New Convoy C8 – Clearly better

What color is your driver’s board?

There are some potential issues with the new layout of the new style driver. Tom E has found that pin 5 of the MCU is connected to the ground pin of the 7135 and this is interfering with some timing control. I had re-flashed my MCU with firmware direct from ToyKeeper and while it was somewhat better due to the proper fuses being used in the flash code, it still wasn’t perfect. I then found out about Tom’s findings and opened my Clear C8 back up, sliced (carefully) that trace to eliminate ground to pin 5, and now it works perfectly.

You might want to look at this Y shaped trace coming from pin 5 and address this issue, see what that does for your lights operation.

Edit: For confirmation, this is on the new red board with Bistro Mini (or Biscotti as it’s being called)

Edit II: This information is being looked at by Simon and his people, the holidays are probably going to slow down any response, but corrections, if found necessary, will be made as soon as possible by Convoy.

I think the Clear C8 from GB and BG have the old/traditional Convoy firmware, which isn’t surprising given the overclocking issue with the new firmware.

Green, with an all-lowercase “convoy” printed on it.

Oh, bugger. Are there any stars/jumpers/traces/anything to get rid of the damned blinkies, or should I just throw in a 105C and be done with it? I bought the beastie just so I wouldn’t have to do anything like this, and…

I’m okay with 1-mode, so I might just hardwire the damned enables on the 7135s to do just that. Just charging a 30Q to go into it, might have to use a protected panny-B instead, then (no LVP anymore if I do the hardwiring).

Doubt it (green board). Unless that comes with blinkies enabled by default, too. Haven’t tried diddling with the switch to try to get into mode-selection, if all bistroed out.

Bah. I think I might just have to go to hardwired 1-mode to fix, well, everything.

So much for buying off-the-rack…

It’s a black one, think it was one of GB’s flash sales or regular sales or something. I just wanted a NW L2-HI to play with, and didn’t want to reflow a HI, screw around with seeing if it’d focus in a non-HI reflector, blah blah blah, all the usual nonsense. For once — ONCE — I just wanted an off-the-rack C8 that I wouldn’t have to build from scratch to get rid of the usual annoyances.

For some reason, I thought C8s used a regular 105C, and that the different UIs were just in firmware.

Didn’t even look to see if there’s any AS5 or anything under the star. Would be just my luck to have it mounted dry… :stuck_out_tongue:

That’s it. I’ve died and gone to Hell…

I thought so too (well, 105D) until people started getting the new silver C8s they ordered. Wasn’t expecting the driver hardware to change, just the firmware. Hopefully the next version will fix the little issues we’ve found — grounded pin 5, components too close to the MCU, and some slightly skewed fuse and PWM values.

Actually, I’ll take back what I said. Just received a clear C8 to review from GB and it has biscotti firmware - yay!

Hey all. I’ve been away for several days. Tom, thank you for doing the extra investigating. I already reported the first finding (wrong fuse values) and now I see I have at least one more issue to report with the grounding of pin 5 on the MCU. The components being too close was reported a while back. These things should all be corrected “soon” but as busy as they are right now it probably won’t happen until after the holidays.

Btw, I found the close components did not interfere with the blue clip. The problem all along was the grounding of pin #5. At least this is what I found with the 5 boards, and two types of clips, blue and black, I have here. All 5 boards now have the pin #5 trace cut and all 5 can be make a connection with the clip, regardless of the closeness of the cap and diode.

I still see two problems:

  • on 1 of the 5 boards, it's consistently acting as next mode memory, even after re-flashing a known good firmware version of Biscotti. Replaces the C cap - no change. Bad part of this is it trips the 10 quick taps for config UI easy. I can only assume it's this one MCU -- the "noinit" storage lass of "fast_presses" doesn't seem to degrade to 0xff.
  • after programming the low byte of the fuse to 0x75, all 5 boards now turn on the LED at a very low setting when clicking thru to enter the config UI. I believe this is a direct result of using the slower clock, lower PWM, and keeping the PWM in FAST mode when setting the PWM output level to 0. 0 value still causes the LED to go on at a level lower than the moon level. This isn't a bad issue, just behaves different than expected.

Tom, one of the things I hoped to fix (improve) during this troubleshooting was a lower moon if possible. Maybe you or TK could chime in on that. Now that you’ve corrected the driver issues does is seem possible to safely lower the moon level? Perhaps the dim config mode could be tweaked up to a level more easily seen or is that a non-issue?

Dunno because getting a decent low moon is so dependent on the 7135's used, there could be variations batch to batch, and you also may have issues when the cell drops to lower voltage levels. We like to think firmware is separate from hardware, but many times, specially in the Chinese electronics world, they can't be so easily separated. Using one FET, or a two channel approach, where 1 channel is a single 7135, then you can achieve a more reliable lower moon mode.

I could test this lower moon mode level out further (FAST phase, 0 value for PWM), and see if it works ok on lower level cells down to 3.0V. If it tests ok, we could in theory use it as long as the 7135's don't change again. I think the LVP logic drops modes down, expecting our set moon mode to actually work in a low voltage condition, so it's kind of important for LVP.

I know a little about this only because TK explained it to me when she was writing Biscotti. The reason I mentioned that about the moon is that the older Convoy (green) drivers had 7135’s that were very slow to activate. This is why she had to get tricky with the moon mode and set it higher than she originally planned to. I’ve heard that the 7135’s used on these red drivers are significantly faster to activate so I thought maybe it could be tweaked. When she was writing the firmware she had only the older green drivers to test it on. ToyKeeper if you’re reading this perhaps you could clarify the details. I don’t fully understand this side of things. I get your point though Tom about possible variations from batch to batch so perhaps it’s better to play it safe and leave the firmware as-is? You or TK would know better than I. That said I can tell Simon to instruct his driver maker not to switch the component types, particularly the 7135 chips going forward and he’s usually pretty good about controlling that type of thing.

Lightbringer the green boards are indeed the older boards with the standard 3/5 mode setup. The good news (maybe) is that Biscotti works perfectly on the old green drivers provided you have a way to flash it or someone willing to do it for you.

Like those annoying non-star 105s? Hmmm. Didn’t notice the blink-on-low yet.

Once this cell finished charging, I’ll pop it in and try it, tnx!

(The only reason I ordered non-star versions of bare 105s is to intentionally flash them to what I wanted, or even just to hardwire a 2.1A or 2.8A 1-mode tac-light. Figured I wouldn’t need the stars to config anything.)

And it is a 3/5, kewl! At least I got rid of the blinkies, now will see about hardwiring it wide-open…

Mine has a red board. I’ve only programmed it once with no blinkies. What are the symptoms other than possible timing problems with blinky modes? It’s probably been answered 100 times already, but I haven’t read every post. I will say, the programming instructions weren’t very clear, but after several tries, I did it.

I can’t keep those straight in my head… hang on…

105C - with stars, 4 7135s bunched up at north/north/east/west

105D - without stars, 4 7135s equidistant at north/south/east/west

Ah, that’s right. You need the 105D so the retaining ring will fit, else it’ll bonk into the corners of probably all 4 of the 7135s on the spring-side.

Yay, another revelation… :smiley:

When you’re trying to program it the blinks are almost too fast to deal with. It’s doable, but tricky. When properly done there should be plenty of time to respond in the configuration mode.

Yeah, you have to be really quick to program it since the flashes are running double-time. It may take a couple tries because of that but as long as you aren’t looking for a mode with blinkies there isn’t much downside really. It’s just that because of what TK surmised and Tom confirmed is an incorrect fuse setting the clock speed is doubled. You won’t notice anything out of the ordinary until you try to use a blinky mode or config mode. It’s still a very nice driver even now if you don’t want disco modes.

So, what is the verdict? I’d like to buy one of these new clear anodized lights, but I don’t want to be disappointed and I don’t want to have to repair it. Are they all bad, or just a few? Should I wait weeks or months and probably lose interest? Are all of the lights with faulty drivers still sitting on the store shelves? Are they being recalled/repaired? If they are being sold with faulty drivers, why are they still asking full price?

Any advice?

If you dont want to reprogramm the driver it works fine.

Aside from everything working faster, and low modes being a bit brighter than expected, I have no issues with the new firmware on my C8 and three S2. The new firmware is a huge improvement over the previous C8/S2 user interface.