Tue 15-jan-2019
LimeSDR transmit check.
Fired up old HP 8614A signal generator at 1880 Mhz, and let it stabilize for 1 hour.
Level measurement after adjusting ALC, 0dBm = -0.4 dBm as checked with spectrum analyzer Checked level with HP L423A probe.
Check of LimeSDR mini output. Output as seen on spectrum analyzer -1.0 dBm. This correlates with LMS7002 datasheet with a possible 1 dB loss between chip and analyzer, including SMA-BNC coax adapter, 2 m foam insulated RG58 size cable, BNC - N connector.
So, for practical purposes the LimeSDR mini provides 0 dBm ( +/- 1.4 dBm )
LimeSDR receive check.
Receive antenna rigged with a 1/10 dB splitter, -10 dB port connected to spectrum analyzer, -1 dB port to LimeSDR.
Connect single Nokia Handset, and make a call from osmobsc to fixed phone, so that only one mobile is using TCH. Move phone to get different uplink levels and correlate these to spectrum analyzer. Since carrier is TDMA, use peak hold on analyzer.
Active subscriber connections: conn ID=31, MSC=0, hodec2_fail=0, mgw_ep=rtpbridge/1@mgw BTS 1, TRX 0, Timeslot 1, Lchan 0: Type TCH_F Connection: 1, State: ESTABLISHED BS Power: 10 dBm, MS Power: 10 dBm Channel Mode / Codec: SPEECH_V1 No Subscriber Bound IP: 192.168.1.75 Port 16384 RTP_TYPE2=0 CONN_ID=0 Conn. IP: 192.168.1.70 Port 4110 RTP_TYPE=3 SPEECH_MODE=0x00 Measurement Report: Flags: MS Timing Offset: 0 L1 MS Power: 10 dBm, Timing Advance: 4 RXL-FULL-dl: -51 dBm, RXL-SUB-dl: -51 dBm RXQ-FULL-dl: 0, RXQ-SUB-dl: 0 RXL-FULL-ul: -54 dBm, RXL-SUB-ul: -54 dBm RXQ-FULL-ul: 0, RXQ-SUB-ul: 0
Level on uplink -54 dBm, indicated on spectrum analyzer -53 (-63) dBm.
OsmoBSC# show conn Active subscriber connections: conn ID=31, MSC=0, hodec2_fail=0, mgw_ep=rtpbridge/1@mgw BTS 1, TRX 0, Timeslot 1, Lchan 0: Type TCH_F Connection: 1, State: ESTABLISHED BS Power: 10 dBm, MS Power: 10 dBm Channel Mode / Codec: SPEECH_V1 No Subscriber Bound IP: 192.168.1.75 Port 16384 RTP_TYPE2=0 CONN_ID=0 Conn. IP: 192.168.1.70 Port 4110 RTP_TYPE=3 SPEECH_MODE=0x00 Measurement Report: Flags: MS Timing Offset: 0 L1 MS Power: 10 dBm, Timing Advance: 4 RXL-FULL-dl: -69 dBm, RXL-SUB-dl: -69 dBm RXQ-FULL-dl: 0, RXQ-SUB-dl: 0 RXL-FULL-ul: -59 dBm, RXL-SUB-ul: -59 dBm RXQ-FULL-ul: 0, RXQ-SUB-ul: 0
Mobile moved to another location within the lab.
Level on uplink -59 dBm, indicated on spectrum analyzer -60 (-70) dBm.
Conclusion:
For such a simple low cost SDR, the accuracy seems within a couple of dB. Instruments used were are not "professional grade", at least not w.r. 2019, but I estimate I can measure within +/- 2 dB or so with this crude config. Power output is 10 times lower than I expected, I was probably mislead by some review on the web, and had to go down to the actual scematic and datasheet to understand that the expected level should be around 0 dBm. (thanks for pointing this out, Harald)
For practical use outside a lab output is far too low. My plan is to add a small simple booster to bring level up to +15 which would allow me to cover my 2 acre lot without breaking any rules.
Else, I am impressed, good value for a low price tag. I had expected far worse than indicated accuracy.
Best Regards, Gullik
Hi Gullik,
thanks a lot for your extensive measurement and testing, as well as the report here to this list.
On Tue, Jan 15, 2019 at 04:13:29PM +0100, Gullik Webjorn wrote:
So, for practical purposes the LimeSDR mini provides 0 dBm ( +/- 1.4 dBm )
Yes, that's more in-line with what I remember. Depending on unit and frequency, I think the expected level is roughly between 0 and 4 dBm.
Level on uplink -54 dBm, indicated on spectrum analyzer -53 (-63) dBm.
So you're saying that the level displayed in the 'show lchan' actually was almost exactly the measrued level (-53 dBm)? That's a very big surprise to me. As far as I understand, the LimeSuite API never reports any absolute values, but rather only in dBFS (dB to full scale). So it's a bit of a mystery how you can get such an accurate reading?
For practical use outside a lab output is far too low. My plan is to add a small simple booster to bring level up to +15 which would allow me to cover my 2 acre lot without breaking any rules.
We made some osmo-bts-amp / osmo-bts-pa designs and devices some time ago, but they only have 10dB gain. They were made to go from +23 to +33 dBm, and not for such low input signals, where higher gain stages are required.
Make surey you carefully look at your spectrum and use transmit filters to suppress the harmonics that your PA will be generating.
Else, I am impressed, good value for a low price tag. I had expected far worse than indicated accuracy.
Regards, Harald
Hello Harald et.al.
On 2019-01-17 23:18, Harald Welte wrote:
Hi Gullik,
thanks a lot for your extensive measurement and testing, as well as the report here to this list.
On Tue, Jan 15, 2019 at 04:13:29PM +0100, Gullik Webjorn wrote:
So, for practical purposes the LimeSDR mini provides 0 dBm ( +/- 1.4 dBm )
Yes, that's more in-line with what I remember. Depending on unit and frequency, I think the expected level is roughly between 0 and 4 dBm.
Level on uplink -54 dBm, indicated on spectrum analyzer -53 (-63) dBm.
So you're saying that the level displayed in the 'show lchan' actually was almost exactly the measrued level (-53 dBm)? That's a very big surprise to me. As far as I understand, the LimeSuite API never reports any absolute values, but rather only in dBFS (dB to full scale). So it's a bit of a mystery how you can get such an accurate reading?
Yes, I am surprised myself, but I documented what I saw. Note also I checked at two different levels. I will probably redo these tests as I get a more output and a "better test range".
For practical use outside a lab output is far too low. My plan is to add a small simple booster to bring level up to +15 which would allow me to cover my 2 acre lot without breaking any rules.
We made some osmo-bts-amp / osmo-bts-pa designs and devices some time ago, but they only have 10dB gain. They were made to go from +23 to +33 dBm, and not for such low input signals, where higher gain stages are required.
Make surey you carefully look at your spectrum and use transmit filters to suppress the harmonics that your PA will be generating.
Yes, currently I use a linear amp designed as a preamp, but output level is a bit below 1 dB compression point, so we are operating linear, current output is 8-9 dBm, which enables extending the test range. I am limited to 100 mW indoor, +20 dBm.
Else, I am impressed, good value for a low price tag. I had expected far worse than indicated accuracy.
Regards, Harald
-
Gullik Webjorn -
Harald Welte