comfychair-inspired quick+cheap+lazy single-sided 17DD FET-driver (poor man's nanjg92)

Maybe somehow reducing the gate voltage to the FET would be a better way. In either case something is going to have to drop the voltage and dissipate ~2W. You would have to mount the FET better; it would get really hot. Do all your FET drivers have the FET upside-down?

Here’s how you estimate the resistance to add. First estimate the resistance in your circuit now. 30Q has ~25mOhms, tailswitch has ~15mOhms, wires have maybe 5~mOhms, and FET has maybe 5mOhms. So with a fully charged cell the voltage left at 5A would be 4.2V-0.05ohms(5A)=3.95V. The 1mm WF Vf is 3.35V at 5A, so you need to drop 0.6V at 5A, so add a resistance of 0.12ohms.

Your last pic looks good but EasyB has tried a similar resistor mod and is saying the resistor wattage would have to be 2 watts or more. Thats a lot of wasted energy as heat.
.
There’s alot of other factures that are coming in to play here but using a online calculator with a battery voltage of 4 volts (battery voltage 4.2v will drop once load is applied 4.0v or less) and a led vf of 3.4v comes out to 120mohm with a 3 watt rating. https://www.amplifiedparts.com/tech-articles/led-current-limiting-resistor
You still have other resistances that are not calculated in that figure. Guessing 100mohm at a 3 watt rating would work close. Might be best to use 3 1206 stacked in parallel at 300mohm to get to 3 watts if using the pads. If your solder it directly to the FET you could use a much larger resistor (2512) and only one needed. https://www.digikey.com/product-detail/en/bourns-inc/CRA2512-FZ-R100ELF/CRA2512-FZ-R100ELFCT-ND/1775059
I would order a few values up and down to see which gives the best results. EasyB might be able to add some more info as what he did and used.
.
Edit: EasyB beat me to it. Since the battery is only at 4.2v for a fraction of second you still figure that as the supply voltage?

Depends what you want. If you want to limit the peak current then use fully charged voltage of 4.2 or 4.15V or whatever it is. Or you could use a little smaller resistance and have it be a bit over 5A at first. It’s probably going to require some trial and error anyway.

Isn’t that drawn resistor connecting batt+ and batt- ? Instead of being in series with the leds??

Or was the idea cutting the trace at the red line and use it as an isolated solder pad?

Edit: never mind, that was indeed the idea :person_facepalming:

The driver design is in the first post of this thread, it will explain why all the FETs are upside-down :slight_smile:

Luminarium, given that you have hundreds of these to do i would look at Moderator’s suggestion of finding the right FET first, which would be like having a FET with the resisitor already built in which would save you SO much time and effort if you find the right one. You could then add a heat sink to the back of the FET so it copes better with the heat, much less fiddly than cutting traces and soldering resistors, and still as robust as your previous drivers.

Hi guys,

Yes Djozz. To cut the trace at red line… I really must do something so I appreciate any good idea.

Guys don’t worry about FET turned upside down. This is picture of Djozz diy basic “how to” work. Djozz thanks for the pics! Thank you for this driver! :beer:

Mine FET is permanently epoxied with serious 2 component epoxy from bottom side plus it is potted with super 2 component cold weld compound from above so there are no any exposed parts plus this driver does not generates much heat at all.
The copper mcpcb and LED above it generates heat which is dissipated by pill, flashlight body so potted driver as a pill part just perfectly fit into heat dissipation.

Hundreds of drivers done without any issue at first at harder way by de soldering 7135 chips but thanks to Simon that managed to get me Ak 47 C1 without any soldered AMC 7135 chips to make my work easier and faster.

If you order larger quantities and have right ingredients it can be done for 1.8$ :+1:

Finding right FET could do. I would like to try both methods (resistor and different FET) but I will obviously have problem of choosing, ordering and getting right components for the job.

Here is a FET with 0.1 ohms on-resistance.

For the resistor method, you could sink a 2512 size resistor to the MCPCB.

Thanks! :sunglasses: Can you find me FET with legs? Sorry for bothering you :frowning: but it will be almost impossible to improvise something with FET without any legs into this diy driver.

EDIT:

Found resistors source here

Should I order 0.12, 0.13, 0.15 ? Could this 3 types be enough for the job or should I try with lower values like 0.1 or 0.11? Cost like peanuts…

Size could be a problem 6.5mm…

You could do the trace-cut around the Q3-pad instead of before it, a bit more work but it gives you an extra 2 or so mm. Getting close to the edge though (possible short with pill).

Or put the resistor in the negative path by simply glueing and soldering it on top of the FET.

Or order large resistor and solder one end directly on a FET (direction of FET like in picture above) and other end left exposed for soldering minus black wire directly on it? It should hold especially if it will be epoxied?
Opinion?…

See my edit :smiley:

:+1:

Axial resistor vs SMD resistor

If axial can be used; one end could be bent and soldered directly to exposed fet side while we could solder minus wire to other exposed side of axial?

But if SMD resistor is better for some reason than I don’t want axial… Opinion?

You can cut off the legs and scrape off the paint on the ends and use them as SMD components.

Which one of those 2 kind can take heat better?

Edit: I bought smd resistors. 2515 package (0.1, 0.11, 0.12, 0.13, 0.15R) plus some ceramic cement resistors ones that looked good to me but guy has only 0.1 & 0.15 versions.

So that is total 270 resistors for 10$ :slight_smile:

Now I’ll try to search for higher resistance MOSFET with legs… Thanks.

Another edit: Those ceramic I bought seems to be larger than a whole driver. I saw that on one of user feedback.

However that ceramic one surely looks like it could be more thermally stable than regular smd one… IT would be great to find smaller version of that. There are no leg MOSFET with high resistance on ali.

Years ago, I bought some AK-47C1 drivers for this mod. Finally applied it today, but then I couldn’t find info on star configuration for the driver.
Found this pic on a Russian site, reposting here in case someone else needs it:

text version of the pic:

default: 100/30/5%/Strobe/SOS
stars:

  1. does nothing
  2. 100/5%/Strobe
  3. 100/30/5%
  4. 100/10%

I would expect these drivers to sell out at some point, I’m fairly sure that they are not made new anymore. But you still found some :slight_smile:

Edit: I see you mention: years ago

Bikedude,

Where did you find them? It is almost impossible to find them now…. It was so impossible for me that I decided to invest some serious $ for developing of my own OEM driver…

But this AK47 C1 DIY driver is extra… Thanks Djozz 1000X times!

for the mod I used the ones I bought many years ago at banggood.
but the picture is from this russian site, maybe they still sell them: Драйвер 3*AMC7135, 1050mA, 17мм, Nanjg AK-47C1 5 режимов купить Минск