Inside Convoy T2

23 posts / 0 new
Last post
Kame Sennin
Kame Sennin's picture
Offline
Last seen: 2 months 4 days ago
Joined: 09/24/2018 - 14:38
Posts: 228
Location: France
Inside Convoy T2

Today i received my first Convoy T2 !

 

It looks and feel exactly as i expected from a Convoy ie : simple and robust with a clean finish

The only thing that ennoyed me somewhat once the "new toy" effect started to settle down was the feeling that the mid is not really enough (to my liking) for some serious work while the high (and the associated battery drain) is too high, wich was enough of an excuse to decide to open it and see if i could mod it  Wink

 

Pictures an schematic will follow asap but to make a long evening short and after verifying several time i have a question :

Is it common for that kind of flashlight to drive the led with a constant voltage (about 3.25V) through a low value resistor (0.2 ohms) with NO current feedback (or anything else that could make a decent current source) ?

I figured that while you need a boost converter and decide to add some kind of u-controller it wasn't that much work to have them drive the led in current mode ... but i might be wrong !*

 

*Long ago i have built a flashlight with cc mode driver (maybe i'll show you pictures of the poor ugly thing one day) but that was a single mode (and low power) one so it was no big deal

Kame Sennin
Kame Sennin's picture
Offline
Last seen: 2 months 4 days ago
Joined: 09/24/2018 - 14:38
Posts: 228
Location: France

Adding pictures and schematic for the driver (updated schematic according to HarleyQuin's feedback)  :

Links :

 

Convoy T2 driver bot

Convoy T2 top

Outer diameter of the PCB : 17.1mm but it doesn't "fill" the inside diameter of the head (threading ID is 18.0mm)

Inner diameter of the retaining ring : 14.7mm

Phlogiston
Offline
Last seen: 13 hours 1 min ago
Joined: 10/27/2016 - 16:57
Posts: 772
Location: Scotland

Thanks for the schematics and photos, those are always interesting to have Thumbs Up

That driver configuration looks normal for a PWM-controlled LED.

Djozz has found the PWM frequency to be 744Hz, which is unfortunately low and makes the PWM flicker perceptible to a fair number of people. Djozz’s post with his impressions and test results is here:

http://budgetlightforum.com/comment/1378223#comment-1378223

HarleyQuin
HarleyQuin's picture
Offline
Last seen: 8 months 2 weeks ago
Joined: 03/29/2013 - 04:47
Posts: 580

@Kame Sennin
Thanks for the effort, that are very good pictures of the driver, and a schematic to boot Smile

Kame Sennin wrote:
Is it common for that kind of flashlight to drive the led with a constant voltage (about 3.25V) through a low value resistor (0.2 ohms) with NO current feedback (or anything else that could make a decent current source) ?

I’m not sure how this driver handles it, but I can say it’s not ‘common’. The PAM2803 based drivers are indeed current controlled with a sense resistor.
If the driver is in fact not current controlled but voltage controlled, it might be adapted for a certain LED. So R3 (R200) could be a current limiter, protecting the FET, or burning off excess voltage to adapt the output voltage of the boost circuit to the desired Vf of the LED. But I haven’t identified this T2 controller IC yet (I also presume the SOT23-5 with D3 on it).

A defining characteristic is the use of an external switch (8205S) in the boost circuit. I’m curious if the external switch would allow for a higher current than the PAM2803 drivers, although it’s not relevant for 1AA (1.2V) use. Here the cell is the limiting factor. But might be interesting for 2AA.

My hopes for an Attiny based driver are gone, of course, the MCU pinout is definitely not Attiny13A and it can’t be easily swapped. Strange to see the sub-par PWM and modes, it’s not typical for Convoy to use visible PWM and a disco mode in a single mode group. That’s what makes me think he just bought the driver and did not put own development into it. Or he sells more lights to muggles and they demand it… But although I don’t consider this driver suitable for use, it’s interesting to study it.

Some annotations to the schematic:
- C2 has the marking 476, which means 47000000pF (6 zeroes), so it’s 47µF.
- there’s a third ceramic capacitor on the board that’s not in your schematic
- R5 is probably “01D” (zero one D) which means 100k in this code form
Again, I’m very grateful for the information you provided.

Beer

The T2 looks great, I consider it a great host, and I’m on my line of boost drivers again, which were a bit neglected in the past.

Oshpark Boards:
HQ ProgKey: Universal Driver Programming Key . Boost: HQ 15mm/17mm programmable boost driver with ATtiny13A
46mm Triple-Channel: BLF SRK FET v3 . 17mm Linear: HQ10D / HQ4D / HQ4S . Contact Boards: 22/24/26mm

Kame Sennin
Kame Sennin's picture
Offline
Last seen: 2 months 4 days ago
Joined: 09/24/2018 - 14:38
Posts: 228
Location: France

Thank you very much for your feedback !

You are totally right about the 47uF, missing ceramic resistor (near the red wire) and "01D" resistor, i just  updated the schematic.

It took me too much time to reverse engineer the schematic because i was looking for a current drive like the PAM2803 and couldn't have it to fit what was on the board Wink and i was expecting an Attiny too Sad

From what i have seen, without a current mode drive (a higher voltage with a higher value serial resistor is a decent current drive at low power) the led output might change pretty much from one led to the other even in the same batch (U/I curve irregularity) and is more temperature sensitive but that is certainly not a real concern for that kind of use Smile

 

kikkoman
kikkoman's picture
Offline
Last seen: 4 days 33 min ago
Joined: 05/03/2012 - 08:08
Posts: 219
Location: Austria

HarleyQuin wrote:

[…] But I haven’t identified this T2 controller IC yet (I also presume the SOT23-5 with D3 on it).

A defining characteristic is the use of an external switch (8205S) in the boost circuit. I’m curious if the external switch would allow for a higher current than the PAM2803 drivers, although it’s not relevant for 1AA (1.2V) use. Here the cell is the limiting factor. But might be interesting for 2AA.

[…]

I’ve seen this style of marking a number of times before, it took me ages to find out who makes them. Big Smile

It’s a Torex XC9104.

Quote:
Output Current | More than 400mA (VIN=1.8V,VOUT=3.3V)

The parallel 8205S and rather small inductor certainly help its current capabilities but efficiency probably sucks. I’m a bit skeptical about that “800mA”.
I’d really love to see an output/runtime graph on High…

Kame Sennin
Kame Sennin's picture
Offline
Last seen: 2 months 4 days ago
Joined: 09/24/2018 - 14:38
Posts: 228
Location: France

kikkoman wrote:
I've seen this style of marking a number of times before, it took me ages to find out who makes them. Big Smile It's a "Torex XC9104. ":https://www.torexsemi.com/products/step-up-dcdc-converters/series/?name=...

 

Nice find !! I have watched many small boost IC datasheets but couldn't find one that was totally fiting !

 I am somewhat short of time at work right now but i hope to be able to take this driver to the bench soon and see what it can do

 

So now we only miss the MCU ... i have seen that "gnd at pin 8" on some cheap driver's picture already but don't know that kind of IC well enough to guess what it could be. Idea anyone ?

kikkoman
kikkoman's picture
Offline
Last seen: 4 days 33 min ago
Joined: 05/03/2012 - 08:08
Posts: 219
Location: Austria
Kame Sennin wrote:

So now we only miss the MCU … i have seen that “gnd at pin 8” on some cheap driver’s picture already but don’t know that kind of IC well enough to guess what it could be. Idea anyone ?

Could be one of the countless 8-bitters that Holtek makes. Cheap and generic. I’ve found quite a few that have GND on pin8 but pin3 would be reset I assume… no luck so far. Maybe a PIC12Fxxx or a clone, VDD and GND match too.
Does it really matter though? It’s going to be far easier to shoehorn in an Attiny.

HarleyQuin
HarleyQuin's picture
Offline
Last seen: 8 months 2 weeks ago
Joined: 03/29/2013 - 04:47
Posts: 580

Thumbs up for a link to the boost controller, I’m pretty sure I hadn’t seen that one.

I’m interested to see some performance values of the driver. I’ll buy me a pair of T2 at Fasttech, 12.47$ -5% BLF code = 11.85$ ist the best offer around. Banggood, Gearbest, all more expensive despite some code offers.

In the meantime, I need 2 measurements: The outer diameter of the driver (roughly to 1/10 of a mm) and the inner diameter of the retaining ring.
Much obliged.

I’m already tweaking my Oshpark boost drivers for the T2, I’m pretty sure I’ll build them myself as I’m not that fond of shoehorning MCUs… any more Big Smile

Oshpark Boards:
HQ ProgKey: Universal Driver Programming Key . Boost: HQ 15mm/17mm programmable boost driver with ATtiny13A
46mm Triple-Channel: BLF SRK FET v3 . 17mm Linear: HQ10D / HQ4D / HQ4S . Contact Boards: 22/24/26mm

Kame Sennin
Kame Sennin's picture
Offline
Last seen: 2 months 4 days ago
Joined: 09/24/2018 - 14:38
Posts: 228
Location: France

PCB OD : 17.1mm

Retaining ring ID : 14.7mm

Schematic updated, links and  dimensions added

HarleyQuin
HarleyQuin's picture
Offline
Last seen: 8 months 2 weeks ago
Joined: 03/29/2013 - 04:47
Posts: 580

Thanks a lot! Smile

Oshpark Boards:
HQ ProgKey: Universal Driver Programming Key . Boost: HQ 15mm/17mm programmable boost driver with ATtiny13A
46mm Triple-Channel: BLF SRK FET v3 . 17mm Linear: HQ10D / HQ4D / HQ4S . Contact Boards: 22/24/26mm

Kame Sennin
Kame Sennin's picture
Offline
Last seen: 2 months 4 days ago
Joined: 09/24/2018 - 14:38
Posts: 228
Location: France

Trying to measure the current drained from the battery whith an ampmeter did some very weird things at max brightness. I guess that the voltage loss in the measurement shunt was too high and made the driver to enter a very fast ON-OFF loop.

So i measured the voltage across a (very) low R shunt and got  3A max with a freshly charged eneloop.

Changed the 0.2 ohms limiting  resistor in the driver of a second T2 (stupid me who didn't think of measuring this one before modification for reference !) for a 0.5 ohms, got about 2A max and then 1 ohm wich was 1A max

Lux measurments are from a calibrated Minolta T-10A using a 20cm polystyrene foam ball setup according to Djozz's "fast and cheap integrating sphere". As i don't have a calibrating light for now i have no good way to convert from lux to lumen but a Convoy S2+ with xm-l2 T6-4C at 1.4A (4*7135) in mid (40%) mode (should be about 200lm ?) reads 924 lux with the same setup.

 

  T2-a 0R2 T2-b 0R5 T2-b 1R0
  A Lux A Lux A Lux
Low 0.023 13  / 11 0.014 8.6
Mid 0.260 152  / 113 0.133 86
High 3.0 1150 2.0 848 1.05 585
Jack Kellar
Jack Kellar's picture
Offline
Last seen: 15 min 52 sec ago
Joined: 03/09/2014 - 14:21
Posts: 1693
Location: Throwing bolts at anomalies

Anybody got any idea as to when this will start being sold at the Convoy store?

Kame Sennin
Kame Sennin's picture
Offline
Last seen: 2 months 4 days ago
Joined: 09/24/2018 - 14:38
Posts: 228
Location: France

You mean T2 ?

You can buy T2 from Convoy store on aliexpress since the mid august iirc (mine was from a 08/29 order)

Jack Kellar
Jack Kellar's picture
Offline
Last seen: 15 min 52 sec ago
Joined: 03/09/2014 - 14:21
Posts: 1693
Location: Throwing bolts at anomalies

I’d never seen it before there, and it sure is. Cool beans, getting that (alongside an S2+) ASAP.

kikstyo
Offline
Last seen: 3 months 3 weeks ago
Joined: 10/10/2018 - 06:31
Posts: 9
Jack Kellar wrote:
I’d never seen it before there, and it sure is. Cool beans, getting that (alongside an S2+) ASAP.

Get the T2 from their Taobao store instead. https://item.taobao.com/item.htm?spm=a1z10.1-c.w4004-16307498356.2.7c0b2...

Taobao Description:
(2) Low voltage protection: When the battery voltage is between 2.9 and 3.1V, the flashlight will alarm, showing a weak brightness with 2 flashes per second.
(3) Reverse battery protection: the battery will not light up, but there will be no abnormalities, no need to worry about burning the battery or circuit

It has low voltage protection and reverse battery protection while the one on Aliexpress store doesn’t have them. Pretty disappointing as the Aliexpress store is direct from Convoy. Anyone contacted Simon about this?

Jack Kellar
Jack Kellar's picture
Offline
Last seen: 15 min 52 sec ago
Joined: 03/09/2014 - 14:21
Posts: 1693
Location: Throwing bolts at anomalies

Thanks, but nope, don’t intend on getting it from TaoBao, even because I don’t have 14500 cells to toss in it so the LVP is moot and I don’t think they even ship to my country. If anything, I’ll wait for Simon to have his stock updated/upgraded.

2100
Offline
Last seen: 9 months 1 week ago
Joined: 05/28/2011 - 10:28
Posts: 4512
Location: SINGAPORE

Oh shoot, 774Hz PMW freq. What the…..

HarleyQuin
HarleyQuin's picture
Offline
Last seen: 8 months 2 weeks ago
Joined: 03/29/2013 - 04:47
Posts: 580

Got 2x Convoy T2 from Fasttech. I’m very pleased.

The moment I held one in my hand I thought:

  • I need more
  • as host please!!!
  • do these come in colors???

Really nice look and feel. I’m a sucker for the S2+, and this is absolutely its little brother.
Greatest asset is that a driver of 17mm does fit. Makes DIY much easier, and the trade-off of a slightly larger diameter is really insignificant.

Lineup:

From left to right
SK68 clone, Ultrafire SK58, Convoy T2 (/w clip + sling), Thorfire TG06, Uniquefire G10, Utorch UT01, Convoy T2, Manker T01, Thrunite Archer 1A

Stock mode spacing is already really good, waaaaaay much better than most, especially the expensive AA lights. But doubleclick-to-strobe and PWM are a downside. Leaves room for improvement.

Wink

Oshpark Boards:
HQ ProgKey: Universal Driver Programming Key . Boost: HQ 15mm/17mm programmable boost driver with ATtiny13A
46mm Triple-Channel: BLF SRK FET v3 . 17mm Linear: HQ10D / HQ4D / HQ4S . Contact Boards: 22/24/26mm

Sillen
Sillen's picture
Offline
Last seen: 1 week 1 day ago
Joined: 08/01/2018 - 06:39
Posts: 375
Location: Melbourne, Australia

Got my Convoy T2 last week. Haven’t had much time to play with it. First impression, big and heavy for a AA only

If you like the smaller lights, I recently took delivery of a batch of small Sofirn lights – SF10, SF12, SF14v2.0 & SP10v2.0. The SP10 comes in black, red and blue.

HarleyQuin
HarleyQuin's picture
Offline
Last seen: 8 months 2 weeks ago
Joined: 03/29/2013 - 04:47
Posts: 580

Sillen wrote:
If you like the smaller lights, I recently took delivery of a batch of small Sofirn lights – SF10, SF12, SF14v2.0 & SP10v2.0. The SP10 comes in black, red and blue.
I really like Sofirn lights, but not the AA ones so much.

Speaking of size, their 3 clicky lights (SF10, SF12, SF14) are all larger than the T2.
And the SP10 is a momentary light (that comparison would really be apples and oranges) more in the ballpark of the UT01. I eyed the SP10 for some time, but refrained because of the don’t-drop-it-or-it’s-broken problem. And it’s much more difficult to mod them. Most lights I buy, I see them as hosts, which happen to have stuff inside… Smile

Oh, and color, the Sofirn red is too pink for me. The Convoy red is good. Not as nice as the Maglite red, but good. I’d take blue and green as well…

Oshpark Boards:
HQ ProgKey: Universal Driver Programming Key . Boost: HQ 15mm/17mm programmable boost driver with ATtiny13A
46mm Triple-Channel: BLF SRK FET v3 . 17mm Linear: HQ10D / HQ4D / HQ4S . Contact Boards: 22/24/26mm

alberto7
Offline
Last seen: 2 hours 50 min ago
Joined: 08/17/2014 - 06:04
Posts: 202

Anyone know the size of the led mcpcb?

gchart
gchart's picture
Offline
Last seen: 13 hours 40 min ago
Joined: 03/19/2016 - 11:57
Posts: 1630
Location: Central IL

Sorry to dig up an old thread, but I was poking around about the driver in the T2 and ran across this. Great schematic!

And yeah, that’s PIC-like layout, not compatible with our usual attiny’s. But the new ones (attiny412) would work, as long as you don’t mind the new architecture. The one downside I see… no provision for voltage readings via a voltage divider. And since the voltage coming into the MCU is boosted, internal vref readings are out of the question. So no driver-based LVP? Hmm… at least it’s a 17mm driver.