Sorry, my time got captured by higher priorities. But better late than never.
I asked you to present a review of Techotrend DVB product offerings, since you marked to live in Germany, and
TT Budget S2-1600 is reported to be excellent budget card in similar price range as the cards others mentioned here (I like a lot both Prof and DVB World brands), which also employs ST0903 demodulator, newer tuner and PCI chipset models, ...
The TT-1600 may be an excellent unit, but not because of its 'newer' tuner. I believe we have beat this horse before, but the tuner chip inside is the STV6110, which is unquestionably newer than the previous STB6100 model. However the design objective for the STV6110 was lower power, not performance. This make some sense for a USB tuner, but those with a background in solid-state RF will likely recognize the trade-offs - poorer phase noise is the apropos facet here. The STV6110 levels are such that this is not likely to be much of a detriment for 8PSK, but a nasty headache for 16PSK and 32PSK. The older, higher power STB6100 is more capable in this respect.
I believe there is some question as to what extent the STV0903 has been validated against these higher order modulations, but the Linux driver does enable these features. I have played a little with this and have had some success in partially locking what may in fact be a 16PSK signal with the Prof tuners (STB6100 + STV0903). We need to start getting prepared for this.
... and no RF noise emitting high end voltage multiplier chip to supply clean power to LNBs (very important for catching sensitive DVB-S2 signals) and also to a motor if you have one. Most of such card users would not have a separate motor and LNB power supplies, but instead would usually power them all from the card, making some DVB-S2 signals reception impossible due to extra RF noise generated by some cards' power circuits.
I believe Zamar is making a veiled reference to reports in Russian forum(s) that claim certain species of the Prof tuners have a marginal power supply that requires a modification for maximum LNB performance. I have read crude translations of these posts and developed a rather bad case of skepticism, as it was readily apparent that the claims were made by people with thin backgrounds and insufficient equipment to support their assertions. But since bad ideas never die, I decided to investigate this earlier today before I got trapped in other matters.
I ran the tuner inputs for both a Prof 7500 and DVB World 2104 tuners separately to my spectrum analyzer, terminated in 75 ohms. These units are both USB, which seems a reasonable comparison. I 'tuned' both to both polarities and with/without 22 kHz tones. None of that made a lot of difference in the plots. I ended up going with 18 V polarization and no tone.
I have plotted both 0 - 5 MHz and 0 - 1.5 GHz for both tuners in the following files attached:
7500-5MHz.gif
2104-5MHz.gif
7500-1.5GHz.gif
2104-1.5GHz.gif
These are interesting results with no clear victor. In the 5 MHz BW, the Prof tuner has a naughty spur around 300 kHz, much larger than anything coming out of the DVB World unit. This does look like an insufficiently filtered switching power supply, but absolutely child's play to handle at that frequency and level. Any switch or LNB deriving power from this without a sufficiently sized capacitor is going to be vulnerable to a lot more problems than this. In the 1.5 GHz BW, the DVB World has misbehaviors at around 500 MHz, while the Prof has none. Kind of a draw in my mind, but no smoking gun given the levels of the spurs.
I then hacked a LNB into the equation to see how a DC load and IF source affects the power supply's noise spectra. I did my best to bridge the SA so it did not cause much of an impedance mismatch, but at these frequencies, this is only a best effort attempt. I then fashioned a quick 'anechoic chamber' for my Invacom Ku quatro LNBF with a waveguide and terminator. Nothing perfect here either, but both tuners got equal treatment. I tuned them up as before and took the same noise spectra in the following files:
7500-5MHz-LNB.gif
2104-5MHz-LNB.gif
7500-1.5GHz-LNB.gif
2104-1.5GHz-LNB.gif
Now we're starting to see some separation and it's not looking good for the DVB World unit. The Prof has lots of spurs rising above a reasonably low noise floor on the 5 MHz plot. I would have liked to have seen better. But look at the DVB World 5 MHz plot - the floor is above virtually the whole Prof spectra, and I'm not even considering the peaks. Tsck, tsck. Both units seem to have little influence on the 1.5 GHz plots, which is dominated by the Invacom's noise floor. No spurs of any obvious consequence. In the end it looks like both are decent for FTA use. Some warts but nothing that would presume detrimental effects. I'd personally give the Prof slightly better marks.
Sorry zamar, I think your Russian buddies need to reconsider their theories. If their LNBs are so bad that the Prof tuner's noise power spectrum is causing reception problems, they are only seeing the tip of the iceberg (no pun intended).