Custom 17mm 5Amp PWM-less linear driver-preliminary

Can’t wait for the review and the sales thread :party:

I just want to tip my hat to this ;). Keep up the good work.

Any help and education is appreciated by me. But I guess what I was trying to say was that this would be good in a low voltage situation so as to not have any PWM whine. :slight_smile: Whatever it takes to put it together, ie, with a buck, or boost driver, whatever it takes to complete the module.

Excellent, looks like the perfect driver!

Actually I need something like this, but I need it to a bike light with 7 XM-L2 in series (the stock driver delivers up to 650mA with 2S, so ~1500mA would be perfect).

How many leds can it handle? Will you make a bigger driver that can handle more LEDs?

Again, you need a boost driver as the voltage of seven LED in Series is higher than a 2S. I really need to make that thread. :wink:
This is a linear driver and can only yep down the voltage. It is perfect for one cell one emitter.

You do need to make that thread. You’ve probably already realized this, but there is actually a lot to explain if you extensively cover only the application considerations of buck/boost/linear/direct, assuming you include things like zener modded 7135 and FET drivers. If we had a clear post describing the topologies I’d definitely refer people to it, but an explanation of the effect of zener or [currently non existant] vreg mods on the 7135 and FET drivers is really essential for a lot of the questions we run into. Take a look at my post #14 in HiTiT’s thread. There was still some confusion after that, which I thought was a decently clear explanation. And if you write to many words, we’ll never get people to read it. Definitely a double-edged sword.

Ok, I’ll do it. I’ll start by explaining Buck, Boost, Linear, Direct Drive. I can always add things later (like Zener diodes, etc.).
I think I’ll do a first part “In a nutshell” and below that a more detailed version with graphs and things for people that want to understand how it really works. I’ll also insist on efficiency.
I only have two 7135 based drivers (I have 2 flashlights, I’m a noob! :slight_smile: ) so I’ll have to buy a buck and boost driver if I want to plot some graphs. That will take time.
When led4power releases his driver I’ll try to include it. Unfortunately I didn’t win at the giveaway so that’ll have to wait. :smiley:
I worked on a high end Buck driver used in Smartphones for 6 months as an intern, so I should be able to give some accurate infos, but I WILL make mistakes so you’ll have to check what I say. :slight_smile:

Sounds good. Please PM me so that I don’t miss the thread.

Better yet, he could post a thread link here. I’m sure there’s a lot of interest!

Led4power, does the driver have low voltage warning or protection? What about something to monitor cell voltage? Thanks

+1 I’m interested too!

Yes,it has.If voltage drops <3V,current is limited to ~150mA.You still can use modes lower than that (moonlight and low it this case),but if you switch to mode with >150mA,current drops to 150mA.If voltage drops <2,8V drivers goes into sleep (~50uA).

Last minute decision: board will be 4-layer for better thermal performance.Now question about pcb thickness:will 1.2mm cause any problems?

I tested driver in roche f8 host,first without any potting,and I noticed one (expected) problem: driver temp!=flashlight temp.

This is because thermal contact driver-flashlight isn't so good(potting improves that and it's highly recommended for this driver),so small flashlight heats up much faster than driver(and opposite,driver temp is higher at 1Amp,than at 5Amp,because more power is dissipated).

This means over temp protection doesn't work well if you don't have separate sensor for LED/flashlight body.But I partially solved this problem with two temp thresholds: one driver overtemp threshold(for ex. 100C),that can't be changed by user and it's always on.It prevents driver from failure caused by overheating.Second,lower overtemp threshold can be set/reset by user(reset value is equal to driver overtemp threshold) as described in OP.It's active only on highest mode.Not ideal,but for now it's best solution.

this thing looks awesome!! I’ll definitely be interested in some of these.

Liking what I see. can’t wait to try one or two or….

It seems nobody reads long posts,so I'll repeat one very important question:

Is 1.2mm pcb thickness OK?

Thanks for the response. I highly doubt a thicker pub would be an issue for anyone. Not a problem for me.

It's all relative.. 1.2mm compared to what? What was it before? I think Nanjg/Qlites are 1.5mm, so 1.2mm is still thinner? As long as you are close to typical stock drivers, it should be fine. Thinner than Nanjg's by that much should be still an improvement.

Retaining rings might not make contact with the ground ring in some instances. Not a bit deal as it’s an easy fix.

So I don’t think it’d matter. 1.2 should work fine unless I’m just missing something…again.

The driver has arrived.

And I has rigged it for testing:

I hope to test in the weekend.

I will be testing it with a SST50 led, I have not decided if I will risk an XML2 led on it (Maybe I need to get some XML2 on copper). On the other hand my current XML2 test led has already survived 5A, the risk is probably not that great.

You should be all right with a 5A driver. :wink:

I’ve run one direct from every type of cell I have, including Sony C4 and C5 and Samsung 20R and 25R, as long as the XM-L2 is on copper it handles these cells just fine even in a direct wired set-up.