Where can one find one of these shadow sl3’s as a host or for under $50? This is a neat little light.
Cereal_killer, you’ve got some silkscreen text over copper. That isn’t going to print and OSH Park warns of unpredictable results. In my experience all they mean is that they can’t tell you exactly how far the fab will “auto clearance” silkscreen near the copper.
Also (I haven’t checked this) don’t forget to make sure that you make the lines thick enough or they’ll be auto-fattened. OSH Park claims that the silkscreen is done by what is effectively a 200DPI printer, so I treat the minimum thickness as 5mil.
Hey guys got some progress made, instead of piggybacking a driver on the stripped factory one which the switch mounts to I’m making a driver specific for this build which the switch mounts directly to for a direct fit, no need to piggyback onto the stripped factory board anymore. (BTW I already have the 17m PIC RGBW driver ordered to play with)
How do I get the “wings” so they’re part of the outline?
I decided to switch to the little baby fet because the AOD3400A Comfy found is capable of 5.7A, I figure 15A+ will be plenty, no need to use the big ones and since I’m using the smaller components (and I dont need the voltage divider circuit either cause the 85-20SU can do it internally) I went ahead and added a white channel to so this driver can fit a wider range of lights.
Do what I do…use the SOT-123 diode instead of the SOT-323, but that is cool driver, oh and de-stop those vias off the attiny Tools>DRC>Masks>Limit (set to something high like 50mm or whatever measurement metric you use) but make sure to go back and stop those VIAs you want to solder to by clicking the via and checking the stop box.
But that is a wicked cool driver man…like how you fanned the small FETs out to the side
To make the wings you will have to draw a milling ARC and attach it to the milling layer, then take the wing from one side, copy, mirror and attach to the other side
I find putting up a grid in view helps make sure I put stuff on the same level thru the driver
wight knows better than me…still an amateur on Eagle
I think you’re going to torch those FETs. Worst case scenario: you learn by doing.
EDIT: WarHawk-AVG, don’t be mistaken. I’m an absolute amateur at Eagle. I only started actually using it after Mattaus posted his Knucklehead V3 and shared the files with me. Before that I’d installed it but never learned how to use it. [I did my battery carriers using an incredibly laborious combination of CorelDraw, Kicad, and PNG importer/resizers.]
That said… yeah, 4 arcs is the best thing I can think of.
Why/how so?
These aren’t current limiting devices. You (and sometimes others here) seem to think of them as regulators. They are absolutely not regulators. If I use a 5.7A max device and pull 10A through it, the device will be “unhappy”.
Current will be based on Vf, total system resistance, and battery voltage sag.
How is what I’m doing not the exact same thing as a single channel BLF driver? I dont understand, why would it “pull 10A threw a 5.7A max device”? Not only would a single battery never support that draw (that would mean it was pulling 30A+ from a battery) but I fail to see how this is any different that the current BLF xxDD FET drivers that we’re all using, both my EDC lights have FET based/attiny drivers… Yes I understand a FET isnt a current regulator, its a relay, the point of this is to be direct drive, these huge emitters can handle it, in my other “puke light” (my RGB TK45) I run a 7135 based driver for real regulation to the little XP-E2’s, the whole point of this driver is to run at direct drive currents but have each channel individually addressable, otherwise I’d just use a single channel 17dd with the emitters in parallel.
v1.3, gate / pull down resistors added (with thin traces in case resistors arnt needed, cut if they are needed)
note bottom most resistor should read “BLUE_R2”
new version 1.4 now with wings
This is why:
The current you will pull through the FET is determined by the input voltage, the LED's Vf, and the FET's on-resistance, and has nothing to do with how much current the FET package can withstand before letting the magic smoke out.
I thought the whole point of moving the cap before the diode was to prevent needing the gate resistors (or was that just for ATtiny based drivers)
I was told by comfychair I believe that a FET acts more or like an electronic light switch, a pure wire, on or off, it helped me understand the point of using them…
I discussed that with DrJones since at first I’ll be using a MCU he sends me I decided to go with his way, he didnt specify to use the gate/pull down resistors but did say the cap should go after the diode in VIN (the old way) so i figured I’ll add the pads for the resistors and see if it needed them, if not the trace is there so you dont need a jumper, if it acts strange I can add them in. Plus who knows how it’ll behave without the voltage sense circuit, I decided in the end not to rely on testing on the 17dd and screw myself later having to air wire them in, in general its always a good idea to have the resistors on a MOSFET’s gate so I went with that.
I think you've missed just about every discovery in the 'scope images' thread!
The Nanjg's circuit design with the cap after the diode is inadvertently creating a boost circuit. It runs the MCU right up near its 6v limit in al the PWM'ed modes (effect goes away in the 100% mode), and if (when) there's a tiny spike up above that 6v limit, the MCU shuts down and the light does things it's not supposed to. Moving the cap breaks the inadvertent boost circuit, and the MCU runs at battery voltage minus the voltage drop through the diode, like you would expect it to. This boosting effect happens even on a completely unmodified 8x 7135 Nanjg 105c running any firmware. But adding a FET in place of the 7135s presents a greater load for the MCU's PWM output, and the boosting effect is increased slightly, which is just enough to push things up into the unreliable range. Adding the gate resistor softens the load the MCU has to drive, so the boosting effect is lessened slightly, which has the roundabout effect of lessening the voltage boosting on Vcc. It's a patch, not a cure.
The gate resistor does not-nice things to the waveform sent to the FET gate, it slows down the turn on/turn off times and gives the FET a semi-sine wave signal instead of a sharp clean square wave signal. FETs don't like sine waves, because that leaves them in the bad neither-on-nor-off range for too long. There's a way to stay away from sending the FET a sine wave signal, that is to relocate the cap and get rid of the gate resistor.
DrJones doesn't work with these direct drive FETs, the 7135s will run reliably with the Nanjg's faulty Vcc design without being unreliable. When driving a FET that design no longer works.
Thanks!
As a point of reference, stock FET drivers (that don't use a separate gate driver or 7136 chip) that have gate resistors, those resistors are always in the 10-15 ohm range. That's a very different prospect compared to the 100 or 130 or 200 ohm needed to patch this voltage boosting issue. A 10 ohm will ease the load on the MCU a little, but yet not drastically alter the shape of the signal sent to the gate. Big 100+ ohm resistors have a dramatic effect on the waveform and are not a good solution to this problem. If these drivers went from not working to working by adding 10-15 ohms at the gate, I'd have absolutely no issue with using them.
Comfychair, your post #71 is informative for those who haven’t been following the conversation in that thread, but I do not think it is accurate on one point. My understanding is that the big gate resistors we are using do not help by “lowering the load on the MCU”, rather they help by softening the hard cutoff to the FET. The boost circuit requires a hard cutoff (so the field will suddenly collapse in the inductor). So with the soft cutoff, you get a much less effective boost circuit / MCU’s Vcc does not get as high.
As you point out in #71 here, the soft cutoff is not a good thing in any other respect so we’d like to avoid it.
I can’t find scope images of Vcc on a driver which has gate resistors in your thread.
Doesn't a stock Nanjg 105C turn on/off at the same speed and work OK without any resistors in the Vdd line? It only goes bad when the 7135s are replaced with a FET.
First post, first two pics. That was before I had the dividers/other settings figured out so take the voltage levels with a grain of salt, but nothing was changed between the two pics. The shape of the signals are directly comparable.
edit: Oh, of Vcc... I don't think I did that. Wasn't really interested after the capacitor move fixed the boosting issue and the drivers started working correctly. I do have pics of a totally stock hardware 105C w/7135s that shows the same boosting issue, just not rising to the level it freaks out the MCU.
I’ve looked at that one. I’m pretty confident about the stuff I wrote in #74, all I meant to say was that I wasn’t just ignoring all those results you posted in the other thread - there simply wasn’t an appropriate image posted for me to point to.
It’s not really the same thing AFAIK. I don’t think the 7135’s are capable of the hard turn-off we get with the MOSFET.
One way to solve the debate on a cap before or after the diode is to eliminate the diode with small signal pfet, like a bsp250 for reverse polarity protection. This board has plenty of room for it.
More debate on where the cap should be? Then here is fellow BLF member who has experienced similar problem that I have seen, only it is well documented in this thread My Medusa. Because different people have different experiences then board layout and component selection could be to blame?
The relevant original post describing why the driver wasn’t working and the fix:
…and it didn’t work. Well it did, but as soon as I screwed the battery pack in it just started flashing. I tried different batteries and sometimes it worked, whils other times it’d just go back to flashing. It drove me nuts! Spent 2 days trouble shooting:
I swapped out the MCU
I tried every battery I owned
I rewired the switch several times
I removed and re-soldered the contact plate several times
Sanded the contact plate down just in case it was a contact issue
I swore. A lot.
I eventually figured out that the driver did not like fully charged batteries, but worked fine when the cells were at their nominal voltage (~3.7V). I eventually emailed tterev3 with my issue and he suggested that the current drain was so high on fully charged cells that it was pulling the voltage too low and resetting the MCU constantly, hence why it worked on more depleted cells and not on fresh cells. The fix was to add a small capacitor between VIN and GND after the reverse polarity diode. In my original circuit I had it placed before the diode. Annoyingly enough this is the first time I’ve ever done that trick and normally place my caps right on the VIN pin of the MCU!
I happened to have already added a small pad for the GND connection very close the VIN pin of the MCU (for programming the PIC) so I soldered straight to that. Lucky!
Now it functions exactly as described here, just with a MUCH brighter white.
So let the debate continue.
P.S. having pads for a cap across the momentary switch inputs would be a plus. Software can handle a switch press several different ways, but mostly polling and interrupts. With interrupts it is nice to have the cap there for debouncing, it just gives you more options and save a further revision.
Well who really knows what happens inside a 7135. The datasheet sure isn't any help.
Oh, just remembered, that scope image of the 105C is showing Vdd and B+, not Vcc... but on all the other drivers with the stock cap/diode arrangement, whatever Vdd (or gate/pin #6) rises to is the stable voltage Vcc is running at. On that one the Vin was well below 4v (something like 3.8v IIRC) yet Vdd is up around 4.5 or so. So it's safe to say a stock 105C has 4.5v on pin #8 when B+ is way down at 3.8v, and much much higher with a fresh cell. It just doesn't go quite high enough to make it unreliable.
How does this work? I mean I understand the principal of how it provides RPP but how does the circuit operate, what triggers the FET, Batt+ directly?