ProSoundWeb Community

Please login or register.

Login with username, password and session length
Advanced search  

Pages: 1 ... 11 12 [13] 14 15 ... 55   Go Down

Author Topic: brain storm an optimal human safety system for back line  (Read 105166 times)

Scott Holtzman

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 5608
  • Ghost AV - Avon Lake, OH
    • Ghost Audio Visual Systems, LLC
Re: brain storm an optimal human safety system for back line
« Reply #120 on: January 19, 2015, 12:08:25 pm »

Good news- bad news. Good news I am coming up with a workable plan for my smart outlet.

#1 test for floating ground.  (do not close relay?)
#2 test for reverse polarity   (do not close relay?)
#3 test for bootleg connection  (do not close relay?)
#4 test for current flowing in ground (open relay).

The bad news is coding the microprocessor brain involves running the processor in an emulator connected to a computer. I am very apprehensive about connecting my computer to a hot chassis design. Especially one capable of being connected to RPBG to detect the faulty wiring. 

I may have to rig up a floating transformer isolated prototype but this still seems a little dicey. I just researched buying a formal opto-isolated emulator rig from the microprocessor company I use and they want over $400  :( . A generic JTAG opto interface is probably $50 but then I would have to discover all the secret handshakes to get it working.

arghhh...

JR

That's a step backwards and I have nothing to offer.  Sounds like a good water cooler discussion, will bring it up at work today and see what everyone thinks.

I assume you have a ribbon cable to a card in the computer and then some type of processor emulator that plugs into your design board?

Logged
Scott AKA "Skyking" Holtzman
River Delta Audio is now:

Ghost Audio Visual Solutions, LLC
Cleveland OH
www.ghostav.rocks

John Roberts {JR}

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 16408
  • Hickory, Mississippi, USA
    • Resotune
Re: brain storm an optimal human safety system for back line
« Reply #121 on: January 19, 2015, 12:43:39 pm »

That's a step backwards and I have nothing to offer.  Sounds like a good water cooler discussion, will bring it up at work today and see what everyone thinks.

I assume you have a ribbon cable to a card in the computer and then some type of processor emulator that plugs into your design board?

Modern micro emulation is performed with a small hardware dongle that plugs into a PC USB port and then provides a (generic?) JTAG programming/emulation interface with the micro. Microprocessors with flash memory can be directly programmed this way, and run one clock step at a time for in circuit de-bugging.

The hardware dongle with an opto-isolator between it and the micro does not seem that exotic. It might run a little slower than a direct digital connection, and programming the flash may take a little more effort than just passing digital states back and forth. I ASSume this has all been sorted but not widely used, so a dedicated solution for my preferred micro family costs hundreds of dollars.

For that much money I am tempted to develop off-line, but this is harder, since I can't easily check the internal A/D measurement results, etc. Of course I could externally vary input stimulus and empirically determine thresholds.

JR   
Logged
Don't tune your drums half-ass. Listen to what a properly "cleared" drum sounds like.   http://circularscience.com/

Mike Sokol

  • Moderator
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 3356
  • Lead instructor for the No~Shock~Zone
    • No~Shock~Zone Electrical Safety
Re: brain storm an optimal human safety system for back line
« Reply #122 on: January 19, 2015, 01:41:16 pm »

Good news- bad news. Good news I am coming up with a workable plan for my smart outlet.

#1 test for floating ground.  (do not close relay?)
#2 test for reverse polarity   (do not close relay?)
#3 test for bootleg connection  (do not close relay?)
#4 test for current flowing in ground (open relay).

The bad news is coding the microprocessor brain involves running the processor in an emulator connected to a computer. I am very apprehensive about connecting my computer to a hot chassis design. Especially one capable of being connected to RPBG to detect the faulty wiring. 

I may have to rig up a floating transformer isolated prototype but this still seems a little dicey. I just researched buying a formal opto-isolated emulator rig from the microprocessor company I use and they want over $400  :( . A generic JTAG opto interface is probably $50 but then I would have to discover all the secret handshakes to get it working.

arghhh...

JR

As you have surmised, there will be numerous mis-wiring issues that you'll need to create on purpose to test your design, and a bunch of them would create dangerous currents back-flowing into your properly grounded computer. Of course, you could run this on a laptop with battery power. However, there will be several outlet fault conditions (RPBG being obvious) that would create a hot-chassis condition on your laptop. Probably would NOT blow up the laptop, but WOULD be deadly for anyone touching metal on the laptop and anything else earth grounded. I would considering using a plastic laptop and keep it isolated from everything else including no RJ-45 connections or USB printer hookups. Still, you'll need to be aware of the potential shock hazard to you if you get in the fault path. I would probably risk it myself, and you (JR) might be willing and knowledgeable enough to do it safety for testing. But I don't want anyone else on this forum creating a hot-chassis condition on purpose because of the risks involved.   
Logged
Mike Sokol
mike@noshockzone.org
www.NoShockZone.org

Stephen Swaffer

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2253
Re: brain storm an optimal human safety system for back line
« Reply #123 on: January 19, 2015, 11:29:49 pm »

Even RPBG can be simulated on an isolated workbench with a circuit fed by a GFCI breaker.  Off the cuff, I'd say 90% of your testing could be done that way using a battery powered laptop though optical isolation would be a nice touch..  Final verification might best be done sans GFCI protection-but at that point you could reasonably do it wearing voltage rated gloves.
Logged
Steve Swaffer

Jonathan Johnson

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 3004
  • Southwest Washington (state, not DC)
Re: brain storm an optimal human safety system for back line
« Reply #124 on: January 20, 2015, 12:18:14 am »

As you have surmised, there will be numerous mis-wiring issues that you'll need to create on purpose to test your design, and a bunch of them would create dangerous currents back-flowing into your properly grounded computer. Of course, you could run this on a laptop with battery power. However, there will be several outlet fault conditions (RPBG being obvious) that would create a hot-chassis condition on your laptop. Probably would NOT blow up the laptop, but WOULD be deadly for anyone touching metal on the laptop and anything else earth grounded.

Connect the laptop to the network via WiFi (no CATx ethernet needed). Then use Remote Desktop or VNC (or some other means) to access the laptop's desktop remotely in order to do the programming. Then you don't have to actually touch the laptop when the intentional faults are present.
Logged
Stop confusing the issue with facts and logic!

John Roberts {JR}

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 16408
  • Hickory, Mississippi, USA
    • Resotune
Re: brain storm an optimal human safety system for back line
« Reply #125 on: January 20, 2015, 11:13:23 am »

The last time I fired up my old laptop it would not boot-up so I would have to fix it to use it, and then load up the emulation software and drivers...

I now have an alternate plan. Phase one I fire up my prototype with a low voltage floating power source to confirm the different test modes with faked stimulus. Phase two, I finish development running it hot chassis. If I use one of my old tuner PCBs as a prototype test platform I have two 12 LED meters that I can use to display different measured results so I can easily see what levels the microcomputer is getting from different faults, so I guess I can't use that as an excuse.  8)

Tests as I see them are

#1- is ground floating or bonded to neutral? With micro referenced to neutral drive a couple k ohm resistor into input side ground with relay open and read that ground voltage with one A/D input. If the ground is floating it will be easy to drive it up to 3V by the micro. Note: this A/D input is one of two used later to sense ground current.

If ground is floating outlet can still be used thanks to the GFCI but I probably won't. Make them find a grounded power drop so indicate open ground fault and stop there. 

#2  test for reverse polarity. With micro referenced to apparent input neutral, sniff the neutral connection on the still floating output side with an A/D input. If what is supposed to be a quiet input neutral is actually line (hot), the quiet output side neutral should register a not very quiet voltage wrt the micro that is referenced to hot. (this will work better with equipment plugged in, but may get a usable reading from outlet wiring, since the micro A/D input is relatively high impedance.  This is my DIY NCVT.     

Again reverse polarity by itself is not immediately dangerous but I might as well indicate the problem and prevent connection. In a price is no object product, another relay could correct the polarity while this gets a little hairy to keep the micro powered during the switch-over. If price is no object I don't need to make it hot chassis but the market drops exponentially with added cost so I will just indicate and refuse to connect power if any fault is found.   

#3- Test for bootleg ground. ASSuming we already passed tests one and two, so polarity is correct and ground is bonded, now switch in a couple amp load between mains line and neutral. To avoid dissipating hundreds of watts, this load can be pulsed on for only a few mSec at a time. This load should cause a small voltage step in both the line and neutral wiring. If the ground is bootlegged to neutral, it will move in lock step with the neutral so there will be no voltage step seen in the ground relative to neutral. If the ground wire is a separate run back to the panel, it will be unaffected by the pulsed load so there should be a measurable step voltage seen between neutral and ground (because neutral is the one changing).   

A bootleg ground is not immediately a shock hazard but wrong to leave it unchallenged, so again we can make the operator search out a proper outlet with proper ground. If the distro wiring is extremely low impedance this pulsed voltage artifact could be small. A tradeoff may exist between sensitivity of this measurement and protecting the A/D input from over current during a fault. 

#4 After passing tests one, two, and three, we are ready to switch on the power relay and connect the mains power to the GFCI outlet and output side. In this operational mode we will connect another A/D input to the output ground with a modest resistance between the input ground and output ground. The micro will look for a voltage difference between the two ends of the resistor to compute a current flow. When the arbitrary current threshold is exceeded the relay will latch open and indicate the ground current fault.

The GFCI will also trip and protect if any of it's line current does not return properly.

Note a hot output side ground will have a path to the input side neutral through the micro A/D so this input will need to be current limited with suitable large resistance to keep and fault current down in sub-lethal mA range.

JR

PS: Now I need to stop wanking get back to my day job...
Logged
Don't tune your drums half-ass. Listen to what a properly "cleared" drum sounds like.   http://circularscience.com/

Robert Lofgren

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 825
Re: brain storm an optimal human safety system for back line
« Reply #126 on: January 20, 2015, 12:49:52 pm »

Modern micro emulation is performed with a small hardware dongle that plugs into a PC USB port and then provides a (generic?) JTAG programming/emulation interface with the micro. Microprocessors with flash memory can be directly programmed this way, and run one clock step at a time for in circuit de-bugging.

The hardware dongle with an opto-isolator between it and the micro does not seem that exotic. It might run a little slower than a direct digital connection, and programming the flash may take a little more effort than just passing digital states back and forth. I ASSume this has all been sorted but not widely used, so a dedicated solution for my preferred micro family costs hundreds of dollars.

For that much money I am tempted to develop off-line, but this is harder, since I can't easily check the internal A/D measurement results, etc. Of course I could externally vary input stimulus and empirically determine thresholds.

JR   
There are so many creditcard prototype boards that run linux and like with all the gpio you'd want. You should be able to take a $35 raspberry pi (arm-based) and then use native gcc and gdb using SSH over wifi. Depending on how you write the code it should be portable...

While you are coding you could even monitor the mains voltage with what you probably already have in your design and any noise on the line and log this.
Logged

John Roberts {JR}

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 16408
  • Hickory, Mississippi, USA
    • Resotune
Re: brain storm an optimal human safety system for back line
« Reply #127 on: January 20, 2015, 01:37:52 pm »

There are so many creditcard prototype boards that run linux and like with all the gpio you'd want. You should be able to take a $35 raspberry pi (arm-based) and then use native gcc and gdb using SSH over wifi. Depending on how you write the code it should be portable...
I recognize about half that jargon...  Are you suggesting a wifi enabled prototype could be run in emulation mode wirelessly?

I am not interested is starting from scratch in a completely new development environment using an unfamiliar (to me) high level language. I generally code embedded applications using machine language. I have too many years and brain cells invested in my current comfort zone.

If it is actually that easy tell us how it works out. I drew you a map for how the different faults could be isolated and measured (hypothetically). 
Quote
While you are coding you could even monitor the mains voltage with what you probably already have in your design and any noise on the line and log this.
Yes I could monitor and meter line voltage while that was not in my original design brief that might be a merchantable extra feature that customers would value more than it cost to implement (a few LEDs).

JR
Logged
Don't tune your drums half-ass. Listen to what a properly "cleared" drum sounds like.   http://circularscience.com/

Frank Koenig

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 872
Re: brain storm an optimal human safety system for back line
« Reply #128 on: January 20, 2015, 07:10:40 pm »

I'm not saying the the Raspberry Pi is necessarily the best choice for this application, but I'll give this supportive anecdote. About 8 years ago a buddy and I built a little telemetry system for a remote home -- more-or-less a home-brew home-automation system. We used Windows XP running on a superannuated laptop and an industrial Ethernet data acquisition box. This works great to this day but we wanted to get away from the old laptop and its power issues and also get away from Windows.

The original code is all Java and my buddy ported it to the Raspberry Pi with remarkably little effort. We were up and running on the Pi in hours. I'm now in the process of designing a simple analog and digital interface board for the Pi using its 40 pin GPIO connection. Once I get that going we can ditch the Ethernet data acquisition system, too.

The Pi originally was developed for educational use but now has been widely adopted by all kinds of folks. A friend who works at an electronic hardware manufacturing company told me they use many in their production test fixtures, for example. The only thing that sucks about the Pi, in my opinion, is the the lack of any ordinary engineering documentation. It's as though its creators wanted the world to reverse engineer it and talk about it endlessly on Internet forums.

Anyway, Raspberry Pi is a solid, cheap, powerful little platform. I just ordered two more -- one might replace another old Windows PC as my "Internet radio receiver".

Best,

--Frank

PS: Forgot to say that we're using "Raspian" Linux on the Pi.
« Last Edit: January 20, 2015, 07:20:19 pm by Frank Koenig »
Logged
"Nature abhors a vacuum tube." -- John Pierce, Bell Labs

John Roberts {JR}

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 16408
  • Hickory, Mississippi, USA
    • Resotune
Re: brain storm an optimal human safety system for back line
« Reply #129 on: January 30, 2015, 04:35:13 pm »

Back on topic, I just got in the special security screw-head driver so I could take apart the commercial GFCI power strip/adapter. No surprises inside, I can see the current difference sense transformer and a pretty heavy duty electro mechanical circuit interrupter(?) . Two pcbs using cordwood construction (top and bottom PCB with parts between them.)

The big honking ground wire is a pass through and not connected to the GFCI boards so I should be able to break the ground without defeating the GFCI protection.

I am tempted to add a ground lift switch with a stinger cap in the lifted ground path. I don't recall hearing much feedback from the several interested guitar players on this thread about how effective the stinger cap is at reducing hum, but i should probably just test this myself. I was planning to give this away to some lucky local guitar player/singer after I am done with this particular investigation.

I may modify this one and drop in on a local bar band gig... Should take all of a couple minutes to confirm if the stinger cap works to manage mains noise. There are no FOH power drops in the small bar gigs I'm talking about so this is probably the optimal solution for them. They can keep using their dicey old guitar amp and not die.

The big dog relay and blinky lights is a premium product offered for sale to unappreciative cheap bastaads... (pearls before swine). This stinger GFCI could be cheap and easy (practical, sellable, etc).

JR 
Logged
Don't tune your drums half-ass. Listen to what a properly "cleared" drum sounds like.   http://circularscience.com/

ProSoundWeb Community

Re: brain storm an optimal human safety system for back line
« Reply #129 on: January 30, 2015, 04:35:13 pm »


Pages: 1 ... 11 12 [13] 14 15 ... 55   Go Up
 



Page created in 0.081 seconds with 22 queries.