Review/Overview: LumensFactory Seraph P7 M Series Turbo Head

[stienke]

This was supposed to make my M6 "day-to-day useful" (currently my M6 is 1185), but as it turns out I lucked into a 3x18650 Megallenium (arriving tomorrow I hope, maybe Thursday) so this is now permanently installed on my M6 body with 33x17670.


Have you picture's from this "rare" M6 body???

[Justin Case]

A better graph to see how well the driver functions in terms of regulation is to graph power vs voltage. When you do that, the curve looks horrible. In comparison, this is what a tightly regulated buck driver delivers for power vs voltage. I guess this convinces me not to get the LF Seraph P7, or at least I would get rid of whatever driver they are using. It is highly doubtful that the P7 is getting 2800mA between 9V-13V input. The input power is only about 8W-9W for that voltage range. Even if we assume a relatively low Vf of 3.3V at 2800mA, that means that the P7 draws around 9.24W at full power. Since drivers are not 100% efficient, the power in has to be greater than 9.24W. Since that is not the case from 9V-13V, I don't think that the P7 is running full tilt if you drive the Seraph with 3xLi-ion.

It may be that at lower input voltage, the P7 gets full power. But the ugly power vs voltage curve doesn't inspire a lot of confidence. Thus, I wouldn't assume full power at Vin from around 5.5V-7.5V unless I actually measured that to be the case.

[mwaldron]

Originally Posted by stienke

Have you picture's from this "rare" M6 body???

It's a BEAST!

Yeah, sorry that was a typo I missed. It's a standard M6 body with a FM 3x17670 holder.

Originally Posted by Justin Case

A better graph to see how well the driver functions in terms of regulation is to graph power vs voltage. When you do that, the curve looks horrible. In comparison, this is what a tightly regulated buck driver delivers for power vs voltage. I guess this convinces me not to get the LF Seraph P7, or at least I would get rid of whatever driver they are using. It is highly doubtful that the P7 is getting 2800mA between 9V-13V input. The input power is only about 8W-9W for that voltage range. Even if we assume a relatively low Vf of 3.3V at 2800mA, that means that the P7 draws around 9.24W at full power. Since drivers are not 100% efficient, the power in has to be greater than 9.24W. Since that is not the case from 9V-13V, I don't think that the P7 is running full tilt if you drive the Seraph with 3xLi-ion.

It may be that at lower input voltage, the P7 gets full power. But the ugly power vs voltage curve doesn't inspire a lot of confidence. Thus, I wouldn't assume full power at Vin from around 5.5V-7.5V unless I actually measured that to be the case.

I never thought about looking at it from that perspective, I just took them at their word, but you're right as soon as you graph power as a function of voltage there is obviously something wrong. It's not really possible they the P7 is being driven at 2.8A through most of the voltage input range.

It's still a very impressive head and I'm glad I was able to get one.

For those who don't feel like retyping the data, here is the graph Justin is talking about:




I find it interesting that no where in that range did I ever notice a change in output, but my guess is it was never a significant enough step for my calibrated human eyeball to detect, except for the increase at 11.4 that I mentioned earlier.

[mdocod]

Interesting test results... Man, I need to get a bench power supply!

I just tested my module on 3 17670s charged to ~4.1V each. 0.95A consumption. Under that load I know for a fact that the voltage is ~3.7V per cell or better, meaning 11+V@0.95A or ~10.5W or better (probably closer to 11W).

I'm wondering if the drivers are that inconsistent from unit to unit, or if this is a combination of factors including the Vf of the emitter.

Assuming a Vf of 3.6V and a driver efficiency of 90%, we should expect ~11W of power consumption at all input voltages to maintain ~2800mA to the LED. My demo unit seems to behave in this fashion given the tests I have the capacity to run.

[mwaldron]

I'm using an old BK Precision 1635 (have a pair of them) and they're great. You can pick them up in working condition around flea markets and hamfests for about $100. They're built like tanks, but the disadvantage is they do not have the current capacity for most incandescent bulbs.

If you have a pair they do support parallel operation to double the current, but it complicates things quite a bit.

For my tests, I let the LED "warm up" at 8V for 2-3 minutes before I dropped the voltage down and started taking measurements. I ran the current through my Fluke 29-II meter for measurements, and used the readout on the supply for the voltage.

[Justin Case]

The problem is that you don't have any measurement data to support your contention that driver efficiency is 90%. There are infinite sets of Vf, If, and efficiency combos that can give the same results. Efficiency could be 90%. But the apparently poor driver QC and the terrible power in vs voltage in curve from mwaldron certainly doesn't inspire any confidence.

The way to resolve the issue is to measure the drive current and Vf. You don't need a bench supply to do that. True, the efficiency calculation will be for one specific Vin/Iin point. But if the driver is any good, it ought to be stable and show consistent behavior across its rated voltage. On the other hand, IMO this driver's stability and consistency are open to question.

[Justin Case]

The last time I saw a Pin vs Vin graph vaguely similar to this one for the Seraph was when I tested some DX11836 drop-ins. Same odd dip and rise at the higher voltage end of the graph.

Seraph


DX11836 #1


DX11836 #3

[Mikey V]

Originally Posted by digitaldave

To be honest, I was planning on getting the 2 cell version with extenders, as that would still give the 3 cell capacity option (or as you have the 2x18500), but I could also use it as a 2 cell light with a D26 or D36 module. But if I'm likely to be using the turbo head on it, then the three cell option makes more sense I guess.

How is the performance on 2x18500, in terms of output and run time? And have you used a D36 head on it?

Thanks,

Dave.

Dave, I haven't tried the D36 head. I ordered the turbo head together with the three cell light as it was my intention to use as a turbo light all long, and I just chucked the regular D26 head back in the box. I already have tons of P60 size lights. I think you would be disappointed with the balance of a two cell version hanging that big head onto it. I can't see how the grip and balance would work. It feels nose heavy and rather short on the grip department even using the three cell tube. With extenders, it would work, but "Rube Goldberg-ness" increases. Consider the two-cell light only if you really need to switch off to a light in that size, as you'll probably want to leave the turbo fully assembled once you try it. The 18500's are outstanding - as bright as a pair of 18650's. They of course will not run quite as long as 18650's, but then, I don't have to swing a real long tom around. (For that, I have a Leef 2x18650 tube with a MC-E Turbo KT-4 Clone head and Surefire tailcap- basically a digital M4 Devastator.) I haven't tested run time, but they are still on the first charge cycle and I have been playing with them for a few weeks. This Seraph modular light 2x18500 setup is a bit smaller than a Surefire M3T if you need a reference.

[mwaldron]

Originally Posted by Justin Case

The way to resolve the issue is to measure the drive current and Vf. You don't need a bench supply to do that. True, the efficiency calculation will be for one specific Vin/Iin point. But if the driver is any good, it ought to be stable and show consistent behavior across its rated voltage. On the other hand, IMO this driver's stability and consistency are open to question.

I thought about this, but I'm just not prepared to remove the LED from the driver to take readings, so I'm going to have to bow out of assisting on those numbers.

The next time some nice warm P7's come up available I'm going to grab one and at that point I'll take the numbers since I have to disassemble it anyway.

[Justin Case]

I don't have one of these heads, so I can't give you specifics on measuring drive current. But in general if you can separate the pill from the reflector, then all you need to do is de-solder say the LED+ wire from the LED, insert your DMM in series at that point, and then power the pill (either with batteries, or in your case with your BK power supply). Read off the drive current from the DMM.

For forward voltage, if you have access to the LED's terminals (e.g., you can remove the reflector to expose the LED), then just power the light while measuring voltage across the LED's terminals using a DMM.

[mdocod]

Originally Posted by Justin Case

I don't have one of these heads, so I can't give you specifics on measuring drive current. But in general if you can separate the pill from the reflector, then all you need to do is de-solder say the LED+ wire from the LED, insert your DMM in series at that point, and then power the pill (either with batteries, or in your case with your BK power supply). Read off the drive current from the DMM.

Easier said than done on this unit. Things are packed up pretty tight....

[Illum]

Originally Posted by uhsodium

would it be great with a 2-stage tailcap?

the A2's "two stage" part involves a wee little resistor, at 1.5A I'd expect that resistor to fry in short order

I'm contemplating 2x17500 + P7 head on my M3T body...but... $130 is a bit

[rokspydr]

will this head fit on the megalennum body?

[mwaldron]

Originally Posted by Illum

the A2's "two stage" part involves a wee little resistor, at 1.5A I'd expect that resistor to fry in short order

In another thread AW has stated that his multi-stage incandescent M6 switch works well with teh LF P7 head. I do have that switch and I can verify it does work, but I have some concerns about longevity issues. I'm concerned that PWMing (what the AW Switch does) a regulator (contained in the LF head) might not be good for the long term reliability of the regulator.

I will qualify the above by saying that I have absolutely no proof, and I made the statement in hopes someone with experience designing buck regulators to explain why I'm wrong and why PWMing their power source won't hurt them.

Originally Posted by rokspydr

will this head fit on the megalennum body?

In fact it fits well, although I prefer the aesthetics of it attached to the M6 body. The megalennium is where my AW switch is installed so that is where the above testing occurred.

Putting the Seraph on my M6 body with 3x17670 after getting my Megallenium allowed me to have 2 awesome lights "of m6 pedigree" and they're both guilt-free lumens, the best kind!

[wacbzz]

Just ordered mine.

Hopefully I'll get it by Thanksgiving so I can one-up the Father-in-Law in yet another of our ongoing "war" catagories - - brightest handheld flashlight.