17mm QX5241 5Amp 1-4 LED 2-4s Li-Ion Buck driver, WIP - I'm back in the saddle.

Here’s a driver I’ve been wanting to do for a long time and never did. It’s based on the Chinese QX5241 buck controller, same controller used in the old DX SKU 20330 and the new-version DRY driver. As far as I know it’s also the same one used for a bunch of other drivers such as one IOS previously carried and this 5A one from LCK-LED as well as a few others.

All of the above drivers other than the DX one seem to employ a second FET wired in a way that I do not understand. It must be there for a reason, but I don’t know what. Here is where I describe the way the second FET is connected.

A Nanjg 105c and a new-version DRY driver together will contain nearly all the needed components to assemble this driver. It requires the torroid from the new-version DRY driver, so that makes this driver quite tall, the top of the torroid is about 11mm above the bottom of the PCB. This driver requires a 5mm diameter by >2.4mm tall spring or spacer due to the large diode on the bottom. As you can see, the combined height of the board may be 14-15mm thick overall. I haven’t calculated sense resistors, but since I only put a single 1206-sized pad you’ll probably need to pick that out for a special order.

LED+ connects to one leg of the torroid, the other leg attaches to the via next to the Q1 marking. LED connects to the body of the FET. I forgot to add some 1μF capacitors on the VIN and VCC lines for the two microchips respectively. I’ll need to go back and add those. We may be able to cram in a second 1206-sized sense resistor.-

Update: Current progress is in post #26.
Update: Current 17mm design progress is in post #42 and post #59. Since that wasn’t working I’ve switched to a 20mm design for now as a sanity check, see posts #76 / 77 / 78. TLDR: 20mm buck section works.



You have omitted a 1uF which is connected to the pin 4 of QX5241 and the ground.

Thanks Microa, I am working on that now :wink: I also left out the 1uF connected to pin 3 and ground.

The Cin at the pin 3 (Vin) should be at least 47uF. The smallest size X7R MLCC I can find is 0805 47uF25V.

are those standard 0603 or 0805 solder pads for the resistors or capacitors? Not sure if they just look really close together to me or not. Most cool though, is this a buck type setup? ATtiny driven?

P.S. Whip up a 20mm when you make the 17 as well :wink:

Really? Why? Only the SM5241 datasheet shows 47uF, the QX5241 shows 1uF and the LY5241 datasheet doesn’t seem to mention it. To me it doesn’t seem like you’d need that much capacitance just to smooth out the 5mA the chip draws.

They are 0805 pads. It’s a buck setup. I used the standard pin configuration for the ATtiny13a w/ one of our normal firmwares. This is for clicky only. It’s going to look at little more packed in a minute or two here.

What is the voltage range of that driver? If it can buck 3x in series down to run a single emitter…I have a home depot defiant I can upgrade to a 26650 rather than C cell :smiley:

I expect this driver to perform like a less-efficient new-version DRY driver. The voltage range should be something like 2-3s or 2-4s. Remember, this thing will be huge with the torroid on top.

Don’t be confused. Cin is 47uF which is connected to pin 3 (Vin) and the 1uF which is connected to pin 4 (Vcc) which is 5V regulated output.

QX5241 application circuit shows 1uF for pin3 (Vin). I don’t know if that’s correct.

I have revised the driver to add 0805 caps for both Vin on the QX5241 and Vcc on the ATtiny. I also added a second sense resistor in parallel. I did not do a very clean job and I will probably improve the layout in the C1/R4/R1 area at least, as well as some unattractive traces.



OK, past bedtime. I cleaned up the sense resistor area and moved the offtime cap to the top of the board (it was right above the sense resistors).

This does pass DRC w/ 6mil spacing.

Yea, I think they look off too. 0805 should have a pad separation around 0.7mm. Oshpark renders boards using tStop for the copper areas instead of Top for some reason. Makes all pads look 3mil larger in the renders.
Are you using the oshpark DRC? The default eagle DRC has 4mil mask instead of 3.
I don’t know, even if its the eagle default 4mil, pads look off. Course I didn’t sleep last night so my eyes might be off.

Oh so this isn’t a single emitter driver but a multiple in series driver…k gotcha

Cool project. Thanks for starting and sharing this.

I didn’t realize that I could adjust that - I see it now. Thanks. I had it set to the default, 4mil. I’m using the Eagle R0805 package for the divider resistors and C0805 for the caps.

WarHawk-AVG, please read the thread. It’s a buck driver and will go at least low enough for a single white emitter. Results are posted in the short thread I linked. The driver will not operate on a single lithium ion cell. The links in the OP were meant to answer this question.

Thanks, let’s hope we get something useful out of this thing. I’m feeling pretty good based on the lineage (all the other successful drivers based on this controller). The only thing that I’m concerned about is how big an efficiency hit we take from eliminating the weird dual-FET setup I mention in the OP. It’s not a synchronous buck configuration, so I really don’t understand what those other designs are doing with the second FET.

Yup…I shoulda read…thanks wight

I realize that I post a lot of links. :wink:

While searching for boards to rob components from in order to populate this driver, I came across this:

http://item.taobao.com/item.htm?id=17981669490

At ~$2.50/each in China I figure that’s not much better than the new-version DRY driver. Has anyone seen these drivers around on AliExpress or eBay for a better shipped price? I see them on eBay for crazy prices (>$10!!), but no decent prices.