Attiny25/45/85 FW Development Thread

I did some more tests with otc temperature compensation. It now works almost perfectly when bleeding resistor has been removed, and with bleeding resistor in place still much better than without compensation. As Toykeeper mentioned there should be some calibration. This can easily be done when doing calibration for the maximum temperature limit: user has to start this calibration with "cold" light (room temperature), just when started store the (low) temperature into eeprom. For "better" drivers which need less compensation (don't have any) probably decreasing the shift value will help - or use a multiplication or a lookup table instead for more granularity. Look up table might even be the best solution since there are only a few values to look up and optimal compensation is not linear (medium press otc value could even need some more compensation).

My code for banggood driver without bleeding resistor is currently:

#define OTC_SHORT_PRESS 235
#define OTC_MEDIUM_PRESS 200

...
uint8_t thermal_compensation = adc_init_and_read(ADC_MODE_TEMP);
uint8_t thermal_low = eeprom_read_byte((const uint8_t*)EE_ADR_THERMAL_LOW);
if ( thermal_compensation > thermal_low )
thermal_compensation = (thermal_compensation - thermal_low) << 4;
else
thermal_compensation = 0;
...
if ( otc_voltage > (OTC_SHORT_PRESS - thermal_compensation) )
...
if ( otc_voltage > (OTC_MEDIUM_PRESS - thermal_compensation) )
...

Edit:

thermal_compensation must not grow larger than OTC_SHORT_PRESS and OTC_MEDIUM_PRESS, a check for this should be added. If someone calibrates his light at a very low temperature, "OTC_SHORT_PRESS - thermal_compensation" may overflow and result in a too high value which may hinder the user to enter configuration mode with short clicks ever again (at higher temperatures).

Any windows user knows how to flash attiny 25 ?

On attiny13 i used this in a script :

Except the t13 (t25), what else i should change ?

The fuses:
-Ulfuse:w:0xD2:m -Uhfuse:w:0xDE:m

If you don’t change them SPI will be disabled and you can’t program the mcu with USBASP anymore!

Thanks .
I still get an error :

Edit : i use bistro

The fuses have to match up with the firmware as well. Need to know what 25 firmware - the firmware MUST be written for a 25, not a 13A. Bistro is for the 25 for example.

The author of the firmware pretty much always states what fuse values to use. Most of the fuse settings are required to be specific values, to match what the firmware is written for.

The error is NOT related to the fuses. USBASP can’t connect to the mcu, check the connection. In case of wrong mcu type setting you will get an other error message.

Edit:
if you’ve programmed the ATtiny25 before with the fuse settings for Attiny13 you can’t program it anymore without a special programmer. In this case the above error message (2nd) may apply. It happened to me once but I can’t recall the specific error message.
The 1st error message is not important and can be ignored.

Got it working by desoldering the attiny from the pcb and flashing it alone .

My latest order from oshpark (fet+1 drivers) seems to have something wrong … Built today 4 drivers : one with attiny25 , and it works like a one mode driver (like the led - and the ground is connected) , and 3 with attiny13a that don’t even light up…. Checked all 4 for shorts , but there aren’t any….

Sharpie i posted on the drivers thread , seems one pcb (the one with tiny25) had a short from the factory .
The other 3 drivers i don’t know what’s the issue .

Anybody tried the SSN( 105c) version of those controllers? Cause currently my biggest issue is the overheating, on theory they shouldbe the same

Give a clue or link - what are you referring to? What is "those controllers"?

Talking about the 105 degrees C version of ttiny 13 and 25: i see they are awailable at Mouser, at least the 25v SSN

Dang, thought you were talk'n bout Nanjg drivers. Hhmm, never heard bout over-heating issues. What happens? Has anyone else said/noticed this? Are the 25's more prone than the 13A to overheating? What 25 boards are you using - those Astrolux ones?

I use wight’s boards from the development thread , v013 . i build them from scratch .

Sorry I was responding to Mitko. But anyway, what are doing to handle the voltage spikes that cause the 25 to reset? Are you scoping the signals? Generic 13A boards and parts usually result in flaky problems with 25/45/85's -- all noted earlier in this thread.

BangGood (or Manker) that developed the 25 driver for the X5/X6's (runs Bistro) used a 12 uF cap, not the standard 10 uF cap -- that seems to fix the problems on that board (well they think it fixed it mostly), But I've done more.

Yes TOm, the controller is overheating and starts acting wierd, happenes too often in smaller lights like S2( mostly) and C8s with longer turbo run time

Nope, Astro drivers are bad - bad caps, bad soldering work too
I use fet +1 v 09 and now for the first time i used V14 boards

I dunno - I either forgot about this heat problem or never heard it before. Has the topic ever come up? I've run lights to the untouchable level, but do't recall seeing anything wierd.

Edit: But.. You got me think'n. Maybe when they are that hot, I just turn them off and don't play with modes, etc. Guess it depends on the nature of the weird problems - flaky while just running in a mode, or flaky when operating (changing modes, etc.).

Try making an offer on this one - you should get it for ~15$ (AU)

I feel like this might just be a big development. :beer: :+1:

I recommend the cheap ones from ebay, there are several vendors, e. g.:

You’ll need more than one if you are going to use them often, this clips - also the expensive ones - don’t last forever. I used the expensive clips of 3M and Pomona before, they have not been better than the cheap ebay ones.
As you probably know you will have to re-wire the connector.

I agree with Mike.

I have flashed over 30 drivers the last 3 months with the cheap soic8 clip from ebay without any issue.