Murphy rays?
Peter Jakacki
Posts: 10,193
Have you ever had one of those days when nothing seems to go right. I was out at my friend's workshop setting up a new Prop based project and I had a couple of systems interconnected with one testing the other unit and running Tachyon, so I was communicating to them other serial USB. The power for the units was either the USB for testing or a 24V switch-mode feeding an isolated 5V DC-DC converter proving power for all the logic. Just to complete the setup I had serial output feeding into another second laptop.
The time came when I had to plug in my laptop to the mains and I always expect to have little grounding problems but the Prop unit's supplies were isolated in themselves. Well things worked okay, sorta, but I kept thinking I had a bad USB serial cable (similar to a Propplug), so I tried various ones and isolated the isolated PSUs and tried all kinds of stuff over a couple of hours until I relented and removed the mains from my nearly depleted laptop, and it all came good. Back in again, plays up etc. This never played up on the bench in my workshop, I was starting to suspect Murphy rays if there were such a thing then they would be responsible for my ills. Time for a coffee.
Right then, time to connect the scope but first I isolated that too with a small inverter, I didn't want the grounds interfering too. When I probed the 5V rail on the units which comes from the USB I found volts of noise on there. Unplug the laptop PSU and it was clean, plug it back in again, noisy. But it all works on my bench, why not here? It finally dawned on me that the only difference was the laptop PSU, I leave the original PSU fixed in my workshop and have a universal one in my laptop bag. This laptop tolerates this but won't charge the batteries with it, a fairly common feature these days, but the PSU gets quite hot and has that "too hot smell".
Looking at the burst noise it's primary frequency is around 35kHz but it only occurs in small bursts 100 times a second, that is double our 50Hz mains here. My guess is that the high-voltage filter cap has dried out from the heat and it's loss of capacitance is noticable around the zero-cross regions where the switcher must struggle with line regulation resulting in this burst noise on the 5V line. The laptop seems to be fine but the spikes on the 5V USB feed are definitely causing the Prop to reset or falter.
So I haven't been able to prove that Murphy rays exist yet as it turned out to be something quite mundane instead, but there are other times I would like to be able to test for them, that's for sure..
The time came when I had to plug in my laptop to the mains and I always expect to have little grounding problems but the Prop unit's supplies were isolated in themselves. Well things worked okay, sorta, but I kept thinking I had a bad USB serial cable (similar to a Propplug), so I tried various ones and isolated the isolated PSUs and tried all kinds of stuff over a couple of hours until I relented and removed the mains from my nearly depleted laptop, and it all came good. Back in again, plays up etc. This never played up on the bench in my workshop, I was starting to suspect Murphy rays if there were such a thing then they would be responsible for my ills. Time for a coffee.
Right then, time to connect the scope but first I isolated that too with a small inverter, I didn't want the grounds interfering too. When I probed the 5V rail on the units which comes from the USB I found volts of noise on there. Unplug the laptop PSU and it was clean, plug it back in again, noisy. But it all works on my bench, why not here? It finally dawned on me that the only difference was the laptop PSU, I leave the original PSU fixed in my workshop and have a universal one in my laptop bag. This laptop tolerates this but won't charge the batteries with it, a fairly common feature these days, but the PSU gets quite hot and has that "too hot smell".
Looking at the burst noise it's primary frequency is around 35kHz but it only occurs in small bursts 100 times a second, that is double our 50Hz mains here. My guess is that the high-voltage filter cap has dried out from the heat and it's loss of capacitance is noticable around the zero-cross regions where the switcher must struggle with line regulation resulting in this burst noise on the 5V line. The laptop seems to be fine but the spikes on the 5V USB feed are definitely causing the Prop to reset or falter.
So I haven't been able to prove that Murphy rays exist yet as it turned out to be something quite mundane instead, but there are other times I would like to be able to test for them, that's for sure..
Comments
They tend to clump together into large Murphy Clouds.
I've been stuck in the middle of one such cloud for a couple of months now...
C.W.
Funny you should say that as I originally suspected this and mentioned Murphy particles to my friend but I later changed my mind when the problem persisted and I had already swept thoroughly in the recommended fashion for such particles.Being heavily shielded where we are I had discounted the Murphy rays possibility, but these were probably tightly beamed from a parallax universe to break through the shielding and focused on the poor Props,
What you need is a mutt that will insist on a nice long walk whenever the rays/particles are in abundance. The only solution is avoidance to prevent undue damage.
Loopy Byteloose,
You do have a valid point "...a nice long walk whenever the rays/particles are in abundance"
The key to success is:
1) Stop learning ( put down the book )
2) Think about what you know ( take a long walk )
3) Create ( Apply your knowledge to what you have learned )
4) Repeat ( If you don't continue to learn you will stagnate )
Al
Murphy particles are what make up the magic smoke which you never want to smell or see.
IMO Murphy was an optimist.
Tim
re: so I was communicating to them other serial USB. The power for the units was either the USB for testing or a 24V switch-mode feeding an isolated 5V DC-DC converter proving power for all the logic. Just to complete the setup I had serial output feeding into another second laptop.
Would something like this USB Isolator help your testing?
USB Isolator
Avoid interference and earth loops
Published in issue 430, October 2012
USB Isolator
If your USB device ever suffers from noise caused by an earth loop or
if you want to protect your PC against external voltages then you need
a USB isolator. The circuit described here offers an optimal electrical
isolation of both the data lines as well as the supply lines between the
PC and the USB device.The project itself is a spin-off from an earlier design for
a portable oscilloscope for Android.
http://www.elektor.com/magazines/2012/october/usb-isolator.2266679.lynkx
Once they made them into 14 pin ICs I modified a lot of remote terminals by adding a socket for them on the TX and RCV lines. They would not stand up to a lightning strike but most times when loss of communications was the reported problem all I had to do was pull the cover and replace those two isolators. I was working for a third party provider and had to maintain a lot of remote sites.
I haven't kept up and have no idea if they are developed enough for very high speed applications or if they are now cost prohibitive.
Tim
Optoisolators are alive and well with a wide array provided in DIP for easy replacement. The latest are MOSfet devices... some that can conduct AC or DC in either direction.
My own feelings is the USB was a boon to clearing up desk top cable clutter, and that is about all. It made it easy for a big company to maintain 100 or so computers.
But, it is a lousy, overly complex interface for asynchronous serial that wants to extend any distance. The much ignored RS422/RS485 asynchronous serial is the real winner for simplicity and reliability with a realistic view of what happens when wire is extended over 1000 meters. And of course, RS235 is the grand-daddy legacy clunker that just won't go away, but at least it has generally defaulted to 8N1 with Tx and Rx without flow control in most modern context.
As far a Murphy rays/particles, I have suspicions that they are attracted to beer, nagging bosses, and weariness towards the end of the day. But I've no proof.
In quantum mechanics the double slit experiment is often used to demonstrate particle wave duality. With Murphyons the half court short in basketball is considered the gold standard demonstration. Research has found that sinking a half court shot is significantly easier when no one is watching than during a game. Unfortunately they haven't been able to publish their results. This is because attempts at peer review cause Murphyon wave function collapse, which in turn makes the half court shot impossible.
Well Loopy, I love RS485 too, and I love opto-mosfets, and digital isolators too. Here's one of my boards with them all rolled into one:
The opto-mosfets control a dynamic termination resistor while the digital isolators do what they do best, handle high speed signals as well as provide isolated power to the bus side. There are of course quite a few other parts involved but mainly for lightning and fault protection.
AIUI there is a constant flow of both positive and negative Murphyons and different entities are surrounded by constantly shifting Murphyon fields. That is why you never want to board a plane with an individual known to be carrying a huge negative Murphyon field. AISI it is the polarity of your Murphyon field at the time you drop the toast that controls whether or not it lands jelly side down.
In the question of the half court shoot the outcome is a function of the interaction of the Murphyon fields present at the time shot is taken. If the Gym is empty the shot is more likely to be made. Once the Gym fills to capacity there is an increased likelihood the shooter will make a very loud fart.
For several days now we have needed to replace the headlight in our car and two of the turn signal lights.
Tonight I finally bought the lights before dinner but did not install them. ...After dinner my wife needed to go out for something, and guess what?
She gets pulled over.... fortunately since the bag from AutoZone was in the front seat along with tonight's receipt, she only got a warning.
"Thou shalt not tempt Murphy in any way, shape or fashion."
Several times I have had a poster with those words in large type pinned to the wall above my work bench.
It is a bit worrisome that so far in this project I have not yet released any Murphy particles.
There is a worrisome thought in the back of my mind about the next such release being when we are having our first demonstration to potential customers and/or backers.
Every time you do a dry run you have just clicked the MTBF counter one more notch. ]8-I>