Satellite Spectrum - perhaps it is time to use the right tool

**VMA** · August 24, 2026 at 1:06 AM

Satellite Spectrum – perhaps it is time to use the right tool

Hello,

I have been following the continued discussions about obtaining increasingly detailed spectrum information from DVB tuner cards and satellite receivers.

I find this slightly ironic.

Satellite enthusiasts are prepared to spend considerable amounts of money on dishes, feeds, LNBs, positioners, tuner cards and receivers, and then spend an enormous amount of time trying to overcome the limitations of a DVB tuner in order to obtain something approaching a proper spectrum display.

A DVB tuner is fundamentally a receiver and demodulator. Its spectrum display can be extremely useful for finding transponders, checking reception, blindscan and identifying signals. I have worked with these techniques myself for many years, and it is remarkable how much useful information can be extracted from comparatively inexpensive tuner hardware.

But there are physical and architectural limits to what such hardware can provide.

If the objective is really to examine the RF spectrum - weak carriers, adjacent signals, transponder shape, interference, noise floor, bandwidth, shoulders, spurious signals, etc. - why not use an instrument designed for that purpose?

A proper spectrum analyser is the obvious solution, and used instruments are often surprisingly affordable. Today there is also another very accessible option: an SDR.

Devices such as a HackRF, PlutoSDR or similar wideband SDR can provide a much more detailed and flexible view of the spectrum than the relatively low-resolution sweeps produced by most DVB tuners.

This does not make tuner-based spectrum displays useless. Quite the opposite. They are excellent tools when the objective is receiving and analysing DVB signals.

But spectrum analysis and DVB reception are two different jobs.

For DVB parameters, lock, modulation, FEC, MER, constellation, transport streams, blindscan, etc., use the tuner.

For examining the actual RF spectrum, use a spectrum analyser or an SDR. It will show you ALL transponders, no matter how low the SR. You will be able to determine the frequency and can THEN try to lock the transponder with a specific blindscan on that frequency.

There comes a point where spending more and more effort trying to make a DVB tuner behave like RF measurement equipment becomes less practical than simply adding the appropriate instrument to the setup.

For serious satellite experimentation, I think the ideal solution is to have both. They complement each other extremely well.

Just make sure to use a DC BLOCKER when connecting the LNB output of the receiver (or a T), to prevent burning the input stage of the spectrum analyzer/SDR.

omna573 · September 20, 2026 at 9:03 PM

I agree with the principle, and I would add that it does not have to be one or the other. I run both, so instead of arguing I will just show the same signals through both instruments, taken minutes apart on the same dish and the same polarisation.

The tuner is a TBS-6903x. Worth noting for this discussion: that card has no RFScan at all. Its capability list reads FFTScan, BLScan, BLScan2, IQScan, SignalInfo, DiSEqC, and no RFScan, so the point by point sweep everyone knows from the 5927 does not exist on it. The only path to a spectrum is the undocumented FFTScan() in StreamReader.dll, which returns a block of bins per call. On the settings in these pictures that is 512 bins across a 60 MHz block, so 117 kHz per bin, and a full pass over 10700 to 12750 MHz takes around 27 seconds. The SDR is a Fobos with AGILE firmware, driven by a sweeper I wrote for it with close cooperaation with CaludeAI: 34 segments of 30 MHz, 1.53 kHz RBW, 655360 bins per sweep.

Picture 1: both programs on the same band, Turksat 42E, vertical, low band. The transponder shapes and positions agree, which was the first thing I wanted to know.

Picture 2: the same pair zoomed to 11490 to 11535 MHz. This one makes your point better than I could. The tuner has locked the carrier and can tell me it is 11515.703 V, SR 8000, 8PSK 5/6. The SDR shows the shape of that same piece of spectrum at 1.53 kHz resolution, and shows what is beside it. Two different jobs, exactly as you say. Note the axes are not comparable: the SDR plot is dB above noise, the tuner plot is in raw FFT units.

Picture 3: the TBS 6903x tuner side on its own, full Ku band with the waterfall underneath.

The limits of the tuner path are real and I would rather state them myself. The FFT values are uncalibrated, so the vertical axis is raw units and not dBm: there is no conversion anywhere in the DLL and I have not fitted one yet. This card is known for 2 to 3 MHz of frequency error. And FFT mode on the STiD135 is exclusive with tuning, so sweeping and locking cannot be interleaved, you have to leave FFT mode, tune, and come back. For weak carriers, low SR, shoulders and spurious signals, the SDR wins and it is not close.

Where the tuner still earns its place in my setup is the time axis rather than the frequency axis. What I am building around it is an archive: every sweep stored with its timestamp, an event log of what appeared and for how long, and a catalogue that groups repeated occurrences across recordings. The two cursors in picture 3 are measuring one such burst. An analyser will show me that burst far better than the card does. It will not tell me that the same carrier comes back the next day at the same time. That is a database question, and it needs a receiver that can be left running for weeks.

So the split ends up being the one you describe: SDR to see everything including the low SR carriers, tuner to confirm what they actually are, one log over both. The reference overlay already works on the tuner's own recordings, and feeding it from the SDR is the next step.

One thing the pictures will have given away: everything above is low band only. The SDR is fed through a QO-100 bias tee from DX Patrol, which has no 22 kHz tone, so I am stuck below 11700 MHz on that side for now. LOF2 and the switch frequency are set to zero in the tuner window for the same reason, to keep both instruments looking at the same thing.

Which brings me to a practical question for you. What do you use as a DC blocker when tapping the LNB output with a T? I would rather get that right before putting anything expensive in parallel with a powered feed. And if anyone has a tidy way to add tone switching ahead of an SDR, I am listening.

Joolz · September 20, 2026 at 10:16 PM

For DC blocking on the input to an SDR I use a low cost F type splitter with DC pass only on one port feeding a TBS DVB card (picture below)

Thank you for sharing your fantastic work, I have SDRs and DVB cards in operation but have never managed to get them displayed in tandem as you are showing. Using the DVB card and syncing the SDR to the Hi/Lo V/H would create excellent visibility of the full LNB spectrum.

**VMA** · September 20, 2026 at 10:58 PM

I use this: https://www.televes.com/en/catalog/pro…r/category/281/

Note that using an SDR like B210, HackRF One, Adalm Pluto, etc., will give you a FFT based spectrum without calibration.

A more precise spectrum (in terms of frequency and amplitude) can of course be obtained using a proper spectrum analyzer.

omna573 · September 21, 2026 at 12:16 AM

Agreed, and I should have been clearer: I am not claiming an absolute amplitude anywhere in that setup. The Fobos SDR plot is labelled dB above noise (in Slovak language) on purpose, and the TBS 6903x tuner plot is in raw FFT units for the same reason. Neither is dBm and neither pretends to be.

There is a small irony in it though. The only absolutely calibrated level in my whole chain comes from the DVB tuner, because the AGC to dBm conversion lives in the demodulator vendor's LLA layer under the DLL rather than in any application. On a 5520SE that comes straight out of RFScan, so the sweep itself is in dBm. On the 6903x there is no RFScan at all, so the spectrum has to come from FFTScan in raw units, but SignalInfo still reports RFLevel in dBm once a carrier is locked.

Which gives me a way to fix the axis without buying anything: sweep the same dish and the same band with both cards, one giving dBm per point and the other raw FFT values, and fit the conversion from that. Not traceable calibration, obviously, but enough to put a meaningful scale on the FFT display, and it can be checked against locks afterwards.

An analyser with a calibration certificate remains the right instrument if the number has to be defensible. For what I am doing, knowing a carrier is 8 dB stronger than it was last week is usually the question, and that only needs the scale to be linear and stable, not absolute.

**VMA** · September 21, 2026 at 11:50 AM

You are doing it right.

This thread was mainly created for people that do not understand that any program based on CrazyCat's streamreader.dll is limited to what streamreader.dll can do. As a consequence, all different software can achieve the same, if programmed correctly. You can do more individual BLScan() calls or less, try to implement whatever algorithm to optimize these calls, but at the end of the day, there is no magic app that is "better" in borderline reception than other apps, because the limitations are within tuner, driver and streamreader.dll.

So, instead of people wasting hours on one blindscan to find some exotic transponders with low SR and very close to a transponder with high SR, my suggestion was: get a spectrum analyser! Then you will "see" any transponder instantly, determine it's frequency and can use any of the apps to blindscan that specific frequency. In fact you can even use my streamreader.exe manually!

And of course, a spectrum analyser is expensive. But when I see what people have in terms of dishes, I conclude they have the resource to buy a spectrum analyser. What does a 4m dish cost, considering its purchase price (even second hand), transport, installation and maintenance?

For sure there is a budget of 1.000-2.000 Euro to get a decent spectrum analyser, or at least 200-300 Euro for a SDR.

**superdish66** · September 21, 2026 at 12:10 PM

Thank you VMA for the clear analysis of streamreader.dll and the cards.

We agree. The limit is the tuner, the driver and the DLL, and no program goes past it.

Between that limit and what a user really gets there is still room, and that room is software. How the card is called, what is measured, what is shown, what is recorded. That is where we work.

A spectrum analyzer is a fine tool, and we leave it to antenna installers. We are hobbyists. We do what we can with the card we have, and the search is the part we enjoy.

Thank you again, kind regards.
Superdish66

**VMA** · September 21, 2026 at 1:30 PM

"Between that limit and what a user really gets there is still room, and that room is software."

Of course there is room for different software: different presentation, logging, automation, filtering, analysis, user interfaces, etc. I have never disputed that.

But this should not be confused with obtaining more information from the tuner.

With the streamreader.exe I released, it is already possible to access essentially what the tuner/driver/streamreader.dll combination makes available. Once you reach that boundary, another application cannot magically see signals or information that never passed through that interface.

So, when comparing applications, I think the interesting technical question is very simple:

What information does application A obtain from the tuner that application B, using the same card, driver and streamreader.dll, does not?

If the answer is “none”, then we are comparing processing, presentation and features — which is perfectly legitimate — but not superior RF reception or some deeper access to the spectrum.

Regarding the spectrum analyzer: I also don't agree that it is something only for professional antenna installers. Some users here operate several large dishes, positioners, multiple LNBs, DVB cards and dedicated computers. These are already substantial installations.

For somebody spending that amount of time and money searching for unusual feeds and difficult transponders, an SDR costing a few hundred euros — or even a decent used spectrum analyzer — is hardly an absurd investment. More importantly, it shows directly what is actually present at RF level, rather than repeatedly asking a DVB tuner to discover something that may be outside its acquisition capabilities.

Software can do a great deal with the information available to it. But software cannot manufacture information that the hardware and driver never supplied.

**superdish66** · September 21, 2026 at 2:04 PM

Thank you VMA.

On lock, two measurements with the same card, driver and streamreader.dll.

1.0W, 11073 V. Tuned with no input stream filter, the card locks and delivers 984388 bytes per second. Tuned with a filter on input stream 1, the lock falls and the stream is zero. Same tuner and same DLL, the call decides the lock.

QO-100, a 332 kS/s carrier. The card spectrum shows no peak there. A fine scan with narrow windows finds it, locks it and records it.

The limit is the tuner, and we agree. How close to that limit a user gets depends on how the tuner is asked, and that can be measured.

From our side, the next answers will be measurements. Thank you again, kind regards.
Superdish66

Delta Star · October 3, 2026 at 9:26 AM

Auf dieser Website wimmelt es nur so von Software.
Ich suche nach einer Software, die dasselbe kann wie mein
Combolook Color HD, damit ich das im Breitbildformat habe.

**VMA** · October 3, 2026 at 1:47 PM

Du musst erstmal sagen, welche PC Hardware du hast.

Du brauchst einen Tuner der mit CrazyScan's streamreader.dll kompatibel ist, sonst geht gar nichts.

Und dann braucht du nur zwischen diesen Anwendungen wählen:

- CrazyScan
- EBSPro
- IQMonitor
- 4T2 Content Analyser (nicht gratis, aber mit CrazyScan streamreader.dll gratis nutzbar - ist aber wohl ein Hack/Crack)
- TSReader Pro (war nicht gratis, wurde nach dem Tod des Autors aber nun gratis veröffentlicht - funktioniert auch mit anderer Hardware)
- VMA Video Analyser Package (nicht gratis)
- PanthaSat

Dann gibt es noch Hardware/Software von Dektec (nicht gratis).

Über "zu viel" Auswahl hat sich eigentlich noch nie jemand beschwert...

"Ich suche nach einer Software, die dasselbe kann wie mein
Combolook Color HD, damit ich das im Breitbildformat habe."

Das gibt es so nicht, weil der Combolook Color HD das Spektrum mit einer sehr hohen Refresh-Rate darstellen kann, einen tollen Blindscan für DVB-S/S2 hat und auch das Konstellationsdiagram in Echtzeit darstellt, ohnen einen Lock zu brauchen. Damit sind alle mir bekannten PC-basierten Tuner überfordert.

Für Breitbildformat kannst du einen 16:9 Fernseher anschliessen (Composite Video), doch ich denke der Combolook Color HD is wie mein Satlook Color HD und kann kein MPEG4 darstellen.

Delta Star · October 3, 2026 at 4:13 PM

TBS5927

Participate now!