Author Topic: Adding E Field Probe - Startup and Operation - not for construction (Read 56498 times)

Cutty Sark Sailor · « **Reply #25 on:** May 25, 2014, 11:47:56 AM »

Don, that's exactly the reason I suggested turning all autos off, and getting the feel for the new addition, not even upgrading my firmware.
We had to 'feel out red' and the competition was a riot.. folks are gonna have to 'feel out E field' in the same way, the characteristics are different, the sensitivity is incredible, the filtering magnificent, and folks are NOT going to remember that the design intent is for signals under 500km, You put too much probe on this thing, you'll be out of whack before you know it.

One thing about the earlier "TOGA" detection... Egon would have that as far back as possible, and I think he did make some server adjustments to move it down the Ground Wave signal, and slightly back in time...

About the 'time delays'.. you'll note looking at the E field signals, each channel has a very specific time delay, determined by the filters. There is virtually no tolerance on those delays at the server level. Any diddling with component values and design is gonna mess with those, At one point the designers, if I remember correctly, considered getting the boards prebuilt, and encapsulated to keep us from diddling with stuff, as we're prone to do.

Thank goodness that was quickly vetoed...

Anyway, as usual, somebody'll do that very thing, then wonder why signals are rejected, things don't work, etc.
I already made one error, failing to consider the specific reason for the specific power LED in the preamp. It didn't occur to me that is had a more important function than just lighting up, and saved us a few bucks.

dfroula · « **Reply #26 on:** May 25, 2014, 11:52:43 AM »

Oh, you never used the external LED on the probe housing? I liked that idea!

What is the LED used for, a voltage reference?

Don
WD9DMP

Cutty Sark Sailor · « **Reply #27 on:** May 25, 2014, 12:00:27 PM »

Quote from: dfroula on May 25, 2014, 11:52:43 AM

Oh, you never used the external LED on the probe housing? I liked that idea!

What is the LED used for, a voltage reference?

Don
WD9DMP

yes, in effect a rail splitter, more or less 2.5 v on the op amp. I would not be surprised if any inherent noise in that particular LED is used to introduce some dither, also... who knows???
BTW,... a visible power light isn't needed... you'll know if the preamp isn't working...
I also decided that in effect adding the wires to the LED might affect the tight filtering... so I changed it back.

dfroula · « **Reply #28 on:** May 25, 2014, 01:05:39 PM »

Hmm...confused again!

Alternate channel mapping and auto-noise adjust enabled, yet both Status and Signals page showing DIFFERENT thresholds.

Don
WD9DMP

Cutty Sark Sailor · « **Reply #29 on:** May 25, 2014, 01:40:54 PM »

Quote from: dfroula on May 25, 2014, 01:05:39 PM

Hmm...confused again!

Alternate channel mapping and auto-noise adjust enabled, yet both Status and Signals page showing DIFFERENT thresholds.

Don
WD9DMP

LEAVE BRITTANY ALONE !!!!!!!!!!!!!!!!!!!!! ARRRRGGGHHH....

Cutty Sark Sailor · « **Reply #30 on:** May 25, 2014, 05:36:17 PM »

Quote from: Cutty Sark Sailor on May 25, 2014, 01:40:54 PM

Quote from: dfroula on May 25, 2014, 01:05:39 PM
Hmm...confused again!

Alternate channel mapping and auto-noise adjust enabled, yet both Status and Signals page showing DIFFERENT thresholds.

Don
WD9DMP
LEAVE BRITTANY ALONE !!!!!!!!!!!!!!!!!!!!! ARRRRGGGHHH....

While server automatic control is still beta, and with both filters engaged, found a lot to be desired, as storms moved into 300 mi range with moderate longer range activity. Had to return to manual, and lower gains to get out of interference and stay there... This is with all 5 channels running. Running 'borderline' now, if you follow me, in full manual, with channel mapping.

Cutty Sark Sailor · « **Reply #31 on:** May 25, 2014, 06:59:57 PM »

***addenda: To initial post... added e.g. spectrum signals comparing E field only to H field only.

Cutty Sark Sailor · « **Reply #32 on:** May 27, 2014, 08:16:50 AM »

Adding to main post 5/27/14

Filters: (somewhat tame feral assumptions)*
The coax and it's shielding is effectively part of a
High Pass filter > 5 KHz (preamp, coax, amp combined)
First stage HP filter is in Pre-Amp, third stage in amp.
50,100,150,200,250 Hz is effectively attenuated
Similar with 60Hz and harmonics
23.4 KHz (18 KHz-30 KHz)
50 KHz (I believe centered at 44KHz)

± effective channel selectivity
A = 5KHz ± —— 23KHz ±
B = 18KHz ± —— 44KHz ±
C = 5KHz ± —— 50KHz ±
* "Somewhat tame feral assumptions" = "Slightly wild guesses"

Cutty Sark Sailor · « **Reply #33 on:** May 27, 2014, 11:44:59 AM »

5/24/14
added stroke filtering comparison image from 2 stations, one E+H, with alternate channel mapping
vs.
station running H only.
Same strokes for each station.

W7DRM (old call was W3DRM) · « **Reply #34 on:** May 27, 2014, 11:59:21 AM »

NOTE TO WXFORUM MODERATORS/ADMINISTRATORS: I just tried to print this thread (using the PRINT button at the top of the thread) and discovered that while it will print the text, the images are represented only by the URL's. That really shoots down the feature because most of the time we want to view images along with the text. Not sure if there is some kind of different setting or if I'm doing something wrong here...

Cutty Sark Sailor · « **Reply #35 on:** May 27, 2014, 12:14:56 PM »

Try printing as PDF...

Those images and videos are coming from a totally different place server. PDF may work, but you might old off. I've requested a 'proof reading' from development... may have to correct some stuff... never know.

W7DRM (old call was W3DRM) · « **Reply #36 on:** May 27, 2014, 12:29:48 PM »

Quote from: Cutty Sark Sailor on May 27, 2014, 12:14:56 PM

Try printing as PDF... Those images and videos are coming from a totally different place server. PDF may work, but you might old off. I've requested a 'proof reading' from development... may have to correct some stuff... never know.

Thanks for the suggestion Mike. But, it didn't work either. I am sure it's simply the way SMF interprets a page it is going to print. If I do a simple Ctl-P then the entire page prints along with the images. The only problem is that if there are multiple pages in the thread then you have to print each page separately.

I'm trying to get my head around your dissertation on the thread subject matter and I like to see things in print so I can make notes as I go along - guess I'm just old-fashioned...

W7DRM (old call was W3DRM) · « **Reply #37 on:** May 27, 2014, 12:43:53 PM »

I just tried to get into the SMF website and nothing comes back. Says there is some kind of problem on the server.

A Google search shows some comments about this printing without images being a bug with the SMF software. I can't verify that since I can't get to the SMF website...

Further Google searching seems to indicate there is a mod to make SMF print images rather than just the URL to them. Again, with the site apparently down, I can't check it out any further...

Cutty Sark Sailor · « **Reply #38 on:** May 27, 2014, 12:48:54 PM »

I know the feeling, Don,... A whole lot of the confusion will be solved once you build and begin using the E-field... I thought maybe putting up the images and videos could help simulate that a bit.
I couldn't get my PDF to print, either... but you're right, with the images, it wants 29 pages. remember, I believe admins have the print function set for text only, that would be standard procedure, using the thread 'print' button.

Take topic one, for example, highlight it all, and copy, past into Word or similar... see if that'll work... then you could edit it a bit.
Here it is raw pdf....
http://ourspecial.net/misc/Adding%20the%20E%20Field%20Probe.pdf

W7DRM (old call was W3DRM) · « **Reply #39 on:** May 27, 2014, 12:53:25 PM »

Quote from: Cutty Sark Sailor on May 27, 2014, 12:48:54 PM

I know the feeling, Don,... A whole lot of the confusion will be solved once you build and begin using the E-field... I thought maybe putting up the images and videos could help simulate that a bit.
I couldn't get my PDF to print, either... but you're right, with the images, it wants 29 pages. remember, I believe admins have the print function set for text only, that would be standard procedure, using the thread 'print' button.

Take topic one, for example, highlight it all, and copy, past into Word or similar... see if that'll work... then you could edit it a bit.
Here it is raw pdf....
http://ourspecial.net/misc/Adding%20the%20E%20Field%20Probe.pdf

With Chrome (and perhaps other browsers too), you can highlight the section you want to print and then when you get to the actual print screen, you can tell it to print only the selected areas. Works great but takes a bit more work on the users part to make it happen. That's what I am about to do right now...

Cutty Sark Sailor · « **Reply #40 on:** May 27, 2014, 05:53:00 PM »

ADDENDA: 5/27/14 (evening) Richo provided Gain and Filtering specs... added to main post, with minor data correction to previously posted info.

Cutty Sark Sailor · « **Reply #41 on:** June 01, 2014, 05:31:05 PM »

Updated: 6/1/14 :qualification sentence modified under "suggested initial settings" regarding filters
Mike

Cutty Sark Sailor · « **Reply #42 on:** June 10, 2014, 07:23:33 PM »

updated: 6/10/14 : Initial Post: added my settings return to full manual operation, with no extra filters, after running in Automatic with both filers for a period of time.

miraculon · « **Reply #43 on:** June 11, 2014, 03:24:23 PM »

I have mine on full automatic. Since I went with 20cm for the antenna length and raised it up higher, it works really well. It also seems relatively immune to noise, which is surprising to me. All those LPF (low pass filters) really do the job, since all of my detectors are in scary proximity to overhead power lines. I am amazed.

This was a seriously good addition to the system!

Greg H

dfroula · « **Reply #44 on:** June 16, 2014, 09:36:15 AM »

Once an E-field amp is added, setting the gains and thresholds is no longer as straightforward as before!

In my opinion, enabling "Alternate Channel Mapping" is an absolute must. The attached images attempt to show why.

Keep in mind that overall "relative gain" of a channel is affected by two things - the gains set on the op-amps (either fixed or adjustable) and the threshold settings. The apparent or "relative" gain can be increased by by increasing the gain on an adjustable op-amp stage or by reducing the threshold on that same channel. Conversely, the apparent or "relative" gain can decreased by decreasing the gain or increasing the threshold.

Gains are set on the amp board(s). Thresholds are set on the Analog to Digital Converters (ADCs) on the STM32F4Discovery board. The STM32F4Discovery has 3 ADCs, each with two channels. After the E-field amp is added, those six available ADCs need to be mapped to the five total channels of the two amps in a way that allows the channel thresholds to be adjusted for one amp without affecting the other. Compromises need to be made. Only the "Alternate Channel Mapping" option allow threshold settings of one amp to be made without affecting the settings of the other amp.

This is particularly important if the "Auto Adjust to Noise" feature is also enabled. That feature does fast relative gain adjustments in response to local noise by varying the thresholds. A slower adjustment is done by changing the gain of the adjustable op-amps on the channel experiencing the noise. Standard channel mapping will move the threshold of a channel on one amp, but unless "Alternate Channel Mapping" is enabled, the associated threshold on the other amp will change as well, even though it may not be experiencing the noise!

The same problem exists somewhat even with "Alternate Channel Mapping" enabled, as the thresholds of channels A and B on the same amp are now shared. However, at least the common threshold will now be on the same amp, where the noise source is more likely to be common. This allows the "Auto Adjust to Noise" feature to do a better job at reducing the effects of transient noise.

Best regards,

Don

dfroula · « **Reply #45 on:** June 16, 2014, 09:52:06 AM »

I finished up my e-field amp kit yesterday. All fired up correctly the first time

.

I kept to the current convention of putting the e-field amp on controller input jack #2. I enabled the "alternate channel mapping" on the controller so that independent threshold adjustments could be made for channels A and B of each amp (but now common for channels A and B on the same amp). The channel C threshold is now common across the two amps, but as there is no channel C active on the H-field amp, there is essentially an independent threshold setting for channel C of the e-field. All gains are independently adjustable, except for the first stage of the e-field amp, which is common to all three channels. That gain is set by adjusting the first stage gain of the C channel. Channels A and B follow the first-stage setting once the gain setting is saved.

Currently, I have the preamp taped to a 6 foot length of PVC tubing clamped to a table in a second-story room (yes, indoors!) for testing. I am using a 6 inch piece of 12 gauge wire at the moment for an antenna.

I had some initial problems getting the signals recognized by the server. It turns out adding more than 2 channels to the controller causes the size of the UDP packet that is used to transmit the lightning data to exceed the 1500 byte MTU size used by default by my router (and most other devices on your local network). Most computers and other network devices will not create a packet with an MTU > 1500 bytes, so fragmentation is never needed. Not so for the Red controller! The controller board handles a data packet for > 2 channels by "fragmenting" the packet into two sub-packets. However, my router had a setting to reject fragmented packets from being accepted as a security feature, and was blocking all data from the controller from being sent, if more than 2 channels' worth of data were being transmitted. If you run into this, you may need to turn off the "Block Fragmented IP Packets" security function of your router, if it has such a setting and it is enabled. For reference, my router is a Motorola/Arris SBG6580 (combo DOCSIS 3 cable modem/router/wifi access point).

Once I sorted that issue, with Tobi's help, all started working perfectly. I eliminated a few local noise sources, found a reasonable gain setting, and enabled the "Auto Adjust to Noise" feature, which I have been using for the past few months. I find it works equally well for the e-field as on the h-field amp. With alternate channel mapping enabled, the "Auto Adjust to Noise" must adjust the common threshold for channels A and B together, rather than separately with the feature off. This results in more frequent gain changes to a degree, yet the firmware still does a great job of adjusting for transient noise, normally keeping all five channels at 100% of the set gains.

I found my signal rates and strike ratios have increased significantly. Often either the E-Field or H-field alone will trigger on a strike. When that happens, the signals from the other amp get sent as well if they meet the 10% of threshold criteria. Often, I see all five channels trigger together on a particularly strong strike.

I tried running with just the e-field. The range is quite comparable to that of the H-field. I believe I will need to tweak the gains lower on the e-field once it is properly mounted outdoors in a permanent enclosure.

Regards,

Don
WD9DMP

JonathanW · « **Reply #46 on:** June 16, 2014, 10:04:59 AM »

Don,
Nice explanation of the alternate channel mapping advantages (I agree), and a great writeup on the troubleshooting you did with your router. Fragmented packet rejection is a rather common feature designed to prevent end-runs around packet inspection (among other things), and for most people probably isn't a critical thing. However, in this case it clearly causes problems.

Cutty Sark Sailor · « **Reply #47 on:** June 16, 2014, 11:46:11 AM »

6/16/2014 added some stuff to initial post.

W7DRM (old call was W3DRM) · « **Reply #48 on:** June 16, 2014, 03:55:30 PM »

Quote from: Cutty Sark Sailor on May 25, 2014, 09:57:35 AM

Quote from: dfroula on May 25, 2014, 09:36:45 AM
---snip...
Don's referring to the antenna type settings on your member information page at login on Blitzortung.org:

If E Field is on Amp 2, you'll skip 'input 2' since the C channel on H field has no antenna... yeah, the numbering begins at "0"

I'm going to suggest that the above referenced antenna setup info for the BO database should also be included in your first post as one of the steps in getting an E-Field kit up and running. That will eliminate someone having to search around for those details. I had remembered seeing this "somewhere in the forum" but couldn't remember where. It took me a few minutes of searching to find it. One additional comment about the settings - the dimension values are in millimeters (mm) correct?

EDIT:

Just found this info is, in fact in the first post... Still suggesting that the dimension values be noted as being in mm's.

Cutty Sark Sailor · « **Reply #49 on:** June 16, 2014, 04:52:19 PM »

Quote from: W3DRM on June 16, 2014, 03:55:30 PM

Quote from: Cutty Sark Sailor on May 25, 2014, 09:57:35 AM
Quote from: dfroula on May 25, 2014, 09:36:45 AM
---snip...
Don's referring to the antenna type settings on your member information page at login on Blitzortung.org:

If E Field is on Amp 2, you'll skip 'input 2' since the C channel on H field has no antenna... yeah, the numbering begins at "0"

I'm going to suggest that the above referenced antenna setup info for the BO database should also be included in your first post as one of the steps in getting an E-Field kit up and running. That will eliminate someone having to search around for those details. I had remembered seeing this "somewhere in the forum" but couldn't remember where. It took me a few minutes of searching to find it. One additional comment about the settings - the dimension values are in millimeters (mm) correct?

EDIT: Just found this info is, in fact in the first post... Still suggesting that the dimension values be noted as being in mm's.

Yes, the pop up help on the page gives instructions...

News:

Author Topic: Adding E Field Probe - Startup and Operation - not for construction (Read 56498 times)