LNB unit-to-unit CNR variations

[pendragon]

I ran into some rather odd performance problems on one of my Ku dishes recently and decided to measure the CNR differences for several LNB and feed combinations. I thought this might be of some passing morbid interest to folks here as it shows rather significant unit-to-unit variations on a LNB that is generally well regarded.

Until recently my top performing Ku-band dish was generally a GeosatPRO 1.2m offset with an Invacom QPH-031 LNBF (quad linear/cicular). I also had a Fortec 1.8m prime-focus with a Frankenstein dual-ortho feed where I changed out the Ku section for a Invacom QTF-031 LNB (quad universal) on a custom adapter. Ideally a 1.8m dish should deliver about a 3.5 dB CNR improvement over a 1.2m, but I never saw this in practice. Long ago I did some quick-and-dirty comparisons with a prime-focus Ku feed (Invacom's ADF-120) on the QTF-031 and concluded there were a couple of reasons for the less than ideal gain improvement: (1) the Fortec is a rather flimsy dish and is not a great fit to a perfect paraboloid and (2) the Ku beam patterns for dual orthos are rather sloppy, probably because the feed provides no effective scalar at these frequencies.

Last spring I put a 1.8m Prodelin offset on a Ajak HH180 motor for exclusive Ku reception. I picked up a second QTF-031 LNB and machined an adapter for the rather impressive feed off the BUC/LNB package that came with the dish. I later added a second sidecar LNBF rotated to match 103W's oddball skew. I've been very happy with the performance of this dish, particularly with very weak or out-of-footprint feeds. However recently I noticed the Prodelin was only beating my 1.2m by about 1.5 - 2.0 dB at the center of the vertical band. Time for some fun on the roof!

I decided to take advantage of Dish's transponders on 85W as they are easy to lock and measure CNRs for both polarizations. I just wish they would cover the whole Ku band I peaked both the Prodelin and 1.2m dishes on that bird and started rattling off CNRs. On first blush both polarizations averaged around a 3.25 dB CNR improvement for the Prodelin. The horizontal difference was fairly flat across the transponders, but the vertical difference was all over the map, ranging from about 1.5 to 4.5 dB of improvement. The worst performance was mid-band, exactly where I suspected.

I rotated the Prodelin feed by 90 degrees to see whether this was a dish artifact; it wasn't as the differences flipped. I then swapped the cables on the LNB - there was no change. That pretty much limited the suspects to the LNB or feed. I needed another LNB with a C120 flange and pulled out a spare Superdish FSS bandstacked device. As I have reported in previous threads, this performs very well for the ridiculously low prices they sell for on eBay. I swapped out LNBs on the Prodelin and ran another comparison against the 1.2m. This was startlingly better on the vertical side, averaging about a 4.0 dB CNR improvement for the Prodelin and having a much flatter profile.

It was tempting to leave the bandstacked LNB in place because of its outstanding performance. However I occasionally find some feeds on Atlantic birds well below 11.7 GHz and I really wanted to keep a universal LNB on the Prodelin. So off came the universal LNB on the Fortec and onto the Prodelin. Again I compared against the 1.2m. Overall this LNB measured just a hair worse than the bandstacked, but much better than the universal LNB that I had replaced. The first graph attachment shows the CNR variations between the two dishes as a function of frequency and by LNB on the 1.8m. Invacom 1 was the LNB originally from the Fortec and Invacom 2 is the crummy unit I had on the Prodelin.

The 1.2m and its LNB was the control, and its frequency response is superimposed on the 1.8m responses. One can probably make a pretty good estimate of the 1.2m response by averaging the curves for the different 1.8m LNBs. One can also readily see the droop in the Invacom 2 vertical response at mid-band. To factor out the 1.2m response, I then differenced CNRs for the Invacom 1 and Bandstacked LNBs against Invacom 2 for the second chart. The improvements in CNR at mid-band are very apparent here.

Invacom provides individual test curves for each LNB and of course they look just fine. Maybe I'll examine them a little closer some other time as Invacom only provides gain vs. frequency and NF vs. frequency. Perhaps when translated to CNRs this will explain the variations, but I rather doubt it. As a final twist I was curious as to how well the giant Ku feed that came with the Prodelin actually performs. I did not make a careful comparison, but I wedged a fairly decent Ku LNBF in its place on the Prodelin and again compared the performance against the 1.2m. Pretty much across the band the LNBF showed a 0.5 dB loss in CNR, confirming the original feed is probably the way to go.

I'm kind of happy this has been sorted out, but I'm not that happy to see such a variation in CNR. I don't really care much about LNB gain variation across the band, but CNR is another matter. A 1.5 dB unit-to-unit CNR variation means the 1.8m dish was acting like a 1.5m dish mid-band for the vertical polarization. While this has been fixed, I am curious about how my other LNBs perform.

 

[brentb636]

Very interesting, Pendragon. I always enjoy your adventures .

 

[Corrado]

Thanks for taking the time to document the results Pendragon.

I'm particularly interested because I use the 1.8 Prodelin and Fortec dishes daily. Any chance we could see some photos of the LNBF setups also?

 

[phlatwound]

Very interesting, Pendragon. I always enjoy your adventures .

Ditto here!

Although I have to admit, at my level, reading pendragon's project threads are kind of like me reading Stephen Hawkings' A Brief History of Time, I can hang in for the first chapter or 3, but after that...........

 

[brentb636]

Ditto here!

Although I have to admit, at my level, reading pendragon's project threads are kind of like me reading Stephen Hawkings' A Brief History of Time, I can hang in for the first chapter or 3, but after that...........

I didn't claim to understand everything, but at least I have a technical background, and can guess a little !

 

[pendragon]

I'm particularly interested because I use the 1.8 Prodelin and Fortec dishes daily. Any chance we could see some photos of the LNBF setups also?

Thanks for your support. I often write far more than I should, although I hope the essence is there at the start and finish. Photos and a description of my modified dual ortho for the Fortec dish can be found in this thread.

I've attached a photo of the Prodelin complement here. The Invacom quad universal LNB is attached to the giant Ku feed that came with the dish via a Rube Goldberg adapter, as the C120 flange on the LNB is different from the flange on the feed. At least it has a nice angle scale. The LNBF to the side is skewed for 103W. Because the focal point of the dish is well behind the front of the main feed, I had to offset the sidecar LNBF by much more than I would like to prevent it from being shadowed. But it still beats anything else I have on that bird.

 

[brentb636]

Personally, I appreciate the way you write your findings. They are well written, and minute in detail, allowing someone to track your logic and findings. Unfortunately for me, although I probably have the technical background to follow your work, I lack the equipment to dig as deeply as you into the details of this hobby. I admire your curiousity and certainly hope it continues and that you continue sharing the information with us.

 

[pedrogarcia]

Interesting results, thanks for sharing them Pendragon.

About a year ago I was testing the Invacom Ku universal range and the most interesting feature I found was that the quads and quattros were signifcantly superior toInvacom single or twin output LNBs on both gain and CNR. The QDF 031 quad and quattro were better than the QDH 031 quad and quattro at installation but over time the c120 mount caused problems due to insects. Comparing a batch serial number box of 12 QDF031 there were small cnr differences but I still prefer Invacom for comercial grade LNBs

 

[pendragon]

Pedro: I've had good results with Invacom LNBs, but this was a bit sobering. One expects a certain level of performance given the individual test curves they provide, but looking at these it appears they only run a very few spot frequencies rather than a sweep. Of course bad units will always show up or degrade with time.

I hear you on insects. If I failed to cover my feeds in some manner, I would have the largest wasp concentration in Denver.

 

[Pixl]

Very interesting test. I did get a little lost in the results as to which is the better LNB.

 

[pedrogarcia]

Pedro: I've had good results with Invacom LNBs, but this was a bit sobering. One expects a certain level of performance given the individual test curves they provide, but looking at these it appears they only run a very few spot frequencies rather than a sweep. Of course bad units will always show up or degrade with time.

I hear you on insects. If I failed to cover my feeds in some manner, I would have the largest wasp concentration in Denver.

the strangest thing with this insect nesting, it always blocks the horizontal TPs first and gives you the impression of an LNB or voltage fault.
I'm in Cyprus now and can't use any form of cover to the c120 because even with a 3m dish on Astra 2D's mini footprint I am on fumes and can't take any signal attenuation. It just becomes a maintainance issue. There is no doubt though Invacom quad is the best for this reception.
Regards

 

[pendragon]

I have a Prodelin 1.2m looks allot like yours. I'm still using the original single polarity LNB and feed. I wanted to change to a dual polarity using the original feed horn, but also found the bolt pattern smaller than a C-120. How did you deal with that. Also in your photo the feed looks further back than it is on my dish. Can you describe how you modified the BUC to become the mount in your photo?

I believe your dish is a lot newer than mine. My BUC/LNB package was enormous and probably weighed 30+ pounds. Because of that my feed support arm is made of 2.5" square steel tubing with no side stabilizers. I could probably stand on the feed support and nothing would happen. The feed itself is similar, but the waveguide is quite short and was recess mounted into the electronics package.

This package was in a large rectangular aluminum box with 'heat sink' fins all over the outside. I emptied the box and removed the plumbing. The external feed mounted to an adaptor plate, which itself mounted to the flange on the internal box waveguide.

This internal flange does not match up to C120, but the waveguide diameter was essentially the same. I sawed the internal flange off the waveguide and drilled it to C120 specs with recessed holes for the bolt heads. This flange is bolted to the Invacom LNB and the rest is attached the same as before. Because the electronics package was an integral part of the mount, I had to cut out a square chunk of the front face, which is the ribbed white piece in the photo.

When I aligned the dish, I used the feed diagrams on the Prodelin website to get everything in ballpark positions. I only had to make minor adjustments to hit the peak. The focal point of the dish is essentially at the interface between the scalar cone and the waveguide.

You may be able to cut off the flange on the BUC/LNB side and use this as a coupler to a C120 LNB, somewhat like I did. But you'll have to come up with a different design to get the LNB/feed assembly attached to the arms.

 

[pendragon]

I'm in Cyprus now and can't use any form of cover to the c120 because even with a 3m dish on Astra 2D's mini footprint I am on fumes and can't take any signal attenuation. It just becomes a maintainance issue. There is no doubt though Invacom quad is the best for this reception.

That sounds like quite a challenge. I'm using the 1.8m for out-of-footprint reception, but I'm nowhere near your limits. Nevertheless when I put up this dish I was curious how much attenuation was caused by the feed cover. This was simple to check as it is held on by a handful of screws and easily removed. I wasn't able to measure anything but a negligible loss, probably less than 0.05 dB. I don't know from what type of material this cover is fabricated, but it looks much better and thinner than typical FTA covers. Regardless, in your shoes I expect I'd be cleaning feeds.

 

[pedrogarcia]

That sounds like quite a challenge. I'm using the 1.8m for out-of-footprint reception, but I'm nowhere near your limits. Nevertheless when I put up this dish I was curious how much attenuation was caused by the feed cover. This was simple to check as it is held on by a handful of screws and easily removed. I wasn't able to measure anything but a negligible loss, probably less than 0.05 dB. I don't know from what type of material this cover is fabricated, but it looks much better and thinner than typical FTA covers. Regardless, in your shoes I expect I'd be cleaning feeds.

I have tried a screen from the 40mm invacom on the C120 feed and measured a loss 0.2dB which was strangely enough to pixellate. Tried Cling film (Reynolds wrap?) and that was ok but it lasts only hours in the sun.

I have now just finished installing a 3.8m dish with my neighbour and we are sharing a QDH 031 quad which we have no problems not surprising with the extra 3.6 sqm of dish area. Like we always say there is no substitute for a larger dish.