Anyone doing hydrogen line observing?

[Carl Reade]

4 hours ago, Coto said:

So I am currently trying to find an engineer to craft the feed for me, so everything should be ready soon.

Question: Should the antenna probe N-Type connector plate (GND?) be in contact with the metallic structure of the feed, or should I isolate it with some tape or something? Does it matter at all?

Another question is, how conductive must the feed material be?

Yes it absolutely needs grounded to the structure is the short answer. Regarding material doesn't really matter. Weathering is probably more important.

[Coto]

Didn't know that, thanks. Is it OK that the feed is also in contact with the dish? A threaded rod (1-meter long screw) holds the feed into place which is held by the dish using 4 nuts in total (2 for the feed, 2 for the dish) end. Is that OK, or should I somehow isolate the feed-screw area with another material where electrical contact is made?

[Carl Reade]

52 minutes ago, Coto said:

Didn't know that, thanks. Is it OK that the feed is also in contact with the dish? A threaded rod (1-meter long screw) holds the feed into place which is held by the dish using 4 nuts in total (2 for the feed, 2 for the dish) end. Is that OK, or should I somehow isolate the feed-screw area with another material where electrical contact is made?

Feed should be fine with contact. I've seen many feed horns attached this way.

[Coto]

How come the antenna probe connector's ground plate must be in contact with the feed, but it doesn't matter if the dish is also in contact? Antenna grounding is one of thing I really don't understand - why must the feed be in contact at all?

[Carl Reade]

On 01/03/2019 at 00:16, Coto said:

How come the antenna probe connector's ground plate must be in contact with the feed, but it doesn't matter if the dish is also in contact? Antenna grounding is one of thing I really don't understand - why must the feed be in contact at all?

Probably need to look at wave guides and transmission lines.

The feed is a wave guide or think of it as a filter and you want that filtered frequency maintained within the coax.

The fed doesn't need to be in contact with the dish as most satellite LNBs are held in plastic to the dish.

The attaching of the feed horns and materials is probably more a strength and convenience issue.

Again it's an antenna theory subject which others may know better.

Here's a useful gif.

 

[Coto]

Alright, so the engineers built the feed for me:

Might not be a factory-level construction, but I'm very happy for just €35 (he told me the next one would cost €60 because it came out harder than he thought haha).

 

Question: Is it ok if I give it a good white paint (after a layer of primer paint)? I don't think I'll have any problem with reflectivity or something?

[Carl Reade]

Looks pretty good paint should be fine.

[Coto]

Carl, what's the specifications of your bias tee? Adam's Bandpass Filter [http://adsbfilter.blogspot.com/2015/06/hydrogen-line-1420-mhz-filter.html] (which I'm planning to use as a bias tee) seems to only be able to pass 150mA through the cable, yet, the LNA + in-line amps take up about 500mA. Am I misinterpreting Adam's words?

Quote

Hello,

Yes, this device also be used as a bias tee to supply other electronic components on the chain from the LNA to the SDR. You can us it with max 20V DC and 150mA current. You have a filter and Bias-T on the same board.

Adam

Also, does your RTL-SDR get oversaturated by the high gain the LNA + 2x in-line Amplifiers output? Right now I'm trying to find the right cables to connect everything together. I would use a shielded 5-meter SMA to SMA extension cable but I've only got an unshielded SMA to SMA cable (I didn't know unshielded cables act as an antenna!). So here's my solution to avoid ordering a shielded SMA to SMA cable (which I've only found on ebay) and hold the build for another month or so:

Quote

Feed -> LNA -> SMA Male to SMA Female (10cm, shielded) -> SMA Male to SMA Male adapter -> SMA Female to F-Type Male adapter -> In-line Amp1* -> F-Type Male to F-Type Male (5m) -> In-line Amp2 -> F-Type Male to SMA Female -> SMA Male to SMA Male -> Bandpass filter+Bias Tee -> RTL-SDR

Is it a good idea to have the second In-line Amp after the long 5m cable? I'm mainly using it due to connectivity capabilities, cause I can't think of any other way of connecting the antenna to the SDR which is far away.

*Are your in-line amps also F-Type?

 

Thanks,

Coto

[Carl Reade]

1 hour ago, Coto said:

Carl, what's the specifications of your bias tee? Adam's Bandpass Filter [http://adsbfilter.blogspot.com/2015/06/hydrogen-line-1420-mhz-filter.html] (which I'm planning to use as a bias tee) seems to only be able to pass 150mA through the cable, yet, the LNA + in-line amps take up about 500mA. Am I misinterpreting Adam's words?

Also, does your RTL-SDR get oversaturated by the high gain the LNA + 2x in-line Amplifiers output? Right now I'm trying to find the right cables to connect everything together. I would use a shielded 5-meter SMA to SMA extension cable but I've only got an unshielded SMA to SMA cable (I didn't know unshielded cables act as an antenna!). So here's my solution to avoid ordering a shielded SMA to SMA cable (which I've only found on ebay) and hold the build for another month or so:

Is it a good idea to have the second In-line Amp after the long 5m cable? I'm mainly using it due to connectivity capabilities, cause I can't think of any other way of connecting the antenna to the SDR which is far away.

*Are your in-line amps also F-Type?

 

Thanks,

Coto

Hi forget Adams design as it's based on 5 Volts. Your LNA and line amps require 12 Volts.

The bias tee is rated at 65 V at 1 Amp so can run all.

All coax cables are shielded or they are an antenna including SMA.

So you use satellite coax from the LNA through the the line amps the whole way back to the bias tee. Then from the bias tee to the filter then the dongle you can use SMA 5 cm tails. That way the filter is not at the power side of the bias tee.

[Coto]

It's based on 5V? Why did he say max 20V? And what do you think I should do? The only other bias-tee I have is this, but I'm not sure if it'll do or not: https://www.ebay.com/itm/RF-Divider-Biaser-Bias-Tee-10MHz-6GHz-for-HAM-radio-RTL-SDR-LNA-Amplifier-50VDC/113170701206?epid=21021845279&hash=item1a597fdb96:g:9Q0AAOSwvmxbWqtq

Quote

So you use satellite coax from the LNA through the the line amps the whole way back to the bias tee.

Is the electronics chain I mentioned correct? Someone suggested the coax is probably 75 Ohms and it'll be a mismatch with the SMA cables.

[Carl Reade]

1 hour ago, Coto said:

It's based on 5V? Why did he say max 20V? And what do you think I should do? The only other bias-tee I have is this, but I'm not sure if it'll do or not: https://www.ebay.com/itm/RF-Divider-Biaser-Bias-Tee-10MHz-6GHz-for-HAM-radio-RTL-SDR-LNA-Amplifier-50VDC/113170701206?epid=21021845279&hash=item1a597fdb96:g:9Q0AAOSwvmxbWqtq

Is the electronics chain I mentioned correct? Someone suggested the coax is probably 75 Ohms and it'll be a mismatch with the SMA cables.

I can't see where 20v mentioned? Adam based his system on LNA4ALL which are fed with 5v (or can take 9v)

Your bias tee doesn't mention what current it can supply so cant answer. You need one that can supply the current for the LNA and line amp.

Yes sat coax is 75 Ohm but the lowest loss at these frequencies and with the gain involved the mismatch is negligible.

 

[Coto]

 

Quote

 

I can't see where 20v mentioned? Adam based his system on LNA4ALL which are fed with 5v (or can take 9v)

 

Adam: "You can us it with max 20V DC and 150mA current."

[Coto]

EDIT: I had an issue/question but I figured it out.

Edited March 11, 2019 by Coto

[Coto]

@Carl Reade 2 questions:

1) Is my single 30dB LNA sufficient, or must I necessarily get at least 1 in-line amplifier? Have you had any success without an in-line amplifier?

2) I am now about to paint the feed. Must I not paint the connector region at all, or can I also only have a tiny space unpainted where the 4 nuts are from the inside? In other words, must the entire connector plate make contact with the waveguide, or can only 4 nuts make contact (which will be in contact with the ground plate of the connector)? Is there any issue regarding "how much contact is enough"?

[Carl Reade]

10 hours ago, Coto said:

@Carl Reade 2 questions:

1) Is my single 30dB LNA sufficient, or must I necessarily get at least 1 in-line amplifier? Have you had any success without an in-line amplifier?

2) I am now about to paint the feed. Must I not paint the connector region at all, or can I also only have a tiny space unpainted where the 4 nuts are from the inside? In other words, must the entire connector plate make contact with the waveguide, or can only 4 nuts make contact (which will be in contact with the ground plate of the connector)? Is there any issue regarding "how much contact is enough"?

Hi you should get a result without a line amp but I would use one to compensate for mismatch, cable loss and connectors etc.

Ref the painting there is no reason you can't paint over the connector plate and nuts as long as they are making an electric connection to the tube.

[Coto]

So, this is the feed designed that I used:

 

And a friend with a 3.2m dish (f/D = 0.3) found this:

The question is, which one is good for which purpose? I don't know how to use CST Microwave Studio to try some antenna experiments yet, but I'd appreciate your opinion if you have any idea.

Also, I got an old feed without a choke ring that I had gotten long before, and since it was just sitting there, I decided to plug in the connector, use the LNA and 1420 MHz Bandpass filter to see if I was going to detect anything pointing it towards the Milky Way. I didn't think I had a high chance, but what do you know... I think I actually did it. There seems to be a slight 400 KHz Doppler shift on the FFT spike (which is probably normal for these types of observations), and I think I only get the spike when I point it towards the milky way (as much as I could guess at least). I've attached a few screenshots along with an image of the few things I used.

This one's averaged (sorta):

I also checked with a ham radio receiver at that frequency but got no signal, so I don't feel like this is RFI, and it can't be system temperature noise because it isn't continuous... Does the introduction of the LNA really make such a small antenna so sensitive?

[Carl Reade]

17 hours ago, Coto said:

So, this is the feed designed that I used:

 

And a friend with a 3.2m dish (f/D = 0.3) found this:

The question is, which one is good for which purpose? I don't know how to use CST Microwave Studio to try some antenna experiments yet, but I'd appreciate your opinion if you have any idea.

Also, I got an old feed without a choke ring that I had gotten long before, and since it was just sitting there, I decided to plug in the connector, use the LNA and 1420 MHz Bandpass filter to see if I was going to detect anything pointing it towards the Milky Way. I didn't think I had a high chance, but what do you know... I think I actually did it. There seems to be a slight 400 KHz Doppler shift on the FFT spike (which is probably normal for these types of observations), and I think I only get the spike when I point it towards the milky way (as much as I could guess at least). I've attached a few screenshots along with an image of the few things I used.

This one's averaged (sorta):

I also checked with a ham radio receiver at that frequency but got no signal, so I don't feel like this is RFI, and it can't be system temperature noise because it isn't continuous... Does the introduction of the LNA really make such a small antenna so sensitive?

Hi it looks like a very narrowband signal. You should expect the H line to be around 1420.4Ghz and more of a "hump" in the spectrum.

I have seen this method work but to bring out the H line but it needed the data processed over a period of time using the CFRAD program and then graphed in Excel.

I could not get the program your using to work right on my PC for some reason.

I would think both designs of feed horn will work fine.

You should do a sun test and direct it at the sun to see if the noise rises.

Edited March 24, 2019 by Carl Reade
Adding

[Coto]

I'm not sure if it's RFI, because it only appeared when I point it around the Milky Way belt. I assume I don't see other peaks because the other spires' HI-emission flux densities are too low for this small antenna. The difference between the post you have in mind is that he used 3 in-line amps (noisy!) and I used a single LNA with Gain = 30 dB and NF = only 0.5 dB. I also didn't get any emission from my ham radio receiver at that frequency, so I don't know.

I'll try to run 2 more observations to see if the signal appears when the Milky Way is out of visibility and when it's within visibility. This should help us conclude something.

I am using GNU Radio on Linux (in VM). The .grc project is some project I found online, but it's actually pretty easy to design the flowgraph that gives you a live FFT and Waterfall plot. What I'm currently doing is creating another flowgraph in GNU Radio that allows you to properly average etc., and essentially make advantage of the radiometer equation (higher integration time = higher SNR).

EDIT: The sun just set, so I'll try that tomorrow.

Edited March 24, 2019 by Coto

[Coto]

Turns out you were right, Carl. After trying to demodulate the signal with GQRX, I got some pulsating sound. After further investigations, it turned out it was my dad's ham radio interfering (he wasn't transmitting, but the radio's electronics must've been producing that odd pulsating noise). Disappears when he turns the radio off... Interesting!

[Carl Reade]

21 minutes ago, Coto said:

Turns out you were right, Carl. After trying to demodulate the signal with GQRX, I got some pulsating sound. After further investigations, it turned out it was my dad's ham radio interfering (he wasn't transmitting, but the radio's electronics must've been producing that odd pulsating noise). Disappears when he turns the radio off... Interesting!

Ah good you found the cause. You seem to have an interference free spectrum then which is welcome.

[Coto]

So I tried some (not-so-proper) averaging to get a higher (sorta) SNR with a different wide-beam feed (that'll be used on my dish when it's ready). There's this small peak very close to f_center visible on every plot but I'm not sure what it is. Too narrow to be HI?

With HI filter (1):

 

With HI filter (2):

 

Without filter:

 

What do you think? All of them are with LNA.

[Carl Reade]

34 minutes ago, Coto said:

So I tried some (not-so-proper) averaging to get a higher (sorta) SNR with a different wide-beam feed (that'll be used on my dish when it's ready). There's this small peak very close to f_center visible on every plot but I'm not sure what it is. Too narrow to be HI?

With HI filter (1):

 

With HI filter (2):

 

Without filter:

 

What do you think? All of them are with LNA.

Now that is the bandwidth you should be using.

Yes far too narrow.

I think the spike is the DC spike inherent in SDRs. As I use SDR# program it can be removed using the "correct IQ" option. Not sure if GNU has that somewhere.

[Coto]

Seems like your suggestion is right again Carl. Tested with a simple wire antenna and HI filter inside the house (not at all exposed to the Milky Way) and I still seem to get that spike:

So it's probably the DC Spike you suggested. You can definitely do something in GNU Radio to get rid of it (I'll take a look how to do it soon), but is there any scenario where enabling IQ correction is not beneficial? Is there any reason why it wouldn't be automatically enabled in every software like SDR# etc., or does enabling IQ correction "sacrifice" something in order to get rid of that DC spike?

[Carl Reade]

1 hour ago, Coto said:

Seems like your suggestion is right again Carl. Tested with a simple wire antenna and HI filter inside the house (not at all exposed to the Milky Way) and I still seem to get that spike:

So it's probably the DC Spike you suggested. You can definitely do something in GNU Radio to get rid of it (I'll take a look how to do it soon), but is there any scenario where enabling IQ correction is not beneficial? Is there any reason why it wouldn't be automatically enabled in every software like SDR# etc., or does enabling IQ correction "sacrifice" something in order to get rid of that DC spike?

Think it's all down to mass produced cheap dongles and it was added to get rid of the spike don't think it has any other effects. The more expensive high end SDRs wouldn't have it.

[Coto]

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?