Airspy R2 and SDRTrunk not working at 10MHz

[carnch]

I've been a scanner junkie for years and very recently decided to get into the SDR scene so I bought two of the cheap $20 RTL-SDR dongles from RTL-SDR.com and was able to get them working in SDRTrunk. I've learned a lot about these and how they work by reading the forums here and through Google searching. I primarily listen to our local 800Mhz P25 phase 1 county services (LE/Fire/EMS) and I've had good luck decoding our LSM/CQPSK using the two cheapies with SDRTrunk.

The system I am scanning spans about 6MHz (854.0425 to 859.9375) so I decided to pick up an Airspy R2 since it has the bandwidth to cover the entire 6MHz spectrum I want to scan. SDRTrunk has support for it so I assumed there would be no issues switching from the RTL-SDRs to the Airspy. Unfortunately, I'm not having much luck using the Airspy for decoding this system with the 10MHz sample rate selected. If I select the 2.5MHz sample rate it will track and decode the voice channels as expected but of course it doesn't cover the entire bandwidth so calls are dropped/not decoded. As soon as I switch back to 10MHz it tracks the entire spectrum and when I click on the control channel line I see under events each of the talk groups being decoded but I get no voice. I will occasionally get a bleep or a bloop (it sounds kind of like an '80s video game), similar to using my Pro-106 scanner which can't decode our simulcast correctly due to lack of CQPSK support. When I put the RTL-SDR dongles back in I get clear voice. Of course the sample rate for those is 2.4MHz.

My PC is an 8-core Xeon with 16Gb of RAM so system resources are not an issue. I've tried changing the Master Gain and using different USB ports as suggested by others but no luck. I've read where others have had issues with the Airspy R2 stating a defective TCXO chip but before I try to RMA this thing I want to make sure I'm not missing anything. I've run SDR# and I am able to hear the control channel for this system but only if I turn the GAIN to its highest level which is 21. If I tune to an FM broadcast channel in SDR# with the sample rate at 10MHz it's very distorted but as soon as I switch to 2.5MHz the broadcast is clear.

Am I missing a setting or something needed to make this work at 10MHz?

 

[br0adband]

Usually attempts to use Airspy at the full (rated) capability of 10 MSPS is frought with issues for many people, most notably older hardware and especially issues with USB 3.0 drivers and hardware - USB 2.0 simply can't do the job so it's not even worth the attempt. The makers of Airspy created some rather simplistic command line tools to troubleshoot issues with using the maximum sampling rate and you can find them here:

https://github.com/airspy/airspyone_host/wiki/Troubleshooting

My suggestion is do the simple test mentioned near the top of the page (the Windows based tools) and see if you get issues with the results. You can have even a Kaby Lake box with tons of RAM and the best processor Intel has to offer these days (or a Ryzen based platform build) and it won't matter all that much, what does matter is getting that data from the Airspy - which is quite capable of 10 MSPS without issues - into the PC for it to make use of.

That's where the issues lie more often than not, in the data throughput from the Airspy to the PC, so those troubleshooting tools should help you get some more info about resolving the issue. I can't imagine the issue(s) you're having would be related to the machine itself but more specifically the USB connection/pathway - if the results of the troubleshooting show issues with it not being able to provide adequate throughput = 10 MSPS solid and sustained without dropouts then you'll probably be looking at a few options: better USB cables (they can and do matter, really) with better shielding (yes that matters too, actually), a standalone PCI-E USB 3.0 card (with better driver support and more capable controller hardware for higher bandwidth), and so on.

Good luck...

 

[SCPD]

Usually attempts to use Airspy at the full (rated) capability of 10 MSPS is frought with issues for many people,

No problems here running an AirSpy at 10 msps on USB 2 ports - on multiple computers.

most notably older hardware and especially issues with USB 3.0 drivers and hardware

This is true. USB 3 ports with proprietary drivers on Windows 7 have caused much grief. If you are on Windows 8.1 and later, you should be okay. USB 2 ports will work fine all the way back to XP.

USB 2.0 simply can't do the job so it's not even worth the attempt.

This is not true. The AirSpy is a USB 2 compatible device - at all sample rates. Plugging it into a USB 3 port causes that port operate in a USB 2 backward compatible way.

Pointing the OP to the troubleshooting page is an excellent idea.

If the AirSpy is plugged into a hub, consider moving it to it's own USB port. If you have problems on a USB 3 port, move the AirSpy to a USB 2 port.

 

[fn_form]

Thanks for bringing this up, I need to finish troubleshooting my own R1 setup.

Is the Airspy R2 working normally with Unitrunker?

 

[SCPD]

Is the Airspy R2 working normally with Unitrunker?

Short answer is yes but please start a new thread for any Unitrunker questions.

 

[carnch]

Usually attempts to use Airspy at the full (rated) capability of 10 MSPS is frought with issues for many people, most notably older hardware and especially issues with USB 3.0 drivers and hardware - USB 2.0 simply can't do the job so it's not even worth the attempt. The makers of Airspy created some rather simplistic command line tools to troubleshoot issues with using the maximum sampling rate and you can find them here:

https://github.com/airspy/airspyone_host/wiki/Troubleshooting

My suggestion is do the simple test mentioned near the top of the page (the Windows based tools) and see if you get issues with the results. You can have even a Kaby Lake box with tons of RAM and the best processor Intel has to offer these days (or a Ryzen based platform build) and it won't matter all that much, what does matter is getting that data from the Airspy - which is quite capable of 10 MSPS without issues - into the PC for it to make use of.

That's where the issues lie more often than not, in the data throughput from the Airspy to the PC, so those troubleshooting tools should help you get some more info about resolving the issue. I can't imagine the issue(s) you're having would be related to the machine itself but more specifically the USB connection/pathway - if the results of the troubleshooting show issues with it not being able to provide adequate throughput = 10 MSPS solid and sustained without dropouts then you'll probably be looking at a few options: better USB cables (they can and do matter, really) with better shielding (yes that matters too, actually), a standalone PCI-E USB 3.0 card (with better driver support and more capable controller hardware for higher bandwidth), and so on.

Good luck...

I've downloaded the Airspy tools and ran the airspy_rx and the results came out as

Streaming at 10.035 MSPS
Streaming at 10.007 MSPS
Streaming at 9.994 MSPS
Streaming at 9.986 MSPS
Streaming at 9.982 MSPS
Streaming at 9.980 MSPS
Streaming at 9.980 MSPS
Streaming at 10.059 MSPS
Streaming at 10.020 MSPS
Streaming at 10.000 MSPS
Streaming at 9.989 MSPS
Streaming at 9.984 MSPS
Streaming at 9.980 MSPS
Streaming at 9.980 MSPS
Streaming at 9.982 MSPS
Streaming at 10.044 MSPS
Streaming at 10.003 MSPS
Streaming at 9.992 MSPS
Streaming at 9.985 MSPS
Streaming at 9.981 MSPS
Streaming at 9.978 MSPS
Streaming at 10.077 MSPS
Caught signal 0

User cancel, exiting...
Total time: 23.0535 s
Average speed 10.0034 MSPS IQ

so it does drop below 10MSPS but not by much. I've tried using a separate USB2.0 PCIe and a USB3.0 PCIe and both work but with a lot of drops and I get a lot of these in the SDRTrunk console window:

16:14:54.629 DEBUG s.t.u.USBTransferProcessor - Airspy 644064DC-23288FCD - buffer overflow - temporary pause until processing catches up
16:14:56.582 DEBUG s.t.u.USBTransferProcessor - Airspy 644064DC-23288FCD - buffer overflow cleared - resuming normal processing
16:14:59.735 DEBUG s.t.u.USBTransferProcessor - Airspy 644064DC-23288FCD - buffer overflow - temporary pause until processing catches up
16:15:02.024 DEBUG s.t.u.USBTransferProcessor - Airspy 644064DC-23288FCD - buffer overflow cleared - resuming normal processing
16:15:05.267 DEBUG s.t.u.USBTransferProcessor - Airspy 644064DC-23288FCD - buffer overflow - temporary pause until processing catches up
16:15:07.207 DEBUG s.t.u.USBTransferProcessor - Airspy 644064DC-23288FCD - buffer overflow cleared - resuming normal processing
16:15:11.235 DEBUG s.t.u.USBTransferProcessor - Airspy 644064DC-23288FCD - buffer overflow - temporary pause until processing catches up

I've tried in three machines (Xeon E5530 and i5-3210M) and all three do the same regardless of whether I use the internal USB ports or the PCIe ports. Oddly enough it works perfectly on my Early 2011 MacBook Pro which has USB 2.0 ports.

Is there a way in SDRTrunk to enable the bit packing feature of the Airspy? That may help. I reqlly only need 6MHz anyway.

 

[DSheirer]

I've tried using a separate USB2.0 PCIe and a USB3.0 PCIe and both work but with a lot of drops and I get a lot of these in the SDRTrunk console window:

16:14:54.629 DEBUG s.t.u.USBTransferProcessor - Airspy 644064DC-23288FCD - buffer overflow - temporary pause until processing catches up
16:14:56.582 DEBUG s.t.u.USBTransferProcessor - Airspy 644064DC-23288FCD - buffer overflow cleared - resuming normal processing
I've tried in three machines (Xeon E5530 and i5-3210M) and all three do the same regardless of whether I use the internal USB ports or the PCIe ports. Oddly enough it works perfectly on my Early 2011 MacBook Pro which has USB 2.0 ports.

Overflow normally kicks in when the processing workload is too much for the computer and sample buffers are filling up. How many channels are you decoding simultaneously? Are you just decoding the control channel(s) and letting sdrtrunk dynamically allocate the traffic channels, or are you decoding all of the traffic channels at the same time?

Which OS are you running? Is it running in a VM?

Is there a way in SDRTrunk to enable the bit packing feature of the Airspy? That may help. I reqlly only need 6MHz anyway.

Not currently.

Denny

 

[carnch]

Overflow normally kicks in when the processing workload is too much for the computer and sample buffers are filling up. How many channels are you decoding simultaneously? Are you just decoding the control channel(s) and letting sdrtrunk dynamically allocate the traffic channels, or are you decoding all of the traffic channels at the same time?

Which OS are you running? Is it running in a VM?

The first box I tried running it on was a Windows 7 workstation and the second Windows 10. Both were physical boxes. I'm not sure what you mean decoding just the control channel or all at once. I have one channel configured which is the control channel but I'm listening and streaming multiple voice channels simultaneously.

This weekend I picked up a Dell M6600 Precision laptop with an i7-2960XM processor in it and low-and-behold I able to run the Airspy with SDRTrunk and have 8 streams running through Icecast all on this one machine with no issues at all. It appears that the quad-core Xeons just weren't able to keep up. I'm not sure why since the core clock speeds of the Xeon and i7 are almost the same (2.6Ghz vs. 2.7Ghz) and both chips have similar L1/L2/L3 caches on them. The i7 hums along happily with no buffer overflows. Problem solved!

Great piece of software Denny! I travel a lot for my job and it'll be nice to be able to listen to my home town while on the road and be able to listen to specific talkgroups at that.

Thanks everyone for the replies and pointing me in the right direction.

 

[DSheirer]

That's great news. I'm glad you got a setup that's working.

I'm not sure what you mean decoding just the control channel or all at once. I have one channel configured which is the control channel but I'm listening and streaming multiple voice channels simultaneously.

It sounds like you have it setup correctly.

cheers,
Denny

 

[jbeen]

Having a similar issue

So, I ran into a similar issue this past week.
Monitoring a system between 854-860Mhz with an Airspy R2 in 10msps mode.

I have Unitrunker setup with VCO-1 as the control channel, which has typically been in the 854Mhz range.

Recently, the control channel switched to the 857Mhz range.
When this happened, I started getting static on the 858 & 859Mhz voice channels.

What I've determined is that when the control channel frequency was in the 854Mhz range, it kept the Airspy's center frequency below 858Mhz.
Once the Airspy's center frequency moved to the 858/859 range, it wasn't working properly in the 10msps sample mode.

Like carnch, if I switched the Airspy to 2.5msps mode, it performed as expected.

For now, what I've done as a work-around is use Unitrunker's "Park" feature to park VCO-1 at 854MHz (where there is no signal).
This forces the Airspy's center frequency to stay around the 856/857Mhz range and everything is happy.

So, carnch, I'd be curious if you tried a similar technique on your original system that wasn't working and see if this works for you.

Josh