openrcforums

Another strange situation that already happened twice is the radio entering the bootloader at startup withou pressing any trim and no usb cable connected.. I have an extension to the battery bay with a JST connector, and another on a USB cable. could that be the problem? The extension is quite long as it is just temporary. I would say 20 centimeters.. Could some noise lead to false detection?

João

Here is the video Mike..

João

https://www.youtube.com/watch?v=Zt0PvM4_izk

Mailed!

Mike, it happened again.. This morning I tried to make a video but the camera didn't film.. So I waited about half hour and tried again.. the radio started again in bootloader mode.

Here is the video.. Next I will try disconnecting the usb cable from the radio an see if this could be caused by noise??

João

https://www.youtube.com/watch?v=iAsYH3hafDk

I think this is a "diode leakage problem". I vaguely recall something like this. Diode D2 feeds the USB power to the radio supply. If the radio supply leaks back through this diode, it can cause a USB detection via the potential divider R50 (27K) and R51 (47K).
I just checked back through e-mails, we have seen that before. Brent changed R50 to 2K7 and R51 to 4K7 to fix this on later boards.

Mike.

Thanks Mike. That means very fine soldering again, right?? Those are the 2 resistors just to the right of the USB header, right?
Also does that mean that all the people using a rev B1 board will have this problem with the bootloader?

João

Before the last change to the bootloader, this problem would show up by powering on and getting USB (double size) showing in the middle of the display.

Mike.

hmm, ok that means some soldering on my future.. What size are the resistors? Is that 0603??

Thanks..

João

I'm fairly certain you can also fix the problem by removing D2. This removes the leakage path. It just means you have to power the Tx on to use the USB.

Mike.

well, I think I always power the tx on to use USB anyway?
But what would happen if I had to do a chip erase? Would it still work?
I was looking at removing those 2 resistors and maybe replace them with a bigger size by connecting somewhere else? I have those values in 1206 and normal through hole 1/4W resistors..
João

Yes it would still work with a chip erase, just need to power it on!

Mike.

Edit: R50 and R51 are a potential divider across the USB power and ground connections. These are available on J9 (PWR and GND).
You should be able to find a via that is the junction of R50 and R51, so it should be possible to wire normal 1/4 watt resistors in using this places.

Yes, the via seems to be just above R51.. Would just have to find the GND and PWR near by.
But Mike, if it works with a chip erase then we don't have any situation where we would need the USB with power off, or do we?
What would you do? resistors or remove D2?

Thanks
João

Probably the resistors. As I said, PWR and GND are on J9, just left of R50 and R51.

Mike.

Thanks.. Will do.. I was trying not to remove that board again.. Had a bit of a trouble to put it in with the encoder..

João

Changed the resistors.. Guess what? The first time I turned the radio on I was greeted with the bootloader..

Let's see if it happens again..

João

As I remember I removed the diode on my B board to reflect the C boards USB action. Too many programming rules to remember for all the radios I have.
I wanted to make all my sky boards the same.
Also I'll have to review the B boards power regulator circuit. I only use a 2 cell lipo in it due to a max voltage restriction at the time. Works fine and I never went back to worry about it. The 2 cell lipo was a sweet spot voltage for the regulator and I had no use for a higher voltage excess. I seem to remember it was a heat sink area being too small for the extra cell voltage to be dissipated. Brent used a switcher regulator on my C board and I use a 3 cell life04 battery for it.

What are you using for a battery voltage and regulator? Linear or switcher? Any added heat sink area?
I should update/refresh my B notes as a few were sold to local club members here.
SM

The board was upgraded to the TO220 regulator with a heatsink mounted on it.

Mike.

Yeah, 1A reg is fine. At the moment I am using a pack made of 6 eneloops..

Ok, replacing the resistors definitely did not fix it Diode next?

So you removed R50 and R51 and replaced them with a 2K7 (from PWR) and a 4K7 (to GND)? Just checking!
Can you check the voltage at the junction of the 2 resistors when power is on?

Mike.

MikeB wrote:The board was upgraded to the TO220 regulator with a heatsink mounted on it.

Mike.

I should probably do that so I can use a 3cell life ,lipo or 8 cell Nimh pack at some point.
I'll have to dig up a low headroom 220 size 5v regulator in case I use a 6 cell Nimh pack. I never know where it'll end up. Maybe a new owner sometime. KIIS.

Part required: AP1117T33L-U.

Mike.

ok, Checked resistors are correct..

tx on, No USB connected. 149mV
TX on and USB connected. 3.1V

João

MikeB wrote:Part required: AP1117T33L-U.

Mike.

I may have bought several for the DSM hack project or 9x reverse polarity repairs I made last year? 3.3v, somehow I thought I needed a 5v one.
It was all covered in the mods at the time so I guess I'll have to search way back and read the how to's. I just used the board as I got it with the recommendation to use a 2 cell lipo and I did. The ver C came to me shortly after that and instead of removing the B board, I did a new 9x build. It already had the switcher and upgraded regulators from Brent.
Just thinking out loud here as I read up on João's progress.
I'll read more, comment less.
SM

João: Those voltages look perfectly OK.
Does it start in the bootloader always or just sometimes?

Mike.

Just sometimes.. Mostly after I leave the radio off for a while..

João

Just left the TX alone for a while.. It happened again..

João

It may be detecting the two trims, rather than the USB then. I did have some problems detecting them reliably originally. There are some "debounce" capacitors across the trim switches. These have to charge up at power on to detect both trims actually OFF. It may be that when you switch off, with the capacitors charged, they slowly discharge. If you switch back on soon enough, they are not fully discharged, so charge back up in time. Leaving it a long time, maybe they then do fully discharge and then cause the incorrect detection.
I'll check the timing in the bootloader.

Mike.

Thanks Mike.. Shame that probably wasn't needed to replace the resistors, but hey, someone has to test so others don't have the same problem

João

The original problem was found by someone when the Tx got quite hot in the sun. This would increase the leakage current through the diode.
Just posted a new test version. I've extended the time waiting for the trim switches to settle in the bootloader.
I've also sorted the baudrate setting for the HC-05 BT module, added a possible 'zero' setting for the throttle (instead of -100%) and added a few more items to the log files.

Mike.