MOSFET selection for Wight DD+7135 driver

Flashlight Modding and DIY Parts
1

Summary: The driver for the BLF Manker A6 has issues with extra flashes when switching from turbo to moon. I determined that this is due to the MOSFET selection. I am looking for a partner to fund and test drivers with 5 different MOSFET parts.

I built many of Wight’s DD+7135 driver and never encountered the turbo-to-moon flash issue. Comparing that with my home-assembled driver, I used a different MOSFET: PSMN6R5-25YLC instead of the PSMN3R0-30YL used in the Manker driver. I swapped the MOSFET in the Manker driver and the flicker disappeared.

Doing some research, I suspect the flicker issue is due to Cdv/dt induced turn-on.

When the FET is on in turbo mode, it causes the battery voltage to drop. When the FET is turned off in moon mode, the battery voltage recovers. This causes the change in drain-to-source voltage that triggers Cdv/dt turn-on. The white paper mentions that if the gate driver can only pull the gate down to 0.7V, then the Cdv/dt induced voltage is added to this 0.7V to make turn-on even more likely. I checked the ATtiny13A data sheet and the output low voltage is only guaranteed to 0.7V.

What makes matters worse is that the Vgs or gate turn-on voltage decreases when the junction temperature increases. So the longer the light is on turbo, the more likely that it will flash when switched to moon.

The PSMN6R5-25YLC that I use has higher on resistance and so the turbo current will be lower. But will it still be acceptable? However, I don’t have the equipment (top batteries, thick probe wires, lumens meter) to do accurate measurements.

I am proposing to purchase five different MOSFETs and build two sets of five DD+7135 drivers for testing.

Here is the parts list from Mouser. With the circuit board from OSH Park, the total BOM cost is about $36 shipping included.

I am looking for a partner with the proper equipment to help fund this project and conduct tests. I will hand assemble the boards and keep one set of 5. The partner would chip in $20 to cover the parts and shipping of the other set of 5 driver boards and is expected to perform tests and post the results here.

Priority will be given to a member that already have 4 or more of Wight’s DD+7135 boards ready to ship to me. I already ordered 12 boards from OSH Park, but it will take them time to make.

2

Since it looks like there is no interest in testing, I might turn this into a WTS after the drivers are assembled.

3

I'm sorry that there's no one yet to partner you up here, it sounds like finding more suitable Mosfets would be very useful.

I am able to do high current and lumen measurements but I'm not much into building these drivers up and not at all able to flash MCU's. And to be honest it is just not really my cup of tea. I hope that there is someone who will pick this up!

4

Have you tried the $0.01 solution of adding a 10K-30K pulldown resistor on the gate? That has solved the problem for me, where present (I don't think it's a big deal anyways).

There is also something that happens when the driver switches from phase to fast PWM that can cause a small blip.

5

Some boards are tight for space. I’m always looking to pack as much as possible into small boards.

6

I'd love to help but I'm always guilty of over-committing and am on a mission of doin BLF catch'n up now. Actually got unread/unresponded pm's down to 10 (was over 30). I am interested in this though. I'm gonna be building up a few FET+1 boards shortly with 2 different FET's I have on hand - I could test/post results on them?

7

Good idea and probably the best approach to the problem. Would require anyone wanting to fix their driver (after identifying a suitable replacement per your proposal) to reflow a new FET.

It's common for many drivers to have a resistor in series to the Gate and a resister going to ground (pull down). Adding a pull down resistor (as RMM suggested) is pretty easy. Adding a series resistor would be harder as you would have to cut a trace feeding the Gate and insert the resistor. Not sure off hand if there is space for that in that driver.

8

I am volunteering to build and program the drivers in case that is not clear. I just need someone who can do rigorous testing. If we can find a suitable MOSFET, then Manker or anyone else can substitute a less expensive part, without any other change, and get rid of the annoying (to some) flash.

Let’s be clear. The difference in on resistance between the MOSFETs I’m testing is a few milliohms. To put this in perspective, a spring is in the tens if not hundreds of milliohms.

9

Just use an 0402 resistor if you're really tight for space, or just solder a resistor from the gate pin to one of the ground pins on the FET (this is really easy on the LFPAK56 FETs). I think that changing to an FET with a higher turn on point and therefore lower performance in our applications just to get rid of a small blip is silly when you can fix it by adding a simple $0.01 resistor. I don't think that a momentary blip on a 0.1 lumen moonlight after blasting 1400 lumens is a big deal anyways, but that's how you fix it.

A series gate resistor will not fix the problem.

You need to look at the FET performance at the desired gate voltage and drain current, and gate-source voltage differential. In a single-cell application we never get the specified 4.5v gate voltage.

10

Richard - can you go as low as 10K ohms? I thought there was concern about losing amps? Wondering because with 20K, I still see the flash intermittently. Haven't experimented/tested as much as I'd like.

11

Richard brings up the real reason I am looking for rigorous testing. I need to convince Richard and other hot-rodders, like myself, that a FET replacement would not result in sacrificing noticeable lumens. The benefit beyond getting rid of the “blip” is significantly lower cost. The current FET costs $0.99 whereas a replacement might cost just $0.64. While that might not seem like much if you are building a few drivers by hand, but it can translate to $1 for a mass-produced item if you figure in the markup for profit.

If the popularity of the BLF Manker A6 will lead to future GBs using the same driver, I think the vendor would appreciate the cost reduction and maybe even pass on the savings to us.

I’m not advocating getting the cheapest FET, but one that sacrifices as little performance as possible while giving blip-free behavior without additional parts or firmware change with a bonus modest cost reduction.

12

I am unsure if such effort is worth it…
I didn’t even noticed it before someone pointed it out in the A6 thread.

When does this preflash disturbs you? I use the 7 mode with memory and it is no problem to see such a blip after coming from turbo.
It’s another thing if it would happen when switched on in moonlight mode, but I just tested it and I couldn’t see it.

I have other fet lights but they are either eswitch with an off mode between turbo and moon or don’t even have moon.
All have pulldown resistors and use the “old”70n03 fet.
I always wanted to build some of the fet+1 drivers but the fet is only available from expensive sources….
I always use gate pulldown because I fear that the tiny will fail and then the light would be on turbo, a failsafe you could say.

Try the 10k gate pulldown and if that does not help you could play a bit with the firmware, maybe a safe handling between turbo and moon would also help: turbo,off for a short delay, moon……

13

Have checked the price of shipping from mtnelectronics.com? Maybe even email RMM to see if he could reduce the shipping. I mean a few FET are small, should be able to fit them into a small padded envelope.

14

I got the parts order from Mouser and built one A17HYBRID-S driver with a PSMN2R4-30YLD MOSFET with tk’s driver with A17HYBRID (not A6) calibrations. The host is a blue Convoy S2+ that just came in. The LED is an XM-L2 U2 1A on Noctigon. The cell is a fresh laptop pull. I was only able to see a glitch switching from turbo to moon once out of more than 30 tries.

The datasheets show that the PSMN2R4-30YLD has slightly lower RDSon than the PSMN3R0-YL in the A6 driver and better dynamic characteristics such as rise and fall times (tr and tf).

I am waiting for an order of 30Q cells to perform some rigorous testing. I also haven’t had time to build the other 4 drivers with different MOSFETs.

So I have shown that proper FET selection can eliminate the turbo-to-moon glitch. I still need to prove that performance is not sacrificed. But from the datasheets, I don’t see why the PSMN2R4 can’t supply as much current if not more than the PSMN3R0.

15

Keep up the good work :-)

I have no real clue what is happening, but as an end user I will be happy with a better (and cheaper, and well available?) FET.