The regulated version provides a constant output for the entire battery life and runs a little bit longer. The pre-set output voltage you choose is firmware encoded and cannot be changed, meaning you can only use it on specific bulbs with the set voltage. The unregulated version works with a wide voltage range ( 6V - 30V ) and you can use different bulbs with different battery configurations you choose anytime. However, output brightness will sag along with battery voltage drop during use. A good remedy is to use batteries with a flat discharge curve such as LiIon, LiFePO4 ( A123 , Saphions ).
AW's Going Regulated ! W00T !
- Thread starter LuxLuthor
- Start date
Help Support Candle Power Flashlight Forum
One other advantage once you know the optimal regulated voltage to be delivered to a particular bulb, it will prevent the bulb from flashing (dead) because your batteries were overcharged to a higher voltage hot off the charger.
Very excited about this! Is there any limit on the size of the differential between vBatt and vBulb? For example, could one of Lux's 2D 15.6V battery packs be stepped down to 7.1 (or whatever is optimal for a 5761; there seems to be some discussion on this point) -- thereby optimizing runtime?
Lunal_Tic
Flashlight Enthusiast
So in the example above, 3x C li-ions + WA1185, there wouldn't really be that much of an advantage or would it be easier on the batteries; i.e. less chance of over discharging?
Wouldn't that already be taken care of by the soft start?
-LT
Northern Lights
Flashlight Enthusiast
I realize the discussion is mostly threoreical, but I thought it might be helpful to throw this in and bring some of our conjectures up to date. Earlier I reported that it appeared that some 5761s new out of the box on soft start were flashing at 7.1Vbulb. Jim Jones and I have been working on some more 5761 ideas and he has accumulated some more experiences down those lines. Although those bulbs very well met the OEM specifications of the manufacturer, they do not always work in over drive. Many individual 5761 have worked well at 7.1Vbulb and some above but just yesterday Jim ran into a few that on soft start and experimental control on the bench were flashing 7.1Vbulb or below. I had similiar experiences with one shipment of 5761. 6.9Vbulb may be the high percentage winner for the 5761.
I am assuming because AW is not using the more delicate FET to regulate the power, that your example would be valid.
Let's magnify the example, and ask AW if he knows:
AW, let's say you want to put in a 24V battery pack, and have it stepped down to power a Philips 5761 with regulator set at 7.0 Vbulb. Would it hold up?
All batteries have some sagging once you get to the end of the rated performance, but the Lithium cells have a flatter discharge curve than NiMH. The older Li-Ion cells (vs. new LiFePO4) do not have higher current outputs that some bulbs need, but if you match the total Vbattery with the ideal Vbulb as in your 3 x C Li-Ion & 1185 example that should be relatively good.
The 1185 bulb wants to have 10.8V delivered to it for optimal (1200+ bulb lumen) output. 3 freshly charged Li-Ions will start with 4.1V x 3 = 12.3V which if not reduced by the resistance of the parts in the flashlight, will kill (instaflash) the bulb. Then as the batteries drain, they will go below that needed 10.8V which is demonstrated with this run time graph from AW. The 1185 uses about 3.3 Amps, so look at the red line.
No. The "soft start" blunts the initial millisecond spike of higher voltage that is required to initially heat up a cold bulb filament. Once it heats up (and starts glowing), there is less voltage required to keep the current flowing.
If in the example above with the 1185 is used, even if there is a soft start feature to blunt the initial spike, the bulb can still instaflash if 12V is being delivered to the bulb for less than a second.
The reason you may not have seen this happen in one of your existing setups is that there is enough resistance in the tailcap spring, stock switch, stock bulb holder, etc., that you just don't have all 12.3V delivered to the bulb.
.
Last edited:
I haven't built a working model to test this value yet but it 'll work on theory - with a max. input voltage of 30V ( unloaded pack V ) down to 6V. However, I have to point out that with this setup which I assume you are using 12 pcs NiMH AA cells. You 'll over discharge all of your cells without you even noticing while the light is still running in regulation. The low voltage warning will set in only when the light falls out of regulation so if I set the warning value to be 6.9V in this 5761 setup, you 'll see the warning when you have discharged you NiMH cells to 0.575V !!! When deciding a battery configuration, you 'll have to take this into consideration. I would say a 7 -8 cell pack for the 5761. I have answered the same question in post #33 as well.
Last edited:
Yes, I understand that separate issue.
I was more trying to get at the nice hardware design - higher tolerance features which could be illustrated with the extreme Vbatt vs. Vbulb in my preposterous 24V battery in 5761 bulb example.
My point (for others) is what I asked you about earlier where your design is not using the FET to absorb the brunt of the power like another HotDriver design we know about.
No, my regulation model does not use the FET to dissipate power. The regulated Vout is PWM controlled ( with a higher efficiency ) as well.
My point (for others) is what I asked you about earlier where your design is not using the FET to absorb the brunt of the power like another HotDriver design we know about.
Lunal_Tic
Flashlight Enthusiast
I knew about the battery sag under load and figured that was what was "saving" my bulbs. I've never used the C cells though and so had no idea about sag relative to load. Thanks for the graph and info. Guess I'll hold off on an 1185 3xC version until the regulated one comes out then ask for the 10.8V setting as you mention. I guess there shouldn't be too much trouble bleeding off the extra juice.
Cheers,
-LT
petrev
Flashlight Enthusiast
Hi Lunal_Tic,
Unlike the other regulator type (LDO-Hotdrivers) the PWM Regulator design doesn't "burn off" excess battery power. The FET doesn't get hot from acting as a variable resistor, it is used to switch the flow on and off very rapidly.
If protected cells are used then the supply can exceed the required Vbulb by a large margin (within parameters of the circuit Vmax) to give longer regulated runtime using the cell protection to prevent excessively low cell voltages.
Packs can be designed with any runtime/size desired and the SS-Regulator will deal with providing the correct Voltage to the bulb.
Pete
.
Unlike the other regulator type (LDO-Hotdrivers) the PWM Regulator design doesn't "burn off" excess battery power. The FET doesn't get hot from acting as a variable resistor, it is used to switch the flow on and off very rapidly.
If protected cells are used then the supply can exceed the required Vbulb by a large margin (within parameters of the circuit Vmax) to give longer regulated runtime using the cell protection to prevent excessively low cell voltages.
Packs can be designed with any runtime/size desired and the SS-Regulator will deal with providing the correct Voltage to the bulb.
Pete
.
Thanks for the info, guys. So, doing some back-of-the-envelope calculations, it looks like one could run a 64430 (or its apparent successor, the 56580) on 11 NiMH cells with Vbulb set at 9.9V, and the light would drop out of regulation when each of the cells had been discharged to 0.9V. (Presumably actual Vbatt per cell would be slightly higher, since the driver's low voltage warning is based on Vbulb? Or am I wrong about that?) Similarly, one could run a 64625 on 14 NiMH cells with Vbulb set at 12.6V -- again, the light would drop out of regulation when each of the cells reached 0.9V.
I don't have any experience with protected Lion cells, but I am intrigued by Petrev's point that one could use up to a 30V pack (AW's stated max input voltage) and drive basically anything -- including a low-voltage bulb like the 5761 -- and rely on the batteries' protection circuits to cut the juice before overdischarge (rather than watching for the driver's low voltage warning). Creating such a large differential between Vbatt and Vbulb makes sense with a PWM driver, because excess voltage won't be wasted as heat. If that's all true, does anyone know what Lion battery configuration would provide the maximum amount of juice (and by that I mean runtime) to a 5761 in a 2D host?
I don't have any experience with protected Lion cells, but I am intrigued by Petrev's point that one could use up to a 30V pack (AW's stated max input voltage) and drive basically anything -- including a low-voltage bulb like the 5761 -- and rely on the batteries' protection circuits to cut the juice before overdischarge (rather than watching for the driver's low voltage warning). Creating such a large differential between Vbatt and Vbulb makes sense with a PWM driver, because excess voltage won't be wasted as heat. If that's all true, does anyone know what Lion battery configuration would provide the maximum amount of juice (and by that I mean runtime) to a 5761 in a 2D host?
petrev
Flashlight Enthusiast
Caution :
Need to check with AW but 30V is probably an absolute maximum so may need to use 4.2V as cell voltage to avoid problems with Hot-off-the-charger cells !
As the 5761 runs at about 7V the highest 2D charge density would probably be 2x DLion 7500mAh if they ever become available.
Physical space in a 2D means that for 7V a 5000mAh D is still better than any other multiple that I can think of that fits. AW D-regulator would warn correctly for unprotected cells.
With higher voltage bulbs in larger hosts the possibilities become more numerous . . .
:thinking:
Cheers
Pete
Need to check with AW but 30V is probably an absolute maximum so may need to use 4.2V as cell voltage to avoid problems with Hot-off-the-charger cells !
As the 5761 runs at about 7V the highest 2D charge density would probably be 2x DLion 7500mAh if they ever become available.
Physical space in a 2D means that for 7V a 5000mAh D is still better than any other multiple that I can think of that fits. AW D-regulator would warn correctly for unprotected cells.
With higher voltage bulbs in larger hosts the possibilities become more numerous . . .
:thinking:
Cheers
Pete
Pete, in your reference to using the 4.2V hot....did you mean if you had a scenario where you had 8s Li-Ions (8x4.2=33.6V) which would put you over the 30V cutoff....and similar with 21 NiMH cells (which I am using in my Hyperblaster)....where hot they may be over 30V?
I have a harder time imagining being able to use 8s Li-Ions (including Saphion/Emoli) in terms of their size and PTC limit in the case of AW's protected Li-Ion....but maybe you meant something else. :thinking:
If Plasmaman's 7500mAh D doesn't work, then wouldn't 2s3p 17500 would be higher capacity (2200mAh x 3) than the 5000mAh D cell, (but the PTC is close to being tripped at the 5.5A the 5761 wants)?
AW, when you said in post #33 that "the low battery warning will be set at a few percentage point below when the driver runs out of regulation" is that also a setting we could specify when ordering? (i.e. Vbulb set for 7.0 and low warning set at 6.8V) ?
petrev
Flashlight Enthusiast
Hi Lux,
That is the question that AW will no doubt clarify soon - some components may see the resting voltage, before there is any significant current draw and voltage sag.
This voltage may be the limiting factor that could kill a component.
That's what I meant - 7 Lion may be the series limit. If one were using say 4,5,6 or 7 series CLions to drive an 1185 you would get huge runtime and never get near the protection limit. Somebody may want to use 8 series cells for some reason ? and this may not be possible.
I personally intend to try a 6s E-Moli 64458 (6D+74) and agree that a 7s E-Moli may be a bit unwieldy but that doesn't mean someone won't make it - or even wish for an 8 - Stranger beasts have been made !
Easier to think of 3p2s configuration - 1100mAh 17500s
This equates to 3300mAh battery - 2 in series.
Another way to look at it is to work out total "pack" capacity
7500D-Cells => 3.6x7500x2 = 54000 (54WHr)
5000D-Cells => 3.6x5000x2 = 36000 (36WHr)
17500-Cells => 3.6x1100x6 = 23760 (24WHr)
Back to AW again
Cheers
Pete
That is the question that AW will no doubt clarify soon - some components may see the resting voltage, before there is any significant current draw and voltage sag.
This voltage may be the limiting factor that could kill a component.
LuxLuthor said:
That's what I meant - 7 Lion may be the series limit. If one were using say 4,5,6 or 7 series CLions to drive an 1185 you would get huge runtime and never get near the protection limit. Somebody may want to use 8 series cells for some reason ? and this may not be possible.
I personally intend to try a 6s E-Moli 64458 (6D+74) and agree that a 7s E-Moli may be a bit unwieldy but that doesn't mean someone won't make it - or even wish for an 8 - Stranger beasts have been made !
LuxLuthor said:
Easier to think of 3p2s configuration - 1100mAh 17500s
This equates to 3300mAh battery - 2 in series.
Another way to look at it is to work out total "pack" capacity
7500D-Cells => 3.6x7500x2 = 54000 (54WHr)
5000D-Cells => 3.6x5000x2 = 36000 (36WHr)
17500-Cells => 3.6x1100x6 = 23760 (24WHr)
Back to AW again
Cheers
Pete
Last edited:
Okay, here's where I'm confused: Suppose one were to build a 5761 in a 2D host using Lions in 2s configuration (e.g., two D cells, or six 17500 cells 2s3p), with Vbulb set at 7.0V. Wouldn't the light drop out of regulation when each of the cells reached 3.5V (and the low voltage warning get triggered shortly thereafter) -- which is well short of their 3.0V maximum discharge limit? And wouldn't that compromise runtime quite a bit?
That's what led me to suggest kicking up the voltage as high as possible (e.g., by using six 17500 protected cells 6s, for 25.2V hot off the charger), and relying on the batteries' protection circuits to prevent overdischarge.
Or am I overlooking something obvious, like Lion cells in 6s configuration can't deliver the current that the bulb requires? Or perhaps the discharge curve for Lion cells is such that when they reach 3.5V they're actually very close (in terms of runtime) to being fully discharged?
:thinking:
That's what led me to suggest kicking up the voltage as high as possible (e.g., by using six 17500 protected cells 6s, for 25.2V hot off the charger), and relying on the batteries' protection circuits to prevent overdischarge.
Or am I overlooking something obvious, like Lion cells in 6s configuration can't deliver the current that the bulb requires? Or perhaps the discharge curve for Lion cells is such that when they reach 3.5V they're actually very close (in terms of runtime) to being fully discharged?
:thinking:
Lunal_Tic
Flashlight Enthusiast
Thanks for the info. I think I've got a better understanding of it now. Just have to wait patiently for the C-reg to come out before I build out that 1185 rig, patiently yeah right.
-LT
Pete, sorry for confusing things. I meant 3x17670 which are 1600mAh, but I picked up one of my 18650 which was where I got the 2200mAh from that label (and which won't fit in a tri-bored D Mag)....so the 3x17670 at 4800mAh are close to the 5000mAh D cell.
I still think of these configurations as 2s3p, but there is no real confusion as long as we mind our "P" and "S" numbers. I consider either 2s3p or 3p2s as correct. :grouphug:
Well, you are close to having the right idea, but in general you do not want to go below 3.5V with a Li-Ions. By the time you have gotten down to 3.0V you have likely damaged the cells to some degree...and to a large degree if you go below 3.0 (hence the PTC discharge limit).
But let's look at a couple real examples so everyone understands this low cell voltage AW caution first stated in his post #33.
Let's say you have 6 Li-Ions powering a 5761 bulb that AW's D Driver is set to deliver 7V to the bulb. That means that each cell will be providing 1/6th of the power.
So let's say that AW has the low battery warning set to start at about 3% below Vbulb, which would be about 6.8V.
For the 6 cells to be exhausted down to that 6.8V level, they would each only be providing 1.13V (6.8V/6 cells). Obviously the first of the six AW Protected cells that hits the PTC 3.0V limit (assuming you were smart enough to be doing this light with protected cells) would break the circuit before the D Hotdriver's low battery warning would occur....BUT.....again the point is that it is not good for a Li-Ion cell to run down to 3.0V, which would happen in this setting. There has been much written about this in the battery section of the forum.
If you had 2s Li-Ions that had sufficient current output performance (could be 2s2p or 2s3p), the drain would trigger the 6.8V low battery warning of the D Driver when both of them hit 3.4V....and this will keep the Li-Ions from being damaged at the 3.0V level.
Even if you have 3s, the 6.8V low alarm would work out to 2.22V/cell....again having the cell's PTC kick in first at 3.0V, and doing some damage.
There would be a similar scenario with using many NiMH cells, which should not be discharged below 0.9 to 1.0Volts. Let's say you used 10s NiMH to have 12V pack. When the 6.8V limit triggered the low battery warning for these 10 NiMH cells...they would be drained down to 0.68V per cell....and you can likely say goodbye to those cells.
Finally, if you use a higher voltage bulb like the WA-1166 and tell AW to set the Vbulb at 12.3V, that would likely result in the low battery warning being set at about 11.9V (3% below Vbulb). Now if you have 3s Li-Ions, the D driver low warning would come on too soon at a per cell voltage of 3.97V (11.9V/3).
That last example is why I was asking if we can tell AW the Vbulb we want, and the Vlow-warning we want. In the WA-1166 bulb example being powered by 3s, I would want the low alarm set to trigger at about 3.4V/cell or 10.2V rather than 11.9V.
DM51
Flashaholic
With NsNp Li-Ion set-ups, you would be relying on not only the individuals cells' low-voltage protection, but their high current protection too. You would get a temporary surge through one or more cells – how temporary would depend on the accuracy/tolerance of the protection circuit settings. The high current protection is there principally to prevent damage from a short rather than to prevent currents that are beyond what is good for the cell. The low-voltage protection may also be tripping at a point which is lower than ideal for the cell, and there are bound to be variations in the individual circuits.
When the first cell trips its low-voltage cut-out, current will then be drawn from its parallel neighbor(s) but at a higher rate, as demanded by the bulb and the other stacks in the pack, at the precise time when this would be undesirable from the cell's POV. This would continue until the next cell's low-voltage protection kicked in. If it was a 2p configuration that second trip would obviously cut all voltage from the pack, but if it was 3p there would be an even bigger current surge through the remaining cell until the high-current protection cut it off.
This would be happening when the cells were nearly depleted and the last thing one would want to do to them would be to suddenly start draining them even harder, even if only for a brief period.
It is placing quite a heavy reliance on the cells' individual protection circuits, and for this reason as well as others already given above, it would probably be unwise to use a configuration where Vbatt is much higher than the Vbulb setting.
When the first cell trips its low-voltage cut-out, current will then be drawn from its parallel neighbor(s) but at a higher rate, as demanded by the bulb and the other stacks in the pack, at the precise time when this would be undesirable from the cell's POV. This would continue until the next cell's low-voltage protection kicked in. If it was a 2p configuration that second trip would obviously cut all voltage from the pack, but if it was 3p there would be an even bigger current surge through the remaining cell until the high-current protection cut it off.
This would be happening when the cells were nearly depleted and the last thing one would want to do to them would be to suddenly start draining them even harder, even if only for a brief period.
It is placing quite a heavy reliance on the cells' individual protection circuits, and for this reason as well as others already given above, it would probably be unwise to use a configuration where Vbatt is much higher than the Vbulb setting.
