Oscilloscope Troubleshooting Tutorial

[jgcharlotte]

Is a scope the best way to troubleshoot comm issues?  My system is about 95%, but I occasionally have a failure.  I think I've filtered everything I've installed recently.

There are a number of inexpensive scopes on Amazon, I just don't know if it's the best way to troubleshoot noise issues.  If so, is there a recommendation for a scope?

 

[larryllix]

An inexpensive scope will not likely have the tools to zoom in and capture the Insteon signals well. You will need delayed trigger and wave memory for detailed analysis of a fleeting signal. You may get a good picture of noise and be able to find the source better though..

[jgcharlotte]

Really, that's all I'm looking for is noise so I can track down the source.  I wouldn't know know what I'm looking at in the Insteon signal anyway  

I would think noise on the line would show up pretty well even with an inexpensive scope, but I don't want to waste money if it doesn't.

[Brian H]

I searched here and may have missed it or it was on an other Insteon related site.

Ela posted a Scope Adapter schematic. It isolated you from the power lines and had a switch controlled diode clamp on the output. The ACT CP000 coupler is discontinued but I have used an X10Pro XPCP. Yes both are originally an X10 coupler but Insteon power line signals are easily passed. X10 is 120KHz and Insteon 131.65KHz.

 

 

[Brian H]

If you are respectful of the power lines. A scope maybe all you need. Though some scopes have the ground side of the probe connected to the chassis. So you would have to be real careful the line was NOT on the ground side. Most likely tripping the breaker as scopes have a three prong grounded plug.

[jgcharlotte]

I'm looking at a battery powered multi-meter/scope, so isolation from ground should not be a problem.

https://smile.amazon.com/gp/product/B071F1H3PG?pf_rd_p=c2945051-950f-485c-b4df-15aac5223b10&pf_rd_r=7Q6FMA54JP27QB1TX93V

 

 

[mwester]

 Except that even with a battery-powered scope, plugging the probes into an outlet may result in the knobs, screws, buttons, or even the case being an electrocution hazard.  Shop carefully.  Even better, use some device like the one Brian posted above to remove the high-voltage AC and isolate the Insteon/X10 signal.  Exercise the appropriate amount of caution based on how much you value your life.

[jgcharlotte]

Well its a multimeter also, I would think that wouldn't be an issue.  But good point, thanks.

[MWareman]

If you are respectful of the power lines. A scope maybe all you need. Though some scopes have the ground side of the probe connected to the chassis. So you would have to be real careful the line was NOT on the ground side. Most likely tripping the breaker as scopes have a three prong grounded plug.

I’m not sure I’ve ever come across a scope that internally connected one side of the probe to ground. Usually there is an external ground post on the panel of the scope and the probe has a ground lead that can be used to ground the screen of the probe... At least the several different scopes I’ve used over the users all did that...

The manual for the specific scope should probably be consulted to confirm though...

[Brian H]

My post may not have been too clear.

On my 40MHz dual trace scope. The BNC connectors shell is connected to ground. If you accidentally connect the ground wire on the probe end to a line connection. It will be connected to the power plugs ground pin.

[Brian H]

8 hours ago, jgcharlotte said:

I'm looking at a battery powered multi-meter/scope, so isolation from ground should not be a problem.

https://smile.amazon.com/gp/product/B071F1H3PG?pf_rd_p=c2945051-950f-485c-b4df-15aac5223b10&pf_rd_r=7Q6FMA54JP27QB1TX93V

 

 

I looked at the link. It does look like it would be isolated from you touching anything nasty.

You may have trouble seeing the Insteon signal. As it is around 3.2 volts into 5 Ohms at the zero crossing of the AC wave. That will be 120 volts on the screen.

[jgcharlotte]

I absolutely remember in my first industrial electronics job, they had a scope that had common tied to ground and they cut off the ground prong.  The whole chassis was energized since we were working on 480vac 3ph circuits.  Very dangerous.  They told me 'don't touch the chassis'.  Then a new guy came in and got a Tektronix scope with dual differential inputs, the right way to do it.  $10,000 scope at the time.  This was mid-seventies.  'Back in the day'.  

[Brian H]

I saw a similar thing with an oscilloscope adapter ACT made for their service technicians.  Backwards engineered schematic was found in places. If hooked up with the  Line and Neutral reversed. The scope case was hot if it was isolated with a ground adapter or the ground pin was cut off.

I had the opportunity to us a dual differential input scope. Nice but expensive as you said.

One other thing that could be causing issues. Many electronic manufacturers use a rated for across AC power capacitor. From the Line to Neutral power input. Effective in keeping internal noise off the power lines. It also will absorb {Signal Sucker} Insteon and X10 power line signals. Some also used a coil between the Line input and the capacitor. The added coil or common mode choke. That would  kept the capacitor from being a signal sucker.

I had a X10 system before Insteon was used. My XTB-IIR Repeater Coupler hits the power lines with over a ten volt signal. With my XTBM X10 meter. I can see about a two volt drop in signal when my VIZIO LED TV is connected to the outlet. My low end multi meters  capacitor function. Can read a capacitor from Line to Neutral. Along with one from Line to ground pin and one from Neutral to ground pin.

[IndyMike]

22 hours ago, jgcharlotte said:

I'm looking at a battery powered multi-meter/scope, so isolation from ground should not be a problem.

https://smile.amazon.com/gp/product/B071F1H3PG?pf_rd_p=c2945051-950f-485c-b4df-15aac5223b10&pf_rd_r=7Q6FMA54JP27QB1TX93V

 

 

That's a rather nice battery operated meter with some scope capability.  Unfortunately, it doesn't have anywhere near the bandwidth or memory for viewing Insteon signals at 130 KHz.

Noise that affects insteon should be in the 100 - 150 KHz range.  The analog bandwidth or the meter is quoted as 10kHz and the sample rate as 200K samples /sec.  That's at least 2 orders of magnitude off what you need.

When ELA and I were trading scope traces years back, I was using a Owon DS6062 with the ACT CP000 adapter.  The OWON was 60MHz bandwidth, 500 M samples/sec, and had 10 M memory depth.  I pushed this devices to it's limit looking at insteon signal "anomalies".

If you truly have noise issues, I do not believe the meter you are looking at will help identify them.  

On the subject of noise - I have come around to the thinking that MOST problems are related to signal absorption (low signal levels).  I'm a bit out of date on what tools are available to measure signal to noise levels at a location.  It should not require a scope.

 

 

Meter Specs

Specification:
• Analog Bandwidth: 0~ 10kHz
• The maximum real-time sampling rate: 200ksps
• DC voltage: 0mV ~ 1000V ± (0.75% rdg + 10dgt)
• AC voltage: 0 mv ~ 750V ± (1.0% rdg + 10dgt)
• DC Current: 0mA ~ 20.00A ± (1.2% rdg + 10dgt)
• AC current: 0mA ~ 20.00A ± (1.5% rdg + 10dgt)
• Resistance: 0W ~ 20.00MW ± (1.0% rdg + 5dgt)
• Capacitance: 0nF ~ 200.0μF ± (2% rdg + 10dgt)
• Frequency: 0kHz ~ 200kHz ± (1.0% rdg + 5dgt),
• Diode test: open circuit voltage of about 1.5 V, the • maximum test current of about 1.5mA.
• Continuity Test: determine resistance: about 30
• Remote control Detection: 38kHz carrier frequency
• Transistor Magnification: 30 ~ 1000 (annex)
• Fuse rating: Φ5 × 20,250mA / 250V

[jgcharlotte]

Thanks.  That's the kind of stuff I'm not that familiar with.  The Insteon signal and the type of noise that would affect it.

I guess I'll invest in some more filters.  It really isn't that bad, but there are still some devices that are not filtered and the problems are not repeatable enough to troubleshoot just by unplugging stuff.  Case in point, I opened the door to the garage yesterday and the garage lights did not come on.  Immediately closed it, then opened it again and they did come on.  I can't believe anything changed a whole lot in that 5 seconds.  I'm probably just on the hairy edge somewhere.  I didn't check the log to try and figure out if the failure was with the o/c switch or switchlinc.  Everything works 95% of the time

[paulbates]

31 minutes ago, jgcharlotte said:

Thanks.  That's the kind of stuff I'm not that familiar with.  The Insteon signal and the type of noise that would affect it.

I guess I'll invest in some more filters.  It really isn't that bad, but there are still some devices that are not filtered and the problems are not repeatable enough to troubleshoot just by unplugging stuff.  Case in point, I opened the door to the garage yesterday and the garage lights did not come on.  Immediately closed it, then opened it again and they did come on.  I can't believe anything changed a whole lot in that 5 seconds.  I'm probably just on the hairy edge somewhere.  I didn't check the log to try and figure out if the failure was with the o/c switch or switchlinc.  Everything works 95% of the time

Likely the signal is a little marginal there and the garage door motor starting up injects noise. You can add a filter like the XPNR across the hot and neutral right at the opener's wiring box. 

Assuming that this is an iolinc, that's a single band device, no RF. Making it more vulnerable to line noise like a motor. Another option is to plug a range extender, or spare lamplinc, right in the back of the iolinc, that will add RF bridging to the circuit there.

Paul

[jgcharlotte]

The garage door I'm referring to is the entrance to the garage, not the actual garage door.  I just have an o/c sensor on the door to turn on the switchlinc.

As far as the garage door itself, I have it plugged into a filterlinc, anticipating motor noise (it plugs into a receptacle in the ceiling).  I used an iolinc for awhile for garage door activation and open/close detection, but it became unreliable, so I replaced it with an o/c sensor and on/off module with a dry contact.  No problems since then.

[larryllix]

5 hours ago, jgcharlotte said:

The garage door I'm referring to is the entrance to the garage, not the actual garage door.  I just have an o/c sensor on the door to turn on the switchlinc.

As far as the garage door itself, I have it plugged into a filterlinc, anticipating motor noise (it plugs into a receptacle in the ceiling).  I used an iolinc for awhile for garage door activation and open/close detection, but it became unreliable, so I replaced it with an o/c sensor and on/off module with a dry contact.  No problems since then.

I get the occasional ioLink missing comm from my garage door monitoring. I never connected the operation of it to the ioLink. I have good neighbours.
I always blame it on being on the far side of my PV inverter system as I have some occasional problems with a KPL, right above it, used a a combination lock, also. That uses four keystrokes in rapid sequence.

I have been thinking about trying an extra CAO Wireless Tag on my second garage door and see if it is worth the effort. Cloud based? Yuk.

[ELA]

On 11/16/2018 at 6:28 AM, jgcharlotte said:

Is a scope the best way to troubleshoot comm issues?  My system is about 95%, but I occasionally have a failure.  I think I've filtered everything I've installed recently.

There are a number of inexpensive scopes on Amazon, I just don't know if it's the best way to troubleshoot noise issues.  If so, is there a recommendation for a scope?

 

Hello jg,

 I have always felt an oscilloscope is an excellent tool to troubleshoot comm issues.  It does require a learning curve and lots of investigations.

Many thanks to IndyMike for introducing me to the CP000 for isolating the scope. You must suppress the 120VAC(60hz) using this or some other device in order to be able to better "see" the lower level Insteon signal. 

   It does not require a real high bandwidth scope to take some basic signal amplitude measurements.  You are more concerned with relative amplitudes for Insteon troubleshooting than you are with viewing details of individual bits.

As IndyMike said, I have always believed that the majority of Insteon issues are insufficient signal levels due to signal suckers.   I created a 9Khz (high pass) filter to allow more accurate "noise" measurements with my scope.    That is required because noise is often very low in amplitude and cannot be seen without filtering out the 60hz to a very high level.

This filter allowed me to "see" noise.  In most cases the noise did not interfere with the Insteon signal unless the communications were already very marginal due to a very low Insteon signal level ( that was often due to signal suckers). 

I am attaching a few of the many scope shots I captured years ago in my attempts to improve my network reliability.    At that time I also created a very powerful tool that emulated an oscilloscope and allowed for the ability to measure and quantify signal suckers. ( called the ELAMONTOR).

One picture shows the relative signal strengths of two different PLMs. I tested a new, right out of the box, PLM only to find it was inferior in signal strength  and returned it.

The other picture shows a terrible LED- bulbs noise.   I tested many different LED bulbs and most did not put out much noise.  That one stood out as terrible.

With an Oscope + isolator, Test PLM and and Filterlinc you can build an isolated test network.   You can then plug in suspect signal suckers into that Isolated test network to view, measure and quantify to what level a signal sucker might affect the Insteon signal level.

Best of luck to you once you start down the rabbit hole.   I am of the belief that the Insteon protocol does not always perform "Simulcasting" as well as it is advertised to.  When that is happening you might find very strong Insteon signal levels, yet experience poor reliability at some nodes in your network.     Those can sometimes be overcome with a newer or dual band device.

 

[jgcharlotte]

Wow, lot of good information, thanks!