Moonbounce on 1296

[VK2OMD]

Your IC910 has specified sensitivity of 0.11µV for 10dB S/N. Lets convert it to NF assuming that the effective noise bandwidth is 2kHz. Using the Receiver sensitivity metric converter, we find that the IC910 NF should be around 4.8dB.

(Click on the image to see full size.)

Thanks to some help from Rex in capturing a recording of noise from one of his '910s, I can tell you that the effective noise bandwidth (ENB) of his receiver was 1970Hz, close enough to our assumed 2kHz to not matter.

Note that the ENB is not the same thing as the -6dB bandwidth, that happens to be 270 to 2600Hz or 2330Hz BW for the particular receiver.

So Dave, unless you measure your own receiver, you could use 1970Hz for its ENB, or 2000Hz which is within 0.1dB. Using the Receiver sensitivity metric converter, an IC-910 with 0.11µV for 10dB S/N and 1970Hz ENB has a Noise Figure of 4.9dB.

Owen

[VK2JDS]

Hi Owen, thanks for the calculation, will go and plug that figure in to the spreadsheet. Do you know if Rex had performed the IF gain increase adjustments , or was this as per out-of-the-box?

This morning I got up at midnight and worked till 2:30 am local dst, unsuccessfully trying for numerous cw stations - who Howard later explained want to see your reply on their return signal freq, so we have to make expected doppler calculations on the fly to figure it out rather than the way JT65C operations work. Anyway, I got through on JT65C to K1JT who was cq'ing on 1296.069, then followed by a call from K1RQG both on the east coast side of USA. After a snooze it was up again at 5:40 am local dst and same again with unsuccessful cw contacts into europe. The morse was a bit more readable looking to the west and I heard VK4AFL, missed the call of a vk3, then heard big signals from OK1CA, F2TU and several DL's. DL4SHF was the strongest. I did manage to get Howard G4CCH England and Viljo ES5PC from Estonia on JT65C , both about the -10 mark so a 2.2 metre dish or larger should copy them ok. I think a early night tonight is in order, but its been fun.
73's

[VK2OMD]

Hi Owen, thanks for the calculation, will go and plug that figure in to the spreadsheet. Do you know if Rex had performed the IF gain increase adjustments , or was this as per out-of-the-box?

Rex recorded the audio noise output, and I analysed the spectral content to determine the Effective Noise Bandwidth (ENB).

ENB is important to understanding how much noise power at a give noise power density will be captured by the receiver. If you put a dummy load at 290K on the receiver input, the power captured from the dummy load (alone) in Watts is K*290*ENB. (K is Boltzmanns constant.)

I worked the noise figure out using the specification sensitivity. Your receiver should be a little better than that, but until you measure the receiver noise figure, or measure its sensitivity, 4.9dB is as good a figure as any.

If a receiver is tested to have a S/N ratio of 10dB with a signal of 0.11µV, we know that the total noise power is 10dB less than the signal power (0.11µV in 50 ohms), but we need to know the bandwidth to calculate the noise contribution from the signal generator (K*290*ENB), then by deduction from the noise implied by the sensitivity measurement, to calculate the receiver internal noise and then the receiver equivalent noise temperature and Noise Figure.

Of course, you could measure the Noise Figure directly using a calibrated noise source and the Y factor method, in fact it is the better method for low NF receivers.

Owen

[VK2OMD]

...Do you know if Rex had performed the IF gain increase adjustments , or was this as per out-of-the-box?
...

I will comment separately on the IF gain 'improvement'.

I have seen a document at http://www.kb2ljj.com/data/icom/Icom%20 ... tivity.pdf that suggests that the IC-910 sensitivity is improved by increasing the IF gain.

I am left wondering why Icom would in a competitive environment, sell a radio that was short of its capability.

But then the article above doesn't actually state how much improvement in Noise Figure or sensitivity was actually achieved by the adjustment.

Since the article does not identify its author, it isn't possible to ask, and its credibility is lacking.

I do know from some tests that I did with Rex, that in the case of his radio, reducing RF gain control hardly reduced the NF until the S meter read somewhere upscale of S7. This suggests that an increase in IF gain is unlikely to yield a significant improvement in NF... but I would take notice of tests on the above mod were they published... show me the numbers.

Of course, increasing IF gain increases S meter deflection, and some people might mistakenly think that is evidence of significantly increased sensitivity.

Receiver sensitivity is a statement of the level of signal required to achieve a given Signal to Noise ratio.

It is about understanding receivers, and that is really useful for weak signal work.

Owen

[VK2OMD]

Another perspective on the effective noise bandwidth figure's use...

Receiver Noise Figure can be defined as the reduction of S/N ratio due to the imperfections of the receiver.

That means that if you have a receiver with a 5dB NF and the input to the receiver has a 20dB S/N, the output will have 15dB S/N, a 5dB degradation in S/N.

So, if you were to measure the sensitivity of your receiver at 0.11µV for 10dB S/N out:

1. what is the S/N of the generator output (receiver input);
2. what then is the receiver Noise Figure?

The generator signal level in dBm is 10*LOG(0.11e-6^2/50)+30dBm = -126.16dBm. The generator noise level in 1970Hz ENB is 10*LOG(1.38e-23*290*1970)+30dBm or -141.03dBm. The generator's S/N ratio is -126.16- -141.03dB = 14.87dB.

Since the receiver input S/N is 14.87dB and the output is 10dB, the receiver Noise Figure is 14.87-10dB = 4.87dB.

Owen

[VK4CDI]

Great stuff Dave!!
Hole dug, concrete in, support pole bolted down, prop pitch refurbished, Linear actuator ordered, mounting hardware welded and painted, phew.......

Hopefully won't be too much longer...
73
Phil VK4CDI

[VK2JDS]

Here's a couple of cut and shrunk captures of the signals off the moon , the first shows the return pings from myself with 3 seconds of transmit (the blank bits with just a white trace) and 3 seconds of receive ( the white trace with the noise either side of it). This is what you will see Phil with 100 watts into the feed. Doppler is far more critical on 1296 than on 70cm as you can see, there is doppler shift occurring in the picture.

The second pic shows 1296.010 and the cw signals that were active there. It was like the cw end of 40 metres during the ARRL EME contest weekend. My 2 pings are top left as I was attempting to net them. There were stations busily CQ'ing from 1296.005 to 1296.029.

these pics are converted to be really small so even dialup users should get them ok.
73's de Dave

[VK2JDS]

Back again with some research findings.

Firstly getting NFM working which is still a work in progress : the laptop i chose to do the measuring generated a lot of noise for such a new machine, without even connecting it the noise level rose significantly just carrying it out to the radio hut while running ( the change was noticable on spectran, you wouldnt hear any difference), also finding a suitable audio isolation transformer for the interface.

Secondly I wanted to explore the possibility of a 0.3dB system noise figure improvement by trying a dual stage vk5eme preamp as suggested by Owen vk1od many posts ago. I know this is a small difference and might not be visible and at the time i did convert a single stage to a dual stage as a test without any sort of proper result. So I got 2 more new kits and began the testing.
With 5 preamps now to experiment with i decided to compare same models using spectran and the dural beacon as a test signal. The identical single stage preamps were the same except one had reverse protection diodes, as were the dual stage preamps, one had reverse protection , the other didnt. They are all in identical diecast cases with plug and play dc sockets to make change-overs really fast if something goes wrong.
The 5th preamp was an identical single stage preamp as the other 2 but with a scrounged front end gasfet out of an old discarded white austar lnb to prove that if you do blow a fet with a t/r switching fault accidentally applying power up it you can get back on air by scavenging parts from the paytv setup. (xyl mightnt approve though...)
So, how did it go. Using the newly fixed Dural beacon from Sydney@178kms over the bluemountains (!) on the 4.6m dish offset a bit to reduce signal the results are:
The noise level out of the 2 stage preamp took some taming with a reduction in the digital gain on spectran to achieve 0dB on its scale - thats the first picture.
Second pic is the single stage preamp which has worked many countries.
Third pic is the scrounged gasfet from the lnb.
Fourth is the IC910H without a preamp at all , just about 15 metres of ldf4-50 to the coax relay straight into the septum.
these are shrunk into jpegs and the noise level was zero'd for each one in spectran with digital gain adjustment.

[VK2JDS]

Had a bit of a malfunction on the upload of the no preamp shot, so lets try again. IC910H barefoot :

notes: the breaks in transmission are when the beacon is identing, also the return pings from the moon were clear, white and audible on all the preamps except when using the IC910 by itself.

[VK2OMD]

Back again with some research findings.

Hi Dave,

Some good experiments in progress from your description.

A few points:

PCs can be noisy... the switched mode power supply is often the worst component, and optional... ie you can switch the AC power off during each measurement interval. I NEVER make measurements with the power supply running. The noise problem has to be solved no matter what software you use.

Elimination of hum/noise loops by use of an isolation box such as Passive DI box or Passive Direct Injection Box is often useful. I NEVER make measurements without such a device.

The interface unit that you need for sucessful WSJT operations is equally good... just make sure that it is working effectively.

I cannot really interpret your graphs. You mention Spectran, but show WSJT graphs.

In my limited exploration / validation of WSJT and related tools I have not developed confidence that it they are a good tool for measurement.

I have also used Spectran, Spectrum Lab, and Spectrogram, spent considerable effort validating them, and have much more confidence in those tools. I have also validated my own tools, FSM and NFM.

Spectrogram v16 is now freeware.

My article Noise Figure inferences of spectrum analysis of weak signals shows examples of use of Spectran and Spectrogram on a weak beacon signal, and how to calculate S/N ratio. The article might give you some ideas, and demonstrates how a conventional spectrum plot is much more useful than a waterfall display for these purposes.

Measuring the S/N ratios achievable from the beacon with the different preamp configurations indicates which is better for that application, but the magnitude of the difference doesn't imply the same magnitude on EME because the noise environment is quite different. For example if you have two receiver systems with Teq say 100K and 500K, and you consider them in two ambient noise environments of a terrestrial shot (~290K) and EME (~60K), on the terrestrial shot the noise ratios (N2/N1) will be (500+290)/(100+290)=2, and on the EME shot (500+60)/(100+60)=3.5. The better receiver makes less difference on the terrestrial shot where its performance is masked somewhat by the high ambient noise.

The consequence of that is the 0.4dB difference with the 2 stage unit that you might expect to see on the moon, will be much less on a terrestrial shot (~0.1dB by my calcs).

Hope this helps.

Owen

[VK2JDS]

Still working on this, got sidetracked with preamp measurements with the tracking gen and spectrum analyser. Interesting to note the vk5eme preamps have a sinewave spectral plot of gain from about 500 Mhz right through to 1350 where it starts to roll off significantly. Changing the length of the input loop moves the dips and humps, while just lifting and lowering it did virtually nothing on the display. Next is a return loss bridge circuit to measure the feed by sweeping it. I have some resistors, box,N plugs and cables - need to make a balanced coupling transformer to connect to the spec-an input.
I think i need a dummies guide to hooking up the soundcard and configuring NFM ; the basics for the audio input part of NFM Owen.
Edit: noise LO and noise HI are push buttons ! just figured it out, thats how to make the measurement. doh

No luck with a transformer so will use it the way i have directly connected it via a 10:1 divider into the old toshiba laptop. (its the quietest). Now i have to get a handle on how to setup the NFM software to read the soundcard input, get an idea of input level and try and make a measurement of millivolts using the line ouput from the IC910.... not sure how to see a level indication in NFM (hopefully this weekend)
The 2 stage preamp is going back up and I will try some moon pings with it again, its still quite noisy even aimed at cold sky and I will have to add some more attenuation on the cable to the input of the laptop to keep the noise level under control.
73's Dave

[VK2OMD]

...
I think i need a dummies guide to hooking up the soundcard and configuring NFM ; the basics for the audio input part of NFM Owen.
...

Dave,

Any issues you have with effective use of NFM due to audio input problems will also be compromising your JT65 performance.

Sorting this should improve your JT65 performance from both the RF angle, and the demodulation angle.

Drop me an email and lets see if we can't solve it.

Owen

[VK2JDS]

7dB sun noise rise with nfm... hmm station not so good. but;

Today was momentous, after all the overseas contacts, today was my first vk contact. Doug vk3um was a massive signal on cw, and we even managed a truly difficult ssb contact.
Wow, and yesterday was Russia on jt65c.
Its going to be a good year fellas, get those dishes and yagis up. Hone and tune the preamps.
73's de Dave

[VK4CDI]

Finally..........
Isolation relay causing a problem, so runnin g direct to preamp (no TX) 13dB sun noise.
Moonset Sat night was like the low end of 20Mx (during sunspot low!!) several CW sigs...then PI9CAM at S6. G4CCH at -12 on JT
Guess RX is FB!!
Sun arvo fixed relay, 12.5dB sun noise, ready to go........
Only problem is PP motor makes a terrible racket at 3am!!!!

Phil VK4CDI

[VK2JDS]

This mornings qso attempt was difficult from this end Phil, I was seeing my own return pings but being through trees is very lossy and would make it very difficult for you to see me.
Today I decided to strip the ldf5-50 out of the dish feed and install a flexible ldf4-50 heliax equivalent for the transmit side.(with a resulting loss in power at the feedpoint)
The septum feed was moving around while moving the dish due to the inflexible nature of the heavy ldf5-50 causing problems at the receive side too. Broken connectors and the new preamp being fitted with sma connectors forced the decision.
Now sma conectors are all good up to 18GHz, but are useless when it comes to any sort of mechanical stress. N to sma adapters just pop open randomly. These sort of connectors are well suited to installation in shack equipment, but are poor in a harsh moving environment out on the end of a dish feed.
again the dural beacon was my friend as i made measurements of what sort of signal i could get from 200 kms away behind a ridgeline (lots of ridgelines actually). i attach some pictures of the various configurations, best was the vlna straight into the septum, 2nd best was the N relay with vlna on a adapter with a flexible heliax to the main cable. the pic of the hardline with small microwave relay was the worst, and the heliax with adapter straight in the output port failed after 5 minutes of rotating the dish back and forth. its due to the rotational force on the heliax even through the clamp. i dont want to compress the cable so its supported in 3 places, but still things break.
in the first picture the heavy black cable from the lower termination on the coax relay goes to a dummy load so the preamp can be easily switched from feed with cold sky to a 290K load.
power measured coming out of the receive port when dumping about 100 watts into the transmit port was about a quarter of a watt so you dont need much of a relay here as long as its extremely low loss.
73's

[VK2JDS]

Heading into winter now and the weather has been very poor over the last 3 weeks or so. Some equipment failures in the hut due to the damp and poor visibility of the moon. Digital elevation meter has been really handy during this weather resulting in contacts into Russia, Bulgaria, England, Netherlands, Japan, Canada, Brazil and California. The local highlights last week have been Doug VK3UM on cw (thanks for putting up with my poor cw rx) and Phil VK4CDI at -30.
Yay Phil, I think with some fine tuning of the dish alignment system there it will improve, but your signal was straight and steady due to that nice new radio.
I have been doing some more sun noise rise measurements and adjustments using NFM. Its a gradually improving station here. Get hold of that program Phil , it will allow you to make direct sun noise rise measurements using the same computer you will be using to decode the signals off the moon. Cant get better than that. It makes it easy by proper measurement what improvement or degradation you are making with each adjustment. I am using it to set the fine adjustment of optimum noise figure by adjusting the input coils of the g4ddk vlna with it all mounted at the feed. :lower the dish on motor drive to stepladder height, open vlna , adjust slightly, close vlna raise up to sun and make a new measurement.
also the pins i am using are about 39mm in the septum and its now achieving 30db return loss on both ports.
Another thing to add is a skirt of about 30cm of bird netting around the outside to reduce spillover.
Keep us posted of your progress Phil
73's de Dave vk2jds

[VK4CDI]

Good stuff Dave, I will check out that software.....
Last weekend, rewelded the El pivot arm, it wasnt mechanically correct. The power supply feeding the 36V linear actuator sags badly,40V down to 22 on load, so need to fix that. I think I have found my low Tx problem! Not enough drive to the SSPA, which is mounted behind the dish. 1W from the TS2000x, through a 20db pad, to give 10mW to the PA. Trouble is a) 20 Metres heliax adds a dB or so, b)TS2000 1W is not exactly one watt.... end result is I am getting 8Watts at the feed. No wonder you were barely copying me.....HP power meter and step attenuator from work should get that spot on!! Az encoder mechanics sorted, need to buy a small Gilmer belt and will have full auto tracking. Hopefully I will be all ready to go for the EOA event.
A couple of piccies on http://www.usq.edu.au/users/moat/photo.htm

Cheers all
Phil VK4CDI

[VK2JDS]

Things are still progressing. Many stations to work on 23cm eme jt65c, every week or so there seems to be a new one or 2, then it goes quiet for a little while as the moons offset 28 day orbit shifts coverage to a different spot.
The most recent interesting station was Felix DP1POL on the German base in Antarctica. He decoded here at -24 quite reliably even though its a 65 element horizontal yagi and 500 watts but he wasnt able to decode me though.
Then Felix had a cable failure as his heliax broke when moving the yagi at -40 degrees C !, now thats cold.

During this coming month there is a moonbounce competition so expect to hear the big boys back on air.
To give you an idea of the signal strengths I worked HB9HAL Christoph at a huge -4 this morning , and he got me at -5. HB9Q is even bigger! and claims he will easily work single yagi stations. Thats a s9 signal on the IC910H - bearing in mind with the G4DDK vlna i have a S5-S6 s meter of noise into cold sky. It was pumping the agc.

It seems nearly every time I turn up in the HB9Q logger on a favourable morning Phil VK4CDI is also there.
His gear is going well and he's working plenty of dx.
We worked a dogpile of sorts last tuesday morning, the Netherlands stations were congratulating themselves in the logger for working us VK's and there were stations spread up the band each 5KHz from 1296.060 to 085. Quite a choice of European stations to choose from on JT65C.
I got a few CW contacts this week as well, so get your speed up a bit as they all like 12+ wpm. (but they tolerate my slower morse hihi).
Matt VK2DAG is gearing up for 13cm and has done some ping tests (which i listened for) so I am hoping he will document his path to eme on 13cm here in the vklogger forum for all to read.

I had a stuck relay and sent 10 watts up the back end of the G4DDK vlna, and also a single stage NE325 backup vlna until i sorted out the problem. they are back on air with new fets and a reset of the bias current and cold sky/290k dummy load measurements over and over tweaking the coils little bit by little bit to get them back honed to exact match for the septum feed. Sun noise rise is now 12.4dB on my 4.6metre on the dual stage and 11.7 on the single stage, 11.3 on the backup-backup single stage(you cant ever have too many preamps ! hihi). If you are in the middle of a great DX contact to some remote part of the planet and a preamp dies , you want to be able to swap onto a good one fast! . Silver plated N bulkhead connectors with vlnas in diecast aluminium boxes are robust and easily swappable and waterproof mounted right at the coax relay on the side of the feed.
all good fun. go and find yourself a unwanted C band dish from someones backyard.

73's from Dave VK2JDS

[VK4GHZ]

When I finally get enough spare time, I want to be able to sit down, concentrate, and read this topic through end to end.
I really want to give 1296 EME a go, and with the IC-910 purchase earlier this year, I feel I'm part the way there.
70W amp to come next.

Thanks Dave for posting, as your enthusiasm for this has been an inspiration, and I reckon 1296 EME will definately relieve the boredom in between 6M F2 DX!

[VK2JDS]

Hi Adam, I hope you do have a go at it. Its something really astounding when you hear your own carrier reflected off the moon, still amazes me.
7:30pm on the first Friday night of November I am giving a talk at the Orange and districts Amateur radio club (ODARC) about how to go about getting into 23cm moonbounce from a practical point of view. Taking the mystery out of it all, explaining the options about the gear you need and the software, how to run it all and the operating procedures etc. Basically everything I have learnt in the last 5 years condensed down to give others a running start into the mode using dishes you dont have to spend 10 years building from scratch.

This weekend was the ARRL 6,2,70 and 23cm moonbounce competition and wow was the 23cm band alive! There were oodles of high powered CW stations to have a tune past and attempt contacts with.
As usual for us Aussies after the Americans and Canadians loose aquisition there is a 'dead zone' when the moon is a bit either side of north, You get some JA's etc but as it progresses west towards Europe the signals appear again and between 1296.005 and 1296.035 there were CW signals every 5KHz or less. Some with quite noticable drift too. There wasnt much JT65 activity though.
This means for us here with long yagis or small diameter dishes good for terrestrial work the chance to hear dx off the moon, or at least see it with spectran or the waterfall display in WSJT6.
Most of the overseas big stations were active, and of the locals VK3UM, VK5MC VK4CDI were in the thick of it. I had several great contacts on saturday night into sunday morning on CW and JT, but had to work monday and missed the morning European window. But it was K5,W6,VA7,OK1,DL0,SP6 stations I worked besides the locals , missed G4 as the moon set though and there were heaps I didnt get to work as there were dogpiles that I just couldnt get onto due to my power limit of 130 watts.
Doug VK3UM puts up with my 8 to 10wpm straight key CW as usual... I am improving, slowly. hihi
This meant I have now passed 50 different 2way eme station contacts.
You can do this too.
73's de Dave vk2jds