FatFlex Prototype: Part 2

Continued from here.

Will there be a special edition with the bicycling specific programming? (UBC I think(

Hello Geoges,

Email sent for two MaxFlex converters :naughty:

Does this mean there is a waiting list or they're available again? Where do I send payment!

Re:maxFlex Prototype: Part 2

Ok - after a bunch of testing the rework appears solid and comes down to removing one capacitor and soldering a replacement across the uC.

I've run the board with 4 LEDs and driven 17W through it (1A to the LEDs) and it no longer resets/crashes - mystery finally solved & fixed.

I have ported the current UIF/UIP/UIB firmware from the nFlex/bFlex platform over to maxFlex and have started testing it and all is looking good.

I have the UIB2 firmware ported, but not tested but will look into that after I've got a few panels reworked and programmed with the UIF/UIP/UIB firmware.

UIF/UIP/UIB will be much more appropriate for flashlight use and for those that like the UIB for bikes.

UIB2 is pretty well bike specific and is the latest version that is shipping with nFlex/bFlex.

There is no 'waiting' list or pre-order list. I have (I think) around 150 maxFlex drivers on hand, so more than enough to take care of initial demand.

cheers,
george.

Re: maxFlex Prototype: Part 2

georges80 said:

Ok - after a bunch of testing the rework appears solid and comes down to removing one capacitor and soldering a replacement across the uC.

I've run the board with 4 LEDs and driven 17W through it (1A to the LEDs) and it no longer resets/crashes - mystery finally solved & fixed.

I have ported the current UIF/UIP/UIB firmware from the nFlex/bFlex platform over to maxFlex and have started testing it and all is looking good.

I have the UIB2 firmware ported, but not tested but will look into that after I've got a few panels reworked and programmed with the UIF/UIP/UIB firmware.

UIF/UIP/UIB will be much more appropriate for flashlight use and for those that like the UIB for bikes.

UIB2 is pretty well bike specific and is the latest version that is shipping with nFlex/bFlex.

There is no 'waiting' list or pre-order list. I have (I think) around 150 maxFlex drivers on hand, so more than enough to take care of initial demand.

cheers,
george.

Click to expand...

Awesome! So how much and when can we place orders?

I am ready! When and where?

When they are ready for order they'll be on the order page of my website. Link is below.

Pricing will be there as well.

Just ran the 1200mA range and it is stable and happy.

15.28V out at 1190mA
11.74V in at 1.66A

93.3% efficiency. Not bad at all - due to the relatively close input/output voltage and keeping the input current below 2A.

I've seen down to 85% efficiency with bigger input/output voltage difference and closer to the 2A level.

I'll publish up some efficiency numbers in the maxFlex tech area over the weekend.

cheers,
george.

George,

Have you had time to run the MaxFlex on a NiMH pack yet, that's where I experienced the flakey problems.

I still have two from the first run, one in a Tri-LED Mag using 2 Li-Ion cells and it's 90% stable. I'm only experience problems when I try and latch to the brightest setting sometimes it turns off in a few seconds but you can turn it right back on.

The other is driving (5xLuxIII's) from a 35Ah SLA, that set up has been rock solid stable from day one.

I don't know what was up with the NiMH packs, it still has me wondering as I tried several different holders and variations :shrug:

Later
Kelly

Sway said:

George,

Have you had time to run the MaxFlex on a NiMH pack yet, that's where I experienced the flakey problems.

I still have two from the first run, one in a Tri-LED Mag using 2 Li-Ion cells and it's 90% stable. I'm only experience problems when I try and latch to the brightest setting sometimes it turns off in a few seconds but you can turn it right back on.

The other is driving (5xLuxIII's) from a 35Ah SLA, that set up has been rock solid stable from day one.

I don't know what was up with the NiMH packs, it still has me wondering as I tried several different holders and variations :shrug:

Later
Kelly

Click to expand...

Hi Kelly,

No - I've run on a power supply - the issue isn't the battery type - it's the voltage/current/LED combination. Certain combinations create slightly more transients on the supply and due to the location of the bypass cap for the uController it was causing the controller to either crash or reset.

I had a rock solid set of crashing/reseting conditions on the bench and when I fixed the capacitor issue that board has been 100% solid at many voltage/current combinations.

That's all that was wrong with maxFlex from the get go - basically a transient on the supply crashing/reseting the uC. Now that it is resolved the board is solid. I've ported the latest UIF/UIP/UIB firmware from the nFlex/bFlex platform (the same code other than just a handful of instructions) so the firmware base is well proven as well.

And - the latching to high (causing a crash/reset in a few seconds) is EXACTLY the failure mode that this rework fixes. In fact that was the only failure mode...

I totally understand why it was crashing before and why the things I did on the bench to try and fix it actually made things worse. In hindsight it was obvious why it got worse - it just didn't "click" at the time. Basically it was the location of the ground side of the bypass capacitor - a noisy area of the ground plane. When I cut the ground plane on the edge of the PCB (sorta kinda forms a ring around the board) it made it fail 100% - basically ~12Hz strobe rate as it turned on, crashed/reset, turned off, restarted and repeated the cycle. Once I had it failing 100% I was able to probe some areas with my scope and the noise transients on the supply were then very obvious and what was needed to fix it even more obvious.

cheers,
george.

georges80 said:

I totally understand why it was crashing before and why the things I did on the bench to try and fix it actually made things worse. In hindsight it was obvious why it got worse - it just didn't "click" at the time. Basically it was the location of the ground side of the bypass capacitor - a noisy area of the ground plane. When I cut the ground plane on the edge of the PCB (sorta kinda forms a ring around the board) it made it fail 100% - basically ~12Hz strobe rate as it turned on, crashed/reset, turned off, restarted and repeated the cycle. Once I had it failing 100% I was able to probe some areas with my scope and the noise transients on the supply were then very obvious and what was needed to fix it even more obvious.

cheers,
george.

Click to expand...

I'm glad you do because everything you just said just blew right over my head :duh2:.

I actually UNDERSTAND what George is talking about! Does this make me a geek or what?

I ordered a MaxFlex last year and wondered do I need to re-order or will the original order go through? Either way, I want one (or two) of those babies and let my modding begin!

Thanks for all the work on the MaxFlex, always loved the nFlex LuxV Mag and want to build the Quad Cree Mag based on your designs. Good things come to those that wait... I see that now.

I see 10 available at www.taskled.com .....well make that 7 now

:nana:

Later
Kelly

Sway said:

I see 10 available at www.taskled.com .....well make that 7 now

:nana:

Later
Kelly

Click to expand...

Thanks for the heads up!

Make that 5 now .

Now 4!

Email sent for two units

edit : PP sent!

I have about 40 that I'm reworking today and they'll be available for shipping by tomorrow.

So, there should be plenty for folk that want some.

cheers,
george.

In the updated manual, looks like the max current drive is 1200mA instead of 1400mA before.

Could you please confirm and comment other changes that you have implemented?

Thanks!

tino_ale said:

In the updated manual, looks like the max current drive is 1200mA instead of 1400mA before.

Could you please confirm and comment other changes that you have implemented?

Thanks!

Click to expand...

Yes, max current has been set to 1200mA. This gives more margin for the dimming to work over a closer Vin to Vf ratio. It also is more conservative from a heat generating/efficiency perspective - especially when folk want to run multiple emitters.

Imagine 5 LEDs running at 1.2A. That's around 5 x 3.7 x 1.2 = 20W. Even at 90% efficiency that is 2W of heat that must be dissipated from maxFlex. Given the small board form factor, it is challenging even when mounting a heatsink/spreader to the provided solder pad area.

I decided to drop the 1400mA since it will just be even more challenging to dissipate the heat in that configuration.

The only other changes are the porting of the UIF/UIP/UIB firmware base that is currently shipping with nFlex & bFlex.

cheers,
george.

Thanks for the info.

I'm planning to run the maxflex at around 12W of output power, around 1W+ to dissipate... And haven't figured out how to do this in a 2D quad cree mod. I just don't know how to attach the converter inside the light so it's pad is heatsinked??

Some of you guys must have solved this before I guess, any idea welcome