17mm & 20/26/27mm single-sided DD/FET driver release: A17DD-SO8 / A20DD-SO8 / etc

Thanks ToyKeeper, I’ll have to mull this over.

I’m getting a PH-08 eventually for a project, so I should be able to post comparison pics then. However, shipping from China (and everywhere else) is now in crazy-holiday mode so it might be a long time before anything arrives. Comfychair’s thread on Roche F6 hacking provides lots and lots of details though, if you’re curious.

I followed the thread pretty closely, although I didn’t participate at the time. I’d say that I do not find either flashlight particularly aesthetically pleasing, but it’s easy for me to assume that the Roche is built to a higher standard. $9 extra worth? I’m not sure about that. Also the PH-08 has the advantage of a separate button board.

5 pages of posts and I am still looking for 2S performance :slight_smile:

I guess that with the zener removal, we’re back to 1S solution for this driver too.

The regulator does the job of the Zener with less leakage current. So this is a 2s solution.

Zener performance doesn’t need to be tested. Due to the high gate voltage (which is assured by the high Vcc provided by the Zener) the FET’s performance is already guaranteed. [In 1s applications we do not have a regulated gate voltage, so it’s possible to get a low enough gate voltage to put the FET into a less efficient state. This was never a concern for Zener builds.]

I don’t think I understand the second half of your post. Maybe RBD has answered your question? If you are under the impression that the LDO version is for 1s applications it’s because you haven’t clicked through and read the OP of the thread about it. |(

one thing that I notice in this thread, is that DBCstm talks a lot :D

anyway.. does anyone wants to build these and sell them?

A17DD-SO8 that is

Wight, do I understand correctly that the new design can be used normally for a single cell light without the LDO or as a 2 cell (like a zener but with the LDO instead) with the LDO?

Just making sure the driver works the same without the LDO, I’d like to do this but again my lack of knowledge leaves questions and uncertainties.

Down to 4 of the first version, need to order some of these and make sure I still have LDO’s from my Knucklehead experience. And which ones they are if I do…

what I said

No, not correct. I placed some strike-through where I quoted you for the benefit of those that skim. The LDO version requires the LDO. This thread is really still about the AxxDD-SO8 drivers which do not require an LDO (but are are Zener ready). v024 is the current version of the A17DD-SO8, here it is on OSH Park.

:wink: RMM asked about selling them, I think he’ll be doing that fairly soon.

It’s now known all the way to Japan that I talk too much! :stuck_out_tongue:

And to think, I was about to put together one or two of these for my friends in Japan, pity…

I have the Ver 024 built and installed in an X6 Triple, waiting on emitters…

I am willing to make me these boards, I have the components but not the FET. And Mouser and Digikey are too expensive. I found these on Ebay, will they do the job? :

http://www.ebay.com/itm/PSMN9R0-30YL-NXP-MOSFET-N-CH-30V-61A-LFPAK-10PCS-LOT-/281514838554?pt=LH_DefaultDomain_0&hash=item418b978a1a

I think it’s probably OK for most single cell single emitter applications, but none of the specs are really that good compared to the one I specified in the OP. I think that besides being targeted at a different application this is probably an older FET (based on the use of LFPAK instead of LFPAK56?). PM sent.

EDIT: Changed my mind and added strike-through. Rds(on) is just not looking good here, and nothing else is that good either. It might work acceptably or it might not. It’s definitely not a good choice.

Oh goodies are coming, my boards will be delivered sometime this week.

i took the liberty of adding a little personal touch on the silk screen to the insides, while the credits given where its due, Thanks again Alex for making these beautiful pieces.

I have ordered them in 2Oz Copper instead of 1Oz, Immersion gold, 1.6mm FR4, black soldermask. to my liking.

I am also missing the FET , for now the best (economical) option was the aliexpress seller but It still a slightly lower spec’d FET not sure if this would be an exact replacement for the PSMN3R0

Its PSMN6R0-30YLB
http://www.aliexpress.com/item/PSMN6R0-30YLB-115-MOSFET-N-CH-30V-LFPAK-PSMN6R0-30YLB-NEW-Semiconductors-6R0-30Y-PSMN6R0-30Y/32223717242.html

Crap, I had no clue that zener mods were that bad with e-switch lights.... Yikes! Sure it was mentioned (I guess?) but totally missed it. Thanks wight for starting that thread and new board design!! I'll be sure to susbscribe and follow. Hhmm... Means the Y3 MT-G2's had that issue? Oh boy... Deleted - no harm intended Tuff to keep up with all the driver goin's on's...

This looks very suitable to me. Do you see any spec you do not like?

I moved my reply over to the other thread.



Not in particular, but some vary.. have a look.


PSMN3R0
/ PSMN6R0
Drain to Source Voltage (Vdss) ------------------- 30V / 30V
Current - Continuous Drain (Id) @ 25°C ---------- 100A (Tc) / 71A (Tc)
Rds On (Max) @ Id, Vgs -------------------------- 3.1 mOhm @ 25A, 10V / 6.5 mOhm @ 20A, 10V
Vgs(th) (Max) @ Id --------------------------------2.2V @ 1mA / 1.95V @ 1mA
Gate Charge (Qg) @ Vgs --------------------------46.4nC @ 10V / 19nC @ 10V
Input Capacitance (Ciss) @ Vds------------------- 2939pF @ 15V / 1088pF @ 15V
Power - Max ---------------------------------------91W / 58W
Mounting Type Surface Mount (same)
Package / Case SC-100, SOT-669, 4-LFPAK (same)
Supplier Device Package LFPAK56, Power-SO8 (same)

When you look at the Vgs graphs the Vgs advantage of the FET you selected (PSMN6R0-30YLB) becomes even more clear in my opinion. AFAIK lower gate charge is desirable, also lower input capacitance sounds like a good thing. Dissipation (the 58W number) doesn’t really matter in this context, we should not dissipate anything significant in the FET. The Rds(on) seems like it may end up 10-15 milliohms higher in the Vgs range we actually use. That may not be ideal, but I think it’s acceptable.