Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[JeffVolp]

Some of you have been following my work on the forthcoming X10 signal meter.  I thought I'd put up a photo of the prototype and tell you a few of its features.  Here is the photo:

http://jvde.us/xtb/XTBM_prototype.jpg

While I originally planned to just replace the ESM1, this will be a "poor man's Monterey".

The readout (updated to match the new photograph) indicates:

X is the X10 signal voltage for the most recent command
F is the received X10 carrier frequency
N is the background noise level
P06 is the last unit code received (P house code, 6 unit code)
DIM is the last command received

There is still a lot of firmware work to do, but the target is this fall.  It will be available in kit or assembled form.

Question, do you prefer the black case or a bone color case, like the XTB?

EDIT NOTE:  The photograph has been updated as the design has evolved.  The noise level is now working correctly, and the Q has been replaced with an actual frequency measurement.

Jeff

[BaBaLou.]

Awesome Jeff!!. Look forward for this be available.

Humm. Bone or Black. Which one will be cheaper Jeff. I choose that one.

[Brian H]

Thanks for the update and teaser. I maybe also in line when it is available.
I personally like Black.

[JeffVolp]

Humm. Bone or Black. Which one will be cheaper Jeff. I choose that one.

I wanted "metallic blue" to match the LCD, but Polycase doesn't make the windowed version in that color.  Black, bone (and gray) are the same cost.

Jeff

[steven r]

...I wanted "metallic blue" to match the LCD, but Polycase doesn't make the windowed version in that color...

Metallic blue would of been cool but bottom line function is more important.
The prototype looks really cool so if most people are saying black then that's fine with me.

[Brian H]

Too bad on the Metalic Blue. That would top my list for sure. 

I do like all the data on the display. Kind of like a TesterLinc's only all on one display. Except maybe the store function that held the last 50 [I think] messages received. The voltage part of the display is a great feature.

[bitman]

That is just plain sexy.
I'll want a kit. I'll need to get it cheaply as possible. 

Black. Beige shows our dirty little fingerprints.

[Boiler]

Hi Jeff,

I can see you haven't been resting on your laurels this summer.  With both the ESM1 and Testerlinc off the market, this device will be a boon to people maintaining their X10 systems.  It looks like a nice improvement over those devices as well.

I do have a few questions:

1) Quality count : Which window is used for this measurement- The transmission window (1ms) or receive window (0.6ms)?  Is there a minimum quality count for decoding a logic 1?  As a reference, my testerlinc routinely shows a Q in the 130 range - Interesting since a 1ms burst of 120Khz would yield a max of 120 counts.

2) Noise N .80V : That's some pretty stout noise.  Is this measured using a bandpass around the X10 frequency, or a highpass?  I'm curious whether this could measure noise in the Insteon carrier region.

3) Displayed X10 Codes: Will this device be able to decode/display X10 extended codes/data?

Sorry for all the questions.  Once again, you've piqued my interest.

Boiler

[JeffVolp]

The quality count is the number of cycles seen in the X10 reception window.  X10 specifies a logic "1" requires a minimum of 48 pulses in a reception window that extends from .250uS to 900uS after the zero crossing.  Actually, my sample window is only 532uS right now, but that will be changed.  It measures the burst voltage at 16 points over the full 1mS pulse length, and displays the average.

As I said, firmware development is in process.  The noise code is not working yet.  There is a bandpass filter with maximum sensitivity just above 120KHz.  Yes, it will certainly measure noise in the Insteon region with almost the same sensitivity as at 120KHz.

Noise is a complex issue.  There can be relatively stable background noise, such as that produced by a single noise source or the unmodulated carrier of a wireless intercom.  Then there can be modulated noise produced by multiple sources summing together.  Differences in the noise frequency cause beat frequencies that modulate the noise amplitude.  Finally, there can be transients, which are high amplitude, short duration events.

While I was going to try to peak detect the transients, their short duration means they really do not effect X10 communication.  My intent now is to display an average noise voltage, and possibly add a descriptive symbol or keyword to indicate whether the noise is stable or varying.

It will display "Xtnd" for an Extended command, but it will not process the extended data portion.  Similarly for X10 Preset Dim commands, where it will display "Pset".

Jeff

[Boiler]

Jeff,

Thank you for the explanation.  I agree with your approach to display average noise (rather than trying to quantify transients) and see the varying/stable tag as being very useful in understanding the noise environment. 

Your explanation did raise another question - will you be averaging noise over the entire 60Hz period or just in a window around the the zero crossing.  Pros and cons for either method.  Standard X10 devices should not be affected unless the noise is present in the transmission window.  However, some devices with AGC could be fooled by out of noise outside the transmission window and decrease their sensitivity to compensate (non gated devices). 

I've always wanted to ask the noise window question for both the ELK and Testerlinc but had nobody to ask.  It makes a huge difference in how you interpret the reading.  In this case, it's really handy being able to ask the designer directly.

Boiler

[JeffVolp]

I am sampling in two windows right now, 1mS before zero crossing, and also during the X10 transmission window itself.  Since that spans the Insteon transmission period, I thought it would allow me to identify an Insteon transmission.  However,  my CFL "noise generator" produces a lot of noise in those two sample windows, which looks like an Insteon signal if I just sample the amplitude.  A recent (untested) change to the PCB allows me to measure the frequency of the signal.  I hope that helps me discriminate between an Insteon command and just in-band noise.  I may sample in a third 1mS window just after the X10 transmission window if that approach alone doesn't work.  That said, there is a limit to how much can be done in a 2KB PIC.

Jeff

[dave w]

I want me one! Don't care 'bout no steeenkeen color.

Actually bone would look good in livingroom or office, as a "dressy" permanent system health indicator. and black or gray looks like a great lab or workshop instrument. I don't think your final color selection will matter much to us hard core "X10ers".

 "Bits" is worried about fingerprints only because he loves peanut butter and jelly sandwiches. Don't worry Bits, only the jelly will show up on the bone. Besides, I bet the bone would look great in your studio 

OK on a more serious note. Jeff,  It looks great. It provides more information that I expected. What is your projected price(s). I haven't "slodered"  a Heathkit in three decades...I'm ready for a nice kit. Can't wait!!

Dave

[JeffVolp]

The cost is dominated by a few components:  PCB, LCD display, case, connectors, power transformer, and AD817A op-amp.  The other 70 odd components are relatively inexpensive.  I only bought bare-bones prototype PCBs so far ($36 each shipped), so I don't have a good number for that.  I'm shooting for the kit to be $70-$80, and the assembled unit maybe $40 more.

Jeff

[dave w]

WOHOO, I got that in a coffee can in the back yard.I'm sure I will have found it by the time you are ready to ship.

[raster69273]

I was hoping for bright yellow so I could find where I last left it.