[WIP] 15mm PAM2803 w/ ATtiny13A rough layout / possibility

I don’t know why they’d even bring up castellations if they aren’t going to get plated. See here - OSH Park Docs ~ Services ~ 2 Layer 2oz 0.8mm Service

LVP, if you gotta have it, just glue the voltage divider on top of the mcu. You’ll still need to choose appropriate new values and add the programming. Do that before you glue and air wire.

For these 15mm boards. Do we know how tall they will be?

depends on oshpark pcb and inductor used
I stated heights in post#84

How will the programming need to differ from other 13A controlled drivers? Would it be possible to swap a 105C mcu and leave LVP unconnected? Build one and see?

The only difference is that we’d disable LVP when building the firmware. From HarleyQuin’s work earlier in this thread I think it’s safe to say that standard PWM freqs are OK. Lower PWM freqs could be required to eliminate buzzing/whining. I still don’t have anything built; my donor drivers have not arrived from FT yet.

I can’t remember = build one and see. Worst case scenario, simply tie LVP pin to Vcc. Leaving the pin ‘floating’ could lead to unpredictable results…

Irlm2502 Rdson 40mOhms at 2.5V vs 40 mOhms at 4V. I’ll have another look at the lfpak 33 but it might be too big.

This lfpak33 sold by RMM seems to have good specs too.
The lfpak33 is similar in size to the others but with obviously a different footprint. I don’t know if that’s advantageous or not.

I used the usual 18 kHz. What bothered me most was that at PWM-1 it still gave 16mA. That’s pretty much for a lowest mode.

.

The FET I used in my first Nanjg110 conversion (Post#24) was a IRLML0030TRPbF, which has (according to datasheet) even a slightly lower Rds than the IRLML2502.

.

Out of curiosity I made a quick and dirty draft of the LFPAK33 in the 17mm board. It’s a mere finger-exercise just to see the part in the board. The LED pads are not on opposite any more, which could be a severe blow if height is a factor. A smaller inductor would give more space on the board, but we might lose more efficiency with a worse inductor than we gain with the LFPAK33. But I’m only guesstimating here.

Here is another layout which stays closer to the Fasttech driver.

The advantage of this design would be, that most parts could be sourced from the Fasttech driver. Only the ATtiny13A and the OTC are missing. Including R2 and R3 and the SS24 diode I could not fit it resonably onto 15mm, so this is a 17mm design. C2 might be not sufficient, though.

It’s an earlier version where I tried to stick to the supersized SS24 diode. I abandoned the idea as far as I’m concerned, as the BAT60A works so well. But perhaps someone else has use for the design.

17mm layout to source most parts from the Fasttech boost driver

Update:

  1. I received the FT boards today. Along with the diodes ImA4Wheelr sent me and the v018 PCBs from OSH Park I had everything I needed to start testing.
  2. First I built the boost circuit on my v018 PCB. This functioned as expected.
  3. Next I populated the MCU and transistor. Since (as HQ keeps reminding me) I never bothered to hook Gate up to MCU Pin 6 like it’s supposed to be I bridged 5 and 6 together for now. Unfortunately this setup has not worked correctly. The driver lights the LED but I have no modes.

Maybe tomorrow I will troubleshoot the situation. The issue is most likely the lack of pulldown resistors. I forgot or did not realize that the FT board uses 0603 sized components. Maybe I will re-do my board with that size of component.

EDIT: Bad news (for me). I was hoping to “trouble shoot” the driver in the sense of “install those resistors and see if it works”. Unfortunately I got motivated just now and installed the resistors. The behavior remains the same. That means that the next step is real troubleshooting with DMM / scope / loupe etc. That is definitely not going to happen tonight, but might happen tomorrow.

wight, At what current does the LED light? Could it be a firmware issue?
HQ- true, not quite diametrically opposed but if OTC were rotated 90^ clockwise towards the via the LED pads would be closer to opposite on the lfpak33 board. I was looking at the Rdson graphs rather than the rating at 10V. According to the graphs the 2502 drops below 40mOhms at 2.5Vgs vs 4Vgs for the 0030 (look at figures 12 and 11respevtively). In any case the proof would be in the pudding or maybe I’m completely misinterpreting those graphs.
If the driver has a spring on it then it seems to me that it might be ok clearance wise to go ahead and put the odd 0603 resistor on the bottom especially if, as with pull down resistors or caps the part is grounded to one of the pads and only one via per part is needed. 3 or 4 such part placements might clear a fair amount of space for a larger inductor and make soldering the darn things easier to boot.

@wight

No modes: I re-checked your v018 and can’t find a difference in the circuit to my boards. So in theory it should work.

.

Which firmware did you use and could it be simplified? This very simple firmware did the job for me.

One idea is the OffTimeCapacitor: the stock values for STAR firmware might not work (combination of OTC-value and ADC_crit/delay in the firmware). If mode change would require only a few ms press, you could hardly do it.
Other parts of more sophisticated firmware might be affected as well (ADC, WDT, delay, Fast/Phase, frequency).

Another thumbs up to ImA4Wheelr for sending you his spare BAT60As and I’m glad that these seem to work in your boost circuit as well.

Pin 5/6: I ‘kept reminding’ you only twice (once per board version), so I consider myself far from nagging :slight_smile:

.

.

@RBD
PM sent.

OTC 90 degree: On the 17mm board this might work. On the 15mm it might collide with the programming clip. On board re-programming was one of my major design goals, so I need space around the MCU. The 0603 cap is higher than the 0603 resistor.

FET Rds: I only looked at the value at 4.5V (typ.) and there the 0030 won :stuck_out_tongue:
I can source the 0030 here for 16ct at the shop around the corner, that’s the reason I like it most.

I avoided components on the spring side at all cost. That’s why I used these nice vias and paths. Most of the lights I plan this driver for have no room to spare in the battery tube.
I converted several Solarforce L2m and clones not only to 18500 but to AA as well, this is tight, I already dented several Eneloops on the minus side. The Mini-Maglite (2AA) has only 3 quarter of a turn from tight to falling off. The G10 lost space in the battery tube when I made them tailstand.
None of these lights have a spring on the driver side, simply no room. I’m extremely happy that Oshpark came up with the 0.8mm boards, which will give some relief.

@all: I used the stock Nanjg-105d firmware from FT to avoid a late night flashing session. FWIW the relevant RDSON figures are those in the VGS range of LED Vf to PAM2803 overvoltage protection.

@RBD: It appeared to be the full amount of current, but I did not put a meter on it. It was past my bedtime.

@HQ: Thanks for rechecking HQ - that jives with my own checking. Even if you only reminded me twice, I probably appreciated it both times. Since I was running the Nanjg ontime firmware I did not install an OTC.

I will now try shorting the output cap, maybe the pulldown is not fast enough.

First off, I appreciate the time you both are investing in this and I hope you two don’t mind my input but please consider me a voice of ignorant curiosity and anything I learn is shared with others. If at any time you wish me to bugger off just let me know and I’ll do so.
With regard to theRdson I just assumed that the Vf is around 3V with an LED current ~1A so that would be the voltage available at the gate and that the ratings at 4.5V and 10V were not as critical as the expected Rdson value at 3V. This might be completely wrong if the boost voltage is spiking but that’s a bit above me. I don’t have a preferred fet but as HQ asked about other possibilities I chimed in along with my reasoning.
Until there are preprogrammed ic’s available for this all I can do is mess around with transplanted ones but I still enjoy the puzzle and the process.
As far as driver space goes, what can I say? I cheat and if I need more room I make it but I can certainly understand the requirement to keep it one sided if possible.
Thanks again guys.

Based on my fidding over the past hour the driver pretty clearly works. I just can’t remember / don’t know how to change a firmware over to properly support this hardware configuration (PWM on Pin 5). I have been able to do this in the past. It should be pretty straightforward stuff, but getting everything setup correctly is just way beyond my patience right now. (It requires knowledge of setting up the AVR’s registers which I have lost.)

I should have fixed the pin5/pin6 discrepancy on the PCB some time ago so I would not have to worry about this.

In the meantime I can turn the light on or off depending on what firmware I flash. :-((

Follow up: Just to be 100% sure that the driver was working normally I flashed STAR (ontime) on there and then bridged pin 5/6 again. Modes worked fine. At this moment I have very limited desire to figure out the right register settings to get output on Pin 5. I’ll simply correct the PCB to use Pin 6. The next steps are really testing how necessary the pulldown resistors are!

You’re certainly not bothering me.

OK, this is the last test before I call it quits for the evening; the results are good I think.

I have removed both pulldown resistors and the circuit seems to operate just fine. I had 3 modes beforehand and I still have 3 modes now. At the moment it looks like we can get away without having them, although using the OTC may change things. OSH Park ~

To get PWM=1 lower it may be necessary to drop PWM frequency.

That’s good news wight.

+1 I’m excited that there is significant progress being made RE: programmable boost drivers!