To sanity check the transmitter, instead of showing the constellation of every sample, show every other sample, i.e. decimate by 2 going into the constellation plot. The two comes from the fact you use 2 samples/symbol. We don't want to see the every sample on a constellation, only symbols. That's why the transmitter block has the grouping it does. Once you've confirmed you see the expected BPSK constellation, you can move on to the receiver.
On Tue, Oct 27, 2015 at 9:08 AM, <mle...@ripnet.com> wrote: > Try cranking your RX gain up/down 5dB and see how that affects your > constellation. > > > > > > > On 2015-10-27 11:55, Washbourne, Logan wrote: > > Ok, that sampling rate mismatch was my fault. The RX spectrum looks better > after that fix. The unpacked to packed bits didn't seem to remove the zero > grouping on the constellation plot. Here are pictures of the setup now. > > > > Logan Washbourne > Electrical Engineering Graduate Student > (Electromagnetics) > > On Tue, Oct 27, 2015 at 10:34 AM, Richard Bell <richard.be...@gmail.com> > wrote: > >> Logan, make sure you're feeding packed bytes into the DPSK Mod block. >> That's probably why you're seeing a constellation point at zero, you're >> feeding in unpacked bytes. That might fix everything. >> >> Rich >> >> Sent from my iPad >> >> On Oct 27, 2015, at 8:16 AM, mle...@ripnet.com wrote: >> >> What does the spectral plot look like? It should look very similar to >> the TX side of the house. >> >> >> >> >> >> >> On 2015-10-27 11:10, Washbourne, Logan wrote: >> >> Thanks for the info guys! They are both using antennas btw. I came in >> this morning and turned the RX gain up pretty high, roughly 45dB, and the >> constellation plot on the RX side is starting to look pretty square, which >> I think is a bad sign. My thoughts are that with the increase in gain, >> there is too much noise coming through. >> >> I've been toying with adding in a frequency offset on the TX side. I've >> been adding and subtracting up to 30kHz from the center frequency and I >> haven't seen any improvement in the constellation plot on the RX side. >> >> Another thing that is confusing me, on the TX constellation plot, before >> it gets sent to the USRP, there are 3 groupings, -.5, 0, .5. I'm not sure >> where the 0 points are coming from. Since this is DBPSK, there should only >> be -.5 and .5 groupings right? >> >> Should this exercise be pretty straightforward? >> >> Logan Washbourne >> Electrical Engineering Graduate Student >> (Electromagnetics) >> >> On Mon, Oct 26, 2015 at 3:12 PM, <mle...@ripnet.com> wrote: >> >>> Consulting now, the book of armaments, errr, UHD manual: >>> >>> >>> >>> http://files.ettus.com/manual/page_dboards.html#dboards_rfx >>> >>> >>> >>> The RFX series has no TX RF gain control, output magnitude is entirely >>> dependent on baseband magnitude. >>> >>> There's 70dB of gain range on RX, however. So, crank up the RX gain. >>> >>> Also, across 1m, the path loss is about 40dB at 2.5GHz, your TX is >>> probably putting out 5dBm, so at the RX end, that's maybe as much as >>> -35dBm. If your RX is set properly, should be enough to see the signal. >>> >>> >>> >>> >>> >>> >>> On 2015-10-26 15:46, Marcus Müller wrote: >>> >>> Hi! >>> >>> * looking at your constellation, relief comes setting in: It looks >>> pretty circular to me; notice how the axes are scaled differently. So no >>> catastrophic IQ imbalance. >>> * the fact that it's circular and not a line is probably the result of >>> the constellation sink not being able to achieve timing synchronization, >>> i.e. it doesn't "know" when a symbol starts >>> * You're right, RX SNR is terrible. However, RX power is very little >>> indeed >>> * Rule of thumb: if(not clipping && in doubt) increase gain; >>> * What exactly is your "RF channel": antennas, direct cable with >>> attenuator? >>> * The mistake I do every few months: Did you connect the antenna/cable >>> to the wrong SMA port? >>> >>> Best regards, >>> Marcus >>> On 26.10.2015 20:33, Washbourne, Logan wrote: >>> >>> I looked at the spectrums of both sides and the RX side looks to be >>> inundated with a lot of noise. I'm not sure if its my environment or my >>> USRPs, but I really expected to see some resemblance of a signal. I am >>> posting links to screencaps of the TX and RX sides. >>> >>> http://imgur.com/GBTHw8T- TX >>> http://imgur.com/DBy8gqs - RX >>> >>> Logan Washbourne >>> Electrical Engineering Graduate Student >>> (Electromagnetics) >>> >>> On Mon, Oct 26, 2015 at 1:31 PM, Marcus Müller <marcus.muel...@ettus.com >>> > wrote: >>> >>>> Can you compare the TX spectrum with the RX spectrum? Maybe it'd >>>> worthwhile to first look at at e.g. 500kHz of RX spectrum to see whether, >>>> due to frequency offset, some of the wanted signal simply gets cut off; >>>> admittedly, with 2S/sym that's not too likely... >>>> >>>> Best regards, >>>> Marcus >>>> >>>> >>>> On 26.10.2015 19:16, Washbourne, Logan wrote: >>>> >>>> I am using 250k! >>>> >>>> >>>> >>>> Logan Washbourne >>>> Electrical Engineering Graduate Student >>>> (Electromagnetics) >>>> >>>> On Mon, Oct 26, 2015 at 1:05 PM, Marcus Müller < >>>> marcus.muel...@ettus.com> wrote: >>>> >>>>> Hello! >>>>> what's the sampling rate you're using? >>>>> >>>>> Best regards, >>>>> Marcus >>>>> >>>>> >>>>> On 26.10.2015 18:38, Washbourne, Logan wrote: >>>>> >>>>> Hello all, >>>>> >>>>> I'm trying to take a step back from my previous postings and try a >>>>> simple approach to try and understand working with over the air >>>>> communications with the USRPs. Let me know if it would be better to post >>>>> this in the USRPs-Users list. >>>>> >>>>> I'm trying to have a simple TX and simple RX side so I can eliminate >>>>> and unnecessary complications. >>>>> >>>>> The TX side has the following flowgraph: Vector >>>>> Source(preamble+data)->DPSK Mod(DBPSK,2samples/symbol)-> Multiply >>>>> Const(.707) -> UHD:USRP SINK(2.448GHz center freq, 10dB default gain) >>>>> >>>>> RX Side: UHD:USRP Source(2.448GHz center freq, 10dB gain default) -> >>>>> DPSK DEMOD(DBPSK, 2 samples/ symbol, rest default values) - > file sink >>>>> >>>>> I'm checking the progress of the communication by looking at a >>>>> constellation plot that is connected directly to the USRP SOURCE block on >>>>> the receiver side. I'm getting a very elliptical shape, more spread out in >>>>> the horizontal direction but not by much. I'm also looking at the bits >>>>> from >>>>> the receiver side in matlab and my preamble is not showing up. >>>>> >>>>> These were the same problems I had when trying to use a correlated >>>>> system for OTA communications. I feel like I'm missing something really >>>>> simple. >>>>> >>>>> Does anyone know of a simple TX, RX grc setup that I can use to test >>>>> my USRPs? >>>>> >>>>> Again, I really appreciate the help you guys give. >>>>> >>>>> Logan Washbourne >>>>> Electrical Engineering Graduate Student >>>>> (Electromagnetics) >>>>> >>>>> _______________________________________________ >>>>> Discuss-gnuradio mailing >>>>> listDiscuss-gnuradio@gnu.orghttps://lists.gnu.org/mailman/listinfo/discuss-gnuradio >>>>> >>>>> >>>>> _______________________________________________ >>>>> Discuss-gnuradio mailing list >>>>> Discuss-gnuradio@gnu.org >>>>> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio >>>>> >>>>> >>>> >>>> _______________________________________________ >>>> Discuss-gnuradio mailing >>>> listDiscuss-gnuradio@gnu.orghttps://lists.gnu.org/mailman/listinfo/discuss-gnuradio >>>> >>>> >>>> _______________________________________________ >>>> Discuss-gnuradio mailing list >>>> Discuss-gnuradio@gnu.org >>>> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio >>>> >>>> >>> >>> _______________________________________________ >>> Discuss-gnuradio mailing >>> listDiscuss-gnuradio@gnu.orghttps://lists.gnu.org/mailman/listinfo/discuss-gnuradio >>> >>> >>> >>> _______________________________________________ >>> Discuss-gnuradio mailing >>> listDiscuss-gnuradio@gnu.orghttps://lists.gnu.org/mailman/listinfo/discuss-gnuradio >>> >>> >>> _______________________________________________ >>> Discuss-gnuradio mailing list >>> Discuss-gnuradio@gnu.org >>> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio >>> >>> >> _______________________________________________ >> Discuss-gnuradio mailing >> listDiscuss-gnuradio@gnu.orghttps://lists.gnu.org/mailman/listinfo/discuss-gnuradio >> >> _______________________________________________ >> Discuss-gnuradio mailing list >> Discuss-gnuradio@gnu.org >> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio >> >> > _______________________________________________ > Discuss-gnuradio mailing > listDiscuss-gnuradio@gnu.orghttps://lists.gnu.org/mailman/listinfo/discuss-gnuradio > > > _______________________________________________ > Discuss-gnuradio mailing list > Discuss-gnuradio@gnu.org > https://lists.gnu.org/mailman/listinfo/discuss-gnuradio > >
_______________________________________________ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
