ERSKY9X Coding

[ozphoenix]

Mike,
Correction to Item 4, above - I do get a switch warning screen if the 'F' switch is in the 'enabled' (forward) position, but only if the 'arrow' for 'F' in the rectangular box is shown as default forward, which is wrong. And, the detection of the switch's position is still reversed (but only the 'F' switch).
regards,
ozphoenix

[MikeB]

1. Switch H is spring loaded, so doesn't need a default position or a warning.
2. It looks like switch F is wired swapped on the production units compared to my prototype, I'll fix the code.
3. Checking.
4. Please try setting "Software volume control" ON in the hardware menu. I'll look into the hardware volume control operation.
5. Something else I need to investigate!

I've not seen "snow lines"!

Mike

[ozphoenix]

Mike,
I created a .bmp (attached as smalltig.zip, but actually it is smalltig.bmp) that I think meets the specs (not sure if it complies with the 24 bit colour requirement). It is a picture of a Tiger Moth, 'centred' in the 128x64 pix box. On the display, the image is 'wrapped' off to the right by about 1/3 of the picture width (which then re-appears on the left of the picture frame) -- the end of the rudder is touching the right side of the screen, the rest of the background 'green' scenery appears on the left hand border of the image frame.
regards,
ozphoenix

[ozphoenix]

Mike,
1. I'd agree, but for completeness and clarity, including it might be worthwhile(??) and simple(?)
2. ok
3. ok
4. Software Vol. is already ticked in the Hardware menu next idea?
5. yes, please - I re-confirmed it with a re-flash to Ersky9x - still hissing (softly, but definitely there and noticeable on my radio).

The 'snow lines' have only happened once, earlier today. I'll keep an eye out for any further occurrences and try to isolate the scenario.

regards,
ozphoenix

1. Switch H is spring loaded, so doesn't need a default position or a warning.
2. It looks like switch F is wired swapped on the production units compared to my prototype, I'll fix the code.
3. Checking.
4. Please try setting "Software volume control" ON in the hardware menu. I'll look into the hardware volume control operation.
5. Something else I need to investigate!

I've not seen "snow lines"!

Mike

[MikeB]

The image file showed it was of type "BMS" internally. I loaded it into "Paint 3D" and re-saved it as type "BMP", now it displays OK.
Here is the modified file (in a real .zip file):

smalltig1.zip

(12.71 KiB) Downloaded 168 times

Still looking at the audio hiss. I tried shutting down the audio amp, that stops the hiss, but takes too long to come out of shutdown.

Mike

[ozphoenix]

Thanks, Mike - everything put away for the night now - will look at these again early tomorrow.

Regards,
ozphoenix

The image file showed it was of type "BMS" internally. I loaded it into "Paint 3D" and re-saved it as type "BMP", now it displays OK.
Here is the modified file (in a real .zip file):

smalltig1.zip

Still looking at the audio hiss. I tried shutting down the audio amp, that stops the hiss, but takes too long to come out of shutdown.

Mike

[MikeB]

I've got a bit of a problem testing the audio. By default I set to software volume control, leaving the hardware volume at full. This may be the cause of the "hiss". I think I need to reduce the hardware volume, however, my prototype board doesn't have the required chip fitted, so I can't change it, or even check the code is working correctly.
I will add something, but I will need to ask you to test it please.

Mike

[jhsa]

So, this means the radio might have a noisy amplifier chip??

João

[ozphoenix]

Thanks, Mike - agreed, that image now works. Strange, because I believe that I used Paint3D to create the original .bmp file, but maybe I didn't. Anyway, thanks again - now, I'm on to other things.
regards,
ozphoenix

The image file showed it was of type "BMS" internally. I loaded it into "Paint 3D" and re-saved it as type "BMP", now it displays OK.
Here is the modified file (in a real .zip file):

smalltig1.zip

Still looking at the audio hiss. I tried shutting down the audio amp, that stops the hiss, but takes too long to come out of shutdown.

Mike

[ozphoenix]

Hmm, interesting - I wonder what they also do differently in OpenTX to prevent the 'hiss'(?!?!)

Yes, of course, happy to test whatever you offer My hearing is gradually getting worse, from years of sticky my head in the back of large-scale computer racks, but this hiss I can hear (or not, if you can get it removed)!

Let me know when there is something to test.

regards,
ozphoenix

I've got a bit of a problem testing the audio. By default I set to software volume control, leaving the hardware volume at full. This may be the cause of the "hiss". I think I need to reduce the hardware volume, however, my prototype board doesn't have the required chip fitted, so I can't change it, or even check the code is working correctly.
I will add something, but I will need to ask you to test it please.

Mike

[MikeB]

If they use the hardware volume, and you have the volume low, that might be why. Also, there is a MP3 device on the board they may be using for everything, but I'm bypassing this and using the processor DAC output, as on other radios (less coding to do to get it working).
I've posted a specific test version for the X12. On the Audio menu, there is an extra option to set the hardware volume, to 0 to 127. To the right of the value you may set I'm displaying the value read back as a check it is being set. Clearly, if you set the hardware volume lower you will need a higher, normal volume setting, but I've always thought the volume was very high.
So, the test is to set the hardware volume low (10-20) and:
1. See if the "hiss" reduces.
2. You can still get adequate volume normally.
but you might set the hardware volume to values throughout the whole range.
Note that while the setting is maintained through a power cycle, I'm only setting the volume chip when you change the value in te menu at present. I suggest you set a voice alert to play a file to test the volume output without powering off.

I've noticed my display backlight seems to be slow turning on, something I don't think was happening!

Mike

[ozphoenix]

Hi Mike,

Firstly, some general feedback from yesterday (not just related to the 'audio' test version) - the model image file display is not consistent across a power-cycle - sometimes the selected image will re-appear correctly in the image box after a power-cycle, sometimes not (says 'No Image' instead). If I go to the General menu the correct file is still named -- if I reselect the file, it re-appears in the box, but is then usually lost on a subsequent power-cycle, even though the name of the file still appears in the General menu.

Now, on to this 'audio' test version:
1. It is difficult to start the radio now - holding the button until it says STARTING does not actually start - goes back to power-off condition. Holding it longer changes STARTING to a STOPPING message, but then the radio continues on to full operation. This has only started to happen on this 'audio' test version.
2. The 'F' switch is now correctly sensed and the test on power-up detects it as correctly/incorrectly set, as required.
3. Choosing 'S1' (for example - I can use S2 or SL or SR, etc) as the volume control seems to work correctly now - I can adjust the volume of a voice file triggered by switch 'H'.
3. The 'hiss' is still there, no matter what value I set Hardware Vol. to, but equally perplexing, the value on the right, on the first power-up after flashing, shows 255 and never changes, no matter what I set the Hardware Vol. level at in subsequent tests. The value I set in the Hardware Vol remains (as set in the menu) across a power-cycle.

I have to go out for a hour or so - if you've had time to look at anything more or want other tests done and you post while I'm out, I'll give you a fast response as soon as I return.

Thanks and regards,
ozphoenix

If they use the hardware volume, and you have the volume low, that might be why. Also, there is a MP3 device on the board they may be using for everything, but I'm bypassing this and using the processor DAC output, as on other radios (less coding to do to get it working).
I've posted a specific test version for the X12. On the Audio menu, there is an extra option to set the hardware volume, to 0 to 127. To the right of the value you may set I'm displaying the value read back as a check it is being set. Clearly, if you set the hardware volume lower you will need a higher, normal volume setting, but I've always thought the volume was very high.
So, the test is to set the hardware volume low (10-20) and:
1. See if the "hiss" reduces.
2. You can still get adequate volume normally.
but you might set the hardware volume to values throughout the whole range.
Note that while the setting is maintained through a power cycle, I'm only setting the volume chip when you change the value in te menu at present. I suggest you set a voice alert to play a file to test the volume output without powering off.

I've noticed my display backlight seems to be slow turning on, something I don't think was happening!

Mike

[MikeB]

I was trying to see why my backlight wasn't coming on quickly, so I may have something slightly wrong regarding the STARTING state.

If the value at the right is 255, then the hardware volume is not working.
Late here so I need to leave it until tomorrow.

Mike

[ozphoenix]

No prob - back online now but figured it was getting late there - will look forward to any progress in your next session - for now, I can work with it as it is (software selected volume knob is working), get on with other things and see where it goes from here.

As for the starting up issue, if I persist with it, I get to time it just right and it comes on every time, it's just a little finnicky

regards,
ozphoenix

I was trying to see why my backlight wasn't coming on quickly, so I may have something slightly wrong regarding the STARTING state.

If the value at the right is 255, then the hardware volume is not working.
Late here so I need to leave it until tomorrow.

Mike

[ozphoenix]

Hi Mike,
On my X12S radio, r222d4 - any model, go to Model Setup, General, go to Thr. Default, press PgUp (Menu) 3 times, look at the Model Image box on the display - flickering like crazy. Second time you do that sequence, image box disappears completely. Also, after each third PgUp, there is a flickering number (usually a 1 or a 2) near the arrow-box after Custom Check.

Only way to get it back to stable (after either first time or second time or more times) is power-cycle or select another model.

And, one time that I did this 'three-PgUp press'-cycle enough times (maybe 10 in a row), the pesky 'snow lines' re-appeared across the screen!

Left it with the lines showing for about 30 seconds, the screen automatically did a reset (but not a full power-cycle of the radio) and the screen snow lines cleared up and a normal Model Main Screen returned and the normal (but words 'No Image', not a .bmp display) image box is returned.
Also, once this evening, while scrolling (with thumbwheel encoder) quickly down the list of model name audio files to fill in the model announce file on the Model General menu, the screen reset to black, the hiss disappeared and then returned when the screen re-painted to the Model Main screen. The radio did not power-cycle, just the screen and audio.

Regards,
ozphoenix

[MikeB]

OK, I've got the fix for the model image flickering.
I've spent most of the day tracking down the flickering number that appears after the throttle check screen appears. Finally found it is to do with the "Idle time" overflowing when the throttle check screen is displayed, causing something to try to be displayed off the left of the screen.
I've tried a couple of things to try to remove the "hiss", but so far it hasn't reduced it.

Mike

[ozphoenix]

Ok, thanks, Mike - must be getting late there - if you post an interim version, I can continue to make some observations during the day(night for you).

But, I think it might be too late already for that. If it's already getting too late, I can continue with the current test version and just work around the known issues for now.

Today would normally be a flying day for me, but the wind has been playing havoc with us recently. Hopefully, I can get some air time in on Sunday, but I'll be using my QX7 for that - still have a lot of setup and testing to do with this X12S, I'm afraid, and not enough time in the day for everything

Thanks for persevering.

regards,
ozphoenix

OK, I've got the fix for the model image flickering.
I've spent most of the day tracking down the flickering number that appears after the throttle check screen appears. Finally found it is to do with the "Idle time" overflowing when the throttle check screen is displayed, causing something to try to be displayed off the left of the screen.
I've tried a couple of things to try to remove the "hiss", but so far it hasn't reduced it.

Mike

[ozphoenix]

Hi Mike,
In the latest test version of eepskye, I cannot explicitly select the X12S in the 'Version to Download' dropdown. Which version of eepskye supports X12S? Is there an alternative selection to use in that list?

EDIT: Think I just saw the answer in one of your recent posts - still to get a version of eepskye to handle the X12S SD data instead of EEPOM so, no(?)

[ozphoenix]

Mike,
This happened again several times today - I think it was exclusively when I was using the scroll wheel to run down a list of files (for example, such things as model selection choices on Model Select page) but the screen I was on and the field I was selecting for varied (several different ones):

Also, once this evening, while scrolling (with thumbwheel encoder) quickly down the list of model name audio files to fill in the model announce file on the Model General menu, the screen reset to black, the hiss disappeared and then returned when the screen re-painted to the Model Main screen. The radio did not power-cycle, just the screen and audio.


On restart, however, the screen returned to the Main Screen (the flying screen).

Something going on with scrolling quickly through lists - maybe you can try to scroll through some lists a few times and see if your video resets to black and reloads?

Regards,
ozphenix

[MikeB]

Sounds like a watchdog reset. You may check this by going UP LONG from the main screen, then LEFT to the "Boot Reason" (12/12) screen. Also, please go to the "SD CARD STAT" screen (10/12, LEFT twice) and look at the number at the right side of the 2nd line. This is the "cluster size" on the SD card, ideally you want 32 there.

I'll see if I can extend the watchdog timeout when looking at file lists. When scrolling these lists there can be many SD card accesses needed as the displayed list only has 16 names in memory at any one time.

Mike

[ozphoenix]

Ok, thanks for those tips. All packed away for today, already - will look first thing in the morning.
Regards,
Ozphoenix

[MikeB]

I've posted an update with a couple of fixes/changes in.

Mike

[ozphoenix]

Got it - thanks - will work with it this morning and report back.
regards,
ozphoenix

I've posted an update with a couple of fixes/changes in.

Mike

[ozphoenix]

Just a quick note:
1. I still get a video restart if I am scrolling a list with the thumbwheel.
2. The model image does not persist, even though the image file name is still in the correct position on the menu page - it disappears when I move between menu pages ('No Image' words are re-inserted) and does not re-appear until I re-select the (pre-existing) file name field and exit back to the main screen. EDIT: Maybe the image doesn't need to persist when on setup screens, but it does need to be present when you return to operational/flying screens.

regards,
ozphoenix

I've posted an update with a couple of fixes/changes in.

Mike

[MikeB]

What is your "cluster size" on the SD card?
I'll investigate the image problem, mine stays present all the time. It is supposed to be read in memory and put on the display image in memory once, then only changed when you change models. The image for the display is put back, if necessary, from the copy in memory. To get "No Image" means an attempt took place to read it from the SD card, which failed.

Some SD cards seem to work better than others!

Mike

Edit: I just added some debug, counting how many times the image is read from the SD card. It stayed on 1, even when I moved round quite a few menus, model, radio and statistics. When, exactly, does you image disappear?

[ozphoenix]

Cluster size is 32.
SD card is new this week, 32GB SANDISK ULTRA HC1 Class 10 high-speed.

What is your "cluster size" on the SD card?
I'll investigate the image problem, mine stays present all the time. It is supposed to be read in memory and put on the display image in memory once, then only changed when you change models. The image for the display is put back, if necessary, from the copy in memory. To get "No Image" means an attempt took place to read it from the SD card, which failed.

Some SD cards seem to work better than others!

Mike

[MikeB]

Just edited my post above!

Mike

[ozphoenix]

Will try a few 'moving around things now, trying to be observant, and get back to you in the next 30 minutes or less.

Just edited my post above!

Mike

[ozphoenix]

P.S. Still tricky to get started:(

Will try a few 'moving around things now, trying to be observant, and get back to you in the next 30 minutes or less.

Just edited my post above!

Mike

[ozphoenix]

I was scrolling through a list of model name .wav files and my screen got the 'snow lines' effect - see attached photo. A restart of the radio is required to remove it.

Still trying to replicate 'lost image' effect and/or video restart.

Will try a few 'moving around things now, trying to be observant, and get back to you in the next 30 minutes or less.

Just edited my post above!

Mike