a PWM controller for the multimodes.

Why are you planning to use DX15880 instead of one of the multimode drivers listed in StefanFS's guide? If you are new to mod'ing, why not stick to the plan as presented by StefanFS instead of already changing a tried and true recipe?

AMC7135 LDO regulator chips can be wired in parallel to obtain the desired drive current. That's how those chips are wired already, for example, on boards such as the single-mode 4xAMC7135 DX board.

If you want 3.8A out, wire two of the single-mode 4xAMC7135 DX boards in parallel to get 2.8A. View this 2-board sandwich as if it were a single board, and connect it to one of the 1000mA multimode boards as described in StefanFS's guide and you will have your 3.8A driver. If you want a 4.5A driver, wire two 1400mA AMC7135 drivers in parallel, and add another 700mA AMC7135 driver in parallel. Then add one of the 1000mA multimode drivers.

The capacity for 3xD NiMH does not add. Voltage adds in series. You will still have 12000mAh capacity.

At 3.8A drive current, an SST-50 could have a Vf of about 3.4V based on the datasheet (see the forward current vs forward voltage graph on P. 9). At 4.5A drive current, the LED could have a Vf of about 3.5V.

The key is to know the performance of your 3xD NiMH cells under load. What voltage can they hold at 3.8A vs 4.5A? And can they hold at least Vf+0.120V (where 0.120V is the dropout voltage of an AMC7135)? If the 3xNiMH can't hold that voltage, then you won't get full drive current out of the sandwich driver. For guidance, see the NiMH shootout.

You can estimate the emitter lumens from the graph of relative output flux vs forward current, also on P. 9 of the SST-50 datasheet. Pick off the relative output factor for the drive current of interest (either 3.8A or 4.5A). Then multiply that factor by the lumens value for 1.75A in the flux bin table on P. 3, for the flux bin (e.g., WG) you have for your SST-50. As an example, I estimate a relative output flux factor of about 1.9 for 3.8A drive current. For a WG flux bin, the estimated emitter lumens is thus 1.9*350 = 665 lumens min, and 1.9*425 = 808 lumens max. Subtract the usual approx 20% losses from the reflector and the glass or plastic window to get an estimate on the so-called out the front (OTF) lumens.

It's not clear to me what the advantage is of the SST-50 over a P7, as used by StefanFS. To equal or exceed the P7's output, you need to pull more amps from your batteries for the SST-50 vs the P7. A milder current load means a higher battery voltage and longer running in regulation when using an AMC7135-based driver. Do you have an SST-50 in a flux bin higher than WG?

 

[TorchBoy]

a multimode (buck/boost) driver of abt 1.0 mA output

Once again I think you mean 1 amp, but you're almost certainly on the wrong track with a boost driver anyway. Quick tutorial:

Current is measured in amps. Symbol for amps is A (capital), and there are 1,000 milliamps (mA) in an amp.

Potential difference, or voltage, is measured in volts. Symbol for volts is V (also capital).

Power is measured in watts. Symbol for watts is W (also capital).

These three things are related by the formula P = I * V.
Power = current x voltage.
Watts = amps x volts.

Capacity of batteries is measured in milliamp-hours, mAh (or Ah). It's how long a battery can give a particular current for. Voltage is irrelevant to capacity, so when you put cells in series the capacity stays the same. If you put your three 12 Ah cells in parallel you'd get a total capacity of 36 Ah but the voltage would be a nominal 1.2 V still.

Energy of batteries is measured in milliwatt-hours, mWh (or Wh). It's what you get when you multiply the capacity by the voltage. There are pros and cons to using energy rather than capacity.

The way a boost driver works is that is bumps the voltage up, obviously, but to do that it has to draw more current from a battery with a lower voltage than the LED wants for the current it is to be driven at. If a boost driver is, say, 80% efficient the power out to the LED will be only 80% of the power drawn from the battery. The rest would be dissipated by the driver as heat. If you drive an LED at 3.4 V with 1 A that's 3.4 W, so the 80% driver will have to draw 4.25 W from the battery. If your battery is a nominal 1.2 V (say, three D NiMHs in parallel) then that's about 3.54 A drawn from the battery. To drive an LED at 4.5 amps, that would be about 16 amps from a 1.2 V battery. To drive enough AMC7135 chips to get that LED current you'd have to boost to at least 0.12 V higher still, which means at least 16.5 A from a nominal 1.2 V battery.

Which is all very interesting, but as Justin says, with three D NiMH cells in series you should have enough voltage even under a moderately high load to be able to step down to the voltage the LED wants, not boost up to it. So just use a multi AMC7135 driver with a PWM controller. Yes, you can put as many AMC7135 chips in parallel as you like.

 

[ma_sha1]

To use 3D NiMH to drive SST-50, I don't think you need a driver.
Wire it up for direct drive, use the new 10A d2flex PWM for 5 levels would be
the most efficient. The hard part would be modding the switch to momentary only, but Will has a great DIY on momentary mod.

 

[ThumperIII]

Thanks for the replies. Sorry about the delay in response, hope you didn't think I was trolling. Had bad multiple power surge and hours of power outage over sizable portion of county. Even with computer on ACP power supply and surge protector, it's been unstable since then and crashed blue screen several times. I will now try to cover your questions:


Justin Case:


Why use a 15880? I didn't have to use this driver, I was just asking about the possibility of using the 15880 driver simply because that is the multimode driver that I happened to have on hand (no additional cost or wait time) that provided selectable levels of brightness control I wanted. The match of the 3 NiMH batteries in series is very close to the voltage I need for the LED, so I thought that although I would not benefit much from a buck/boost type driver, I did not see that it should hurt anything either and might keep the voltage up when the batteries get near depleted. Also (I thought) having the buck/boost type driver might allow me to wring out some additional run time if I had to use 3D alkalines in an emergency. I know the 15880 driver is not a 7135 based driver like the ones in the other threads I mentioned and uses a different circuit output control design. But it had a IC with PWM controlled output. The drivers in StefanFS's post had onboard 7135s and the other 7135 boards were just added parallel to the existing ones. Got that. Simple if you have a driver with onboard 7135s. What I didn't know, without actual test wiring and the high risk of burning up components if I do it wrong, was if the circuit design of the 15880 driver would also allow tapping at the output leg of the PWM in a manner similar to that done to the PWM of drivers with existing onboard 7135s. (i.e. Would installing some 7135s controlled by the PWM be compatible or incompatible/interfere with the rest of the circuit of the 15880)? Per your comments, that driver and 7135s are not compatible and it will not work. It appears that I will need a different multi-mode driver with onboard 7135s for compatibility with paralleling additional 7135 boards.


7135 stacks: Yes, I could just parallel some number of 7135 single mode boards to arrive at various different set voltage outputs (nominal 350mA per 7135), but that by itself does not yield the adjustable modes I want without some PWM source added into the circuit. I had mentioned the different output levels of different counts of 7135s so I was somewhat on track there at least. And it does not appear plausible for a SST-50 -- given the limits of the batteries.


I don't have any excuse why I flubbed the battery amps (series additive voltage & parallel additive amps) just not thinking it through.


Battery sustained output capability: This was and is still an unknown and I don't know how to determine the capability prior to actually wiring up the circuit and testing it. Since I didn't know if the circuit components I had available were compatible, I inquired here.


P7 vs SST-50: I expect to use a P7 LED. But as I had both a P7 and a SST-50 onhand I asked for opinions as to using one vs the other. Less light output from the SST-50 balanced with possibly smoother beam than the P7. And whether the batteries would even be capable of providing a higher output necessary to compensate for the lower output per amp of the SST-50. I didn't want a direct drive or a simple resistor curciut with either LED as that doesn't give the modes options.


TorchBoy:


12 amps from the NiMHs: I didn't quite catch your reference and thought we were just miss-communicating at first. After I read thru the calculations I understand now what you meant. It was always there, I just didn't SEE it.


1.0mA vs 1.0A : Yes, I screwed up here (more of a typo than not knowing the difference). Then I missed correcting it when I went back and corrected some other errors with an edit.


Capitalization V=Volt, etc. My mixup had something to do with article on correct grammar, I think it was discussed that when the abbreviated word was used as a noun it got capitalized but if used as an adjective it should be lowercase, or some such nonsense. Never mind, not an issue.


Battery capacity: see above


Boost driver: see above re 15880


Thanks, I will check on the different mode drivers suggested and while I am waiting for the driver order, I'll just stick in the P7 and underdrive it with the 15880 and have 1A high w/ modes. Already tested in a breadboard layout. Modes and memory work and its bright enough for usable light outdoors. At the present battery charge its probably in the buck mode. I'll hold the SST-50 for a later build with some 18650 batteries and more circuit knowledge.


ma sha1:
Thanks for that suggestion, I will read about that new PWM and see if I can fit it into some future light with the SST-50.


Again, Thanks all. Appreciate your patience. Obviously, just because I want or thought something to be, it just might not be so.

 

[TorchBoy]

I don't have any excuse why I flubbed the battery amps (series additive voltage & parallel additive amps) just not thinking it through.

Battery capacity, in mAh.

TorchBoy:

12 amps from the NiMHs: I didn't quite catch your reference and thought we were just miss-communicating at first.

We were miscommunicating! I really thought you wanted to drive something at 12 amps, not use 12 amp-hour batteries to drive something. I'm glad we've got that sorted.

I changed my little tutorial slightly.

BTW, :welcome:

 

[Justin Case]

7135 stacks: Yes, I could just parallel some number of 7135 single mode boards to arrive at various different set voltage outputs (nominal 350mA per 7135), but that by itself does not yield the adjustable modes I want without some PWM source added into the circuit. I had mentioned the different output levels of different counts of 7135s so I was somewhat on track there at least. And it does not appear plausible for a SST-50 -- given the limits of the batteries.


Battery sustained output capability: This was and is still an unknown and I don't know how to determine the capability prior to actually wiring up the circuit and testing it. Since I didn't know if the circuit components I had available were compatible, I inquired here.


I didn't want a direct drive or a simple resistor curciut with either LED as that doesn't give the modes options.

The two multimode drivers suggested in StefanFS's guide both use 3xAMC7135 to provide 1050mA nominal on top of whatever drive current is delivered by the sandwiched single mode AMC drivers. These multimode drivers cost less than $4 each, so trying to save money by deviating from the proven recipe to use some other driver on-hand seems ill-advised if you are a mod'ing beginner. Gain some familiarity first, then freelance.

To estimate battery performance, see the NiMH shootout thread. If your brand is tested there, then you are all set. If not, then you either need to get the right battery analyzer equipment or rig a tedious, manual method to do the piecewise integration.

As ma_sha1 suggested, you can use a TaskLED d2Flex direct drive PWM controller to give you multimodes when using direct drive.

 

[Fulgeo]

It's not clear to me what the advantage is of the SST-50 over a P7, as used by StefanFS. To equal or exceed the P7's output, you need to pull more amps from your batteries for the SST-50 vs the P7. A milder current load means a higher battery voltage and longer running in regulation when using an AMC7135-based driver. Do you have an SST-50 in a flux bin higher than WG?

Yes this is the real question with the SST-50. I have a WJ Flux bin SST-50 on order which can deliver 500-600 lumens at 1.75 amps and supposedly 1250 lumens at 5.0 amps. Based on some test posted here at CPF when the SSC P7 first became readily available it was found they are very efficient. A P7 can deliver up to 65% of its output at 1.4 amps which is pretty good. But to me the clear advantage of the SST-50 over a P7 is its single die construction. This is important when using an aspherical lens to make a max thrower. Nice smooth beam without the "+" effect the P7 throws up with an aspherical lens. Not saying P7s are bad as I am a fan of them just in this particular application have to give the advantage to the SST-50. Happy Mods!

 

[Justin Case]

The use of an aspheric is all well and good, but the OP specifically wrote "I made a heavy pill to heatsink all parts and have DX Alum OP reflector." when focused correctly, a P7 produces an excellent beam in an OP reflector.

 

[Fulgeo]

The use of an aspheric is all well and good, but the OP specifically wrote "I made a heavy pill to heatsink all parts and have DX Alum OP reflector." when focused correctly, a P7 produces an excellent beam in an OP reflector.

Not bashing your post. Just pointing out the one "clear" advantage of the SST-50. I actually am in the process of accumulating all the parts needed to fashion 9 P7 mods for christmas gifts this year. I could have gone with the SST-50s but didn't. Already have about 3/4 of the parts.

 

[Justin Case]

No problem. I didn't think there was any bashing going on. I was just clarifying that the dark cross and-or donut hole that one can get with the 4-die SSC P7 isn't an issue even with an inexpensive Mag-compatible OP reflector from DX or KD.

I'm considering some AA Mini Mag LED mods for Christmas gifts. I have a small supply of the mariposaoyako copper heat sinks.