Anyone doing hydrogen line observing?

[Carl Reade]

3 hours ago, Coto said:

Carl, I see that you used the filter at the end of the electronics chain (right before the SDR). I was told by Adam (my filter supplier) to place the filter between two amplifiers.

What is the best location for the filter? I'll be using an LNA (which has one stage of amplification, then a high-pass filter, then another stage of amplification), then I'll have an in-line amp, and then (in a few weeks when it'll arrive) I'll have another in-line amp. Why does my LNA have the filter between the two stages of amplification, and not after them all like you have? Adam is also suggesting I should follow the LNA's "ideology", but I'm really not sure why.

What is the difference between placing a filter between vs after the stages of amplification?

Hi I'm assuming Adam thinks your using wideband amps which will amplify all frequencies hence the need to place a filter between them as they could get saturated with unwanted signals.

The reason I place my filter last is because anything that is picked up in the system chain ie the coax or line amps is filtered out before it gets to the SDR.

Don't forget your front end LNA is already filtered for the H line it's not wideband.

So adding a filter at the SDR just gives you added protection from signals picked up from the line amplifier or coax.

If your not getting interference from nearby out of band signals you may not even need it. There are no set rules.

Don't over think it. Front end filtered LNA amplifies the signal you want, next the line amp compensates for cable loss then filter (if needed) gets rid of any unwanted signals then SDR.

[gabrielxp46]

Besides the dish and bi-square antenna, what equipments did you use in your design like low noise amplifiers, bandwidth filter and etc? I would like more details because I am wanting to build a radio telescope as well.

[Carl Reade]

10 hours ago, gabrielxp46 said:

Besides the dish and bi-square antenna, what equipments did you use in your design like low noise amplifiers, bandwidth filter and etc? I would like more details because I am wanting to build a radio telescope as well.

Hi heres info from a previous post,

Ok best advice i would say is,

Get the biggest dish you can! You will likely have to make the dish feed yourself either the the one on this thread or the can type (measurements online).

Next is a good low noise amplifier (LNA). Around 30db gain and the lowest noise figure you can get. This noise figure will be the benchmark for the system.

You will need probably two 20db satellite line amplifiers these are cheap and work fine.

You will need filters in line as well to keep out interference. You can get these on the LNA4ALL site.

You will need a power supply 12V and a DC inserter to power the amplifiers.

You will need a receiver, a SDR dongle is fine

A computer and software.

Once you have a working scope then you can improve it as you go along.

Carl

For the first stage LNA I used one from www.g8fek.com. Then a satellite line amplifier and a filter I had already. LNA4ALL do H line filters. Requirement for filters depends on your RF environment.

[Coto]

2 hours ago, Carl Reade said:

You will need probably two 20db satellite line amplifiers these are cheap and work fine.

Just a comment: I decided not to use Satellite in-line amplifiers because of 1) impedance mismatch (50 Ohm vs 75 Ohm) (and you'll also need "special" connectors), and 2) reliability.

Instead, I bought one of these https://www.ebay.com/itm/202594426764 (I'd buy 2 but I already have another cheap LNA from ebay). These are SMA (50 Ohm, no impedance mismatch), and are almost definitely injecting less noise (have a lower noise figure) than traditional satellite in-line amps. They also tend to provide a higher gain than satellite amps (30* dB vs 20 dB max).

*I acknowledge that their noise figure and gain values aren't as reliable on eBay, so maybe not 30 dB, but surely more than 20 dB.

[Coto]

I also tried to use a friend's 3.2m dish who's planning on turning it into a radio telescope. He hasn't got any amplifier, filter etc. yet, just the 1420 MHz feed improperly installed and a 6m cable going to his SDR in the control room. I decided I had nothing to lose, so I wrote a quick script to take an (averaged) sample every 15 minutes for ~12 hours. Here is a gif (as expected, no detection without an amplifier but a 6m cable straight to the SDR):

[gabrielxp46]

I am thinking of buying a lna similar to this:

https://www.ebay.com/itm/Hydrogen-Line-Pre-filtered-Low-Noise-Amplifier-1420-MHz-LNA-32-dB-Gain-LNA/283206609273?hash=item41f06de979:g:gDoAAOSwkcBa83S1

I would like to know what software I can use to make observations besides the radio sky pipe and radio eyes

 

[Carl Reade]

6 hours ago, Coto said:

I also tried to use a friend's 3.2m dish who's planning on turning it into a radio telescope. He hasn't got any amplifier, filter etc. yet, just the 1420 MHz feed improperly installed and a 6m cable going to his SDR in the control room. I decided I had nothing to lose, so I wrote a quick script to take an (averaged) sample every 15 minutes for ~12 hours. Here is a gif (as expected, no detection without an amplifier but a 6m cable straight to the SDR):

Its likely to be in there you would just need the data extracted from it. If you were recording with CFRAD program you could extract it.

[Carl Reade]

18 hours ago, gabrielxp46 said:

 

I am thinking of buying a lna similar to this:

https://www.ebay.com/itm/Hydrogen-Line-Pre-filtered-Low-Noise-Amplifier-1420-MHz-LNA-32-dB-Gain-LNA/283206609273?hash=item41f06de979:g:gDoAAOSwkcBa83S1

I would like to know what software I can use to make observations besides the radio sky pipe and radio eyes

Hi radio skyline probably not much use unless your measuring total power.

Have a look at this site this is where I learned a lot and what I use CFRAD can pull out the H line with the smallest of signals and can be graphed in all sorts of ways.

Carl

 

[gabrielxp46]

Which site do I get more information about CFRAD?

[gabrielxp46]

You spoke of a site where you learned a lot, but I did not see any links, could you tell me?

[Carl Reade]

50 minutes ago, gabrielxp46 said:

You spoke of a site where you learned a lot, but I did not see any links, could you tell me?

Hi sorry link didn't post,

http://parac.eu/projectmk4.htm

[gabrielxp46]

Thank you  very much

[Coto]

Also Carl, any idea what that huge peak around 1421.2 MHz is? RFI at another frequency (harmonics)? Could it vanish with a filter or something?

[Carl Reade]

14 minutes ago, Coto said:

Also Carl, any idea what that huge peak around 1421.2 MHz is? RFI at another frequency (harmonics)? Could it vanish with a filter or something?

I get that at different points sometimes. I would suspect it's either PC or SDR generated RFI. It's recommended the SDR is a good 5 meters from the PC. PCs put out all sorts of RFI.

[Coto]

12 hours ago, Carl Reade said:

I get that at different points sometimes. I would suspect it's either PC or SDR generated RFI. It's recommended the SDR is a good 5 meters from the PC. PCs put out all sorts of RFI.

Using a 5m USB extension sounds like a pretty bad idea in terms of packet loss - it's not even recommended for typical file transfer if I recall correctly... Wouldn't it be better to isolate the SDR from RFI using some sort of Faraday cage or something?

[Carl Reade]

3 hours ago, Coto said:

Using a 5m USB extension sounds like a pretty bad idea in terms of packet loss - it's not even recommended for typical file transfer if I recall correctly... Wouldn't it be better to isolate the SDR from RFI using some sort of Faraday cage or something?

Never had any issues with 5m and used 10m powered for astro cameras all worked fine. Ferrite rings also good to add. Yes metal boxes where you can.

[Coto]

Carl, I got my friend to put a faraday cage over his SDR:

But the RF spike still appears in the FFT. Does the cage need to make electrical contact with anything? Right now it is just "floating" there without making contact with anything.

[Carl Reade]

3 hours ago, Coto said:

Carl, I got my friend to put a faraday cage over his SDR:

But the RF spike still appears in the FFT. Does the cage need to make electrical contact with anything? Right now it is just "floating" there without making contact with anything.

You will probably need to do a bit of detective work to see what it is. With antenna, without antenna etc. But I suspect it's probably the PC causing it so maybe hard to get rid of. Try ferrite rings as well on the USB lead.

[Coto]

Hey Carl, I set the dish up and tried an observation.

 

This is my block diagram (electronics chain):

Single linear polarization feed -> LNA -> LNA2 (in-line amp) -> 3m LMR-400 low-loss cable -> Bias Tee -> 1420 MHz Bandpass Filter -> RTL-SDR V3 -> 60cm USB extension cable -> Computer

 

I observed for about 6 or 7 hours, with a not-so-proper-but-still-OK averaging (it takes about 5 minutes to go from the "noisy" spectrum curve to the averaged smooth spectrum curve, so you can kind of say that it averages over ~5 minutes).

 

I uploaded the Observation on YouTube (I wrote a quick script that takes screenshots every minute for 6 hours and sticked the images together). I also recorded (unaveraged) Waterfall plots, in case I got a very high SNR and things were also visible in the Waterfall plot without averaging.

 

Any idea why there is no (apparent) detection? I'll admit I was quite a bit optimistic this time with a 1.5m dish pointing to zenith, but I don't think I got anything...

 

I have three suspicions:

1) The feed's monopole is misaligned - I set it perpendicular (90deg) to the grid bars, expecting ~100% efficiency (since the bars will act as a polarization filter, I turned it so I got the most out of what's reflected). If I've done this wrong I'm probably receiving nearly 0% of the incident waves.

2) The software can't extract the emission from noise without averaging for longer (not sure how long you've averaged for to get your results but I didn't expect I'd need over 5 minutes of integration time for a reasonably high-SNR detection). Perhaps I'll need to work with different software or write another GNU Radio flowgraph.

3) The primary LNA is not getting any power (I don't think that's the case cause my in-line is having a LED lit up, so it's definitely getting power - the question is if it forwards the DC voltage to the primary LNA, but I don't think that's the problem - wish all LNAs had an LED on them though, it'd be super helpful for knowing if they get power or not).

 

Here is the YouTube video of the Observation:

 

A few pictures of the setup:

 

Edited April 20, 2019 by Coto

[Carl Reade]

Hi first things first, you have just placed a large metal pole straight through your feed horn you cannot do that. The horn needs held to the dish from the outside. You will need to fill the hole back in.

I would remove the line LNA  or check the main LNA is getting 12v with a volt meter. I don't know if the line LNA allows power pass?

I can see the H line live without much averaging.

Try using SDR sharp I never got anything out of GNU radio. Peak hold should give you an indication of the readout has changed.

Edited April 20, 2019 by Carl Reade

[Coto]

1 minute ago, Carl Reade said:

Hi first things first, you have just placed a large metal pole straight through your feed horn you cannot do that. The horn needs held to the dish from the outside. You will need to fill the hole back in.

I would remove the line LNA  or check the main LNA is getting 12v with a volt meter. I don't know if the line LNA allows power pass?

 I can see the H line live without much averaging.

Try using SDR sharp I never got anything out of GNU radio.

Carl, I got the antenna with the metal pole tested with some professional equipment, and it carries 90% efficiency (loss = 10% reflected back). I also asked the guy who helped me carry the experiment (who is kind of an RF professional), he told me that I wouldn't have a problem with the pole - the waves NOT going into the feedhorn would be from near the center of the dish, which is covered by the large feedhorn's diameter anyway, so I don't see any problem. To answer the question of "what happens once the waves hit the feedhorn at the end, do they get reflected off the pole or what?", the experiment showed that only 10% would be reflected off (due to the pole(?)), so I should be receiving 90% of the incident waves (since the same rules apply for both RX and TX).

How could I check if the LNA is getting 12V? I don't know if it'd be easier with a voltmeter but I was thinking of plugging it and unplugging it (with the in-line) and see if the noise level gets affected. If it doesn't then I know it doesn't get any power which could be the problem.

[Carl Reade]

Trust me you will never see a feed horn with a metal pole through it. The can is mathematically sized along with the rod for the desired frequency putting a rod through it makes its whole dynamic has changed and effectively no use. The efficiency of the dish is irrelevant and usually worked out at 60% 70% efficiency.

Measure the input to the line amp with a volt meter and see if there is voltage.

[Coto]

When I say "efficiency" I don't mean dish/entire system efficiency, I only refer to the feedhorn + pole that we tested in the lab. We tested the feed with the pole, and we looked at the S-Parameter plot. At 1420 MHz the return (due to impedance mismatch) was -10 dB (converted to linear, we get 10% loss). This is an acceptable loss for my case.

 

I did check what you said, not with a voltmeter, but using an SDR and switching between amplifiers. Indeed the LNA with the LED doesn't pass the power to the primary low-loss LNA, and this is definitely the strongest evidence as to why I got no signal. I will try as soon as I can to retry another observation (that is, next time the milky way is overhead at zenith, if weather permits).

 

Do you know how I can make it so the in-line amp passes the DC power to the primary LNA?

Edited April 20, 2019 by Coto

[Carl Reade]

I still think your wave guide is no longer a wave guide.

If the main LNA isn't powered there will be nothing. At the moment for testing I would power the LNA separate using its power lug at the side careful of the positive, negative.

I would remove the line amp for the moment just use the bias tee direct to the main LNA.

[Coto]

6 minutes ago, Carl Reade said:

 I would remove the line amp for the moment just use the bias tee direct to the main LNA.

That's what I'm going to try. Was just wondering if you had any ideas (i.e. soldering some wires to the LNA boards or just modding them somehow).