+1.
I need a driver for my 30+ v Soraa 95cri emitters.
30+V. Damn.
Alphanumeric display displays menus, remaining battery voltage, LED drive current, CCA temperature, operational mode. The ability to define the number of presets and the intensity of each preset. A small uP could do this with ease, along with driving the power MOSFETS and other housekeeping chores… Instead of a timed interval for high power operation, have the processor look at heatsink temperature to keep the emitter and drive circuitry safe.
How about the display backlight changing colors to indicate remaining battery voltage?
Wi-fi and an app. Configure your light over wi-fi, using your phone or PC. Even easier, use a micro-USB cable.
Painlessly flash in new firmware
The possibilities are endless.
Cheers
(My emphasis above)
I’m just curious, what is it that makes this type of small boost driver so difficult to produce? I’ve tried to read through this thread, and I think I understand that high current inductors are large and inefficient and get hot. But what is so special about the approaches Zebralight and Nightcore take to solve this problem, for example, with their xhp35 lights running on a single 18650.
Please understand, I’m not being critical at all; I’m in awe of the electronic wizards here on BLF. But I’m genuinely curious. Look at these images of a Zebralight SC600 mk 3. What is the magic here, and why have other companies and individuals been unable to do this?
I stole these images from a cpf thread, thanks Tachead:
Basically it is an unfair comparison. The highest power production single cell XHP35 boost driver I have heard of to date is around 1.5A.
The drivers we are working on are designed to deliver 2-3x that power
A 1.5A boost driver is fairly simple. That is only about 6A input. We are doing closer to 20A input.
Far as what makes it difficult, when you get to these power levels EVERYTHING matters, particularly in such a small space. Electromagnetic fields are created everywhere and play havoc with the circuit.
Plus finding components that can handle the power and fit on a 17mm driver is very hard.
For example the last design works fine if we stick a big capacitor on it but then it won’t fit in a flashlight.
Factor in that each prototype costs around $20 and it adds up quick.
Everything is correct what Texas_Ace said. Developing something with this kind of power density makes it hard. The main problem always was the voltage sag on the input, or not enough output capacitance, both problems can be solved with more capacitance. The problem then is that I don’t have the right capacitors here that could be soldered on the board, and still fit in a flashlight. The 50€ minimum order for free shipping (otherwise I pay 20€ shipping) on Mouser makes it even worse.
Getting input and output capacitors that meet the requirements and that are small enough for the 17mm, isn’t easy. Especially with the high-value ceramic capacitor shortage right now. Basically every capacitor with high capacitance isn’t available. And it seems, that won’t change in the near future.
But of course, the main problem is cost, and time. I paid a lot of money for parts, and invested a lot of time as well. That’s why I changed my approach for this project. Rather than making a lot of prototypes and testing, I now invested just time and no money for a few months. The last months I didn’t have a lot of time for the project, but I finally have a (from my point of view) really good board in EAGLE, and found the necessary parts on Mouser. For me, it’s not budget lightforum anymore, it’s rather candlepowerforum level, and I hope I need just one more prototype. Please understand that this project is pretty expensive, but it will get finished.
I do think that a lower power boost 17/20mm boost driver with NarsilM would be very welcome already.
With the high power one, you would just need to change 1 or 2 resistors to change the maximum current. Maybe in the future I make a lower current driver that costs less though.
I think there will be a good amount of interest in 20 and 22mm drivers if that helps solve the parts problem…
Thank you, Texas_Ace and Schoki, for the detailed responses. I hadn’t realized the power differential between existing boost drivers and what you are working on. We all appreciate the time and money invested into projects like this.
Of course. You’re speaking about an 18W driver vs a 50W driver.
Extremely amazing when you think about it. A 17mm 50W boost driver is unprecedented in the industry.
Hello guys,
Please excuse me for joining this conversation, as I am a newbie in regard of the flashlight electronics topic, but I have a (general) problem, for which I think, a solution would be highly desired, as success in that regard could be perceived as a big improvement for many of us. (Also excuse me for introducing a problem, which does not match the concept of your latest conversation.)
I have read about the linear drivers across some pages and forum topics, which are used pretty much all across the baseline of Convoy products (S2+, C8), and many more similar flashlights.
Because mostly there is a significant difference between emitter Vf values (for given current), and battery voltages, they just convert the difference to heat, which leads to undesired efficiency loss, and worse heat attributes for given current levels.
I was lucky enough to make a few emitter swaps in my favourite headlamp model (Skilhunt H03), and I discovered, that heat is pretty much controlled with even the two highest output settings. I have read, that it has a driver, which converts the battery input voltage down to the Vf of the used LED (certainly, it has a buck driver), so it could stay fairly efficient, regardless the fact, that I used an emitter with considerably lower Vf across its current characteristics (change was from Cree XM-L2 to Nichia E21A).
But, that is not the case with any of the known linear drivers - so using lower Vf emitters with those drivers could mean even more heat compared to the older XM-L2 or XPL emitters, if they are used at the same current. (These emitters would be most new Cree whites, almost all Nichias, many of the monochromes /extremely low for mostly orange, red, and near infrared/ for signaling purposes) - Also, I would not like to direct drive these low Vf emitters.
Generally, I would like to ask for a solution, which could be used in a few smaller popular flashlight hosts, which can only support d = 17 mm driver plates (e.g. Convoy S2, Jaxman E2L), and probably they have limited space to fill vertically (e.g. it would be around <5mm for both of Convoy S2 and Jaxman E2L, if I remember it right).
So, for now, my question would be, that could you design a single li-ion battery buck driver for a d = 17 mm driver plate, or it has too many difficulties? (‘Moderate-to-high’ max. current, e.g. 2-3A for S2+, 4-6A for C8)
If at least one of you would like to, and would be able to put some effort into it, I am willing to donate for its development based on the needed parts list, and other factors/milestones. (I cannot offer more than a few $10, but I hope, others could join the funding of this project as well, if it can start.)
With a little exaggeration, my dream would come true, if I could use nearly any emitters efficiently and with controlled current levels with the above mentioned flashlight hosts (and it could be widely available across the flashlight users/lovers, then).
17mm buck drivers exist in plenty of flavors already.
The problem? Since buck drivers have much higher resistance than FET and linear drivers, in a 1x18650 setup, they go out of regulation quite quickly except if you have a very low forward voltage emitter(E21A quad).
For example, imagine you have a 3.2V forward voltage emitter at 3A. A linear driver will keep constant brightness until the cell goes down to 3.3V. It is less efficient than a buck or a boost, but will keep brightness constant at all times.
Usually, unless you have a very expensive buck, there will be a room of 0.5V to stay in regulation. That means you would fall out of regulation at 3,7V, much sooner than the linear regulator. It is much more efficient though at 3.8-4.2V however.
The easy way to get perfectly regulated brightness, but then, you would need a 2S setup, meaning you would have a light twice as long.
TLDR: Unless you have very low forward voltage LEDs, the best drivers for 1x18650 lights are boost, linear, and the Holy grail, the buck boost.
adam7027. What you are wanting is a buck or boost driver like we have been talking about. These are fairly easy for low currents but get exponentially harder as power goes up.
Also, while the linear drivers are less efficient at low modes then a buck/boost driver, it is usually not enough that you would notice it besides a bit less battery life. It is not enough to make the light noticeably hotter to the hand anyways from just an emitter swap.
As the improved efficiency of the emitter makes up for the worse efficiency of the driver, so the final output is basically the same. You just don’t gain efficiency that you could from the lower Vf LED.
In my testing of linear / PWM drivers, even at 50% duty they were still close to 90% efficient, which is about what you can expect for a buck/boost driver. Only in modes lower then that will there be a noticeable different in efficiency.
Although the linear driver gets more efficient as the voltage drops. So in the long run the final outcome is not nearly as bad as it sounds on paper except for some special use cases (very long term use at very low modes for example, where you battery life is measured in days).
Edit: Bluesword beat me to it.
Thank you all for your helpful answers.
Good to know about the higher resistance, I didn't consider that. Still, if you know premium single cell buck or buck-boost drivers for d = 17mm, I would happily take a look at them. There is a chance, that in some less frequent cases, I would like to use emitters comparable to E21A quad, regarding the Vf (e.g some monochrome reds are of much lower Vf starting around less than 2.5V).
I also noticed the small room for regulated output with the stock Skilhunt H03. It 'lost' its highest setting around 3.8V, while with the E21A quad, it could go down to 3.3V with a plain old Pana NCR18650B for the same effect. Something approximately as good as that driver could be good for me.
Otherwise, if I'd use a linear driver, maybe then it would be better to use some mid-current, capacity optimized cells to make the difference a bit less, and their discharge slope also gets less steeper at the second half of the discharge.
For what current level would you consider a buck (or buck-boost) driver to be fairly easily made?
I found somewhere a driver, which could fit most of my use cases /no blinkies, more modes relatively close to each other in the midrange - I planned to use it with 4-5 pcs of 7135 regulator/, but I found a review, which had discouraging results for me. Based on the fact, that it was made with a Cree XP-G2 emitter, I only expect worse result with e.g. Nichia 219C or 319A - that was the main reason for me to ask for help.
Are there 'good' and 'bad' (i.e. with significant difference) linear drivers, according to your tests?
Thank you, I already looked at that specific driver, just hoped for other solutions as well.
But the fact, that it 'renders' all modes without PWM, could be good.
Buck/boost drivers are reasonably simple to make with up to ~2-3A max on either the input or output, whichever is higher. Once you go much above that it gets exponentially harder to keep everything in check.
The 7135 drivers are fine for what they are, basic and cheap but they do the job.
Just listed Emitters, pads, heat sinks but need driver for
WTS Once in a life time 95CRI 3000K massive emitters
Someone please let me know that we can at least build single mode drivers!! Please
I pointed you at the MIC2298 in that thread, maybe someone here can take a look, see if it’s feasible.