RTL-SDR - an SDR receiver from a cheap DVB-T tuner on a USB connector

I have a temporary cooling system as you can see in the pictures in the topic below and the tuner works very well at frequencies above 1500MHz.
https://www.elektroda.pl/rtvforum/viewtopic.php?p=18678813#18678813
The point is that I am doing a lot of experiments with tuners and it's a shame to buy something more expensive and burn it. That's why I asked about TCXO. Another experiment here. Converting to HF with capacitors, as you can see, the tuner is doing well. Reception on samoróbce 5/8 antenna on 2m without connected ground.
https://drive.google.com/drive/folders/1kCL27eIarPhaitkTCKGKjaVB0w1G0YBv?usp=sharing

I would give a Chinese link, but I have rules, I do not publish links to auction sites, because in a few years they will be out of date. For literally PLN 60 you have direct sampling, you have a bias tee and you have TCXO 0.5ppm and a metal casing and a thermal pad between the plate and the casing. There is decent filtering on the voltage input from the USB. This is not a tip-top solution, because it costs PLN 100. Why bother? I did a lot of experiments 8 years ago, when there was nothing else but a TV and radio dongle. I even made a solution that yes, the casing was a cooling element, but I used copper wire as a thermowell. And so I went down to 35 degrees. I used the box for the TV head as the housing. But when I bought SDRPlay it all turned pale. And so, as I wrote earlier, MRI SDR can be purchased for PLN 150. You can buy RTL SDR with all possible variations. The Chinese have already done such miracles that only the music box in the plug is missing and the stereo speakers. You won't come up with anything new either. Well, unless you design the board from scratch and solder a new receiver with switchable input filters. Only the R820T doesn't have the ability to work with multiple filters. And MRI001 does and is therefore better. However, it seems to me that the refrigerated R820T2 is more sensitive than the MRI001 above 1GHz. But there is no starting point below. I do not know how the RTL system works with the R820T2 on an ordinary generator. As you know, calibration is possible. But calibrating on VHF does not mean calibrated on UHF. The swimming itself can be calmed down by the stabilization of the operating temperature of the system, but ... Unfortunately, the tuner system heats up terribly, it must be cooled. Now cooling should be resolved in such a way that there are no temperature fluctuations, because this causes deviations. After warming up the receiver, you would have stable operation. But I admit that without a decent TCXO, it is rather art worthy of circus performances. I ordered mine for the Malachit receiver. I gave PLN 48 with the shipment. It is a classic cube with dimensions of approximately 2cm X 1cm X 1cm. Even in Malachite there was a problem with soldering something like this and I had to combine with wires soldered to the pads of the old generator. It will be even more difficult on the RTL SDR board. You would have to make a housing with cooling systems and place TCXO next to the board and glue, solder to the whole housing. Then if you had used a similar TCXO module you would have a stability of 0.1ppm (in the more expensive variant). This allows you to experiment with the GPS signal. But as I wrote earlier, without a spin for PLN 60, it will do well, and for grazing PLN 100. Where for a hundred songs, he doesn't sing or dance, and so he has everything, even the preamplifier.

Added after 12 [minutes]:


here you have a board of such a receiver for PLN 100. With TCXO, preamplifier, direct sampling and quite decent filtering of the supply voltage. Murmurs of goodies. The option for PLN 60 is only a bit poorer.

Added after 10 [minutes]:


This is an option for PLN 60, you have no direct sampling in this case. There is also no preamplifier. There is TCXO 0.5ppm and quite decent workmanship, cooling and metal housing. By the way, direct sampling is a lame option, the converter is better, as already.

My problem is that I like to test, combine and do something myself. Once upon a time, a dozen or so years ago, I even devoted my time to AVT sets and made various designs of speech compressors, CB microphone amplifiers, KF AVT-2133 receiver and AVT-115 converter still working. Maybe someday I will be tempted by something better but working up to at least 6GHz and with a width of 10MHz.

Get yourself HackRF. You can get a clone cheaply. It's a TRX but the receiver is poor. As a measuring instrument, it can be quite interesting.

After testing at the frequency of 1546MHz two copies of the tuner in a blue housing, one of which had a large quartz in an oblong housing, and the other a small rectangular one, I can clearly say that the best tuner for listening in that band is the tuner in the older version with large quartz.

Test made on SDRConsole, SDRSharp, JAERO.
I observed the shifts in frequency on the indicator in the lower right corner of the JAERO program. Picture below.

Older version with large quartz:

Tuner without cooling . After reaching the maximum temperature - ppm increased by 5 compared to the initial one, after heating it floats quite stable up to 10Hz up and down. JAERO program decodes data without errors.

Tuner with fan cooling . It reaches the maximum room temperature - ppm quite stable does not differ much from the initial float up to 5Hz up and down. Jaero decodes the data without errors.

A newer version with a small rectangular quartz.

Tuner without cooling . After reaching the maximum temperature - ppm increased by 2 compared to the initial temperature, after heating it is relatively stable but it floats around 30Hz up and down. JAERO program decodes data but sometimes shows data errors, problem with synchronization.

Tuner with fan cooling . It reaches the maximum room temperature - unstable ppm differs significantly from the initial one, it floats around 60Hz up and down. Jaero shows a lot of errors, does not decode data, synchronization problem.

Cooling alone is not everything. There must be no temperature fluctuations. The system needs to warm up. Perhaps after about half an hour, the frequency fluctuation is even smaller.

I have just been doing tests for a couple of days on the newer tuner and the older one. The tuners worked for 8 hours without cooling as you wrote to warm up the systems and unfortunately the older tuners have a jump of about 1Hz and the newer ones up to 6Hz and with the older ones, no matter how it is cooled or closer or closer, it does not float as much as the newer ones. I have a question, has anyone tried to do such tests on the RTL-SDT V3 donglu? Unfortunately, I do not have it and I have no way of comparing his work. I would like to add that the newer ones in the blue casing do better without cooling than with cooling, but unfortunately you cannot listen to them as high as with cooling. on the older version, with cooling, it was possible to listen to a maximum of 1752MHz.
I would like to add that I found another old tuner that was lying in the box untouched and it is also more stable than the new ones, even though it has never been used before.

One little thing. The quality of the used generator or quartz has an influence on the stability of the work. Cooling allows you to receive frequencies higher than perhaps 1300MHz. In order to be stable, the generator should be thermally insulated, because any change in the AMBIENT temperature will have an impact on the frequency of the receiver. The heating process is important because the temperature of the device rises. The plate heats up and this affects the generator or the quartz itself. When everything warms up, there are no thermal movements. However, they will if you are blowing it all. But if you insulate what you need to do and only cool the R820T / 2, this should help. Of course, miracles cannot be expected, because the device was not designed for communication purposes. Even when it comes to the parameters of home devices, they can leave a lot to be desired with these tuners. It's all about selecting components. In these tuners, they were basically absent and the manufacturer went into economy, not quality. Therefore, it is naive to expect a feat from the tuner for $ 8. What's more, the older models can be better, because the focus was still on quality. Later, competition forced the price to be reduced and this led to the use of the cheapest elements. Nobody who buys such a dongle asks about frequency stability. It turns on, picks up, picks up, that's fun. If you absolutely want stability, you have to use TCXO. Without it, you'll always have some swimming, and in the case of these televisions, more bigger than smaller. And you won't find anything here that works as communications equipment, because the $ 10 device is not one, although we wouldn't close our eyes and melt into dreams. There will be no oil in this soil, and you have probably set yourself up for diamonds. There is not even any carbon here, there is only peat and nothing else. Do you want something better? Buy MRI SDR. You want RTL for satellite purposes, for GPS analysis, it must be TCXO. You can't jump higher than your kidneys.

I do not require miracles from these tuners, I only checked the stability of these devices by trial and error and described them here. I do all this as a hobby and out of pure curiosity, and I am not going to throw away money on something that I will not use in the long run. A simple blue tuner is enough for me to test the software, even if I accidentally burned the R820T2, and the RTL is still getting closer, it is always possible to convert it to a KF and give it to one of my friends amateurs. Best regards.

A few words of explanation on the philosophical side. Philosophy is not a marginal science. Even though he despises her. Pay attention to the time. Time is a thing where we only travel forward and at a certain speed. This means that we are traveling in line with the passage of that time. We can't take it back. For this reason, many have found time to be invaluable. Losing time is an irreparable damage. You waste your time focusing on something that doesn't make sense. If you were to convert RTL into some "performance" receiver, there would be some constructive conclusions. Now you can only say that the component with poor parameters sucks from a radiocommunication point of view. Make point cooling, it's not difficult to cool mainly the R820T / 2 but also the RTL system, although the latter does not heat up dramatically - unlike the Raphael. Do you like to play? Use a different generator. Discard the defective quartz and put the generator in. Give up stupid things in life, save for synthesis! What for? And you can check how the system will behave when we change its frequency. Or maybe ... it will be able to work up to 2000MHz? I do not know. The synthesis / source of the signal in oblique eyes costs PLN 140-200 depending on the version. You will use the synthesis later for a HF receiver, whatever. The TCXO will cost about the same as the tuner, but ... when you buy such a generator, you will only have a very stable RTL SDR. Which is not insignificant. Studying the Doppler effect is basically impossible without good TCXO. You can have up to 0.3ppm for less than PLN 50. So what is this all about? About testing the sugar content of sugar?

Does anyone have and tested these sdr receivers? Especially in the 1700MHz range (HRPT NOAA).

I have the lower one, at lower frequencies it does not differ from SDRPlay RSP1. Well above GHz, I do not know, because so far I do not have an antenna. These are receivers on the MSi001 chip. These receivers beat the RTL SDR with one hand, and it's tucked in your pocket. They have much better dynamics of reception. The bandwidth is 10MHz, not 3 as for RTL SDR. The reception is from about 10KHz to 2GHz, dribble, no holes, no need to use a converter. The stability is 0.5ppm. Both receivers have bandpass filters and an input amplifier, just like the SDRPlay RSP1. Both are detected by the software as RSP1, they require the same drivers. I use SDR Console, although the original SDRPlay manufacturers recommend SDR Uno. The latter is less readable and complicated for me to configure, but it is a matter of taste. The bottom receiver is more expensive but worth the price. The upper one can be bought even for something over PLN 150. I saw photos from the inside and I will say that the performance ... is similar to the cheap RTL SDR, which does not change the fact that it is still MSi001. He should walk on short and VHF without making scenes However, my practice shows that a well-cooled Rafael 820T / 2 on the 23cm band works noticeably better. I don't know how it is around 2GHz, at 1.2GHz it is just like that. As we know, the receiver does not have to work linearly and it can be better, but unfortunately it can also be worse. However, I do not remember exactly whether I turned on the amplifier. I took measurements in very difficult conditions, i.e. at home, on the ground floor, measuring the signal of the SR8LHL beacon, which is about 6 km away from me, and there is a forest on the way and there are houses. I used a half-wave loop dipole for reception. As the reference receiver I used the AOR AR8000, which incidentally was the worst of all three in this band. At the same time, it should be emphasized that ... RTL SDR in the 1.7 GHz band may have serious problems because ... due to interference emitted by mobile telephony, reception may be difficult. The receivers you mentioned have a high-pass filter for this band, which theoretically does not eliminate cellular interference, but eliminates the interference from the bottom of the band. Moreover, the MSi001 chip has separate inputs for LW-MW, HF, UHF and UHF L band. It is a fairly advanced "multi purpose" chip. The receiver has two GAS-FET switches at the input, which are used to switch filters. There is an RSP1 diagram on the electrode, because I made it available. It is almost identical to the Chinese receiver. Almost, because the Chinese tried to improve something, they reportedly made a manual bias tee, activated with a switch. Also ... an active GPS antenna can be connected to this. I am looking for a ready QFH antenna plan for 1.7 GHz, because I am also interested in receiving satellites from this band. Unfortunately, the plans are only available for the GPS band and 137MHz. If you had something like that, it would be nice if you shared. Adding an amplifier to the QFH would have very good results with the MSi001 chip. The R820T2 could get wasted and would probably require an additional input filter. In the link you have some information from the manufacturer. Anyway, these receivers are not RTL SDRs, in my opinion a much higher shelf, not the highest one, this is the beginning of the middle one. But the high, middle shelf is also on the MSi001 ...

http://play.fallows.ca/wp/radio/software-defined-radio/mirics-msi3101-chipset-heart-of-sdrplay/

I was wondering how the RTL-SDR V3 would handle the VHF tuning.
I think he's doing pretty well.





The resistor in series is probably unnecessary, but I preferred not to connect it directly to the output of the tube head (the level of about 200mVpp, if you believe the oscilloscope).
How to apply decent FM modulation, it even makes sense on the graph.
In any case, there is no need to tune in this case because it is tuned correctly, which also results from the FFT graphs on the digital oscilloscope.

Once upon a time I used SDRPlay as a spectrum analyzer and I will say that I was positively surprised. I had to check if the heterodyne in the receiver was working. And the program showed the signal, level and ... harmonics nicely. As for tuning IF it is at this point that SDR does for IFR. If you are tuning the output in the head, then definitely. You won't generally tune the IF circuit, because you need a generator for that, in this case 10.7 and possibly further, if there is another intermediate. But as for the signal from the head itself, it will work.

don_viking wrote:

I do not require miracles from these tuners, I only checked the stability of these devices by trial and error and described them here. I do all this as a hobby and out of pure curiosity, and I am not going to throw away money on something that I will not use in the long run. A simple blue tuner is enough for me to test the software, even if I accidentally burned the R820T2, and the RTL is still getting closer, it is always possible to convert it to a KF and give it to one of my friends amateurs. Best regards.

Below is a photo of both older and newer tuners. At first glance, you can see that the newer version is poorer in elements (more economical).

There are no power supply filtering capacitors, so nothing prevents you from soldering them. As already mentioned above @^PITER^ K. the problem is temperature and more precisely its fluctuations. The HC49 housing is physically larger, so it should have greater temperature inertia, in addition, it has better contact with the PCB - greater thermal mass. As the surface of the SMD quartz "cap" is relatively flat, you can try to coat it with thermo-conductive paste and put a nut or screw on it and additionally isolate it with a piece of polystyrene or foam.

I used to browse the pages for a description of what is installed and in what tuners, whether it is xtal or tcxo. I found such a picture on the page below.
Link

I wrote a few posts above about this topic, Has anyone physically checked what is installed in individual tuners? Theoretically, immediately after turning on the tuner with TCXO, the received signal should not flow, but stand still. Any other behavior of the signal proves that it is not TCXO or I am wrong.

Added 4/17/2021

I sold this supposedly TCXO from LV5TDLX to this blue tuner and the difference between the tuner warmed up after 15 minutes of operation and cooled after 30 minutes is only 2PPM. I would still have to adjust the parts at the input of the wrap, because now the initial shift is -86PPM previously on the earlier quartz was + 35PPM, but it doesn't bother me if the reception is stable. During the operation of the cooled set, the stability is about 1Hz.



As you can see in these three photos, the PPM error was stable during the cooling of the set in the 20-minute interval. Below is a photo of the tuner modification.

After testing the tuner, I can say that it is currently very stable and when receiving without cooling, the synchronization is around 1 Hz, it floats around 0.5 Hz up and down. With cooling, it floats a maximum of 2 Hz up and down, i.e. a total of 4 Hz. When cooling the tuner on experimental drivers that were once available on the network, see the description on this website, I was able to set the reception to a maximum of 1770MHz.



I also did tests during cooling on other experimental controllers on the L-Band, the description is on this side. The maximum frequency that I was able to receive was 1809MHz above the tuner stopped receiving, despite the frequency shift, only the image was visible as in the last photo, which was standing in one place.



Below is the link to the drivers if anyone would like to try. I would like to add that if the tuner is cooled down too much, it stops working and the system shows that it cannot detect the device and the tuner is unavailable.
Drivers

And where is the power supply of this TCXO?

I wrote earlier that I mounted it "kind of TCXO". In fact, it results from the fact that it is quartz in a smaller housing. But it is more stable than those newly fitted to these tuners. I measured this "TCXO" with a multimeter before I soldered it and there is a short circuit between 2 and 4, i.e. it is not TCXO. It seems to me that there should be no short circuit with TCXO, since the voltage is to go to 4. The 1 is on the lower left, 4 on the upper left looking at the photo of my tuner.

don_viking can you tell from what satellite position (from which satellite) you are feeding SDR and JAERO?
From the lists that I found on the net, I do not see a satellite to Europe - maybe I was looking wrong.
As for quartz in cheap SDRs, they are quartz in small housings - the same as in radio remote controls. I've played around with it, and often the older quarks are more stable. I also used generators. The problem is the heat released by IC.

There are several satellites, but I listened to Alphasat I-XL the most.

Hi. Some time ago I crashed with SHARP SDR and it keeps showing this info:

I was installing a newer version of SDRSHARP in a new directory - still the same.
The hardware works without problems on SDRCONSOLE - but I miss SHARPA functionality for decoding the METEOR satellite.
Will you help?

And what do you have in Device after opening the gear?

SDR sharp, regardless of version, crashes on libusb-1.0.dll
Sometimes a minute, sometimes a few.
I have tried different versions of the file but there is no change.
Windows7x64
Any ideas what you can do about it?

SP5IT wrote:

Windows7x64

Two different computers running the W7 32bit Pro and I never crashed this file. Used with dongles as well as with this . There may be a collision with other / s USB devices.

device - can't open anything.

Missing / badly installed drivers. Zadig and install

SP5IT wrote:

Zdiag

Zadig

This will not be a problem with the setup because the tuner works on sdrconsole as stated above. What tuner you have exactly, write the name or send a photo, it can also be tuner plates. You probably have a problem with the rtlsdr.dll file. If you have the wrong file, you won't go into sdrsharp. I had that too. I also had the same problem as I connected the tuner converted to quartz via the HUB under the computer and sometimes it did not work in sdrsharp when I cooled it too much, it had a similar message, and on sdrconsole it worked without any problems even during intense cooling. And when I connected it directly through the extension to the computer, it was detected in sdrsharp and it worked flawlessly. Libusb-1.0.dll and librtlsdr.dll have nothing to do with it. There is another option in the sdrsharp program when choosing a tuner, you have the RTL-SDR (820T) option, try it may start.


Jap I sent you my version of sdrsharp to PW, check if it will work for you.