To follow up on the posts I piggybacked on another thread, I thought it would be better to create my own. Sorry for the tl;dr size.
[Images]
After getting lightning fast help from Tech_Support in the other thread (about spurious signals from the LNA) I thought low IF mode would be my cure for everything. Spoke too soon.
Nothing will remove an FM station image in the 2m repeater band except turning off the LNA. Luckily the image sits right in a gap between the receivable repeaters around here. The next FM transmitter tower is 20km away, all channels have 15kW max. and only a few produce fair, not insane signals. But even on shortwave, with ridiculous 1m of wire for an antenna and barely fair condx I managed to get images from the RTTY contest last weekend on top of the same band (20m).
- rsp_images_200k_baseband.png (42.26 KiB) Viewed 10524 times
[AGC]
-- On the airband I noticed that planes from our little airstrip nearby and close airliners can overload the frontend badly, so I had to turn the tuner AGC on. But then the gain reduction goes down to 102dB within 10s - and never ever goes back up, turning the RSP into a pretty deaf receiver. As it turns out, the AGC always does that in low IF mode, regardless of the frequency, or in other words AGC just does not work in low IF mode?
-- I have another issue with the AGC - its action is "stepped" and every step is causing an artifact in reception and hence in the audio. Tuned to a CW station with AGC on - every time the AGC slider moves, I hear (and see) a "ping" noise, which may actually be the DC correction in action? It seems to re-align every time the noise level changes due to AGC action and also seems to add some ringing (when I turn off post-tuner correction the ping becomes an electronic gong), which renders e.g. weak CW signals unintelligible. AM signals suffer from the artifacts as well. On the air band, the DC correction action (if that's the culprit) is causing a decaying hum noise every time a nearby plane triggers bigger action of the AGC (not necessarily on the frequency I'm listening to. On HF AM and SSB, the AGC stepping causes pops, for some reason much more pronounced in HDSDR than in SDR#1361.
-- Just now I finally got a properly loud signal on 49m from Radio Romania. Having the SDR# AGC off and the RSP tuner AGC on, I could hear its reaction on the station ID tones with a pause in between - the noise is coming up in pop-pop-pop steps, then the modulation sets in but the AGC needs another nn milliseconds to jump into action, leaving a deafening loud and distorted volume spike. I tried to record that but the signal is so effed up that SDR only records noise.
So if I leave the AGC off the RSP1 can't cope with the huge variety of signal strengths in e.g. the airband of course and if I turn zero IF mode back on to make it work, I'm getting more images (particularly in the air band), audio artifacts and the IQ imbalance issues on VHF and higher again. And if I have it working, I'm very sorry, but this is by far the worst AGC action I've ever heard, luckily the look-ahead style AGC in SDR# is catching the sheer volume of the attack phase spike but I think I prefer leaving the tuner AGC off on AM and live with the inconvenience of dealing with gain control myself and simply write off the airband.
[Band filters and scanning]
-- Besides the unfortunate fact that the unique Russian scanner plug-in for SDR# does not work in SDR# build 1361 (which is really a huge minus for anyone with scanning needs, this plug-in is just amazing), a filter switch is sitting right in the middle of the band (120MHz), with corresponding gain/noise floor differences between both filter bands and a short gap in reception during the switch. The (slow) "FMSuite" scanner chokes a bit on both side effects. I'm currently inclined to say that the RSP is probably not the greatest choice for the special requirements of the air bands and for scanning bigger frequency ranges (in particular the huge civil and miliary air bands) automatically. If I recall that right, none of the filter bands are oriented at the common frequency allocations and tend to sit in the middle of some band, causing a step in gain/noise. It's nice that they're not in the ham bands but 2m/70cm is dead as a dodo around here and ham radio is currently the least interesting part of playing with radios for me, and probably for many more RSP owners.
[Distortion on HF AM broadcast stations]
-- I just got inexplicable distortion in zero IF mode at 200kHz IF bandwidth (wider bandwidth too) in the SW BC bands. If a strong signal in the band would cause this, reducing gain should fix, right? It doesn't, only switching to low IF mode (and no AGC) does. That's a bit unsettling too. Please hear this series of recordings from RCI: First zero IF, after the gap low IF and then zero IF again, with AGC on and me tuning 1kHz up and down to show the "pings". Click
My main suspect is again DC correction, even though there is nothing visible of it on HF in general, no spike, no gap.
[Glitches]
- I had tuned to 1030 MHz to see radar echoes and then I clicked on a stored FM band frequency in SDR#. The receiver arrived at that band completely deaf - I checked if my antenna fell off the roof, then I checked the ExtIO menu and everything was like always. After I changed the baseband forth and back the receiver came back on. Unfortunately I cannot find an easy way to reproduce this for you. I suspect a bandpass did't switch when I switched the frequency? Another glitch I encountered is HDSDR occasionally opening with extremely choppy audio, I have to switch IF bandwidth forth and back to recover (took a while to figure that out too).
[Get off my lawn]
- Coming from straightforward analog receivers, then using a quite hassle-free RTL dongle (only some occasional baseband switching to get rid of images) I'm surprised how much I find myself fumbling around in the settings to fight issues and how much learning about its architecture seems to be needed. Yes I know about the profiles, but the shortcuts to recall them don't work most of the time in SDR# 1361, only when the menu was opened via mouse-click once it works for a while, so there might be some keyboard focus issue. That the window is modal is nice - except when another modal window pops up behind it and does not allow me to close the configuration window, which in turn keeps me from closing the modal window behind, which happened a few times.
While not being exactly happy about these issues, I also figured out some strengths of the RSP which do make me happy:
[FM reception]
- The RSP is generally performing very well on FM on VHF and up. I recently read a blog post by an FM broadcast DXer, comparing the RSP to his trusty old Onkyo receiver (FM-DXer's #1 or #2 choice) and the RSP could match its sensitivity and SNR and won (due to the general SDR software properties of course) in terms of selectivity.
I should soon notice what that means for me (not an FM-DXer): I'm strangely receiving a measly 5kW FM station from 426km away - 24/7, without the faintest hint of tropo DX conditions in the air. How this is even possible is beyond both me and the engineer of that station I emailed to verify the 5kW. Of course I'm also receiving a few 25-100kW stations 250km away and e.g. a 100W (!) FM station 120km away, but not that steady, it boggles the mind. Anyway, none of my other radios can pick up any of these stations on the same "antenna", which is currently just 1m of thick wire stuck out of the attic window like a vertical rod, coax shield connected to a heating radiator, making the whole affair even more unbelievable. In short, the RSP1 is obviously a beast of an FMBC receiver that seems to match the performance of the best consumer FM tuners ever built to say the least. I'd suspect an increasing number of RSP1 purchases from FM-DXers in the future. Suddenly I find FM Dxing quite interesting.
- Another blogger just hooked up some passive L-band antenna to the RSP and compared it with an RTL dongle and the HackRF One on Inmarsat signals and the difference is quite amazing. Of course the comparison is not quite fair, because the RSP and the HackRF sport an LNA while the RTL stick doesn't. I made my own simple tests, indicating that the LNA might be a big part of what sets the RSP apart from the unamplified rest, but not all of it:
I compared my (standard) RTL dongle and the RSP1 with aforementioned passive rod antenna contraption on VHF Orbcomm and NOAA satellites with the LNA on and off, and when the signals are low the LNA adds deciding 5+dB(FS) of SNR to the signal.
For the Orbcomm sats the signal peaks at 10dB over the noise for the RTL stick, then settles quickly down to 5dB and much less when the satellite approaches the horizon. With the RSP1 the SNR peaks at 15+dB for a much longer time and settles at 10dB over the noise for the biggest part of the trajectory, until the sat leaves the Orbitron radar view. Surprisingly, with the LNA off the RSP seemed to be more or less level with the dongle on 137 MHz.
- Maybe the external noise in that band was limiting the possible results, so I made another, probably more apt comparison - I used (half indoors) a puny AX-31B amplified planar LPDA to get some Intelsat AOR signals, so I could leave the internal LNA off (which is not really "off" as I recently learned) on the RSP1 and this time the RSP1 clearly won the comparison in terms of "SNR without the LNA" as expected. Reception is on the edge with this setup, and the RSP's visibly better SNR gives JAERO a sufficient signal to decode.
RTL-stick:
- InmarsatRTL.png (192.06 KiB) Viewed 10524 times
- InmarsatRSPnoLNA.png (133.62 KiB) Viewed 10524 times
To get an impression of the HF performance of the RSP1 I'd have to wait for warmer days so I can go to my electrically very quiet favorite listening spot directly at the nearby coastline, which also features some 10m flag poles I use to get wires up in the air, or just use one of them as a vertical. In the noise of my house I didn't notice any advantages over e.g. my portables, so I'm curious to check out how it behaves out there, although after getting in trouble with 1m of wire I'm afraid the half-decent wires I'd be going to try will make the RSP burst into flames.
In summary, in my very varied all-purpose usage profile the required fiddling with the settings gets a bit in my way, making me even less happy about the unexpected high susceptibility for mixer images and spurious signals, DC imbalance issues and the terrible AGC and I hope this is not a general property of SDRs altogether. I was expecting that it would simply replace the RTL stick in all of its aspects (which I guess is a not so uncommon wrong expectation) but I found some places where the stick still wins, due to the design differences. I'm still undecided whether or not to send the RSP back, it has some areas where it really, really shines and some other where it only shows a good potential, that has yet to be harvested.
I hope some of the issues can be sorted in the future, and I also hope the unlucky SDR# situation can be improved to its previous state, because SDR# really has a lot of advantages over the other softwares for me (much better ergonomics and visualisation, plug-in architecture, more versatile AGC, better IQ imbalance management, more convenient frequency management and spectrum labels, less tilted to the HF side...).
Ollie
RSP#201215#12#853