new P7 driver 3amps?

[Meterman]

Well here are the measurements in a excel file.

All measurements where made with calibrated meters.

Hi SUBjohan,

Your excel file is a very good job!

I like precise measurements, in German I say "WER MIT MIST MISST, MISST MIST." It's a pity that in translation the play on words gets lost.

How do you measure currents? My new AC/DC current clamp (Chauvin Arnoux K1) has a span from 1mA to 4500mA and very small jaws.

Wulf

[StefanFS]

In my experience, when buying cheap, "made in China" stuff, two units of the very same item produced in China will, many times, have similar, but somewhat different components. That the supposedly same driver, received by two different people, works differently should not come as a total surprise.

My suspicion is that you will not get a clear, understandable answer on this one, if any.

That can be true sometimes. I have been lucky with the stuff from China I have been using so far, but I try to select what seem to be good quality. Buying cheap has different meaning in different parts of the world, I know that it can sometimes mean 'trash' or low quality in the US. Here cheap (billigt) can also mean that you made a bargain. I think there might be different UI versions coexisting and that the seller missed this detail, and the versions are mixed. The driver works well and seem to be very well built, it's the software that differs in the sense that my units have no memory function.
Stefan

[SUBjohan]

Hi SUBjohan,

Your excel file is a very good job!

I like precise measurements, in German I say "WER MIT MIST MISST, MISST MIST." It's a pity that in translation the play on words gets lost.

How do you measure currents? My new AC/DC current clamp (Chauvin Arnoux K1) has a span from 1mA to 4500mA and very small jaws.

Wulf

Yes, in The Netherlands we have the same kind of saying "Meten is weten, maar dan moet je wel weten wat je meet".

I used 4 GossenMetrawatt Metrahit 29S multimeters so could measure everything in the same time.

Greetz Johan

[Meterman]

I used 4 GossenMetrawatt Metrahit 29S multimeters so could measure everything in the same time.

That's really a bunch to have a lot of precision at the same time!

Of this series I employ a Metrahit 27M, specially designed for measuring mainly mΩ and μΩ.

But I'm afraid we go OT.

Wulf

[manne]

Thanks for the usefull measurements SUBjohan, the poor efficiency on med and low looks like they use a improper way to dim, like simply a PIC (12f629 or something similar) that PWM the input voltage per MOSFET oder regulator's enable pin, can you confirm this? So maybe it should be possible to modify it to one mode?

[TorchBoy]

... the poor efficiency on med and low looks like they use a improper way to dim, ...

Add to that, I'm disappointed the medium isn't more logarithmically placed between the other two, but that wouldn't suit everyone either. It would be nice to be able to set them wherever.

[Galiphrey]

That drop of efficiency for medium makes it look like the driver is loosing the same amount of power whether in medium or high mode (instead of loss being as a proportion of the mode).

[spencer]

Mine has a memory function. Just got this driver and so far I like it.

Excellent work on the measurements SUBjohan

[DocD]

So i've been reading the replys and it seem to be a positive towards these driver's but they are not very efficent on med and low ? in the real word does this shorten run time by converting power into heat ? how does it hold up against other drivers ? cheers DocD

[StefanFS]

So i've been reading the replys and it seem to be a positive towards these driver's but they are not very efficent on med and low ? in the real word does this shorten run time by converting power into heat ? how does it hold up against other drivers ? cheers DocD

They seem to work well enough. About efficiency on low & med real world runtimes needs to be performed in addition to the calculations made.

Stefan

[mtnbkr1]

Is there a recommended way to deal with the heat for this KD driver?

[StefanFS]

~10 hours on low is what I got from six fresh eneloops in an mdocod battery holder in a 2D Mag. Not going to bother with medium, it drained the cells to ~0.9 Volt.

[nightstalker101]

Is there a recommended way to deal with the heat for this KD driver?

Ya, I'm not sure either, I ran mine for 15min on high and it was too hot to touch.

[Galiphrey]

Ya, I'm not sure either, I ran mine for 15min on high and it was too hot to touch.

Which area gets hot (or just the whole thing?) ?

[nightstalker101]

Which area gets hot (or just the whole thing?) ?

I'm not really sure which area heats up, however the whole thing ends up getting hot. I'm not sure if thats a bad thing, however its questionable.

[Galiphrey]

>>-Auto-alarm towards high temperature,then switch to low mode automatically

I wonder how hot it has to get before the "auto-alarm" kicks in.

[SUBjohan]

>>-Auto-alarm towards high temperature,then switch to low mode automatically

I wonder how hot it has to get before the "auto-alarm" kicks in.

I left mine running on high at 10V for 30mins and the auto-alarm did-not kick in. The driver was in free air.

Greetz Johan

[Galiphrey]

I left mine running on high at 10V for 30mins and the auto-alarm did-not kick in. The driver was in free air.

Greetz Johan

I've received and installed mine now; I potted it to the underbelly of a P7 heatsink. It DOES "auto-alarm" after about 60 seconds of being on high. That is--it switches to the low mode. (I can begin to feel the heat on the outside of the flashlight body.) Then, apparently after cooling off for some seconds, it switches back to high again, and then repeats this later on... It does NOT do any of that after many minutes of medium-level. And if I'm reading your measurements right, the board itself produces the same amount of heat either way (high vs medium), so the difference then is the heat from the LED and through the heatsink, which (I assume) is being detected by the board. Unless I'm off the deep end now, then it seems like potting it to an LED heatsink the way I did doesn't cool the board, but actually heats it up... (?)

[Packhorse]

What voltage are you running into it?

[Galiphrey]

What voltage are you running into it?

About 11.7 volts. Do you suppose it's more efficient at higher or lower input voltages?

Edit: ((My guess would be higher voltages))

[SUBjohan]

About 11.7 volts. Do you suppose it's more efficient at higher or lower input voltages?

Edit: ((My guess would be higher voltages))

Nope, the higher the voltage the less efficient it gets.
And over 10V the Watt losses go up verry quick
I think (not measured ) that the Watt losses around 12V should be 3,5 Watts wich get turned into heat.
At 6V the Watt losses are 2,1 Watts (I did measure this ) wich is 40% less heat.

Greetz Johan

[milkyspit]

Unless I'm off the deep end now, then it seems like potting it to an LED heatsink the way I did doesn't cool the board, but actually heats it up... (?)

You're right, that's not a particularly good idea. (I learned the hard way some time back!)

As for heat, just taking a wild guess but one source might be the inductor itself, particularly if the inductor is a little less efficient than it could be. That's a lot of power passing through the inductor and things could easily get pretty toasty. Inductors are a pretty simple component, but the subtleties of forming a good one make their construction kind of a black art. The really good ones are few and far between, and typically cost a heck of a lot more than one would expect given the materials involved.

Anyone have an extra piece or two of this board they might be willing to sell? I'll place an order with Kaidomain but would also like to grab a board or two in the near term to play with.

[Packhorse]

OK I just made up a heat sink for my board. Its basicly a 1 inch x 1inch aluminium channel. I drilled a 9mm hole for the inductor to sit in and thermal pasted it in place.
I ran it for 5 minutes off a 12V battery (battery is a little flat and was at 12v) and ran 2 P7s in series off it.


High 1.9amps 22.8watt output 2.6 amps at 6.7v 17.42 watt or 76% efficient (it did drive a single P9 to 2.8amp)
Med .60 amps 7.2watt Output .77amps 3.12v 2.4watt 33% efficient.
Low .21 amps 2.52watt Output.24amps @1.8v .432watt 17% efficient.

Med and Low seem way way inefficent. perhaps I misread. I will remeasure again later.

But the important thing is that I can run 2 P7s off of it at 12v & it did not go into heat protection mode. It got warm but not hot.

[DocD]

So the way i'm understanding this is that the driver, has hope for being able to drive 2 P7's or may more? but as a single driver,ok on high and rubbish on med and low.

[Phaserburn]

I know it's not as elegant a solution, but why not direct drive the P7 with three nimh D cells? No circuit losses or circuit heat issues to deal with. The nimh cells will hold voltage easily and provide around 2.8 - 3A to the emitter (with a vf of 3.25 - 3.5V), while giving excellent runtime to boot. I would think that dimming due to voltage drop in this scenario would be hard to detect by eye over the course of a long burn.

[Packhorse]

For me, I would rather have full brightness for the whole of the run time. Plus the low volt warning.
The 2 lower power levels and thermal protection are just bonuses.
Not to mention the ability to run it off an existing 12volt battery pack.

[1 what]

Re; What gets hot and how to manage it.
Been following this but just too busy to contribute. Last night I wired up my board and used my calibrated finger to see what got hottest first (an ancient and sensitive measurement often used back in the days of "valves").
Result:


Since this side of the board has components of different heights and this is one of the low profile ones it makes sense to build it up to the top level and heat sink to it. A small piece of 3mm Al plus heatsink epoxy works well:


I then epoxied the "new flat profile" onto a dicast box I'm using for a project I'll publish in the next week or 2 and ran it for 15mins on hi with 8V in. There were no problems and I could still touch the other side of the driver at 15 mins.

[Aircraft800]

I have the newer version of this driver, there is two components where you have the spacer, is the heatsinking the same? Has anyone had any experience with the newer driver?

Is this where I would need to heatsink?


Is this the correct wiring? I can use either negative batt in?

[DocD]

I have the newer version of this driver, there is two components where you have the spacer, is the heatsinking the same? Has anyone had any experience with the newer driver?

Is this where I would need to heatsink?


Is this the correct wiring? I can use either negative batt in?

do ypu have a link? cheers DocD

[Aircraft800]

do ypu have a link? cheers DocD

http://www.kaidomain.com/ProductDeta...ProductId=1866

Quote Kaidomain: "Measures 19mm in diameter, and 8mm in height (Note: the diameter changes to 17mm since Aug.28th (shipping day)."

I got the 17mm board, new design.