Direct Drive P7 M@g Experience

[sami_voodoo]

Alan

Just a question, how do you measure the current across the P7 LED? I read on another thread that there is a voltage drop caused by the ammeter which will underestimate the current flow. Is this true? How did you correct for this underestimation?

 

[Alan B]

Alan

Just a question, how do you measure the current across the P7 LED? I read on another thread that there is a voltage drop caused by the ammeter which will underestimate the current flow. Is this true? How did you correct for this underestimation?

This is correct. Actually there are several issues here as we have removed the tailcap and spring and substituted a meter and probes. If the meter and probe resistance is equal to the tailcap then the measurement is the same, otherwise there is some error.

In this situation the voltage across the LED can be measured in both configurations (with meter and with tailcap) and if the voltage across the LED is the same then the currents are the same. If not the LED voltage versus current graph can be used to estimate fairly accurately the change.

I have not done that. I did use a high quality meter, which generally equates to a low drop. However I suspect the probes are more of a drop than the meter. So using a high quality meter (or better yet a current shunt) and heavy short leads is better, or a clamp-on DC ammeter and a very short lead.

If you are operating near the maximum current of the LED it is more important to measure this accurately than where we are operating in the 2 amp area.

 

[sami_voodoo]

Thanks for clearing that up. I was in the process of making a P7 2D Mag. I wanted to run the P7 a bit harder than last time as I'll be using it myself.

Two ways of doing this are by using either one of Taskled's hipCC driver boards, or one of Der Wichtel's buck convertors. In as far as I can ascertain, Der Wichtel's design is a constant voltage source, while the hipCC is a constant current source. Der Wichtel's convertor will thus need a bit of voltage tuning to make it run at 2.8 Amps depending on the voltage bin of the LED. The hipCC I think is sort of plug and play - it'll give 2.8 amps (depending on the sense resistor) as long as you've got above 4.5 volts at the input. Am I right?

Do you think there's a reliable way of going for a DD configuration with an I-bin P7 and two 18650s in parallel? Two 18650s in series for a drive efficiency of 85% with either driver (albeit with flat regulation) seems to be a "waste". What do you think? Would you advise for DD or against it?

Thanks for any suggestions!
Sami

 

[Alan B]

The current will change with temperature with a constant voltage setup. So it is not quite as nice as a constant current regulator.

Direct Drive with a small resistor will probably produce higher efficiency than most drivers, but the current will change with battery condition and temperature.

I think direct drive is very good for a bit less than full power. If you want full power then regulation is a good idea to keep from frying the LED on a hot night with full batteries on a long full power run.

Direct driving the I-Bin will probably require a small current limit resistor, but not a lot of power wasted. This was discussed earlier in this thread if I recall correctly.

 

[sami_voodoo]

Hmmm, I think I'll try to get a CC regulator then.

Oh and I remember everything about the resistor mod! :thumbsup:

Thanks a bunch,

Cheers!
Sami

 

[Alan B]

Hmmm, I think I'll try to get a CC regulator then.

Oh and I remember everything about the resistor mod! :thumbsup:

Thanks a bunch,

Cheers!
Sami

Of course!

 

[Justin Case]

Brand new 3D Maglite
Hardwired switch, 22 gauge, stranded wire
DSWNJ-bin P7
3 x AA Eneloops in Tenergy D adapter shells, cells freshly charged in the NC-MQN06U Eneloop charger
Shiny 1976 copper penny (freshly cleaned with toothpaste) inserted between the last and second to last coils of the tailcap spring to enable electrical contact with the AA Eneloop

2.8A initial current measured at the Mag tail. About 23,000 lux for the hot spot at 1 meter using the stock smooth reflector.

 

[Alan B]

Brand new 3D Maglite
Hardwired switch, 22 gauge, stranded wire
DSWNJ-bin P7
3 x AA Eneloops in Tenergy D adapter shells, cells freshly charged in the NC-MQN06U Eneloop charger
Shiny 1976 copper penny (freshly cleaned with toothpaste) inserted between the last and second to last coils of the tailcap spring to enable electrical contact with the AA Eneloop

2.8A initial current measured at the Mag tail. About 23,000 lux for the hot spot at 1 meter using the stock smooth reflector.

Nice data point!

I take it those AA to D adapters use 1AA per D?

 

[Justin Case]

Yes, the adapters take one AA per shell. The Tenergy D adapter shells are very similar to the Eneloop adapters, except the Tenergys use black plastic with a silver adhesive label instead of the white plastic with blue lettering of the Eneloop adapter.

Before I charged up the Eneloops, I also made some tailcap measurements with the partially discharged cells. I initially measured 1.9A and 18,900 lux hot spot at 1 meter. After 2 minutes of warmup, I got the same current and lux measurements. Unfortunately, I didn't measure the resting voltage of the cells.

For the hot off the charger run, the 2.8A and 23,000 lux measurements were the initial readings, which probably took about 30 seconds to complete. I didn't do the 2 minute warmup measurement, though.

I'd also like to do a longer run of say 10 minutes to see how the Eneloops hold up and also try some D alkalines.

 

[nmos]

I just did my first 2D Mag upgrade last month. I got a P7 (CSXOI) and some 3AA->1D adapters from Kaidomain and a heatsink from H22A. These are series adapters. I thought I might have to use them in series and make some simple current limiter but it worked surprisingly well wired up directly with the battery packs in parallel. Current at the tail cap seems to be between 400-800ma which is pretty tame by the standards around here but it's really transformed my Mag from something that was better suited as a club than a light to the first thing I grab when going out for nighttime walks or checking out bumps in the night.

 

[HarryN]

So power lost in the resistor would be I²R or 2.8²x0.1=0.78 Watt.

Or, in other words, the system is (2.8*3.25)/(0.78+(2.8*3.25)) = 92.1% efficient. Assuming a Vf = 3.25 and If = 2.8 A.

Not bad at all IMHO Now I know what I'll be doing as a future project.

Thing is, until now I was dismissing the resistor mod as a band-aid measure.

Cheers!
Sami

So, after all this time, I was ahead of my time with resistored setups. :laughing: Ok, maybe not really.

I would like to come back to the resistance a bit more, specifically, a realistic amount of resistance that can be reproduced. My DMM of course provides multiple digits for the resistance measurement, but I always just ignore most of them, as it takes very little to move the meter 0.1 ohms.

Is it really realistic to think that a "normal" solder joint on a SMT resistor will really provide the circuit with 0.1 or even 0.01 ohms? Or is this approaching imagination? I would have guessed the solder itself would have that much resistance in a joint, but I am still figuring EE stuff out.

Thanks

Harry

 

[nmos]

Is it really realistic to think that a "normal" solder joint on a SMT resistor will really provide the circuit with 0.1 or even 0.01 ohms? Or is this approaching imagination? I would have guessed the solder itself would have that much resistance in a joint, but I am still figuring EE stuff out.

Maybe .01 ohm but even that sounds high to me. I'd expect the switch, tailcap/spring etc. to pretty much swamp a few solder joints.

 

[HarryN]

Maybe .01 ohm but even that sounds high to me. I'd expect the switch, tailcap/spring etc. to pretty much swamp a few solder joints.

Hi - Thanks for that. I guess I am talking about on a PCB itself for a driver circuit, not the losses in a flashlight body. I guess solder joints are considered quite low resistance - must be less than 0.01 ohms then I guess.

Thanks

Harry

 

[shdwkeeper]

Have you guys tried to Direct Drive an DSWOI P7 with an 18650 Lion? Is that recommended or is the voltage to high. Don't want to fry this LED.

Thanks.

Precisely. Adding a little impedance on direct drive can still produce a very efficient system. Regulators are nice, but not 92% efficient very often.

Using the "J" bin P7 with direct drive does make for a very efficient system, both in terms of electrical loss, and in terms of the LED efficiency going up a bit with lower current. However the Jbin does not quite reach full output with the batteries I have tried thus far. However 2.0 amps is not a bad place to operate, it makes the heatsink's job a lot easier and puts out a lot of lumens...

 

[supasizefries]

Thought I'd contribute my results for my latest flashlight build.

The Light:

• 2C Mag using stock spring *see comment below*
• One IMR 18650, (used as received from Lighthound. ie. did not charge 3.8-4.0V)
• SSC P7 DSWOI bin
• Britelumens heatsink
• Teflon wire used for stock switch and LED
• UCL glass, stock reflector (cut and bored)
• homemade 18650 battery sleeve (ie. ghetto PVC pipe)

Measurements with Radioshack DMM :

• Voltage across LED 3.32-3.35 V
• Tailcap: ~2.4 A (maybe due to resistance of DMM?)

Comment: I wedged the bulb tower spring with the tail cap spring to bridge the gap when using one IMR 18650. This is temporary until I find a permanent solution.

I will try to measure current in series with LED when I get the chance.

This is my first DD build. So far, I'm lovin the simplicity of direct drive!

Edit: Thought I'd add that it's roughly as bright as my Der Wichtel regulated 2D Mag P7 (DSXPJ) 6 x Eneloops.

 

[Alan B]

Sounds like a good build. The resistance of the DMM will lower the current a bit, so you are probably at or over the rating, and this with a not fully charged cell. It may be a bit hard on the LED with fully charged cells. You may not want to reduce the resistance further.

 

[supasizefries]

Sounds like a good build. The resistance of the DMM will lower the current a bit, so you are probably at or over the rating, and this with a not fully charged cell. It may be a bit hard on the LED with fully charged cells. You may not want to reduce the resistance further.

Decided to try a fully charged AW protected 18650. I was measuring about 3.1-3.2 amps at the tail. With the heat sink between my thumb and pointer finger, it got pretty damn hot, almost untouchable hot, inside a minute or so. I also tested with an el cheapo 18650 (fully charged) that shipped with my Solarforce L2 and it was only pulling about 1.8-1.9 A from the tail. I might have to invest in some regulation to keep current in check when using quality batteries.

 

[Alan B]

Decided to try a fully charged AW protected 18650. I was measuring about 3.1-3.2 amps at the tail. With the heat sink between my thumb and pointer finger, it got pretty damn hot, almost untouchable hot, inside a minute or so. I also tested with an el cheapo 18650 (fully charged) that shipped with my Solarforce L2 and it was only pulling about 1.8-1.9 A from the tail. I might have to invest in some regulation to keep current in check when using quality batteries.

It is really excellent to see these experiences. There really are important differences in components for direct drive systems. In your setup a little resistance is needed, either in the cell or you could add some in the light. It doesn't take much to control the current. If the LED is in control things will run away as it heats up, and the Vf of the LED drops a bit, so you want enough resistance to control the overcurrent.

 

[bshanahan14rulz]

What's a good rule of thumb when coming up with a resistor value to DD with? Like, if I am planning on running off of a single Li-Ion cell, do I use 4.2V to calculate my resistor? or 3.7V? Or should I pick a value like 4.0 or 4.1V?

 

[Alan B]

I would use 3.7V. The amount of energy/time above 3.7V is very small and the LED/heatsink should be able to handle that in most situations.