Output difference is visible so what ever it is I was thinking that the mux should take care of everything as current sensing will only be over the sense resistor. I’ll give it a go.
No idea what the lower bandwidth compared to the others does, but yes ADG819 is probably ok.
At any output current ? if it’s at a high output it could be the peak current hitting the switch limit.
Thanks for your input. At least I can avoid SOT23 packages, I have more than enough room for the 6-ball without shuffling components again.
Not the limit, noticeable at “medium range” currents.
Yeah sorry, that was a bit a stupid question. Still, I would check the gate voltage and the voltage drop across the HDR FET to be sure before doing doing another design.
I tested with the spice models of ADG819 (10pA typ, but actually it has way higher max at 85°C, models uses typ) and ADG787 (50nA typ) and while there is a bigger error with the second one it’s pretty much negligible, especially with the small difference we have between the two input (max of maxVsense-0V when HDR FET is off). So it’s probably OK with leakage in the 10s nA or maybe higher but I don’t really know how to calculate for it sorry.
TS5A3159A , maybe that one, smaller in DSBGA, and way cheaper than the analog one.
17mm buck/boost, huh? You’ve got too many good ideas…
If you really know coding write a code for 2 fet synchronous switching, and leave controller behind ;))
I’ll do some measurements. I know I did not get these variations with previous versions without HDR so I can only assume it’s Rdson.
Thanks for checking! I’ll have a look at that one. I managed to fit the ADG819, design is done (not ordered). If this one is a better option I’ll just make another design.
I designed it some time ago and did a quick test, it’s working at least. I’ll check if I can squeeze HDR in there once I get that sorted.
I’ve had a look at that solution, but with AVR 8 bit MCUs the switching frequency quite low and massive capacitors are needed for smooth output at higher amps, at least that’s my understanding. I honestly don’t see anything in that solution that is better other than cost for parts, and that’s not really something that concerns me. Perhaps I’m mistaken, if there is any reason why it is better I’m all ears.
Yeah, need a MCU with high speed PWM (with what size, power consumption), mosfets and gate drivers with charge pumps… switchings IC are so optimized and miniaturised that I doubt it’s possible to develop something remotely equivalent.
What buck-boost were you using ? The TPS63020 used in the lume1 and Zebralights ?
There is the TPS63806, 5A limit in boost mode and 4 in buck, and TPS63027 with 4.5A limit. Both smaller, with slightly lower Rdson switches, but BGA…
There is also an Analog one with 8A limit but it’s big, not very efficient and super expensive
By the way the H600II use a 140mV total Vsense , ~30mΩ sense resistor, ~12mΩ 2x2mm HDR FET, not very optimized for efficiency.
Just as with this boost driver I made two 17mm driver versions based on different ICs: TPS63802 and ISL91110IR (5.4A switch current), both worked fine. I was interested in making a board based on TPS55288 (external FETs) and playing around with IC2 but it’s not something I started, I don’t know if it’s even possible to fit that solution on a 17mm board.
Hm, im just curious. If you know how so why nor to try. Coder is a god those days ;))
Congrats Mike, this is looking great! Subbed.
Been following this closely. I love this open collaboration and hope to see more of it. Great work mike and freeman
A driver topology that I think would be nice to see for very low Vf 3V arrays is a buck+linear design. Buck mode for the practical and long runtime modes, so basically ~4-5A and less. Then have a linear FET channel to offer turbo modes. I think this is especially well geared to small form factor lights as it allows smaller components on the switching circuit, and a cost/space efficient regulated FET channel.
I’ve also had this on my mind. Buck+linear+direct drive or buck-boost+linear+direct drive would be expensive but combine top efficiency with top power density.
Thanks!
About linear and buck or /buck boost, I was thinking about trying to add linear capabilities to my buck boost driver. It’s yet another thing that’s been waiting in line on the to do list for a while.
I’ll admit what I’m really after is something like a 6A buck/boost. Higher would suit triples and quads (and maybe C_LPM1.TG), but 6A is a good peak for a lot of emitters. Sure, SST40 and Luxeon V can hit it for a while on linear drivers, but I also still like XPL-HI and LH351D, which have higher Vf.
I don’t know what size that would fit in (and I don’t care about cost) but it would just about be a grail driver for me.
I’ve also got an assortment of 12v emitters that want to live, and from the looks of three projects, there should be a good option soon. In 17mm is a special treat because then it can go anywhere.
The buck boost IC I use has a switch current of 5.4A, so output in boost is naturally lower. There doesn’t seem to be all that many buck boost ICs that can do 6A out in boost mode, at least without using external mosfets.
Now I’ve built one and tested it. With the exact same type of LED I cranked it up and got about 7 amps to the LED. It does have some flickering just as thefreeman described but he has a working solution for that (5K resistor).
I don’t see the need to have two different versions, I think they are fairly evenly matched. TPS might have better efficiency in the lowest modes but MP3431 can crank it up a little more. I have to choose one so I’m going to settle on the MP3431. I don’t have any reason to continue developing the TPS version.
I’ve designed another version of MP3431, V5. It has a mux to eliminate Rdson from current sensing path, this means LED current will not vary depending on temperature or voltage effects on the FET. I also added support for a single AUX LED, no room for more than one. I have some final functionality tests to do on the V4 before pushing through the new V5 design but so far so good.
Requiring external mosfets: What are the implications of that? I assume from a practical perspective, it’s more stuff on the board and therefore unlikely to fit in 17mm.
All buck, boost and buck boost designs require mosfets for switching. The boost ICs I have used here and the 5.4A buck boost IC I have used all have those required mosfets internally. The TPS55288 reguires four mosfets for switching but as only two of them are internal one will have to add two external mosfets. This could be difficult to fit on 17mm driver size, I don’t know if it can be done but I haven’t had a go at that yet.