Excellent. Great post you linked to there. I will be giving that a shot. I have a 4S build in progress. I should be trying it soon. I'll report back.
Let us know how the 4S goes. The big unknown for me on some of these Chinese boards is what the individual components, especially the capacitors and MOSFETs is voltage rated at. When you take into account the ringing that occurs I would hope to see a 25V rated input capacitor for 16V input use, but who knows? At least ceramics don't explode like electrolytics but I would definitely remove that little yellow tantalum cap. since you're probably not using it anymore, those can burn.
I've already used the driver in free air with 4S driving an MT-G2 at about 10 amps (9.65 per the DMM). Same reading I get with 3S cells. Set up on the driver was
- Two R20 resistors added
- QX9920 buck converter swapped in because I broke a leg on the stock buck converter.
- MCU swappped - Attiny13a with JonnyC's STAR momentary modified by Tom E to include strobe.
Used the stock FET. I did not run it for very long (just a few minutes because it was late), but it showed no signs of distress and worked great. Didn't even get any whine from the MCU.
wight wrote:
It’s been a while since I thought about this driver (and I still don’t have one myself). Glancing over your pics again, maybe it does not have a voltage divider. That shouldn’t be showstopper though, airwiring a voltage divider should be an easy thing. I think I’ve covered the component selection (two resistors) pretty well over here (#917). For such a large driver I’d probably just use standard 1/8w or 1/4w through-hole resistors, you probably have the room for them. Higher values are better, less power draw.
Tried out your above idea using 4S 18650's driving MT-G2 with HX-1175b here with STAR Momentary (Ramp Down at 130 and Sleep at 115) and it works. I need to fine tune it though as it may be stepping down a bit early at this moment. I didn't take notes during the process. So here is what I can remember.
- First tried 110000 Ohms for R1 and 5100 for R2. Ramped down way to early. Like 13.x volts.
- Second tried 110000 Ohms for R1 and 6200 for R2. Started ramping down modes around 11.3 and Sleeping at 9.x Too low.
- Third tried 91000 Ohms for R1 and 4700 for R2. Started ramping down too high again, like around 13.x.
- Fourth, tried 82000 Ohms for R1 and 4700 for R2 and adjusted Sleep in FW up to 120. Haven't tested this set up yet.
I think the Atiny13a used for the above has a high reference voltage. 10% either way from 1.1 volts per the data sheet. The resisters used were 1% accuracy metal films. I just used matched (resistance and capacity) healthy laptop pull cells.
Sorry for the slim data. I need to get a power supply. My take away is that wight's idea works and that these high voltage applications will require fine tuning on an individual basis.
EDIT: fine tuned wording (in bold).
Sounds like a good start. Nailing down the ramp-down voltage first and then tuning the value for sleep makes sense.
I slightly edited my above post.
I'm excited about this as Low Voltage Protection it is much more critical for these high voltage lights. In a single cell light, it gets dimmer up in the 3.9's. Of course, you still have the risk of the light being on while your not aware. So LVP helps with that. But in a 4S light with a single emitter, you can go low while using the light. I'm really glad you shared the knowledge to help protect against that. I really need to find a decent budget power supply.
For the MCU, I guess you all don’t have pinouts, but do you at least know where some of the pins on what you think is the MCU chip are going to? Just curious if that would help identify the MCU?
I’m not aware of an MCU in that package. At all. I think it’s… what, a 10-pin TSSOP? I looked through the product lines for Atmel, Microchip, and I think also Zilog. I couldn’t find anything in that form factor to compare the pinout to! I’ll admit that I had trouble looking through Zilog’s portfolio.
Anyway the point is that step one is to at last identify an MCU that can physically match that landing pattern.
Maybe it’s not an MCU? Maybe something like this:
Does the one you all are thinking is an MCU look like the 10-pin in the pic in this post (on the 266050 driver):
I’m a newb here, just dropping in, but I was thinking that if you could get some traces from the 10-pin chip to wherever they go, that might give you some idea what the functionality of that chip might be?
Just thinking (sorry :)!!)…
Hi Ohaya. The 10 pin chip is definitely the MCU. No clue which one though. The pin functions that I know are as follows (looking at the top of the chip):
PWM Not Used
Not Used Not Used
Ground Not Used
Positive Power Supply Don't Know, but traces appear not connected to anything
Not Used Don't Know, Looks like it could be used for Voltage Monitoring but traces appear not connected to anything
Here is the orientation of the chip for the above pin out:
For the MCU, what about these from Renesas (looks like there are several 10-pin ones, with different memory specs):
http://documentation.renesas.com/doc/DocumentServer/r01cl0004ej0200_78k0_78k0s.pdf
Package info:
http://resource.renesas.com/lib/package/pdf/outdrawing/p10ma-65-cac.pdf
Good call ohaya. Looks like the pinouts match to me - so it’s one of the μPD78F920x. Timer output for PWM on Pin 1, GND and Vdd on 3 & 4, and analog input on 6. That’s all the functions ImA4Wheelr described.
I guess the next question is whether we know anybody who can tell us about flashing these things cheaply.
EDIT: Datasheet link
You all might ping Dr.Jones, to see if he has any firmware that’ll work, but although I only looked briefly, these chips look pretty proprietary, from the old NEC days.
Meh, I’m not actually interested enough mysef. This is the only halfway decent driver I’m aware of which uses this MCU, although IIRC we’ve seen it on at least one other driver, I think it was a DD driver.
Good find on the actual MCU, now we can safely ignore it. 
I have [finally!] ordered two of these HX-1175b drivers. 
Has anyone tried this driver with the XHP70 (6 volt configuration) yet?
Since RMM has done a lot of measuring and experimenting with this driver using MT-G2, I'm hoping he wil drop in with some input. Buck drivers are supposed to have less ripple when supply voltage is closer to Vf. Did the voltage spikes with 3S cells and one MT-G2 appear potentially problematic for XHP70?
No problems with the XHP70. I've only tried up to about 8A with one, but it can handle it just fine. Even with a huge 20% ripple (this one is generally 10% or less) you're still under 10A, which the XHP70 can handle easily.
I really need to update some of my findings in this thread. I got the ripple down to a minimum, but it was around the same time that the new XM-L2s started coming out, so even though the ripple was negligible (~1%) it would still blow them at 7A, because even direct drive the LEDs couldn't handle 7A anymore. I turned them down to 6A and all was good.
Thank you RMM. Good to hear that. I was expecting a different answer based on you earlier reports in this thread.
I really want to design a board with a different MCU that is loosely designed like this board, but I'm still early in the learning curve for Eagle. If you ever go down that path, I would like to try to help with testing. High current bucks are what I am always interested in.
I've been focusing mainly on the small ones and have learned a ton since I started fumbling with this one a year ago. The cheapest route would probably be to just downsize this board then use these for donor parts. Buying good parts individually will end up costing more than what you can buy these boards for! If your goal is to power MT-G2 and XHPs then you have a lot of wiggle room when it comes to the output waveform, but if you are planning on the smaller 3V LEDs then things have to be good to go.
If you are looking at designing one, then I would recommend reading some of the datasheets, especially the TI datasheets, regarding circuit layout. When you start getting into bigger switching converters at higher switching frequencies layout matters a lot--you can't just hook the wires up and expect it to work correctly.
Was playing around with LVP on this driver a couple weeks ago with what appears to be some success on my PS, but that PS is cheap and gives weird results. Using 100K and 5.1K resistors and I don't remember the numbers I used in Star Momentary with 4S cells driving an MT-G2.
Need to test with cells. Been running with 4S cells since I put the light together then, but they still have not drained enough. I'll report back when I have more info.
RMM,
Back in Post 153, you had some great info about ripple and you sound like you have learned more about modding this driver since then. Hope you don't mind a couple questions, if you have the time to respond:
- When you refer to output caps, do you mean just across the LED Leads?
- Regarding inductor positioning and shielding, do you know if one can get some limited shielding using aluminum foil or something between the Leads and the inductor? No biggie if you don't know, I will be googling it soon.
- Do you have any update suggestions for this driver?
EDIT: I accidentally connected a stock HX-1175b to 2S XM-L's on copper last night with 3S cells. The emitters fried instantly.