RA+ locking up Blank White Screen

[00Warpig00]

My RA+ has been locking up allot lately with a blank white screen on the controller.

It has been locking up enough (about twice a day) that I have started to play with things to figure out what is causing it.

Most of the time if I unplug my IO Module from the expansion hub the screen will come back up.

Occasionally when I unplug the expansion hub from the relay box the screen will come back up. Even after I unplugged the IO module and that didn't fix it.

I have systematically tried unplugging all my modules one at a time from the expansion hub. Seems so far the only one that gives any results (the controller unlocking and showing the screen again) is the IO module.

occasionally I can unplug the VGA cable from the controller and re plug it and the screen will unlock (but not normally)

on rare occasion while unplugging and replugging modules nothing will unlock it and it needs to have the power cable removed from the relay box.

Nick

[rimai]

Have you extended the USB cables?
Can you try other ports in the hub?

[00Warpig00]

I have not extended the USB cables. I decided to break out and extend the VGA cable instead and keep my relay expansion module close to the hub and just put the relay box in the other room. I have not done this as of yet though and do not have my relay expansion box plugged in right now even though I have code loaded for the relay expansion ports. Could this be part of the problem? strange thing is I have had my relay expansion module unplugged for several weeks and the lock ups started conveniently when I was out of town the week before last. I did make a code change the week prior to that when I put in the code to display the ATO timeout on the Portal and to turn off my skimmer if the timeout flag had been raised. I will change around a few of the connections on the hub and see what may happen.

Nick

[rimai]

Try moving the usb cables to the top row of connectors on the hub and see if it makes a difference.
I'm wondering if there is any loose connection.

[00Warpig00]

I had the IO mod plugged into the top port to the left end (if you were reading the sticker) locked up again today while at work. I unplugged it from the I/O mod end (small connector) and the screen came right back up.
Funny you mention loose connections, the small end that plugs into the IO mod is almost loose enough to fall out on it's own. I will see if I have another cable that is tighter. I have also moved the big end to the hub on the top row on the top side (while reading the sticker) on the port on the left.

I hate to tell you this, but frankly these are the crappiest cheapest USB cables I have EVER seen. I have already had two of my cables have a wire pop off on the large end that I had to break open and resolder them to get the cable to work at all. You might want to consider better quality cables for people you are shipping new controllers and modules to it might save you and all your customers some headaches. Particularly if loose connections could be causing lockups and non-reliability. After all reliability is the MOST IMPORTANT quality of a controller.

I am getting very frustrated with this whole project so far as I cannot seem to get this controller to be reliable and that is the #1 most important quality of a controller.

Not sure what you are paying for USB cables but if you are not familiar with monoprice.com they are an awesome supplier for high quality cables of all types amongst other audio/video related electronics at very reasonable prices.

http://www.monoprice.com/products/produ ... 1&format=2
http://www.monoprice.com/products/produ ... 1&format=2


Nick

[rimai]

Hi Nick,

I understand your frustration and have noted your comments.
I don't know for sure if loose connections are the problem. I'm just trying to work with you to find out why you are having problems.
I'm in touch with someone that is involved in the core Arduino library coding and he is going to try to help us with this issue.
I believe the problem you are having is related to the signal between the controller and your modules and that's what he is involved with. Would it be possible to remove the I/O module to see if there is any difference in behavior?

[00Warpig00]

Sure I can unplug the IO Mod for a couple days to see if this changes things. I'm also having trouble with two of my three completely submerged (active when sump/rodi topoff/sw topoff is empty) float switches that are connected to it anyway. Two of three of the submerged switches are active even though the float should be fully floating and non active (wires out the top of the switches). I have put these two switches to a meter and they are in fact active (and shouldn't be) as they are fully submerged. If you remember I had you send me a replacement float switch because it seemed as if the one submerged in my sump was defective. I got the new one, installed it and it was fine for about a day and now that switch shows active all the time too even though it is fully submerged. ugg...

I had a pump next to the submerged float switches in my RODI/SW topoff Brutes that was causing the switches to be active but I have since moved the pumps at least a foot away from the float switches. IDK what is going on with these switches. Magnetic field, voodoo, bad karma etc. I have been unable to find an explanation for why most of my submerged switches always show active.

BTW I also uploaded that code about a week ago you suggested to my IO module to enable the watchdog feature, if it matters.

Nick

[rimai]

How is it doing without the I/O?
Here is what I got from Todd:

Well, I’ve been digging through the TWI driver, and I’ve already found three bugs. So it’s a good exercise if nothing else.
First off, there are at least two timing holes in Wire.begin, because it doesn’t initialize its variables correctly. How often does the master poll the IOExpander?
Put the “Wire.onRequest” line before the “Wire.begin” line. That will work around one of them. You’ll also want to move the pinMode lines before any of the Wire calls, since requestEvent could be called immediately after “.begin”.

OK, then I would suggest the changes I stated below.
Next, in ReefAngel.cpp, add in ReefAngelClass::Init
Wire.onReceive(NULL);
Wire.onRequest(NULL);
Before the Wire.begin();
Also, I think you said that you had done this already, but I think moving the WDT enable to the top of ReefAngelClass is a good idea.

Would you like to go through this changes?
I can send you updated files for you to load.

[00Warpig00]

No lockups today.

I would be open to try some changes so long as if there is strangeness with them I can revert to my current config without too much trouble.

Nick

[rimai]

Let's wait another couple days without the module.

[00Warpig00]

no lockups today either

[eds]

Interestingly enough mine did this today.

Wifi and Vortech Module with USB expansion. No reason for it, it just locked up on a white screen.
I only noticed because my phone started screaming it couldnt contact the controller.

I don't have an IO module so not sure if there is something I can try. My code at the moment is
generated by the Wizard.

[rimai]

Does it come back if you reboot?

[eds]

Yeah I just pulled the cable out (the vga) and put it back in and it is running as per usual again.

[00Warpig00]

How is it doing without the I/O?
Here is what I got from Todd:

Well, I’ve been digging through the TWI driver, and I’ve already found three bugs. So it’s a good exercise if nothing else.
First off, there are at least two timing holes in Wire.begin, because it doesn’t initialize its variables correctly. How often does the master poll the IOExpander?
Put the “Wire.onRequest” line before the “Wire.begin” line. That will work around one of them. You’ll also want to move the pinMode lines before any of the Wire calls, since requestEvent could be called immediately after “.begin”.

OK, then I would suggest the changes I stated below.
Next, in ReefAngel.cpp, add in ReefAngelClass::Init
Wire.onReceive(NULL);
Wire.onRequest(NULL);
Before the Wire.begin();
Also, I think you said that you had done this already, but I think moving the WDT enable to the top of ReefAngelClass is a good idea.

Would you like to go through this changes?
I can send you updated files for you to load.

Roberto,

Now that my Wifi Mod seems to be behaving a bit better (10+ days without it needing to be reset) I would like to make some progress on my IO module issues. Earlier in the thread you asked if I would like to make some changes and you offered to send me some files to try. I would really love to get my 6 float switches and my IO module working so I can progress on my auto water change ideas as well as high\low water alarms for my SW\RODI brutes as well as my sump.

NOTE: I plugged my IO Mod back in last night and within two hours my controller was locked up on a blank white screen. I think that is pretty conclusive that the problem is with the IO mod somewhere. The past month I had it disconnected the the controller didnt lock up with a white screen even one time.


Nick

[rimai]

Ok.
Overwrite the ReefAngel.cpp file with the attached one.
The file is located at Documents\Arduino\libraries\ReefAngel

Then, upload the IOExpansionCommands.ino code to the I/O module.

These were the suggestions from Tom.
Let me know how it goes.

[00Warpig00]

uploaded ino to IO Mod.

overwrote ReefAngel.cpp in the Arduino Libraries folder.

Recompiled and uploaded my current running config sketch to my controller.

hooked IO module up to expansion hub hooked T1 T2 T3 both ATO float switches PH probe and WiFi Mod back up to controller. Plugged in VGA cable from Relay box to power up and reboot controller.

I guess we will see how it goes.

BTW are new libraries out (1.0.2) in the past week or so? When I opened Arduino it said new libs were available. I did not update. I figured this was because I replaced the ReefAngel.cpp file. I cannot remember the Libs I was on. It was the most current as of a few weeks ago. whatever libs those were. I looked at the lib update thread that is dated July that has info from 1.0.2 in it but it seems as if I have read that before. Or is the 1.0.2 new in the past week or two?

My android app now reports I am using lib version 1.0.1

Nevermind. A little poking around and it appears as if I did miss a lib update in the past couple of days.

I re read the changelog thread and when reading it closer I dont think I have actually read it before today.
When was the lib update? it might be helpful when you update the changelog thread if you put the date next to new entries since the date of the thread is from back in July and does not seem to update when you make a change.


Nick

[00Warpig00]

Roberto,

Still trying to slay my IO Module dragons...

I upgraded my libs to 1.0.3 and recompiled and re uploaded to my RA+ about a week ago. (I didn't want to get too far behind in the libs). If you recall earlier in this thread you supplied me with a custom 1.0.1 reefangel.cpp lib file that was an attempt to address some IO mod issues. I assume when I updated my libs\recompiled\uploaded to my RA+ I wiped out whatever changes you had made to this custom file. Were these changes included in lib updates 1.0.2 or 1.0.3? If not I may need to put them back...

I believe my IO module is still acting up. I will know for sure in the next couple days. My controller was stuck in a lockup loop yesterday for more than 10 hours. Attached is a youtube link to a video of the controller looping. It looped like this for over 10 hours straight until I pulled the USB cable out of the IO Module. I ordered 10 new 1 ft long USB cables (with ferrites) from monoprice and replaced ALL of my USB cables that shipped with my RA+ tonight. Most of them seem to fit my RA modules a little tighter and the cables seem to be scores better in the quality dept than the shipped RA+ USB cables. If you want to take a look at my video the controller loops until I pull the cable. You cant see me pull the cable from the IO mod but the instant I do the controller starts right up normally.

[youtube]http://www.youtube.com/watch?v=blUBstcC ... P363SBnLCw[/youtube]

hmm youtube links appear to not work using the youtube link button in the editor.

http://www.youtube.com/watch?v=blUBstcC ... P363SBnLCw

[rimai]

Ok, now that you have it updated, let's try updating the module too.
Can you load this into the module?

Code: Select all

#include <Wire.h>
#include <avr/wdt.h>

byte IOports[] ={
  3,5,6,9,10,11};

byte IOOut=0;
void setup()
{
  wdt_enable(WDTO_1S);
  Wire.onReceive(NULL);
  Wire.onRequest(requestEvent);
  Wire.begin(9);
  for (int a=0;a<6;a++)
  {
    pinMode(IOports[a],INPUT);
    digitalWrite(IOports[a],HIGH); //pull up resistor
  }
}

void loop()
{
}

void requestEvent() {
  wdt_reset();
  IOOut=0;
  for (int a=0;a<6;a++)
  {
    IOOut+=digitalRead(IOports[a])<<a; //pull up resistor
  }
  Wire.write(IOOut);
}

[00Warpig00]

Thanks,

I will load it up over the weekend.

Will I need to replace the reefangel.cpp file again or where the changes to that included in the lib updates?

Nick

[javisaman]

I'm experiencing the exact same problem. White screen when usb hub is connected. Was there a library update that fixes this?

Thanks

[rimai]

What libraries version are you running?
Make sure to follow this update thread:
http://forum.reefangel.com/viewtopic.php?f=7&t=3068
Or the best is to just uninstall the RA Installer software and install a fresh one downloaded from the website.

[javisaman]

Using version 1.0.1. I'll try updating the libraries.

[javisaman]

Switched to 1.0.7, seems to have fixed the issue, though unplugging the usb hub and reconnecting might have done it too. I'll keep an eye on it, but is it known what the cause is?

[rimai]

Usually noise coming from somewhere.