RX-888 Unboxing and mini-review

[ChrisSmolinski]

I ordered an RX-888 via eBay October 6th, it arrived today, ten days later, not too long of a wait.

Not a very large box:



Not much in the box, the radio wrapped in some bubble wrap (not taped), and a hefty USB3 cable:







Some pictures of the RX-888:







I got SDR Console v3.0.25 running, details on how to do this and install the Cypress FX3 driver here: https://www.sdr-radio.com/rx-666


No antenna, you can see there are a few birdies:


20 dB  attenuation knocked them down a little:




0-30 MHz, 20 dB attenuation, 58 ft T2FD:




31 meter band, 58 ft T2FD, 20 dB attenuation:




MW band, 58 ft T2FD, 20 dB attenuation:




MW band, 58 ft T2FD, no attenuation:




MW band, 1000 ft Loop On Ground antenna:



The 49m band just after 1900z, stations from Europe are just starting to fade in:



I'm going to use it on the 49 & 48 meter bands tonight checking for Europirates, and compare to my other receivers.

While I have seen reports of it running hot, it is just barely warm.  Maybe the Yak scattered power helps. Oh wait, that's another Chinese product 

[ChrisSmolinski]

Some photos of the PCB:





In addition to the power LED visible through the very small hole in the enclosure, there are three other LEDs, MODE A, MODE B, and OVFL.  OVFL is presumably overflow, I have been unable to get it to light up. Maybe I need more antenna.  One of the MODE LEDs lights very dimly when tuned to 0-30 MHz (maybe it is flickering at a very high rate?). Neither lights when tuned to VHF/UHF.

Is it really a 16 bit ADC? I guess I would need to pry the heatsink off the IC, not really interested in doing that. And even if the IC says it's a LTC2208, who knows what it really is 

When opening the enclosure I did notice the nuts on the SMA connectors wee barely tight, so something that any other owners may want to check out.

[ChrisSmolinski]

The RX-888 is somewhat different than most direct sampling SDRs. Rather than having an onboard FPGA or other circuitry to decimate/reduce the bandwidth and produce an I/Q stream, it always streams raw 16 bit ADC samples over USB to the computer at the full 65 MHz rate. Your computer software is then responsible for performing these tasks, on top of the other signal processing to produce a waterfall, demodulated audio, etc.

This reduces the complexity of the SDR and presumably helps lower the cost.  As computers are significantly more powerful than when SDRs first appeared on the market, this is probably a good trade-off.

Now at 2030z Europe is really starting to come in on 49 meters. So far, performance seems quite reasonable, especially for the price. But I'll need to wait for some DX targets to see how it performs with them.

For those interested, I bought mine from eBay vendor tztmotor

[dewdude]

The lack of FPGA makes me have to walk away. The price is good for a 16-bit unit...but not when I have to do massive hardware upgrades.

[dxace1]

Chris -- can you put some audio up for us on the Slack group....

[ChrisSmolinski]

Now for VHF/UHF reception. 

First, I calibrated the internal ADC clock in the RX-888 (SDR Console has an options window for doing this) by adjusting the PPM value while watching 10 MHz WWV. I found in my case a setting of +0.5 ppm was necessary, obviously your mileage may vary.

There is a separate gain setting for the R820 tuner IC, in my case a setting of 25 worked best, again you will need to experiment to find your ideal setting, based mostly on your antenna. One note - I found the SDR Console did not automatically set this gain in the RX-888 itself when I launched it again, even though the slider was set to 25, I had to adjust it manually to force it to send the setting to the radio.

My VHF/UHF antenna is a discone up 55 ft or so in a tree, with a pre-amp located at the antenna, see this thread: https://www.hfunderground.com/board/index.php/topic,61347.0.html

In the shack this antenna feeds several radios through splitters, which does reduce signal levels somewhat vs just feeding one receiver.

First, here is the RX-888 with no antenna, so you can see the effective bandwidth, about 9 MHz, limited by the R820 tuner IC. There is a noticeable increase in signal levels towards the higher frequencies. This is also apparent when looking at HF signals:



Next looking at local TV channel 8. You can see the ATSC pilot signal at 180.31 MHz. Also up at 186.31 is the pilot for channel 9 which is much further away near Washington DC:



Here's a view of the NOAA Weather Radio stations:


And finally the middle of the FM band:

[kris]

     A few more notes:
- on my RX, the IC in the top right corner has a heat sink like the one on the left
- integrated circuits and voltage regulators get hot> you can't handle it with your hand
(I think about screwing a small fan housing to the heat sink like in computers, 5V power supply derived from the USB3 socket, holes in the housing would be useful to ensure air exchange)
- flexible white thermal conductive foam on PCB is very delicate and its thin layer tears when inserted into the housing.
When doing this, insert a strip of foil between the foam and the casing, and then pull it out.
    Inspired by Chris's example, I installed SDR Console.
I am learning how to use it and I see an advantage over HDSDR.
 - installation is extremely simple and fast
- simpler operation, pleasant design
- richer menu
- probably lower noise level
- better operation of the AM Synchro mode
    I am still checking the phenomenon of aliases, which in HDSDR irritate me a lot on HF. Perhaps Console's algorithms will be better than HDSDR in this physical phenomenon. The creator of the original BBRF103 receiver, Oscar Steila, developed and introduced filters: Notch FM Band, BPF FM Band and LPF 32MHz. You need to check this as well as the LO settings in relation to the band you are interested in.

[4nradio]

Hi Chris, thanks for posting your RX-888 first impressions and the interesting screenshots.

My own RX-888 is showing fewer spurs and a lower noise floor when viewed in the same 32 MHz bandwidth with no attenuation in SDR-Console:



Here is the response with 20 dB attenuation:



I wonder what would account for this difference? Maybe I have some option or setting that varies from your own SDR-Console installation (I'm using 3.0.26, build 2017). I hope it's something like this, rather than quality control in the receiver.

The skeptic in me says "QA? whazzat??" when it comes to small-batch production runs from unknown electronics houses overseas, but I'm hoping it's just a software configuration matter 

73, Guy

[Ray Lalleu]

The RX-888 is somewhat different than most direct sampling SDRs. Rather than having an onboard FPGA or other circuitry to decimate/reduce the bandwidth and produce an I/Q stream, it always streams raw 16 bit ADC samples over USB to the computer at the full 65 MHz rate. Your computer software is then responsible for performing these tasks, on top of the other signal processing to produce a waterfall, demodulated audio, etc.

This reduces the complexity of the SDR and presumably helps lower the cost.  As computers are significantly more powerful than when SDRs first appeared on the market, this is probably a good trade-off.

Really ? With a tad of overhead, that makes a flow to the PC around 1300 Mbits/sec, so what is the line to the PC ?
And how can a PC process 65 Msamples/sec?
What is the equipment of your PC for the RX888 ?

[ChrisSmolinski]

The RX-888 is somewhat different than most direct sampling SDRs. Rather than having an onboard FPGA or other circuitry to decimate/reduce the bandwidth and produce an I/Q stream, it always streams raw 16 bit ADC samples over USB to the computer at the full 65 MHz rate. Your computer software is then responsible for performing these tasks, on top of the other signal processing to produce a waterfall, demodulated audio, etc.

This reduces the complexity of the SDR and presumably helps lower the cost.  As computers are significantly more powerful than when SDRs first appeared on the market, this is probably a good trade-off.

Really ? With a tad of overhead, that makes a flow to the PC around 1300 Mbits/sec, so what is the line to the PC ?
And how can a PC process 65 Msamples/sec?
What is the equipment of your PC for the RX888 ?

I'm running a 3.2 GHz i7-8700.  It seems to handle the RX-888 fine, which only uses a small percentage of available resources and CPU time. SDR Console also makes use of GPU processing.

[kris]

My own RX-888 is showing fewer spurs and a lower noise floor when viewed in the same 32 MHz bandwidth with no attenuation in SDR-Console:

   Hi 4nradio = It looks like you have no antenna connected> noise level below 120dBm and no HF signal. Check the cable!
     You don't get any signal or your computer / program freezes.
Show what you have on a lower bandwidth, but you guess you've done it?
 
    My PC i5 dual core, 8GB RAM, 2.5GHz clock can't handle receiving width above 8MHz.
With a width of 1 MHz, CPU usage is 15-20%


       I was dissatisfied with the noise level of around 85 dBm in the 49m band. I thought it was the fault of the antenna and the environment. Today I just discovered that my 4.62A power supply has a noise floor of 10dBm more than the more powerful 6.7A power supply. In addition, the active speaker power supply adds about 3dBm of noise. Check your power supplies!

         My PC exhibits strange behavior after disconnecting the AC adapter and switching to battery. The CPU load quickly increases to almost 48% and the noise level by at least 10dBm and the data stream processing is stuck.
   Don't panic and wait, this state calms down after a while. The CPU load drops to normal, signal processing is back and the noise level is over 5dBm lower than it was on the power supply.
   In this configuration, you can already understand quite poor signals.

[4nradio]

Hi Kris,

I had no antenna connected to my RX-888 intentionally, as I was showing Chris Smolinski the different noise floor my receiver has, and the fewer spurs. This was for comparison to the similar screenshots he posted of his RX-888 without an antenna connected.

My RX-888 works great with a low CPU & GPU load, both on SDR-Console and HDSDR. If I run the receiver on the full 32 MHz bandwidth, CPU is around 25% as well as GPU. Normally I use 2 MHz bandwidth for medium wave coverage however (about 7% CPU in that situation). I am using a new 17-inch "gaming" laptop, Acer Nitro 5 series.

I'm very happy with it. I like the RX-888 so much I just bought and took delivery of another RX-888 for backup and other uses 

73, Guy

[Jari Finland]

Comparison test.
RX888 vs Perseus.

https://youtu.be/tC50LJg0CrM

I compared RX888 and Perseus in the same time with the same antenna. In night time listening the difference is not so clear. One can argue that RX888 is slightly behind in sensitivity. However, in morning during daytime listening RX888 cannot compete. In spite of in-built amplifier both the sensitivity and dynamics are lacking. Especially clear this is in the case of Utvarp Föröya 531 kHz.

[alpard]

Comparison test.
RX888 vs Perseus.

https://youtu.be/tC50LJg0CrM

I compared RX888 and Perseus in the same time with the same antenna. In night time listening the difference is not so clear. One can argue that RX888 is slightly behind in sensitivity. However, in morning during daytime listening RX888 cannot compete. In spite of in-built amplifier both the sensitivity and dynamics are lacking. Especially clear this is in the case of Utvarp Föröya 531 kHz.

But Utvarp Föröya 531 kHz. is very near from your location?
Only about a few hundred miles?