Seoul P7 direct drive concern

Are the dies on the P7 the same as used in a P4?


If they are, why could'nt we run each die in the P7 at 1 amp? Too much heat buildup I guess? :candle:

Yes, that's exactly right. It uses four P4 dies (dice?) and heat is the reason why it shouldn't be run at 4A

KrisP said:

Will 4C cells fit into a 3D? Does anything need to be modified other than needing a PVC tube?
I'm still worried that 3.8A is going to rapidly decease the LED's life

Click to expand...

Other than the 1" PVC tube, you can either flip the original spring upside down, or what I did was bought a small conical spring from the hardware store for $1 and Dremel'd away the anodizing inside of the tailcap.

KrisP,

As a response to your first post, a large increase in current draw can occur with a slight increase in voltage. Here is a table I made comparing the current draw with change in voltage (Vf-curve). I did this using a CC laboratory power supply (constant current capable). I tested the J-bin P7 here (3.25V-3.5V @ 2.8A):

(click image for larger view)
The columns on the left shows the forward voltage compared to current draw when the Pentium 4 heatsink (which the P7 is mounted to) has a 70mm cooling fan blowing a bit of air on it, while the columns on the right show it using no fan (just natural convection currents).

Just to see the logarithmic curve, here is the above data in a Vf vs. I curve:

(click image for larger view)
This graph also shows the power consumption to show how that line is not perfectly linear.

Do not forget that each die should be getting a fourth of the current (theoretically), so each die should be able to handle over 1A of current, however the thing we question with this Seoul P7 is whether the P7 emitter package can get rid of the heat produced by those dies quick enough. I tested my P7 at 5280mA since that is the maximum current that my CC power supply can handle (it is only rated to go to 5A). Each die was fine with 1320mA, but I have to have a strong fan blowing air across the heatsink so it would not over heat. Each die has a pair of bond wires each going to the anode and cathode, so they are electrically identical to a newer 4-bond wire Cree XR-E. Thermal conductivity is better per die since the dies are bonded directly to a copper slug (zinc plated?) versus the Cree XR-E, where the die is bonded to a copper pad, then a thin aluminum substrate.

If the P7 is bonded well to the heatsink in the Maglite, and the flashlight body is dissipating the heat okay, then I see no problem with running your P7 at 3.8A. In my P7 Maglite (2D with 2 parallel D-sized li-ion cells), I plan for a max current of 4A. I will have George80's D2DIM circuit to allow me to dim the P7 (via PWM) for the bulk of the flashlight use, with the 4A turbo mode (100% duty) for blasting through a field, or for show (with no more than 15 min of consecutive use). Of course this is with DD'ing the P7 with li-ions, which will have a slightly lower maximum voltage at full charge than the 3x NiMH cells.

For best results, I would stay below 3A (longer battery life, less heat). I usually like playing on the dangerous side by "overdriving" LEDs at higher currents than most would. I suppose that if you would shorten the life to a few thousand hours, you would never really notice a drop in output during use after years of use (unless you use your light all night while on duty or something), so why not get some "fun" out of it? For many modders (me included), they will switch out the emitter when a newer and better one comes out before the original goes bad. If you would damage it (leave it on over night next to a heater in the winter or something), then they are not too difficult and expensive to replace. However, if this is for a gift, then keep the currents low to err on the safest side. I guess that I mean that you should not run the P7 at currents that you are not comfortable with. The P7 will not die the instant that you send over 2.8A of current to it, so no immediate alarm. Thermal runaway with DD is a problem unless you know that it will not get out of hand with your choice of batteries or electronics.

I would like to hear evan9162 chime in about heat. I need to get a hold of him to see if I can get him a new Seoul P7 to test fully. Good luck and happy modding!

-Tony

darkzero said:

Forgive me but how do you get to the burst mode? I have a Triflupic running a P7 now but in a small light. This is my first Flupic and I'm not familiar on how to fully use it yet. I only know mode 3. I highly doubt I can get 3.6a out of a R123 anyways.

Click to expand...

darkzero,I posted in your other thread.Your mod caught my attention before I checked this thread.Level 1 and 2 are the ones programmable for output.Level 3 is Burst which is set at up to 1200ma,or whatever the batt/s are capable of,which may be less.

Here is the flowchart if you dont have it already,https://www.candlepowerforums.com/threads/123520

Gryloc said:

If the P7 is bonded well to the heatsink in the Maglite, and the flashlight body is dissipating the heat okay, then I see no problem with running your P7 at 3.8A. In my P7 Maglite (2D with 2 parallel D-sized li-ion cells), I plan for a max current of 4A. I will have George80's D2DIM circuit to allow me to dim the P7 (via PWM) for the bulk of the flashlight use, with the 4A turbo mode (100% duty) for blasting through a field, or for show (with no more than 15 min of consecutive use).

Click to expand...

Gryloc,great chart and information about the P7.
In my current setup I have the emitter bonded to the anodized heatsink with thermal epoxy,the driver is bonded chip side(all connections made on the backside) to an H22A heatsink with the lip removed,in between the switch and the emitter sink.Everything is assembled with a ton of thermal paste,including on the threads between the body and head.The body gets warm to about midway of the switch,the head gets as warm as the body in front of the switch.Head finning would be very beneficial in my opinion.It pretty much stabilizes after short while,but it does get warm pretty fast.I run it mostly around 2a.

Gryloc said:

For best results, I would stay below 3A (longer battery life, less heat). I usually like playing on the dangerous side by "overdriving" LEDs at higher currents than most would. I suppose that if you would shorten the life to a few thousand hours, you would never really notice a drop in output during use after years of use (unless you use your light all night while on duty or something), so why not get some "fun" out of it? For many modders (me included), they will switch out the emitter when a newer and better one comes out before the original goes bad. If you would damage it (leave it on over night next to a heater in the winter or something), then they are not too difficult and expensive to replace. However, if this is for a gift, then keep the currents low to err on the safest side. I guess that I mean that you should not run the P7 at currents that you are not comfortable with. The P7 will not die the instant that you send over 2.8A of current to it, so no immediate alarm. Thermal runaway with DD is a problem unless you know that it will not get out of hand with your choice of batteries or electronics.

Click to expand...

I totally agree,I plan on building some to give to family that will use the drivers with 7135 chips.Output will be limited then to around 2400ma.But in my current set-up the Flupic goes from ~2200ma to ~3600ma,with nothing in between.I would prefer to run it in between those currents,but nothings available yet.

Thanks for the good info.

Michael

Thanks for all the replies... I couldn't resist and ordered a tri-flupic for it

Texlite... I have a tri-flupic in another light and I was sure level 10 and max brigthness were the same. I set interface 1 to give 1, 10, MAX and I couldn't see any difference between 10 and MAX... Maybe my battery couldn't supply the current for MAX or something.

KrisP said:

Thanks for all the replies... I couldn't resist and ordered a tri-flupic for it

Texlite... I have a tri-flupic in another light and I was sure level 10 and max brigthness were the same. I set interface 1 to give 1, 10, MAX and I couldn't see any difference between 10 and MAX... Maybe my battery couldn't supply the current for MAX or something.

Click to expand...

For my sake I hope it was your battery,otherwise I've lost one channel,until it goes into DD/burst.I wont know for sure until I get another put together,or I pull my current setup apart and check.I was wondering that but hadn't asked anyone yet.Thanks for the info.

Michael

Saying an LED "if designed to run at 1A at 4V, give it 4V and it will take care of itself" is completely incorrect and false. If an LED is designed to run at 1A at 4V at Tj of 25C, THEN it is reasonably correct (ignoring production variation).... the problem is the moment you turn it on it starts heating up, and every C increase drops the Vf required to maintain a given current.... so you turn it on, and feed a constant 4V... it heats up 1C, and now it only needs 3.9V to run at 1A.... so now it's running 1.1A... and heats up another C, Vf drops, current rises... heats up more... that is the FUNDAMENTAL rule of LEDs and thermal runaway, running a constant voltage with disregard to current limits and temperature at the junction, and running at the "spec sheet" limit (ignoring the specs CLEARLY state something like "at Tj=25C") is to ignore at your peril. That's what the temperature derating charts on the spec sheet for, so that you can adjust accordingly (and/or run a constant-current driver and not have to worry about it).

TexLite said:

I totally agree,I plan on building some to give to family that will use the drivers with 7135 chips.Output will be limited then to around 2400ma.But in my current set-up the Flupic goes from ~2200ma to ~3600ma,with nothing in between.I would prefer to run it in between those currents,but nothings available yet.

Click to expand...

If you were to build an AMC7135 based SSC P7 driver from one 1A multimode driver with a slave sandwich consisting of two simple 1A AMC 7135 boards you get 3A. If you use one 1A multimode with one 1.4A + one 0.35A you get 2.75A.

These AMC7135 based drivers have similar behaviour compared to the FluPIC design, similar discharge curve etc.

I built a 3A driver that works well:
https://www.candlepowerforums.com/threads/195358

Stefan

I don't like the AMC7135 boards. I only use them for home lighting with 5vDC power packs because in that situation I don't mind them wasting power.

KrisP said:

I don't like the AMC7135 boards. I only use them for home lighting with 5vDC power packs because in that situation I don't mind them wasting power.

Click to expand...

Yes, a lot of people don't like them. That's fine. My point was that it's possible to combine drivers to get different currents, eg. 2.4A, 2.7A or 3A to the emitter. Others find them very useful. They provide a very cost effective solution for driving the SSC P7 at different currents. They don't waste power when used properly within the preferred voltage range with suitable cells. The multimode AMC7135 drivers share some basic concepts with the FluPIC design and to the end user the result is similar(like the discharge curve). Of course, the programming and timing is different. For example, all my FluPICs have a difficult time with forward switches due to the timing.
Stefan

Sorry, I didn't mean to sound rude in any way

I didn't think you were. I just wanted to be clear about my thinking/reasoning about those drivers. It's your thread I'm invading!
Stefan

KrisP said:

Hi,


Can anyone else with a P7 Mag, running direct drive from Ni-MH, please measure their voltage and current?


Thanks

Click to expand...

I built a M@g 3D P7 DD off 3 "D" size Accupower LSD batterries tonight. FiveMega MOP reflector. Fresh batterries off my Maha 808M give me 3.6 volts at the emitter, and 3.61 amps at the tailcap.

It's bright with a huge hotspot. I took it outside and compared it to two other M@gs I've built. The P7 is a little better than my tri Cree 3D M@g (R2's @ 1 amp via a Shark, McR19XR reflectors, and 3 D Accupower LSD's).

It won't get on my penta Cree :naughty:, but I didn't expect it to.

I ran this light for about 15 minutes straight and the head got warm, but not bad at all. I don't think I'll bother with finning the head unless it's for looks.

This is probably about as much light as I can get out of 3 NiMH's. Definitely cheaper than a tri- Cree as well.

The beam isn't perfect, it still has a small artifact right in the center, but it's hardly noticable.

When the Sandwich Shoppe comes out with a dedicated driver for this pill I'll put one in for flat regulation. But I want to stay around 3.5 amps current.

With the tri-flupic fitted now, the current draw at the cells is 3.1A on MAX with freshly charged cells.

would this work?

To get full power and a "dim" mode, could I use http://www.dealextreme.com/details.dx/sku.1886 (2 pieces) run in parallel? so w/ one operating, it would be 1400ma. To get full power (2800ma), connect the 2nd one in parallel. Would this work?

Assuming the switching is arranged appropriately, that should work fine - the 1400mA driver uses 4x 7135 run in parallel, and there's no reason that multiple boards can't also be paralleled.

if I put the switch on the battery side of the 2nd driver and it is off, would there be a problem w/ the led +/- leads getting current from the other driver's led +/-? Or would it be better to switch both the input and output off/on at the same time?
thanks for the help!