Just past 31 seconds is indicative of a power supply issue as well as the symptoms. That's ironic as you clearly took some precautions. Here are some thoughts:

1. Thick wires - good as this lowers inductance. If there's any appreciable wire length, twist them up. The twisted pair won't benefit you with this issue but will make your system less susceptible to nearby emitters as well as be less of a transmitter. This is a good general practice.

2. 100uF bypass near power pins - good to have a nice size bulk cap nearby. You may need smaller caps as well for their higher frequency discharge capabilities. I doubt that is the issue.

3. The lab power supply is where I would point my suspicion. What is the voltage (presumably 3.3 as you're applying directly to device)? What is the peak current capability of the power supply? Does the supply have a current limiting feature and, if so, is it on?

The O2+ has a large peak in required current as the WiFi system comes up, right around 31 seconds. Likely something in your general power supply circuit is limiting charge to the device as peak charge is required.

Best of luck!
Bill

@tjoseph1 Good work! I'd like to re-emphasize for all newcomers that "power supply", as opposed to "THE power supply", is critical. Of course THE power supply must be able to satisfy peak current demand. The overall power supply also includes any cabling which may provide enough inductance that a power supply barely over the minimum required capability may be inadequate at the power pins of the Omega. It is only the available current at the Omega power pins that matter, not what an upstream subsystem can provide.

--Bill

@arianmares I've taken delivery of many units from them and have been very pleased with service. @Lazar-Demin is fairly active on this forum. He can direct you to the most effective path for the service you need.

--Bill

@crispyoz I see it's 15 minutes later and I'm still laughing! However one part of my brain asked if this would be $100k USD or $100k AUD.....and that also cracked me up even more!

Anyone else want in? As you can tell by our stories, magic smoke escape is a serious problem. If you're experienced at blowing stuff up you might be just the kind of person we need on the team!

--Bill

Ultimate magic smoke (UMS) story: The summer prior to my senior year in college I was in a salaried position as Student Engineer at Inland Steel in the No.1 Electric Arc Furnace and Billet Caster (a continuous caster). The prior summer, the 1F cap story, worked in the No. 3 Basic Oxygen Furnace and Slab Caster, the number 1 steel making shop in the world, not just then but also now.

Do a web search on continuous casters and find some YouTube videos if you can. They are fascinating....and potentially dangerous as you're about to find out when a 21 year old has control system schematics and a pencil.

The continuous caster control system was in a room perhaps 15m by 10m with most walls covered by enclosed racks of equipment including a bajillion relays. The schematics are a mess, outdated and staff members were concerned about using them and erroneously "fixing" something resulting in a disaster. My job was to red-line the schematics to bring them up to date. So, I have schematics, a pencil, and a multimeter and I'm in the caster control room. There's a door at both ends of the room with the caster, in operation, only about 2 to 3m beyond one of them. I get to work.....with the caster in operation. What could possibly go wrong here?

I'm frustrated, but need to make progress. I think one circuit could be validated if I close this one relay and I'd see just this one other relay flop in an adjacent panel (can anyone guess where this is going?). So I have both panel doors open and find both relays. Using my pencil eraser I close the target relay while keeping an eye on the other.. Then it seemed like there were several hundred relay closures in the room all around me....and it kept on happening!!!! Then a siren alarm sounded....then I looked at the door at the caster end where a red light was flashing.....then I saw molten steel bouncing off of everything.....then I saw two guy run by the door apparently trying to save their lives....then I saw more molten steel fly around all over the place. Oh...my....I just caused a breakout....a phenomenon created when the caster is sped up such that the steel can't form a cooled shell before leaving the mold, so the internal molten contents started spilling onto the floor. Damn! My 21 year old frustration almost got people killed!!!

So, I didn't let the magic smoke out, but I made a very, very big error! There was some water around and some oil & grease, so there was some smoke and a whole bunch of steam. I'll never forget that and it was in 1980!!

Here's the punch line, and it is crazy. After a few weeks I admitted to my boss that I thought I caused the breakout and explained to him in graphic detail. His response to me? "You know what, Bill? We consider you a virgin until you destroy at least $50,000 worth of equipment." Feel free to jump on-line and find one of those inflation calculation tools to see what 1980's $50,000 USD is worth today. You will be stunned!

There's my best magic smoke story, and I was still a student at the time.

take care all!
Bill

After vacation I went to Chicago's O'Hare Airport, boarded a DC 10, and said a sort of goodbye to my mom, dad, two sisters, and countless other relatives and friends. I landed in Los Angeles and several days later was working as a systems engineer at Hughes Aircraft in Fullerton, CA, USA.

Here I am, a 22 year old that wants to conquer the world, and wants to do it NOW! My first boss, Jim Brennan, probably about 62, was a great guy! I learn that he had worked for Collins Radio in Santa Ana, CA and was the project engineer on the UHF radio system used to communicate from Earth to the Command Module and Lunar Module. I understand that as my dad and I are laying on our stomachs in front of the TV listening to Neil Armstrong say "This is one small step for a man..." that it was my boss that was responsible for us hearing and seeing that. I was 10 years old that day and decided i wanted to be en engineer. To me, Jim was a god, and he was my first boss.

Why did I tell you all of this, @CAP-33? In almost 40 years of engineering, I've let the magic smoke out a few times. Nahhhh......let's be honest.....a few dozen times! Or is it a couple hundred times?

Jim taught me something spectacularly important. He taught me I couldn't solve every problem simultaneously. He taught me professional engineering is methodical...it is deliberate. He taught me you can only go so fast. If you try to go faster, you actually slow down (e.g. all of the recoveries from letting the magic smoke escape). He taught me, sure we hack at things, but to do things "right" we might need to order parts and wait for them to arrive. He taught me to SLOW DOWN!!!

When we let the magic smoke out, we made some sort of mistake. Bottom line is engineering is a lot of hacking and the loss of material must be anticipated. We always need to have a backup plan. Do design well, simulate, breadboard and test, commit to silicon or board and test, and test and test and test. Doing it well is rigorous. For me, after almost 40 years, it's still fun!

Datasheets are your friend! Use a regulator and keep the magic smoke where it belongs....where it can continue to do magic for you!

The folks on this forum are great! Best of fortune to you!

--Bill

@crispyoz I really, really, really wish I could up-vote this more than once! Hilarious!!!

--Bill

@crispyoz @Lino-Amorim Absolutely. I didn't linger on the topic long enough to consider causation. You are likely using dynamic IP address allocation at the router. This results in the "dynamic" that isn't really a problem, if you're comfortable with it. The address lease will expire at some time (typical is 24 hours) and the same address will be re-leased if the client is still connected. If the client is not connected at lease expiration the next connection may result in a new IP address, even if you don't have a new router.

In my home I prefer more stability. As such most of the 20 or so IP addresses in use are fixed, with a few dynamic. You can do this by turning off IP address allocation at the router and begin using your Linux box as a dhcp server. You would pair up a device MAC address and the IP address you desire in the /etc/dhcpd.conf file. There's more to it than that, but not much more.

If you'd like to go down this path, please do a little on-line research and take a crack at it. There's assistance available here. The folks on this forum are great!

--Bill

@Lino-Amorim There's really nothing "wrong" here. I've done this hundreds of times over the years. Simply follow the instruction given - remove the newly "bad" host key using the ssh-keygen command, or hand edit the file yourself (line 1), then go through the process of first connection again.

@Lino-Amorim Outstanding! Congratulations! Any lessons learned you'd like to share?

Just past 31 seconds is indicative of a power supply issue as well as the symptoms. That's ironic as you clearly took some precautions. Here are some thoughts:

1. Thick wires - good as this lowers inductance. If there's any appreciable wire length, twist them up. The twisted pair won't benefit you with this issue but will make your system less susceptible to nearby emitters as well as be less of a transmitter. This is a good general practice.

2. 100uF bypass near power pins - good to have a nice size bulk cap nearby. You may need smaller caps as well for their higher frequency discharge capabilities. I doubt that is the issue.

3. The lab power supply is where I would point my suspicion. What is the voltage (presumably 3.3 as you're applying directly to device)? What is the peak current capability of the power supply? Does the supply have a current limiting feature and, if so, is it on?

The O2+ has a large peak in required current as the WiFi system comes up, right around 31 seconds. Likely something in your general power supply circuit is limiting charge to the device as peak charge is required.

Best of luck!
Bill

@crispyoz Yep....I knew that! ;^)

@crispyoz You fell down? Cause? Fosters?

@Nagarjuna-Reddy Off the top of my head I would say the Onion server can see client meta-info only down to the IP layer (3) whereas the media from client to it's first hop is at layer 2 (link). @Lazar-Demin can correct me if I'm wrong, but I've gotta bet there's no means to tell how clients are connected at that level.

--Bill

@tjoseph1 Good work! I'd like to re-emphasize for all newcomers that "power supply", as opposed to "THE power supply", is critical. Of course THE power supply must be able to satisfy peak current demand. The overall power supply also includes any cabling which may provide enough inductance that a power supply barely over the minimum required capability may be inadequate at the power pins of the Omega. It is only the available current at the Omega power pins that matter, not what an upstream subsystem can provide.

--Bill

@tjoseph1 Good work!

--Bill

@tjoseph1 No mistakes I'm aware of or opinionated about. Really seems to me there have been a number of problems on the forum recently that seem to smell like power supply problems. Keep on keepin' on!

--Bill

@tjoseph1 I'm with you, hence the sarcastic humor. Keep hammering at it and you'll get it! When you need help, we're here for you, then before you know it, you're one of us telling someone about O2+ power requirements or vi is at the center of 'evil'. Dig?

take care!!
Bill

@crispyoz @Scruffy I'll second that as many, many people have been bitten by the current surge required as the WiFi modem comes on line. Capacitance across and near the power/gnd is a good thing and long wires add series inductance which is a bad thing.

Just beware that vi is at the center of 'evil'.