WXforum.net

Weather Station Hardware => Blitzortung => Topic started by: miraculon on September 02, 2013, 05:08:33 PM

Title: The Terminator
Post by: miraculon on September 02, 2013, 05:08:33 PM
I noticed that there were some fundamental differences between Red and "Green" in terms of frequency response, loading resistance and input structure. Aside from the obvious remote PGA control, I was wondering if any of these could be applied to the older "green" boards, specifically 6.8USB.

The Red has:


I found that converting to "single ended" input (defeating the differential) was beneficial for reducing the residual noise. The input resistors on the "green" amp (6.8USB) were 1% tolerance, which means 40dB CMRR at best. I suspect that the common-mode noise with all the gain behind it was not able to reject the noise.

This was posted on the USATOA board:
Quote
Taking a tip from the new system RED, I decided to ground-reference the input on the "older" 6.8USB amplifier. (IA3/IB3 jumpers placed)

The 2nd step was to try reducing the input resistance to 2.2K Ohms. I made a jumper resistor that goes in series with one of the 82 Ohm resistors already on the board. The "single ended" jumper, made this an easy add on.

Here is the summary from USATOA:
Quote
Placed 2.1K resistors at IA-1 and IB-1. This puts 2.1K + 82 Ohms across the antenna. (2182 Ohms).
The VU meters on the Linux box dropped a few dB. I increased the gain from 53 to 70, putting the noise floor about back where it was on the VU.

While these two measures improved the noise rejection, the lack of the response peak due to the 2.2K resistance smoothed out the waveform. The "Green" was now performing poorly on the "Participation" chart on Blitzortung.org.

In order to change the frequency response from 17KHz to 34KHz, I removed the F1, F2 and F3 filter jumpers to change the response.

Quote
I decided to give this a try and open all the filter jumpers. I started getting an error in the plots on Lightningmaps.org for each strike. It seemed that I started getting more participated strikes, though. Then I placed just the F3A/B filter jumpers. At first there were some timing errors, but they stopped and I got the normal waveforms/frequency/Lissajous plots. Then the errors started appearing (>8000usec). I restored the filters back to original, and the errors stopped. See attached for the error.

Today, I tried again and got better results. I was not getting the errors on the waveform/frequency/Lissajous strike plots.

I ran a Bode plot of the filter response which revealed that with the 34KHz filter settings, there is a peak near 10KHz that is a couple of dB higher than the 17KHz filter settings. Today, I tried reducing the gain from 101 to 70x which is a reduction in gain from 40dB to 37dB. (I had bumped up the gain as an experiment)

I have been monitoring both the participation table and the waveform plots on LightningMaps.org are good. In fact, instead of being smoothed out, they more closely resemble the RED waveforms.

The final point, is when I check the box on the Overview plot for "668,706", I see many more strikes showing. I believe that this shows strikes that were detected by both stations. Before the filter change, there were very few that would show with this box selected. Now there are many strikes showing. The individual strike data on LightningMaps.org is also showing both of my stations much more frequently as well.

I'll continue to monitor the behavior with these new settings to check the long-term implications.

Greg H.






Title: Re: The Terminator
Post by: Weather Spares on September 06, 2013, 05:19:56 AM
Yes, please do keep us updated. As a green owner I would be interested in knowing whether these changes long term could be implemented to make some improvements.

Also as a note, my Green board shipped without the 1M resistors by accident. After enquiring Egons response was:

Quote
the 1M rsistor is only necessary if you use an electrical antenna.

Title: Re: The Terminator
Post by: miraculon on September 06, 2013, 07:53:08 AM
So far I am pleased with the results. I am still hampered by ambient noise, although these mods have improved the lower level nuisance noises. I have bought a shielded RJ45 coupler and 7 foot STP CAT5 cable to try moving the antenna/amp to the other side of the closet. I need to try that to see if it helps. The opposite side of the closet seems quieter on the AM radio.

I notice that I have moved up the participant list slightly. I thought it was doing much better than before during closer lightning. I'll have to watch the approaching storms again since I made the filter change.

I am also very pleased with the variable threshold mod. I have since added a volt meter so that I know exactly where the threshold is set. I had to build a differential-to-single ended op-amp stage because the panel voltmeters seem to be all single ended. The common ground pin upsets the biases severely without the diff amp.

I posted the details on the variable comparator in another post previously. (prior to adding the meter)

Greg H.