Some crazy CFv7 issues

[KamiKuzi]

Hey Guys,

I would like to have your opinion/ideas on some issues.

I've got my first saber up and running. (And yes I broke some of your newbie rules ^-^) Most of the functions are fine, but I have some crazy issues.
It's a CFv7 with CeX and one PeX running a RGBW Cree MC-E with a dual 14650 pack. Ch1 is red, Ch2(CeX) is green, Ch3(CeX) is blue and Ch4(PeX) is white.
It has a crystal chamber with RGB LED coming from the Ch1,2,3 control signals, two momentary switches, half pli and 4 other accent LEDs.
All that in a MHS hilt. In the bottom there are the battery pack, CF and speaker. In the middle i have the crystal chamber with the wires running through three copper pipes. In the top are the CeX, PeX, switches, recharge and R.I.C.E port and of course the Cree LED.

I had a serious crosstalk issue between the speaker and the CeX channels. The blue channel shimmered during boot and high pitched sounds. It was gone when the speaker is disconnected. Also I couldn't get a clean red because of the blue and green radomly shimmering sometimes (not reproduceable). I measured everything with an oscilloscope and found some needle impulses and noise on the control signals which triggered the CeX board. I fixed that using a pull-down and series resistor between the CF and CeX for the control signals. Could be because of my setup using very thin wires running through the crystal chamber...?

There might be a bug in the CF when using crystal pulse for the accent LEDs. The accent LEDs do a nice pulse like they're supposed to, but when you ignite the blade all main channels flash up for a short moment with corresponding intensitiy of the accent LEDs in that moment. Example: you push the on switch in the moment the crystal pulse is brightest theres a big flash on all main channels with full current. Push power on when accent LEDs are almost dark: slight flash. I could reproduce that for single main channels by setting the corresponding XTAL values in the config file. Low value - small flash, high value - big flash. Even for single channels, like xtal=0,0,1023,0 causes only Ch3 to flash. Anyone ever had something like that???

I still have a very annoying contact bounce on the Aux switch, making it very hard to browse the sound selection! I wanted to fix that with a 10k pull-up resistor and some nF pull-down capacitor. Question for the experts: Can I use the 3,3V where you can connect the overall power on LED as a pull up point?

Next issue is the PLI. I wanted to have a half PLI and half sequence in idle mode and full sequence in normal and lockup mode. The half PLI+sequence works after booting or selecting the sound bank. After you turn the blade on and off again there is a full PLI over all 8 LEDs or no PLI at all depending on the power down sequence. i.e. if the last step of power down sequence is 0000000 I get no PLI. With the default power down sequence I get full PLI. Looks like the idle.cfg isn't read correctly after the power down. Is that normal?

Another issue is that power on force doesn't work, no matter what I set it to. I always have automatic blade ignition. Didn't want to use it anyway...

Mute-on-the-boot only affects the menu volume, but not the sound volume. The FX volume is always as set in config. Menu volume does change with hilt orientation during boot.

That's all I've noticed for now. But I haven't testet all functions yet.
And let me add that I'm not new to electronics. It's my job, but more in the industrial/automation/switchgear sector.


Now let the newbie-bashing begin... :-)

[STARKILLER]

First, welcome to the forum and to the madness that is this hobby  :)) No one is going to bash you here, and it is obvious that you are no 'newb' when it comes to electronics ;). As to your issues, I would leave their diagnosis to Erv' as I've never encountered them before other than the blue channel flashing, but that was during EARLY testing back in 2011/12 (? :-\) I would hold off on using any work arounds or mods until we can hear from Erv'.

[erv]

channel flashing is mostly due to a wrong grounding of the color extender, which ground must return exactly to the main negative of the CF, and for instance, certainly not to the recharge port.

crystal pulse issue (glitch during ignition) : got it reported by genesis custom sabers, fixed in the most recent firmware.

aux switch : switches already have pull up internally in the board plus software to debounce. You cannot use the 3.3V pad for that as swiches run internally at 2.5V. That's the first time I hear about switch bouncing issues, you must have a very bouncing switch or ground / wiring issue.

PLI issue : I need to have a look to reproduce your situation. As I'm working on v7.5 that's the right moment to address those.

power on force : I'm fairly sure that works. Set the thresh to 0 and that disables automatic power on with a shake of the hilt. Make sure you don't use pof instead of pom, and vice versa

mute on the boot affects indeed the boot sound and menu volume only, that's meant to be that and it's how it's been requested (therefore coded). If the need is to mute the board, you then have mute on the go.

that being said, I'm (as always) opened to suggestions and discussion about features in general

[KamiKuzi]

OK. I must say you guys are really awesome. And really quick.

Let's start with the mute-on-boot. OK. Thanks. I misunderstood the manual.

crystal pulse issue: I got the CF about 2 or 3 weeks ago. So I guess it's the most recent.

The switch is a short 16mm AV white LED from custom saber shop. Maybe I just got a bad one... But I'll check my connections.

And I'll recheck my PoF settings. I'm not quite shure at the moment.

I bet I have a rather unusual ground wiring.
Later I can post an accurate wiring diagram of my setup. Have to get it up to date first...

[KamiKuzi]

Bloody grounding!!!  >:(

Erv' you're a genius!!!  :-*

I REALLY underestimated that. I checked my ground wiring and changed it. The CEx has now a single dedicated wire to the CF. All other parts in the top of the hilt use the old wire. I also removed my improvised resistors.
And while I was at it: added an additional ground wire only for the aux switch.

And guess what?

Blue and green flicker: gone!!!
Switch bounce: gone!!!

I added my circuit diagrams.
I hope they're not too confusing. As I said: I'm used to industrial big stuff, hence the kind of diagram...

V1, before
V2, after


I didn't have time to test the other stuff yet, but I'll get to it in a few days.

[erv]

excellent news !

regarding your other issues with PLI and such, I'll have a look this week. The glitch reported by GCS has been corrected but hasn't been added to the current firmware, but rather to the new 7.5 which is currently (still) beta. I'll double check with your situation that combines half PLI and xtal pulse in idle and eventually fix this.

Once 7.5 is out, you can get your board flashed

[KamiKuzi]

I disabled crystal pulse after I found the glitch. I didn't want to fry my LEDs or my battery Pack. I was on my PSU while testing and discovering it. I have my LEDs set to 700mA max. So if all 4 LEDs flash it's almost 3A. My PSU limiter kicked in all the time and that current would exeed my battery discharge limit anyway. So I disabled it since I'm on battery.

The PLI issue doesn't have anything to do with that! I have that either way. For now I disabled PLI in all banks exept one so I can check power level.
I'll do some checks later to be sure it wasn't my grounding issue...

[KamiKuzi]

OK. Now the results of my PLI tests:
I verified the PLI function by turning down my PSU everytime.
I had following setup. I copied the Bank 1 files and only modified the pli and halfd parms.
Bank 1:
idle: pli=1, halfd=1, sequence programmed
normal: pli=1, halfd=1, sequence programmed, powerup end:00101111 power down end:00000000
lockup: pli=1, halfd=1, sequence programmed
result: half PLI works, but no sequence after reboot. PLI ok but LED6 is on after ignition, during and after lockup. PLI OK and other LEDs are off after shutdown.

Bank 2:
idle: pli=1, halfd=1, sequence programmed, powerup end:00101111 power down end:00000000
normal: pli=1, halfd=1, sequence programmed
lockup: pli=0, halfd=0, sequence programmed
result: half Pli works, but no sequence after reboot. PLI ok but LED6 is on after ignition! Full sequence during lockup. After lockup PLI OK but LED5-8 are on or off depending on the moment the lockup sequence ended. After shutdown PLI works and other LEDs are off.

Bank 3:
idle: pli=1, halfd=1, sequence programmed, powerup end:00101111 power down end:00000000
normal: pli=0, halfd=0, sequence programmed
lockup: pli=0, halfd=0, sequence programmed
result: half Pli works, but no sequence after reboot. Sequence with no PLI after ignition and during/after lockup. It gets strange after shutdown: FULL PLI over all 8 LEDs.

Somehow half PLI + sequence doesn't work. And because of the sequencer not running, LEDs remain in their last state.
Only the last bit (full PLI) doesn't make any sense to me...
Also tested it with the original files only setting idle pli=1 and halfd=1. Same result.


And I have noticed something else. Different topic. Color Mixing. Maybe I didn't get the manual again...?
Case 1:
led=500,0,0,0   fled=500,0,0,500   lled=0,0,0,0   focmix=0001
This gets me a red blade. A red plus white FoC. Tilting the hilt down results in less and less red during FoC.
Case 2:
led=500,0,0,0   fled=500,0,0,500   lled=0,500,0,0   focmix=0011
This gets me a red blade. A red plus white FoC. Tilting the hilt down results in less and less red and more and more green during FoC.
Case 3:
led=500,0,0,0   fled=500,0,0,500   lled=0,500,0,0   focmix=1111
This gets me a red blade. A red plus white FoC. Tilting the hilt down results in red+white and more and more green. The alternating thing mentioned in the manual...

The first two cases are not in the manual, are they? I thought lled is only supposed to be used in all-mixed mode. If it's a feature, it's pretty cool. If it's a bug, leave it! It's pretty cool. You can control fade out or crossfade of main color.


Aaaand I also tested PoF: It works, but it 's veeeeery sensitive. Even setting pom=50 only requires a slight shake.

[erv]

useful test, that will help checking the PLI, half display and sequences, hopefully today

the lockup thing with color mixing is normal. I only described in the manual a FoC color change, but obviously, you can have a full cross fade between colors during lockup, would it be full mixed or "on top" FoC / lockup.

regarding POM : it's not specially "sensitive". Thresholds are on range 0-1023, so 50 is indeed very low.

[KamiKuzi]

Ah. OK. The manual says pof=[0-50]. And that quote: "Values from 5 to 15 are usually fine"

With a range 0-1023 50 is indeed not very much... and it explains my first experience with blade instantly igniting. I had it set to 10.

[erv]

will update the manual

[erv]

OK. Now the results of my PLI tests:
I verified the PLI function by turning down my PSU everytime.
I had following setup. I copied the Bank 1 files and only modified the pli and halfd parms.

I just wanted to address the bug issue above : along with the PRIZM features that have been added to CFv7.5, I've spent quite some times to fix a couple of things, like this one. Now sequence is properly played and it's possible to change between full display and half display for the PLI, on the fly. It's going to be tested thoroughly by beta testers, but so far so good.

[KamiKuzi]

Man, you are really awesome!
Your dedication is very impressive.

Now I'm actually thinking about getting an update even if I have to take half the saber apart to get the CF out again.

I wouldn't mind becoming a beta tester though...