We’re in the same boat there for sure!
Yeah, I saw there was some pretty bold things happening in there (rounding and 
such). I had wondered about sending a 2-D array. 

I will try to do more with the Python next week.

Good news is, At the moment  i have observed only a very small amount of 
instantaneous phase error over a short amount of time. Which reflects the 
graphs in the N321 docs. 
Obviously that doesn’t count for temperature fluctuation or extended runtime 
effects, but it’s a start!

Thanks for the help. 

<end transmission>

> On Jan 21, 2022, at 17:17, Marcus D. Leech <[email protected]> wrote:
> 
> 
> On 2022-01-21 16:52, Paul Atreides wrote:
>> ok. is that just setting up 2 streamers to start at the same time?
>> 
> From looking at the code, I *THINK* that if you send it a 2D numpy array the 
> rows represent the channels.
> 
> The "send_waveform" API is Python-API only, and does a fair amount of 
> dancing-around to "make things just work"
>   even if you only send a 1D wave-form array.
> 
> But, of course, I cannot find any documentation on it, just looking at the 
> code.
> 
> 
>> On Fri, Jan 21, 2022 at 4:50 PM Marcus D. Leech <[email protected]> 
>> wrote:
>>> On 2022-01-21 16:46, Paul Atreides wrote:
>>>> is there a way to incur a phase shift on one of the streams from the 
>>>> in-built UHD examples?
>>>> the python API allows has 
>>>> 
>>>> usrp.send_waveform(samples, duration, center_freq, sample_rate, 
>>>> [0,1],gain) 
>>>> 
>>>> but this sends the same samples to both outputs. I've used this icommand 
>>>> in conjunction with the LO sharing options for the multi_usrp object to 
>>>> confirm the phase between channels
>>>> i'd like to be able to apply some digital phase shift between the 2 to 
>>>> test the level of control, hence all the gr-uhd questions.
>>>> 
>>>> any ideas or references?
>>> You'll probably have to send samples yourself, with phase-offsets between 
>>> the two sample streams.
>>> 
>>> 
>>>> 
>>>> On Wed, Jan 19, 2022 at 9:41 AM Rob Kossler <[email protected]> wrote:
>>>>> On Wed, Jan 19, 2022 at 1:00 AM Paul Atreides <[email protected]> 
>>>>> wrote:
>>>>> >
>>>>> > Ok, so just circling back on this. Ive got a decent handle on the 
>>>>> > Python API. I’ve confirmed the LO signal comes out of the port once the 
>>>>> > splitter output is enabled.
>>>>> >
>>>>> > Rob you’re saying that for sure both channels need to be external AND 
>>>>> > exported? making the settings:
>>>>> > TX 0:
>>>>> > LO source     = external
>>>>> > LO Exported = true
>>>>> > TX 1:
>>>>> > LO source      = external
>>>>> > LO Exported = true
>>>>> Yes, except as you pointed out below, LO Exported for Tx1 is not
>>>>> needed and likely does nothing.
>>>>> 
>>>>> >
>>>>> > Radio #0 TX LO OUT 0  = enabled
>>>>> > Radio #1 TX LO OUT  1 = enabled (for debug)
>>>>> > (When I set These they turn on the lights for the LO outs)
>>>>> >
>>>>> > The front panel cable should connect TX LO output 0 to TX LO Input 1
>>>>> Yes
>>>>> 
>>>>> >
>>>>> > Like you said, According to the block diagram, the TX LO input 1 should 
>>>>> > connect both channels to the 1x2 splitter so I should only need 1 LO 
>>>>> > output turned on.  Id like to point out that in the block diagram it 
>>>>> > also doesn’t look like it’s possible to export channel 1’s LO. The only 
>>>>> > connection in the diagram is internal.
>>>>> 
>>>>> Yes, Somewhere in the Ettus documentation, it indicates that this is
>>>>> the case (I've forgotten where).
>>>>> 
>>>>> >
>>>>> > Am I missing something? Have you gotten this to work?
>>>>> Yes, I have been using this in a 16x16 system with 8 N321 devices all
>>>>> connected by a single LO (arbitrarily chosen LO export channel
>>>>> distributed to both Tx and Rx LO inputs). I've also tested it on
>>>>> smaller systems as small as 4x4 (and possibly 2x2 like your case, but
>>>>> not positive on this).
>>> 
> 
_______________________________________________
USRP-users mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to