6. Batch "TA" 1S or 2S Bistro HD OTSM/NarsilM drivers 15-47.5mm size fit, S2+, C8, H03, MF01, MT03, Q8 Clicky or E-switch

307 posts / 0 new
Last post
ma tumba
Offline
Last seen: 5 days 1 hour ago
Joined: 07/28/2015 - 19:09
Posts: 168
Location: Russia

Lexel wrote:
AMCs usually 2-3

Sorry, what does this mean?
beastlykings
beastlykings's picture
Offline
Last seen: 8 hours 42 min ago
Joined: 12/29/2017 - 17:06
Posts: 186
Location: Michigan, USA
zeroflow wrote:
beastlykings wrote:
Lexel wrote:
I worked today on Ramping IOS and Anduril and got both working to compile a hex without errors in Atmel Studio

Mind if I ask what parameters you used for Anduril?

I wonder, because I’ve decided that’s what I want to flash onto the H03 drivers you’ve sent me.

I’ve been doing some research, and the closest I can find is something in the main Anduril thread where someone said “The TA uses 3 power channels, so the #define for the FW3A might work.”

So I’m assuming that’s a good starting point? Does anyone know if this still holds true?

Sorry, I know Anduril isn’t your thing and you’re just getting started in it. I just love this UI so stinking much I flashed all my compatible lights over to it already.

If you can wait until tomorrow, I can send you my Anduril config for the H03.

In general, the ramping levels are from the FW3A while the button light things are from the D4S.

Oh man that would be great! That makes sense. I can definitely wait, tracking says they should be here in a day or two. Can you send me the hex too? Atmel studio has been giving me fits, and I haven’t gotten around to setting up an ubuntu environment, so just dropping the hex in there is ideal until I get around to customizing things.

Thank you!

beastlykings
beastlykings's picture
Offline
Last seen: 8 hours 42 min ago
Joined: 12/29/2017 - 17:06
Posts: 186
Location: Michigan, USA
ma tumba wrote:
Lexel wrote:
AMCs usually 2-3
Sorry, what does this mean?

I don’t know much about anything, but the AMC is probably the 7135 current regulator.

JasonWW
JasonWW's picture
Offline
Last seen: 2 hours 18 min ago
Joined: 10/22/2016 - 11:41
Posts: 11336
Location: Houston Texas
beastlykings wrote:
ma tumba wrote:
Lexel wrote:
AMCs usually 2-3
Sorry, what does this mean?

I don’t know much about anything, but the AMC is probably the 7135 current regulator.


It is.

I think Lexel is saying that if you use a 1/255 on the AMC 7135 channel the current will be too low and the led will not light up. So lowest he uses for the AMC is either 2 or 3.

zeroflow
Offline
Last seen: 1 week 4 days ago
Joined: 05/23/2017 - 02:15
Posts: 189
Location: Austria

beastlykings wrote:
Can you send me the hex too? Atmel studio has been giving me fits, and I haven't gotten around to setting up an ubuntu environment, so just dropping the hex in there is ideal until I get around to customizing things. Thank you!

 

As promised:

 

Hex with stock Anduril

https://zeroflow.at/flashlight/anduril.latest.H03_TA.hex

  • Stock Anduril
  • Ramping config from FW3A
  • Aux-LED is disabled as there was not enough space with the 3rd PWM channel

Hex with slightly modified Anduril

https://zeroflow.at/flashlight/anduril.zeroflow.H03_TA.hex

  • Anduril, but disabled stepped ramping
  • Lowest moonlight may flicker, in that case, set ramping floor to level 2 to fix it
  • AUX-LED works like in the D4S

Anduril-Source with Atmel Studio project

https://zeroflow.at/flashlight/flashlight-firmware-original.zip

  • Stock Anduril
  • Added files / defines for H03 Texas Avenger driver
  • Added Atmel Studio Project
  • I only tested if it builds
  • Structure is the same as stock Anduril, so if you want to get newer code from ToyKeepers repository,you can merge it.

Anduril-Source with my changes

https://zeroflow.at/flashlight/flashlight-firmware-zeroflow-inofficial.zip

  • Add defines for all blinky / flashy modes
  • Add define for stepped ramping
  • Add define for level 1 = reduced moonlight
  • Add H03_TA config with FW3A ramps and D4S AUX-LED

Guide how to import Anduril into Atmel Studio

https://imgur.com/a/NxB0xuY

  • No grantee for correctness, it just creates a hex file.

 

So now, if you want a lighted button and stepped ramping, you would need to search what else to disable to save the missing 52 bytes of flash.

contactcr
Offline
Last seen: 24 min 40 sec ago
Joined: 05/19/2017 - 18:52
Posts: 2598
Location: US

Looks like the time has come to tweak your 4S Q8 buck driver now that Sofirn has made this Q8 battery carrier extension. XHP35’s here we come?

beastlykings
beastlykings's picture
Offline
Last seen: 8 hours 42 min ago
Joined: 12/29/2017 - 17:06
Posts: 186
Location: Michigan, USA
zeroflow wrote:

beastlykings wrote:
Can you send me the hex too? Atmel studio has been giving me fits, and I haven’t gotten around to setting up an ubuntu environment, so just dropping the hex in there is ideal until I get around to customizing things. Thank you!

 


As promised:


 


Hex with stock Anduril


https://zeroflow.at/flashlight/anduril.latest.H03_TA.hex



  • Stock Anduril

  • Ramping config from FW3A

  • Aux-LED is disabled as there was not enough space with the 3rd PWM channel


Hex with slightly modified Anduril


https://zeroflow.at/flashlight/anduril.zeroflow.H03_TA.hex



  • Anduril, but disabled stepped ramping

  • Lowest moonlight may flicker, in that case, set ramping floor to level 2 to fix it

  • AUX-LED works like in the D4S


Anduril-Source with Atmel Studio project


https://zeroflow.at/flashlight/flashlight-firmware-original.zip



  • Stock Anduril

  • Added files / defines for H03 Texas Avenger driver

  • Added Atmel Studio Project

  • I only tested if it builds

  • Structure is the same as stock Anduril, so if you want to get newer code from ToyKeepers repository,you can merge it.


Anduril-Source with my changes


https://zeroflow.at/flashlight/flashlight-firmware-zeroflow-inofficial.zip



  • Add defines for all blinky / flashy modes

  • Add define for stepped ramping

  • Add define for level 1 = reduced moonlight

  • Add H03_TA config with FW3A ramps and D4S AUX-LED


Guide how to import Anduril into Atmel Studio


https://imgur.com/a/NxB0xuY



  • No grantee for correctness, it just creates a hex file.


 


So now, if you want a lighted button and stepped ramping, you would need to search what else to disable to save the missing 52 bytes of flash.

Thank you very much! Everyone is so helpful here.

I can’t wait for them stinking drivers to get here!

ma tumba
Offline
Last seen: 5 days 1 hour ago
Joined: 07/28/2015 - 19:09
Posts: 168
Location: Russia
JasonWW wrote:
beastlykings wrote:
ma tumba wrote:
Lexel wrote:
AMCs usually 2-3
Sorry, what does this mean?

I don’t know much about anything, but the AMC is probably the 7135 current regulator.


It is.

I think Lexel is saying that if you use a 1/255 on the AMC 7135 channel the current will be too low and the led will not light up. So lowest he uses for the AMC is either 2 or 3.


I did not realize that the numerical levels come off the AMCs, other than one/all7135. Is that so?
JasonWW
JasonWW's picture
Offline
Last seen: 2 hours 18 min ago
Joined: 10/22/2016 - 11:41
Posts: 11336
Location: Houston Texas
ma tumba wrote:
JasonWW wrote:
beastlykings wrote:
ma tumba wrote:
Lexel wrote:
AMCs usually 2-3
Sorry, what does this mean?

I don’t know much about anything, but the AMC is probably the 7135 current regulator.


It is.

I think Lexel is saying that if you use a 1/255 on the AMC 7135 channel the current will be too low and the led will not light up. So lowest he uses for the AMC is either 2 or 3.


I did not realize that the numerical levels come off the AMCs, other than one/all7135. Is that so?

I really dont know, I’m just translating for Lexel.

I think he’s saying that on a FET only driver (1 channel) you have a huge range of brightness levels from many amps down to milliamps and you can use that step 1 of 255 and the led will light up just fine. In fact it may not get dim enough for a true moonlight.

On a 2 channel driver which is using a single AMC 7135 chip the range of amperage is much smaller. Max is only 350 miliamps and if you use 1 of 255 the output is so low the led might not light up.

Does that make sense?

Lexel
Lexel's picture
Offline
Last seen: 1 hour 16 min ago
Joined: 11/01/2016 - 08:00
Posts: 5251
Location: Germany

JasonWW wrote:

I really dont know, I’m just translating for Lexel.

I think he’s saying that on a FET only driver (1 channel) you have a huge range of brightness levels from many amps down to milliamps and you can use that step 1 of 255 and the led will light up just fine. In fact it may not get dim enough for a true moonlight.

On a 2 channel driver which is using a single AMC 7135 chip the range of amperage is much smaller. Max is only 350 miliamps and if you use 1 of 255 the output is so low the led might not light up.

Does that make sense?

We use 8 Bit dimming per channel from the MCU so 256 levels
of course 0 does not count, so we got 255 steps left

FETs are pretty fast in switching so with 20kHz PWM we have 50us for a full PWM cycle
50us/256 we get about 200ns

the AMC is a current regulator and on the lower end the very short PWM cycles do barely switch current through its like a low pass filter, the shorter the on cycle in the PWM is the higher is its frequency change from low to low

JasonWW
JasonWW's picture
Offline
Last seen: 2 hours 18 min ago
Joined: 10/22/2016 - 11:41
Posts: 11336
Location: Houston Texas

Lexel wrote:

We use 8 Bit dimming per channel from the MCU so 256 levels
of course 0 does not count, so we got 255 steps left

FETs are pretty fast in switching so with 20kHz PWM we have 50us for a full PWM cycle
50us/256 we get about 200ns

the AMC is a current regulator and on the lower end the very short PWM cycles do barely switch current through its like a low pass filter, the shorter the on cycle in the PWM is the higher is its frequency change from low to low


Now I need someone to translate for me. Lol

Okay, I think I understand some of it.

So 8 bits is like: 00100010
There are a total of 255 different combinations of ones and zeros. Each of these “steps” represents an amperage and brightness level for that channel.

This is for each channel? I wasn’t sure about that. I thought maybe it was 255 steps for all channels. Anyway…

So the MOFFET’S you use typically cycle on and off 20,000 times per second.

So 20,000 units into 1 second = 50 microseconds.

50 microseconds divided by 255 steps gives you an even smaller 200 nanoseconds.

I don’t understand beyond this. I’ll have to think about it and PWM later.

So I think my earlier translation was right. With only one AMC 7135 and 350mA if we divide this up into 255 sections, the lowest (a 1 out of 255) is not likely to light up the led. Lexel seems to usually only go down to 2 or 3 out of 255 to get that very dim moonlight.

zeroflow
Offline
Last seen: 1 week 4 days ago
Joined: 05/23/2017 - 02:15
Posts: 189
Location: Austria

@Level: For the Q8 4P linear driver: Is it possible to buy the driver with a soldered brass ring or does that need to be harvested off an existing driver?

JasonWW
JasonWW's picture
Offline
Last seen: 2 hours 18 min ago
Joined: 10/22/2016 - 11:41
Posts: 11336
Location: Houston Texas

zeroflow wrote:

@Level: For the Q8 4P linear driver: Is it possible to buy the driver with a soldered brass ring or does that need to be harvested off an existing driver?


Your asking about a linear driver?
BlueSwordM
BlueSwordM's picture
Online
Last seen: 11 min 45 sec ago
Joined: 11/29/2017 - 12:34
Posts: 5190
Location: Canada

He is probably asking about the FET+7135+16×7135 driver, not for a pure linear driver.

Or maybe he has 219Bs and wants to driver them in a regular manner.

My very own high current Beryllium Copper springs Gen 3:
http://budgetlightforum.com/node/67401
Liitokala Aliexpress Stores Battery Fraud: http://budgetlightforum.com/node/60547

JasonWW
JasonWW's picture
Offline
Last seen: 2 hours 18 min ago
Joined: 10/22/2016 - 11:41
Posts: 11336
Location: Houston Texas

I was going to say you could buy the factory driver for like 10 bucks that’s ready to go, but I’m not sure if Lexel is making some kind of special linear driver.

Any of Lexels aftermarket drivers do require swapping the brass ring over.

Lexel
Lexel's picture
Offline
Last seen: 1 hour 16 min ago
Joined: 11/01/2016 - 08:00
Posts: 5251
Location: Germany

16x AMC driver is up to 5.6A regulated, above that the FET comed into play

contactcr
Offline
Last seen: 24 min 40 sec ago
Joined: 05/19/2017 - 18:52
Posts: 2598
Location: US

Lexel, please work your magic when you have time.

C8F 21700 dimensions:

driver cavity = 21.26mm


retaining ring ID = 15.96mm

beastlykings
beastlykings's picture
Offline
Last seen: 8 hours 42 min ago
Joined: 12/29/2017 - 17:06
Posts: 186
Location: Michigan, USA

@zeroflow hey I’ve been trying to get these files to flash and I’m pulling my hair out. I’ve tried researching for awhile now with no success. Everything seems like a bad connection to my programmer, but I used this same config earlier today to flash my buddies D4. Of the two Lexel drivers I got, neither would flash, and I kept getting the error
“programm enable: target doesn’t answer. 1.
Initializelation failed, rc=1.”

So I did some googling and tried using -B 32 to lower the clock speed. No avail. I cleaned the contacts, I reseted connections, I changed from 5v to 3.3v, I did everything.. Eventually I used the -F command to force it to write the hex. Which it did, and then failed to verify, and the driver ceased to function when tested in the light with battery.
I thought it was the hex, so I flashed a known-good D4 hex to it, same results, flash but fail to verify.

I got the second driver out, which I had not soldered to yet, and it gave the same exact initial error. I did not use the -F command on it. I went back and forth between the two drivers, I could not get either to work with any combination of settings. Then, suddenly, the second driver flashed the h03.hex with the stock settings for avrdude,I have no idea why. It verified fine. I have not checked to see if it is functional in a light.

The first one still fails to flash. Now it is saying the device signature is wrong. And indeed, the signature does change every time. This leads me to think it’s a connection problem, but I’ve flashed a half dozen lights with this already with zero issues.

The only other thing I can think, is that I accidentally bridged the negative led terminal by the FET with the negative ring around the PCB. And when I put the light together the light came on full brightness and it took me a few seconds to figure it out. Could this have damaged the MCU? I cleared the bridge, and the stock loaded Narsilm ran fine, ramping and turbo and everything else, so it seemed to me that the MCU was fine. Could something else have burned? I can try soldering on a new MCU, but I will wait for guidance.

Also, if it’s a damaged MCU, then why did the untouched driver also fail to flash at first?? Makes no sense.

Sorry this is so long, I’ve tried helping myself here but I’m getting over my head. Any suggestions guys?

Edit: I’m using the hex files you sent me.

beastlykings
beastlykings's picture
Offline
Last seen: 8 hours 42 min ago
Joined: 12/29/2017 - 17:06
Posts: 186
Location: Michigan, USA

Alright so the MCU that managed to flash.. works perfectly in the light.

So I tried flashing the bad one again and got the same results. So I checked my wiring, pulled apart my taped joint where the solder joints are, it’s all fine. Then I got the idea, if the MCU is bad, then I’ll flash the new one before soldering it in. I get the same error when trying to flash the bare chip. And then I realized I get the same error with nothing attached at all. So something is definitely screwy. I think it’s the flasher I have, good thing I bought two, when I get home on Tuesday I will try the backup flasher.

Any other ideas beyond that?

Lexel
Lexel's picture
Offline
Last seen: 1 hour 16 min ago
Joined: 11/01/2016 - 08:00
Posts: 5251
Location: Germany

beastlykings wrote:
@zeroflow hey I’ve been trying to get these files to flash and I’m pulling my hair out. I’ve tried researching for awhile now with no success. Everything seems like a bad connection to my programmer, but I used this same config earlier today to flash my buddies D4. Of the two Lexel drivers I got, neither would flash, and I kept getting the error
“programm enable: target doesn’t answer. 1.
Initializelation failed, rc=1.”

So I did some googling and tried using -B 32 to lower the clock speed. No avail. I cleaned the contacts, I reseted connections, I changed from 5v to 3.3v, I did everything.. Eventually I used the -F command to force it to write the hex. Which it did, and then failed to verify, and the driver ceased to function when tested in the light with battery.
I thought it was the hex, so I flashed a known-good D4 hex to it, same results, flash but fail to verify.

I got the second driver out, which I had not soldered to yet, and it gave the same exact initial error. I did not use the -F command on it. I went back and forth between the two drivers, I could not get either to work with any combination of settings. Then, suddenly, the second driver flashed the h03.hex with the stock settings for avrdude,I have no idea why. It verified fine. I have not checked to see if it is functional in a light.

The first one still fails to flash. Now it is saying the device signature is wrong. And indeed, the signature does change every time. This leads me to think it’s a connection problem, but I’ve flashed a half dozen lights with this already with zero issues.

The only other thing I can think, is that I accidentally bridged the negative led terminal by the FET with the negative ring around the PCB. And when I put the light together the light came on full brightness and it took me a few seconds to figure it out. Could this have damaged the MCU? I cleared the bridge, and the stock loaded Narsilm ran fine, ramping and turbo and everything else, so it seemed to me that the MCU was fine. Could something else have burned? I can try soldering on a new MCU, but I will wait for guidance.

Also, if it’s a damaged MCU, then why did the untouched driver also fail to flash at first?? Makes no sense.

Sorry this is so long, I’ve tried helping myself here but I’m getting over my head. Any suggestions guys?

Edit: I’m using the hex files you sent me.

My Buck drivers have a capacitor on the LVP voltage divider to avoid problems with ripple current, you have to look at it ist 0603 C3 marked
one side has a solder bridge to a pad, remove this bridge with solder wick to flash it

Lexel
Lexel's picture
Offline
Last seen: 1 hour 16 min ago
Joined: 11/01/2016 - 08:00
Posts: 5251
Location: Germany

As I get often questions like is it already shipped ect. I point again on Post 2 with the Excel table with the Order information
Also its best to check if there is no error in your order I try my best but with that many people and conversations there are always some errors possible

follow here the progress of your order and check if the order contains all you need

Tracking website

https://www.deutschepost.de/sendung/simpleQuery.html?locale=en_GB

ma tumba
Offline
Last seen: 5 days 1 hour ago
Joined: 07/28/2015 - 19:09
Posts: 168
Location: Russia

I am not sure how do I access the excel spreadsheet. Post #2 has a screenshot of the spreadsheet which is obviously a truncated part of the entire thing.

beastlykings
beastlykings's picture
Offline
Last seen: 8 hours 42 min ago
Joined: 12/29/2017 - 17:06
Posts: 186
Location: Michigan, USA
ma tumba wrote:
I am not sure how do I access the excel spreadsheet. Post #2 has a screenshot of the spreadsheet which is obviously a truncated part of the entire thing.

I think that’s the point, or else everyone’s full address and other information would be on the open internet

LichtAn
Offline
Last seen: 4 months 1 week ago
Joined: 03/08/2018 - 02:37
Posts: 322
Location: Germany

It’s a .png file. An image.

Lexel
Lexel's picture
Offline
Last seen: 1 hour 16 min ago
Joined: 11/01/2016 - 08:00
Posts: 5251
Location: Germany
LichtAn wrote:
It’s a .png file. An image.

yes of course full address should be hidden

I had it one time shipping stuff to the wrong address, so better check it there

beastlykings
beastlykings's picture
Offline
Last seen: 8 hours 42 min ago
Joined: 12/29/2017 - 17:06
Posts: 186
Location: Michigan, USA

For future reference of anyone reading this and having the same problem as me: I don’t know why I couldn’t flash the bare MCUs when I tried the first time, but I can now. The problem was definitely just a dead MCU on that driver. I’m guessing I killed it when I bridged LED negative to ground. Use your multimeters folks. Everything is fine and well and running smoothly with my lights, sorry for clogging up this thread haha.

Lexel
Lexel's picture
Offline
Last seen: 1 hour 16 min ago
Joined: 11/01/2016 - 08:00
Posts: 5251
Location: Germany

I had also a few MCUs killed while using the SOIC clip, bad connection while flashing and it went dead

-X3-
-X3-'s picture
Offline
Last seen: 20 min 13 sec ago
Joined: 07/16/2014 - 03:28
Posts: 2708
Location: France, Angers

Received my two LX1 drivers yesterday Big Smile
Finally my 2D Maglite can shine again !
Now, what to do with the second one ?:-p

"-X3-, is there any place in your house without a flashlight ?" 

My Flashlight public album (mods, emitter swaps, eye candy)

My reviews channel (French language, Olight, Thorfire, Sofirn, Lumintop : 60+ lights tested)

My personal channel (including Olight SR mini, S1, S2, S1A and S-mini disassembly)

M4DM4X blog, saves you $$$ 

beastlykings
beastlykings's picture
Offline
Last seen: 8 hours 42 min ago
Joined: 12/29/2017 - 17:06
Posts: 186
Location: Michigan, USA

Hello again. I have a question. I paired one of your drivers I built with the stock emitter from the H03, and separately, with a 219C, both work fine. But when I pair it with an LH351D, it also works fine, but it glows ever so faintly when it should be off. It doesn’t matter if it’s in lockout mode or just off, when the battery is connected, it glows extremely faintly. When the battery is disconnected, the light continues to glow for a couple seconds before going out.

I’ve isolated the issue to the driver itself, it happens on both ones I built. I haven’t been able to test the ones you built, they are in lights that I gave to friends out of town.

I took the drivers out of the light, disconnected all wires but the LED wires. I’ve tried it with several different LEDs of different types, 219C, XML2, and the LH351D.

The light is so faint, I’m sure the standby time is still extremely long, but I can’t imagine this is normal? What could be wrong that’s causing this? What is special about the LH351D that the other emitters aren’t glowing?

As always thanks for all the help.

zeroflow
Offline
Last seen: 1 week 4 days ago
Joined: 05/23/2017 - 02:15
Posts: 189
Location: Austria

beastlykings wrote:
But when I pair it with an LH351D, it also works fine, but it glows ever so faintly when it should be off.

 

Ah, good to hear that I'm not the only one.

I messaged Lexel about that and he gave me the advice of cleaning the driver. Sadly that didn't work.

 

What's pretty clear is, that there is some leakage current somewhere in the path. And considering how the LED is connected, it needs to be somewhere in the negative path as the positive is directly connected.

 

Furthermore, I tried running the driver outside the host with the LED still connected to rule out any leakage current through the MCPCB or the host itself.

In that case, the LED still has a dim glow, so I would rule out the MCPCB and Host.

 

So now, on the driver, the following things are connected to LED-:

  • MOSFET
  • Bank of 7135 chips on the bottom
  • Single 7135 chip on the top
  • (The driver PCB itself)

 

The MOSFET should be a BSC009NE2LS5I which has a diagram for drain-source leakage current in it.

In the diagram, I see that a Vds between 3..4V at room temperature would yield around 30µA of current (If I'm not mistaken)

 

 

That's not much, but maybe that's enough for the LED to light up.

 

Maybe if I have too much time, I will desolder the FET and see if the AMC7135s also have a leakage current.

Another option would be someone with a low value current source who could test at which level the LH351D start a dim glow.

 

 

 

Pages