HL2 on protocol 2?

[Matthew Doherty]

has anyone got their HL2 working on protocol 2? 

[Sid Boyce]

HL2 firmware does not support Protocol 2 as it's based on Hermes Protocol 1.
73 ... Sid.


On 23/01/2021 06:14, Matthew Doherty wrote:
> has anyone got their HL2 working on protocol 2? --
> You received this message because you are subscribed to the Google
> Groups "Hermes-Lite" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to hermes-lite...@googlegroups.com
> <mailto:hermes-lite...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/hermes-lite/26a1c2f0-7f25-472c-b1b2-4c115f017656n%40googlegroups.com
> <https://groups.google.com/d/msgid/hermes-lite/26a1c2f0-7f25-472c-b1b2-4c115f017656n%40googlegroups.com?utm_medium=email&utm_source=footer>.


--
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support
Senior Staff Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks

[Alan Hopper]

Hi Matthew,

as Sid says the current firmware is Protocol 1. There is nothing to stop anyone developing a Protocol 2 version but it would be a lot of work for no obvious gain (especially as Thetis now supports P1). Is there some particular feature of P2 that interests you?

73 Alan M0NNB

[Matthew Doherty]

1.5mhz bandwidth over 192khz. being able to see a whole ham band at once over seeing a small section of it is kinda a big deal. 

[Alan Hopper]

Hi Matthew,

there has been talk of enabling wider bandwidth in P1 and this is more likely to happen than P2. Powersdr does allow stitching multiple 384k receivers to see more and the same thing has been very low on my Spark todo list for a long time but you are the first to express interest in wider bandwidth.  There is also the low res wideband data that some sw will display.

73 Alan M0NNB

[Matthew Doherty]

good to know, just got my HL2 a few days ago. currently using thetis, the wideband feature isnt working for me, and I cant enable 384k. even 384 would be a good upgrade from 192. 

[Alan Hopper]

All other software should give you 384k and I believe Reid is planning a version of thetis that is tweaked for HL2.   

[Steve Haynal]

Hi Slappormatt,

Like Sid and Alan said, the HL2 does not support protocol2. Protocol2 makes demands on the FPGA that just don't work well with the smaller less expensive FPGA used in the HL2. The last I heard, DHCP had to be stripped from the gateware for protocol2 to fit on a ANON 10E which uses a similar FPGA. That was with only 2 receivers.

For visualization, it is often not necessary to send 1.5MHz real baseband to visualize 1.5MHz of spectrum. If you want to visualize 1.5MHz of spectrum on a typical monitor with 1920 pixels width and refresh rate of 30 or 60 Hz, then you could cut down the amount of data to only what is required to provide a reasonable visualization to the eye (1920 FFT bins, updated 30 times a second). We have the bandscope data on the HL2 to visualize the entire HF spectrum, but I would be interested in enhancing that to enable better visualization of portions of the band. I'd like to be able to pick and zoom into about 1 MHz of any portion of the bandscope data and still see a good representation of HF activity. My understanding is that the receive-only Kiwi SDR supports this.

It may be possible to extend protocol 1 and add wider bandwidths, but I'm not sure what the application would be except for visualization. SparkSDR can create many slice receivers for skimming, but even a 192kHz bandwidth is enough to capture all CW,FT8,WSP,JT9,PSK signals on a band. Maybe transverter users will have applications for wider bandwidth. But below 30MHz, all the modulated signals I know of are under 12kHz wide.

73,

Steve

kf7o

[ron.ni...@gmail.com]

Hi Steve,

How fast can a single receiver slice running at 384ksps change its Rx frequency? (latency for the FPGA filters to settle, etc.)  I've been wondering if multiple IQ samples sets at different Rx frequencies could be downloaded from a single HL2 receiver slice in 100 mS or so.  If the Rx frequency can be changed at a high enough rate, one could use 1 slice for Rx/Tx, and a 2nd slice to scan the entire RF spectrum (or subset) for a slow-motion but high-res wideband visualizer.

73,

Ron

n6ywu

[Steve Haynal]

Hi Ron,

That is an interesting idea, to setup one receiver to "scan" frequencies at a regular interval so that we can create wider bandwidth visualizations. The DSP pipeline consists of  CIC decimation filters (2 filters combined to ~10 stages) plus a ~900 tap final FIR filter. These all run at 2*76.8MHz=153.6MHz. A quick probably over estimated guess is that it can take ~2000 clock ticks to flush old state and have reliable new state in the filters. This is on the order of 10 to 15us. At a sample rate of 384kHz, sample spacing is 2.6us. So maybe 5 to 10 samples would have to be discarded between frequency changes. This quick analysis could very well be missing something though. We can also reset the DSP pipeline as that feature was added recently to support synchronous receivers, but the time to prime the DSP pipeline is probably similar.

This seems very doable. Just thinking off the top of my head, if we want to display 1.5MHz of bandwidth, we need to scan 4 384kHz regions. With a 1920 wide display, each region has 480 pixels. Let us round this up to a 512 point/bin FFT. That means we need 4*512  = 2048 "good" samples. Add another 40 samples for the frequency changing and we need ~2100 samples. At 384kHz, we receive 2100 samples about every 6ms. A 30Hz refresh rate is a screen refresh every 33ms. So even if I am off somewhere by a factor of 2 or 3, this should still be possible.

73,

Steve

kf7o

[Reid Campbell]

Yes, I can confirm that Thetis will do 384K with protocol 1. I have it running at the minute, just need to find a block of time to add a few more features like the Tx buffering.

Cheers

Reid
Gi8TME/Mi0BOT  

To unsubscribe from this group and stop receiving emails from it, send an email to hermes-lite...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/9e1efef3-a317-4582-a975-89b4faa72da2n%40googlegroups.com.

[RD Powers]

> I'd like to be able to pick and zoom into about 1 MHz of any portion of the bandscope data and still see a good representation of HF activity. My understanding is that the receive-only Kiwi SDR supports this.

I'm finding this discussion very interesting and don't want to take people off on a tangent, so consider this a FYI.

I have played with other people's KiwiSDRs on line and you can zoom out and see the entire 0-30 MHz bandwidth.  I found this very fascinating the first time I did this.  I'm not sure it's all that useful long term, but it is a kind of cool thing to watch for a while.

If I understood what I read about KiwiSDR correctly, no IQ data leaves the KiwiSDR daughter card.  It's already formatted as video and audio streams and all the Beagleboard host computer does is a control plane function.  

My reading about FlexRadio 6000 indicates the same thing.  Data leaving the radio is already formatted as video and audio streams and is carried using Vita-49 protocol to the client running on PC or mobile device.  Same for audio in the other direction.  Flex 6000 has onboard FPGA and DSP resources, along with ARM cores running Linux that perform the control plane function.

There are a lot of interesting architectures around that are worthy of consideration, IMO.

Regards,

RDP