New SDR radio does not seem to work

[WaveyDipole]

I recently purchased an SDR radio, one of these cheap RTL8232U/R820T2 based gizmos from the far east (see picture below). This one is in a nice aluminium case and I paid a little more for it than for a standard USB dongle. it arrived the other day but unfortunately it does not seem to be working as expected.

Initially I plugged in and the light came on so all seemed good to go.

I then tried several things without much success and have reached the conclusion that is must be a hardware fault. These are the things I have tried:

- initially tried windows software (SDR#, HDSDR, SDR-Radio) but although the device was recognized, none of these programs could communicate with it.

- switched to Linux for diagnostic purposes. Connected device to the computer, installed rlt_sdr and ran rtl_test -t. The kernel driver has been blacklisted as instructed. The device is recognized as 'Using device 0: Generic RTL2832U OE'. The device is detected as idVendor=0bda and idProduct=2838. The product id is rather confusing as it ought perhaps to be 2832. The radio IC is an R820T2, not an E4000. rtl_test seems to identify the R820T2 tuner correctly but then seems to be looking for an E4000 tuner?

- checked the device current usage which is around 100mA. This should be well within the capability of a typical USB port. Tried with no other devices connected.

- tested the dongle on another computer. Same result.

I have now also noticed that the external blue light no longer comes on, or sometimes just flashes briefly. I opened the aluminium case, primarily to confirm the chipset, and found that it does indeed have a T820T2 tuner. The case contains a main PCB and a daughterboard which appears to be an actual SDR board from the original dongle package. There is another LED on this board which does light up and stay on. The case gets only slightly warm during operation but not warm enough - I would not describe it as hot - to warrant concern.

I have contacted the seller to request a refund/return, although sending it back to China will cost nearly 9gpb according to Royal Mail so I'm not sure this would be worthwhile, although I am still waiting for the seller to get back to me and tell me what they are willing to do.

So as already stated, although it is possible I missed sommething in configuration, since this has failed in both Windows and Linux and two different computers, I think it reasonable to assume that this is a hardware fault. However, I also wondered whether it could be something as simple as a dodgy joint, particularly between main board and daughter board, and whether it might be worth re-flowing this using a hot air station? If I don't have any luck with the seller, might that be worth a try?

Below is the output from rtl_test (when it can actually 'see' the dongle):

$ rtl_test -t
Found 1 device(s):
  0:  Realtek, RTL2838UHIDIR, SN: 00000001

Using device 0: Generic RTL2832U OEM
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_read_reg failed with -1
rtlsdr_write_reg failed with -1
rtlsdr_read_reg failed with -1
rtlsdr_write_reg failed with -1
rtlsdr_read_reg failed with -1
rtlsdr_write_reg failed with -1
rtlsdr_read_reg failed with -1
rtlsdr_write_reg failed with -1
No supported tuner found
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
Enabled direct sampling mode, input 1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
Supported gain values (1): 0.0
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -9
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
rtlsdr_demod_write_reg failed with -1
rtlsdr_demod_read_reg failed with -1
WARNING: Failed to set sample rate.
No E4000 tuner found, aborting.
rtlsdr_write_reg failed with -1

[denverpilot]

You broke the rule! “Don’t turn it on, take it apart!” LOL.

Kidding of course. Sorry about the infant mortality... never any fun when new kit doesn’t work right out of the box, and you’re looking forward to having a play!

[RoGeorge]

Did you installed "ExtIO_RTL2832.dll'?

My installation notes for R820T2 based USB dongle (from some years ago):
1. zadig
2. hdsdr.de
3. ExtIO_RTL2832.dll

[WaveyDipole]

You broke the rule! “Don’t turn it on, take it apart!” LOL.

Well yes, I did do it the other way around so I guess I got what I deserved 
I did have to open it up eventually if only to confirm which chips were present.

Kidding of course. Sorry about the infant mortality... never any fun when new kit doesn’t work right out of the box, and you’re looking forward to having a play!

Yes, rather disappointing but even more so when factoring in the the cost of sending it back halfway around the world. Still waiting on the seller to respond regarding that however.

Did you installed "ExtIO_RTL2832.dll'?

My installation notes for R820T2 based USB dongle (from some years ago):
1. zadig
2. hdsdr.de
3. ExtIO_RTL2832.dll

Yes, that is exactly what I discovered also. HDSDR requires that one runs zadig and installs the driver. It also requires that ExtIO_RTL2832.dll is present in the HDSDR folder and loaded. I did follow both of these steps. When I load the DLL manually, the results are rather unpredictable although the program asks whether I want to load another ExIO_RTL2832.dll. Sometimes the program hangs. Sometimes it fails to load. Sometimes it appears to load successfully but the device is not shown as connected. Either way, it doesn't work.

[cdev]

It is a direct sampling unit, so, did you switch it to "direct sampling, q input"  in your software, for use on HF ? 

And use a decent antenna and ground?

Should be in the tuner settings area of your software,

or if you use gnuradio- in the connect string.
You need to use this string.

rtl=0, direct_samp=2 to receive HF (below 24 MHz) Or else your receiver will seem dead.

Look up "direct sampling rtl2832" on youtube to see how that works.

Your device uses has a unique hack that turns a TV dongle into perhaps one of the most widely tunable range receivers around with very few additional parts.

[WaveyDipole]

It seems I may have been blindly following the numerous examples I found in the Internet and using the wrong flag for rtl_test command. Apparently the '-t' flag is specifically to enable tests for the E4000 tuner, which this dongle does not have and the correct command is to run just 'rtl_test' without any flags. However, the poutput of this is still not very encouraging:

# rtl_test
Found 1 device(s):
  0:  Realtek, RTL2838UHIDIR, SN: 00000001

Using device 0: Generic RTL2832U OEM
usb_claim_interface error -5
Failed to open rtlsdr device #0.

I also ran lsmod to check whether the post is being held by the kernel driver, but the dvb_usb_rtl28xxu is not listed in the output so it would appear that it is successfully blacklisted and not being loaded.

I have also had a reply for the seller asking if I can solder the back of the board. Not quite sure what they mean, but they have also offered to send a replacement board if this does not work.

[cdev]

I own the same board you have.. I can help you get it working. Its dongle is no different than any other dongle except that it uses the normally unused pins 4+5 on the RTL chip for HF input.


If I were you, unless you already have, I would not open it's case because the wire connection is fragile and can easily be damaged. And re-soldering it is a PITA. That is what the seller is trying to explain to you. He or she is trying to ask you are you skillful enough to solder those two little wires yourself.

But it sounds to me as if you never were able to switch to the "Q" input. So its likely not broken. Google "direct sampling" RTLSDR.

To use it on HF, you need to select the Q input

Then, with a good antenna, it works surprisingly well, considering that its using the differential input to the RTL chip. The design is actually quite good, it maintains ESD protection and matches the RF impedance. It could probably be optimized using a network analyzer but my own experience with it is that it works fairly well.

I bought it out of curiosity and wasn't expecting it to work as well as it does. I bought it as a kit.

Note that you need to use the HF antenna input and tune it to an HF frequency.

Try to see if you can receive the 20 or 40 meter ham bands.. 14 MHz or 7 Mhz.

If your rtl_* apps are broken I would remove and then reinstall the rtlsdr library.  Thats not the hardware's fault.

Since its on windows, you're on your own there.

[WaveyDipole]

cdev, thanks for your comments.

The plot thickens with this unit. The seller sent me some instructions indicating the 5 solder points he wanted me to re-solder. There are the big 5 points under the antenna jack on the internal board. I refloweed these but it made no difference. The seller has offered to send a replacement board if the soldering didn't work. Since it didn't, I replied to that effect and am awaiting a response.

Later on I had another look at the board under the microscope to see whether there were any bad connections etc. I couldn't see anything obvious, only a tiny blob of solder next to a resistor which was easy to remove. I then cleaned the board with IPA removing any traces of flux. After that I probed with a DMM to check that power was reaching the daughterboard - it was and at this point to my surprise, the blue LED lit up as well.

I now tried the board on the computer - but still the sme problem. I substituted the USB cable once again and I was surprised when rtl_test now retuned a sane result:

$ rtl_test
Found 1 device(s):
  0:  Realtek, RTL2838UHIDIR, SN: 00000001

Using device 0: Generic RTL2832U OEM
Found Rafael Micro R820T tuner
Supported gain values (29): 0.0 0.9 1.4 2.7 3.7 7.7 8.7 12.5 14.4 15.7 16.6 19.7 20.7 22.9 25.4 28.0 29.7 32.8 33.8 36.4 37.2 38.6 40.2 42.1 43.4 43.9 44.5 48.0 49.6
[R82XX] PLL not locked!
Sampling at 2048000 S/s.

Info: This tool will continuously read from the device, and report if
samples get lost. If you observe no further output, everything is fine.

Reading samples in async mode...

Now I had tried that subsitute cable before without success. I also found and tested a third one - still working. Went back to the originally supplied one and the errors came back. Sure enough, the substiitute restored normal operation again. So it seems that I definitely have a duff cable, but why did the substitute one not work before?

After a while the blue LED stopped working again, but the radio continued to work and I was now running gqrx. However, the receiption on FM was absolutely awfull. The signal was weak and getting lost in the hiss and crackle. The signal strength was low regardless of whether I used the supplied antenna, an external one or a length of wire. I couldn't tune into HF as gqrx doesn't seem to allow me to tune below 24.897Mhz. I tried adding 'direct_samp=2' to the device string, but that causes the program to crash for some reason. I also tried HDSDR on Windows. It did work this time, but the reception was even worse than in gqrx on Linux. I couldn't copy any FM stations at all. It was all lost in the noise.

Maybe cleaning it up removed some debris that was shorting something although I am still baffled about the LED. If you have the same box, does the LED stay on all the time?

I still think the hardware is dodgy as the front end of the receiver seems suspect and the LED does not stay on.

UPDATE:
I had a look on the gqrx website and found these instructions to install the latest version.

http://gqrx.dk/download/install-ubuntu

Following these I got the latest version up and running and this now does accept 'direct_samp=2' in the device string. However, I cannot figure out where its tuning to. I don't get any AM stations on LW, MW although I do get various beeps on these bands where AM stations should be. BTW, I am selecting AM mode and using the normal setting for filter width and shape.

[cdev]

I think GQRX has an "ignore limits" checkbox or something like that you need to check..

Try CubicSDR. Or since it sounds like you are on Windows, HDSDR.

[cdev]

You'll have better luck if you use a USB extension and a couple of ferrite common mode chokes to quench the RFI that is an unavoidable rider on the outside of USB cables..


You are using a device that doesnt have a Elonics tuner so you should NOT be using rtl_test -t
 here..  here is the help for the program.

rtl_test -h
rtl_test, a benchmark tool for RTL2832 based DVB-T receivers

Usage:
   [-s samplerate (default: 2048000 Hz)]
   [-d device_index (default: 0)]
   [-t enable Elonics E4000 tuner benchmark]
   [-p[seconds] enable PPM error measurement (default: 10 seconds)]
   [-b output_block_size (default: 16 * 16384)]
   [-S force sync output (default: async)]


Try "rtl_test -p10" or
"rtl_test -p60" or similar...

Make sure your long wire is up and in the clear and that you have a decent ground. I would also suggest wiring up a 9:1 unun if you can..

[WaveyDipole]

I think GQRX has an "ignore limits" checkbox or something like that you need to check..

Try CubicSDR. Or since it sounds like you are on Windows, HDSDR.

I am actually playing with software on both platforms. I am currently booted into Windows and have just had a look at CubicSDR which is quite nice. Same issue with the HF bands though so it doesn't seem to be the software but the radio. I will have a look for that "ignore limits" option when I go back into Linux.

You'll have better luck if you use a USB extension and a couple of ferrite common mode chokes to quench the RFI that is an unavoidable rider on the outside of USB cables..

Something to look at when I get the SDR box working properly - or get a properly working one - whichever comes first!

You are using a device that doesnt have a Elonics tuner so you should NOT be using rtl_test -t
 here..  here is the help for the program.

rtl_test -h
rtl_test, a benchmark tool for RTL2832 based DVB-T receivers

Usage:
   [-s samplerate (default: 2048000 Hz)]
   [-d device_index (default: 0)]
   [-t enable Elonics E4000 tuner benchmark]
   [-p[seconds] enable PPM error measurement (default: 10 seconds)]
   [-b output_block_size (default: 16 * 16384)]
   [-S force sync output (default: async)]

Try "rtl_test -p10" or
"rtl_test -p60" or similar...

Yes, thanks for pointing it out. Someone on the osmacom.org rtl_sdr mailing list pointed it out as well so now I just use 'rtl_test' without any parameters which is sufficient.
 

Make sure your long wire is up and in the clear and that you have a decent ground. I would also suggest wiring up a 9:1 unun if you can..

My external antenna has a proper ground and I did add a self-made 9:1 unun shortly after I put it up. It seems that the dongle is quite fussy about having a proper ground. I also have a proper FM dipole antenna in the loft, but I don't have the correct connector to use it with the SDR dongle. I appreciate that the supplied antenna is not worth bothering with.

I was also thinking of building a discone at some point as I would like to pick up aircraft and possibly other stuff, but only once I have the SDR setup working properly. I might also set it up on a Raspberry PI and set it up to be accessible via TCP.

[cdev]

Try this- $ rtl_test -p100
Found 1 device(s):
  0:  Realtek, RTL2838UHIDIR, SN: 00000001

Using device 0: Generic RTL2832U OEM
Found Rafael Micro R820T tuner
Supported gain values (29): 0.0 0.9 1.4 2.7 3.7 7.7 8.7 12.5 14.4 15.7 16.6 19.7 20.7 22.9 25.4 28.0 29.7 32.8 33.8 36.4 37.2 38.6 40.2 42.1 43.4 43.9 44.5 48.0 49.6
[R82XX] PLL not locked!
Sampling at 2048000 S/s.
Reporting PPM error measurement every 100 seconds...
Press ^C after a few minutes.
Reading samples in async mode...
real sample rate: 2048006 current PPM: 3 cumulative PPM: 3
real sample rate: 2048007 current PPM: 4 cumulative PPM: 3
real sample rate: 2047993 current PPM: -3 cumulative PPM: 1
real sample rate: 2047998 current PPM: -1 cumulative PPM: 1
real sample rate: 2047993 current PPM: -3 cumulative PPM: 0
real sample rate: 2047987 current PPM: -6 cumulative PPM: -1
real sample rate: 2047998 current PPM: -1 cumulative PPM: -1
real sample rate: 2047993 current PPM: -3 cumulative PPM: -1
real sample rate: 2047991 current PPM: -4 cumulative PPM: -2
real sample rate: 2047992 current PPM: -3 cumulative PPM: -2
real sample rate: 2047995 current PPM: -2 cumulative PPM: -2
real sample rate: 2047999 current PPM: 0 cumulative PPM: -2
real sample rate: 2047982 current PPM: -9 cumulative PPM: -2
cb transfer status: 5, canceling...
cb transfer status: 5, canceling...
cb transfer status: 5, canceling...
cb transfer status: 5, canceling...


If your computer's clock is being set via a network time server it should be quite accurate and then the PPM figure you get is likely to be fairly close, after it runs for a few minutes the wandering will stop and it will just stay on the same number.

You'll get a ballpark idea of how far off its frequency settings might be.

[cdev]

Since you have it open could you photograph how they anchor the two little wires to the RTL chip and keep them there? Also be careful very very careful not to knock them off.

I think you'll be pleasantly surprised at the reception. For what you paid its hard to beat.

Although its not quite as sensitive as some of the upconverters, particularly ones that contain a low noise amplifier, the direct sampling is free-er of birdies than the typical upconverted HF output.

All RTLSDRs running on HF need a good ground to work well, and seem quite vulnerable to noise of all kinds. or perhaps a better term is less forgiving of it.

Considering all that they work pretty well.

[Beamin]

Did you installed "ExtIO_RTL2832.dll'?

My installation notes for R820T2 based USB dongle (from some years ago):
1. zadig
2. hdsdr.de
3. ExtIO_RTL2832.dll

I almost threw mine away but this fixed the problem.

I also had issued on windows 10 with the comport being not setup right. There is a program with an installer on the sdr# website.

Be careful not to get the comport wrong.

[WaveyDipole]

I own the same board you have.. I can help you get it working. Its dongle is no different than any other dongle except that it uses the normally unused pins 4+5 on the RTL chip for HF input.

If I were you, unless you already have, I would not open it's case because the wire connection is fragile and can easily be damaged. And re-soldering it is a PITA. That is what the seller is trying to explain to you. He or she is trying to ask you are you skillful enough to solder those two little wires yourself.

Since you have it open could you photograph how they anchor the two little wires to the RTL chip and keep them there? Also be careful very very careful not to knock them off.

I think you'll be pleasantly surprised at the reception. For what you paid its hard to beat.

I am curious about this myself, because I can't see any wires attached! Each pin is simply connected via a capacitor to GND and there are not plated through junctions to the underside. Apparently it is possible to use pin1 was well, but visually this appears to be n/c. From the other end, the HF antenna connector connects to a couple of inductors printed on the PCB and some additional circuitry, but I would need to trace out how the connections to the radio are made. The only obvious connections to the main PCB from the daughterboard seem to be the antenna socket and 3 out of the 4 USB pins. If it was necessary to add these wires, then I suppose they could be connected to the capacitor pads. I have attached a pic of the connections.

Did you installed "ExtIO_RTL2832.dll'?

My installation notes for R820T2 based USB dongle (from some years ago):
1. zadig
2. hdsdr.de
3. ExtIO_RTL2832.dll

I almost threw mine away but this fixed the problem.

I also had issued on windows 10 with the comport being not setup right. There is a program with an installer on the sdr# website.

Be careful not to get the comport wrong.

Beamin, it turns out that you have to run Zadig against each com port. I managed to get HDSDR to work but SDR# stubbornly refused to, until I tried the 32bit version that is. I initially downloaded the 64bit version as I have a 64bit Windows.

[cdev]

Thats the Rafael Micro tuner chip, the HF direct sampling is done by the RTL2832 chip.

We want the other chip, the one with the wires connected to it!

Also, how is it working?

Adjust the RF gain for the maximum spread between the peaks and valleys.  You should be able to get 50 db approximately, with the AGC turned off. Any less and you need to improve the common mode filter on the USB and the boards bypass caps.. There is a bunch of additional improvement which can be done.  A bigger cap across the USB power rails helps, but not so big that it is seen as a short when the power goes on. Then the dongle wont start.

I also added some SMD caps to further decouple the power going to the VHF/UHF antenna. the "bias-tee".

[WaveyDipole]

Thats the Rafael Micro tuner chip, the HF direct sampling is done by the RTL2832 chip.

We want the other chip, the one with the wires connected to it!

Also, how is it working?

Sorry, let's try that again with the correct chip 

I have uploaded a pic of the RTL2832 chip. It looks like pins 4 and 5 are n/c, but there are plated through holes and a pad provided to facilitate a connection which makes sense. Still, regrettably, this does not help you out with a picture of the wires. Looking at the wider pictures of the main board, the blue lines show where the points on the main board connect with two further points underneath the daughterboard which correspond to the two holes topside on the daughterboard. So it would seem that this is where connections are supposed to be made. I verified with a DMM that these points are definitely not connected. I also checked that the larger hole is connected with the smaller one nearer the IC as indicated by the white line. The other connection point has a clearly visible track to the IC. Perhaps its an assembly omission or maybe its up to the buyer to research and make those connections? Since the antenna circuit is disconnected this would explain the abscence of any AM band signals, but then what were the various high pitch noises I was getting? Might they have been PLL noise or birdies perhaps?

My board does not have the ferrite ring that your board has. On the other hand the antenna connector, IR sensor and power LED are absent from your SDR board. I'm not sure really why these were put in since they are inaccessible inside the case and serve no useful purpose.

Adjust the RF gain for the maximum spread between the peaks and valleys.  You should be able to get 50 db approximately, with the AGC turned off. Any less and you need to improve the common mode filter on the USB and the boards bypass caps.. There is a bunch of additional improvement which can be done.  A bigger cap across the USB power rails helps, but not so big that it is seen as a short when the power goes on. Then the dongle wont start.

I also added some SMD caps to further decouple the power going to the VHF/UHF antenna. the "bias-tee".

Thanks. These are all these are good points for me to think about. I'm waiting on confirmation but hopefully I will be getting a replacement board. If they don't want the old one back then I will have a play with it and connect those two points to see what happens. I also found that the power LED that does light up has a 2.2k resistor in series with it from the 5v power rail which looked rather high and sure enough there is only 0.8v across the LED. I reckon it was probably meant to be a 220 ohm.

[cdev]

Wow, pretty cool!

Looks like the various vendors selling them are all now selling enough of them to get their own custom dongle mods from the folks at Newskysz.

I'm surprised that it doesn't work. Are you sure?

[WaveyDipole]

Its getting rather late here now, but hopefully I will be able to do a bit more work on it tomorrow then I should know for sure.

BTW, just noticed your country of origin. Antartica?

[cdev]

Thats in honour of my Aunt.

[WaveyDipole]

Today I connected up those two links with a short pieces of wire and I am now getting AM broadcast signals. I could barely pick up anything on LW. I can usually pick up R4 and a couple of French stations quite strongly, but I just barely got a very weak R4 and no French stations. On MW frequencies the story was a bit better, but it seems I can only pick up strong local stations there as well. I would have to say that sensitivity is quite poor on MW and abysmal at LW frequencies. I'm not expecting much for the cost, but I guess I expected it to perform rather than just being able to pick up a few strong local stations. However, in normal mode FM sensitivity is a bit poor as well. I couldn't pick up much at all on HF frequencies. I should at least be able to pick up the Chinese station. I will wait until the evening when the bands come to life a bit more and then try again.

Incidentally, the problem with the LED turned out to be a faulty intermittent LED as well as the wrong  series resistor being used. Another LED from my bits box and a 500ohm resistor work perfectly OK. I also checked and the current drawn is still around 100mA.

One thing that I don't like about the board I have is that the RTL dongle board is soldered directly on top of the main board. The only insulation is the screen. Not only would this create potential for shorts, but must also add capacitance between RTL dongle board and the main board, the substantial part of this area essentially creating a ground plane. Could this adversely affect performance? Your board on the other hand has a neat cut-out for the TRL dongle board to fit into so this would not be an issue. I have been thinking about de-soldering the RTL dongle board and mounting it on wire 'stilts' to give it a bit of clearance from the main board.

The seller confirmed that he will be sending me a replacement board, so it will be interesting to see how that performs by comparison and whether those HF antenna connections are connected up.

In the meantime, do you have details of the windings required on that ferrite please? I have a couple of rings and can probably find some copper enamelled wire and would like to try adding that to see if it makes a difference.

[Whales]

> gqrx .... bad fm reception

My first experience was the same.  Things barely stood out from the noise and I was quite disappointed.  No prior SDR experience to compare it to.

It turns out I needed to pump up the gain in one of the UI tabs.  Not sure where the gain is implemented on-chip, but boosting it to somewhere in the 20-30dB range increased my overall perception of the performance dramatically.  FM stations were now clear and the world escaped the floor.

There is an AGC mode, but it seemed overly aggressive to me.  Probably suited for the dongle's original purpose (digital TV), not hunting through the ether for quiet curiosities.

[WaveyDipole]

With AM using direct sampling mode, having hardware AGC enabled seems to make a substantial improvement. With it disabled and and slider set to max (49db) weak stations disappear into the noise and stronger ones get very crackly. Even with AGC enabled, the stronger peaks are only at around -60db and there is still substantial noise on these stations. These are stations I get loud and clear on any AM radio. The weaker ones -80 or below. The audio gain, as expected, just makes everything louder including the noise. On the HF connector, connecting the antenna earth actually makes things a lot worse so I leave it disconnected.

With FM, the graphs sits at the -60 make and the peaks reach into maybe -30db, but the signals generally disappear into the noise. With hardware AGC unchecked and the slider set to 49DB, the graph sinks to -90DB and the signals disappear altogether. There are a number of signals I can pick up loud and clear on an ordinary FM radio that are just not present or very noisy on the RTLSDR. On FM via the UV connector, connecting the antenna earth does improve reception, but there is still a lot of noise.

Switching the presents between Fast/Medium/Slow/Off makes no difference.

[cdev]

The antenna you use is all important..

But reading through the stuff online it appears that the design of this device likely could be improved substantially - that they could have done better.

I have the device they were selling before, but I assembled it myself. They didn't assemble it.  And it was a major PITA, so much so that I have avoided doing anything with it since then.

Generally, it holds its own against any of the similarly priced (consumer level) portable shortwave radios - including the quite decent better Tecsun models, with a decent antenna. Which they also need, to be honest.

But from what I am reading it likely could be improve quite a bit.
Here is their doc file..

Maybe this link will work:
http://115.28.16.44:81/file/3156.doc

That is their instructions for it.

There also seem to be quite a few videos online of the earlier device, its construction and operation. Look at how it works in the Youtube videos of it.  It is what it is.
It seems that the bypassing and their choice of parts could be improved a lot.


Start out learning a bit more about rtlsdrs, notably that they are vulnerable to noise and USB connections are inherently noisy already with VHF. On HF they are even more noisy.

I put lots of ferrite cores on my USBs when I use SDR.

Also to get decent reception I needed a long wire, a good ground, and a 9:1 unun.

To be honest with you, despite some moments of decent reception everything RTLSDR have always made for marginal shortwave receptionfor me where I live because its so noisy.

Ive had much better luck with a different SDR probably because it has much more dynamic range and better rejection of noise and out of band signals.

Lacking an antenna tuner, a unun makes all the difference.

That applies pretty much to all my receivers, all of which are SDRs. I live in a very noisy suburban location.

All of that said, likely anybody with any decent amount of RF experience could likely find areas for improvement on the Ba5SBA design.

Russian web pages go into length about this but not having actually trying any of these changes.

I would try using a different RF transformer. I would use a binocular core, or perhaps two or even four side by side instead of a toroid. Thats a better broadband transformer. And I would try more turns on the balanced input side

You should make a post asking for other people's opinions on what the best transformer to use would be. 

Maybe you might also want to experiment with different materials, or maybe even commercial RF transformers..

I bet you could improve the performance significantly. Iam reluctant to on mine because there si a limit to how many times the little wires can be resoldered I am sure. It will become impossible to fix.

[WaveyDipole]

I have now traced out the HF antenna circuit. There is a semicondictor device in the circuit and I'm not sure whether this is a transistor or a pair of diodes so I have attached a images showing both possibilities. The circuit does have some similarity to your hand drawn one in that it has two inductors on the input and a number of capacitors to form a network. However there is no 1:2 impedance matching transformer. Instead we have a 3 pin semiconductor marked J3.  There seems to be no ESR protection on the input either. In addition to the two connection points for pins 4 & 5 of the RTL2832, there appears to be a possibility for a 3.3v power feed taken from the regulator on the SDR dongle board. The points marked 'X' were all left disconnected in this example.

I note that your hand schematic has some additional caps on the USB feed and you also suggested adding clip on ferrites. There is a 22uF cap drawn next to the torroid inductor. Is it in parrallel with it? The primary of the torroid in your schematic seems to have a power feed - was this to drive an active antenna? There is also a capacitor at the top of your diagram, one end of which seems to connect to earth. Was the other connected to the 5v line for filtering?

I would apprciate it is someone could give me some idea what the SMD part marked 'J3' might be? The references I found show this to be a two pin device, either a varicap or other type of diode. The device here has 3 connected pins.