I did modify the board today and used recommended "bypass" (option 2). Works much better now - even with my batteries. So I am happy now on that ... at least servos will not undervoltage and stop board any longer.

I REALLY APPRECIATE YOUR SUPPORT - so I guess, this topic can be closed.

ROOT CAUSE: BOARD VOLTAGE REGULATION

SOLUTION: BYPASS VOLTAGE REGULATION

But still > see new topic:

Frustration level: very high

As I read a similar post, in which you recommend same, that was something I was thinking to do, too. And as I planned to solder pins for ultrasonic anyway I need to dismantle the board. So I will do so and report. BTW, I would choose option 2 from your modification.

The blinking yellow LED indicates the voltage on servos is unstable and I've seen it too. Press the 100+ key (row 4, column 3) on IR remote to switching between 8/16 servos. Allow the servos to rotate freely. 8 is the number of servos involved in Nybble's walking. What's the speed you can get up to? It doesn't have to be the highest speed because the gears will wear out faster at higher speed.

There are two bottlenecks on the servo's performance, and are my biggest concern with the whole system.

1. Battery's output ability. Its discharging ability could be too low, or has protection set at low threshold.

2. The board's step down chip U2 also has a protection at 5A. The battery could be good enough to supply whatever the servos requests, but then U2 will step in, frequently cut/restart its power supply.

See the following video when I test on NyBoard V0_1. Towards the end of continuous discharging, the power supply on servo is going to be unstable. It's like a person hard to breathe after a sprint.

If you have a soldering iron, you can try to bypass U2. With more people reporting problems on the powering, I've decided to separate the usages with NyBoard more explicitly: one regulated/limited low voltage for regular plastic servos, and one direct/unlimited supply for the metal servos in Nybble kit. It should give more tolerance on battery, and reduce the uncertain variance of the system.

So close ... Damn; now I need to get different ones. AAAAaaargh.

So , - tests are done and result is quite frustrating.

Nevertheless; at the end most interesting was fact, that when I did run ModuleTest and activated all servos with IR "1", than all servos moved quite slowly but moved, once I switch to "2" that also worked, but sometimes either on "3" or "4" all servos stopped and yellow led started blinking. Sooooo ... my conclusion right now would be ... Shitty batteries.

OK, actual status:

* I made an external switch, so no wires or servos can harm battery switch any longer

* I did upload WriteInstinct and did start calibration process, but it does fail with yellow blinking led.

I assume my battery protection is the issue.

I am going to retest all servos via ModuleTest.

Hard to say. I did had impression, that FTDI became slightly warmer while I was working on calibration just before the board interrupted, started that noise. I am not really sure, if while calibration maybe a servor did touch that "wiggly" (not good closed) switch lever on battery housing and interrupted the mainpower supply, while FTDI was connected to work on serial monitor. Anyway - no need for a refund. For me it is a lot of learning (quite steep learning curve right now), which I appreciate. Keep on giving such a great support is all I would kindly request for.

So bad! Any notes on the FTDI's breaking moment? Heat/smoke/smell/cat's motion, etc.?

I can send you a refund for the FTDI. Email info@petoi.com.

Done. Transfer is not working. And yes, I did check and update related port in advance. All new ordered FTDI do transfer, the "broken" one not.

Before I proceed working on NyBoard at CatFrame, I will first modify the battery housing and remove this "switch lever". I do not trust that design, as lever is too long (even bended) and can get in touch to a wire on servo while motion. I will place an external switch to it.

Could you give the previous "broken" FTDI another try?

Hi Rongzhong ,

first good news; the FTDI did arrive. Second good news; those are working now and I am able to connect to my backup NyBoard (the 2. and brand new one) and I can upload sketch! All works fine here. I could also receives and see inputs in serial monitor.

I will now test to send a sketch on NyBoard, which is build in CatFrame with servos (that one, which is doing those strange beep noise in video). Cross fingers!

.

.

.

Yeah !! Some more very good news now .... it is working (at least based on first simple test)!

I did send TestInfraredRemote sketch to my "problem-board" and it:

A) stopped that strange beep-sequence

B) showed all infrared remote signals on serial port correct (to simple check input/output)

Seems to be OK as by now. Next step is to check WriteInstinct and servo calibration again.

I am still worried, what caused that FTDI to become broken that sudden (missing clear rootcause).

It doesn't have to be EBL. I have the Chinese factory run test programs on NyBoard. They bought their local brands. It's just a little tricky to find the right one. Any member on this forum know more about batteries and power management? I also hope to make NyBoard more tolerant with different power supply.

It's a good idea to send surveys after the major shipping cycle. I also want to include questionnaires on delivery time, custom fee, etc. to give future users better estimations on the whole process.

Yes - you´ve mentioned that in your describtion. Unfortunatelly here in germany I was not able to find those specific EBL batteries as mentioned by you (what a pitty). Mine are xCell with protection circuit. To bad.

Do you know any kind of practical work around, so I do not have to complete new set of cells?

Maybe an idea for this forum; I kind of "recommendation chart" could be helpful.

What do you think about a survey by us users:

* what kind of battery cell do you use (in which country)

* how capable and stable does your battery cell work with NyBoard

* does it fit to battery housing :)

Nybble consumes a large current when running and booting up. The momentary peak could be even larger. In experiments on my side, it works with the EBL batteries, and the average current is about 3.5A. But when I power it with a DC power supply that claims 6A output, it cannot even boot up. The power supply may mistake the peak current as a signal of short circuit and triggers its protection.

If your battery is longer as the one shown in my photos, it probably has a "too good" protection circuit on top. Mine longer battery doesn't work on Nybble, even though it's rated with a higher discharging capability.

Hi Rongzhong ,

again - appreciate your absolute great and quick support in this forum. That sounds quite logical. I´ll try to test a re-upload tomorrow as soon as I can and will respond to you again. I am still worried about unknown rootcause behind all that. Could it be, that my main power source was accidentely shortly interrupted due to no fully fixed "switch lever" (as batteries in housing did squeese housing and springs to their limit and "switch lever" could be only hardly pressed in between to close circuit), while calibration process. Still strange...

No, this beeping pattern doesn't come from my codes. It could be a result of the interrupted uploading causing the execution overflows to some random pieces of memory stack.

In my previous tests, the board at times generated abnormal long beeps when all servos move at once. I think it was because the voltage drop on chip interrupted the program. But the beep never formed a repeating pattern. And the board could reboot by pressing the reset button. Some bits of the stored program can even be altered during the voltage shock, but it can be fixed by re-uploading the sketch.

Your FTDI is not working currently so you may see nothing through serial monitor. Later, if you make the FTDI work again, upload any sketch to see whether the logic chip still works. Test it without plugging any servos. If you do plug servos, remember to power the board through the battery terminal beforehand. Otherwise at the moment when the servos are getting initialized, the board may pull too much current through the USB port. On my Mac, it occasionally disables the USB hub. I've done limited tests on Windows machines and never observed problems. But it could be a cause in your case.

--

I noticed that in your video the FTDI still blinks. See what appears in the serial monitor. If it's some random characters, re-uploading the sketch should fix the problem. If you cannot upload, press the reset button on NyBoard and click the upload button in Arduino IDE at the same time. Release the reset button after 2 seconds when the program finishes compiling and is about to upload the sketch.

5-short-1-long beep does repeat, correct.

Thanks for info ... I send a different link.

Hi Rongzhong Li ,

first of all I appreciate your very fast response! I wasn´t aware about different cell sizes possible. I assumed those to be standarized - but OK. As I do not like the cheap solution of that "switch" I am going to make a 3D print with an update later on anyway.

I did send a video, which describes current issue much better than all my words, I guess. I hope it will help for a diagnostic from distance. Many thanks.