Just received mine 2 days ago. Works perfectly, if you are aware of your surroundings and critical of the bearings returned. Example. Driving down a road with tall buildings on both sides will likely skew the bearings to be more alligned with the road due to reflections. The ‘navigate to destination’ does not work well in build up areas, take a few bearings from open areas and then plan a course towards the estimated position - then do a 90 degree turn when you get close. That should give you a fairly accurate estimate, box around it and you have a near perfect solution. Doing stationary measurements works too, if location is free from multi-path interference. Wish there was a faster way to switch from driving to stationary, e.g. a single button that would: 1. Change from GPS to manual antenna direction 2. Switch ‘Do not measure when stationary’ Off 3. Prompt you for vehicle (or stationary antenna) direction of element 0 And finally, a quick FRQ entry from the android navigation web page, rather than having to go to settings first. This works som much better than KerberosSDR (like from ‘hardly at all’ to ‘Very well’). Have not tried my Kerberos SDR with the new software, will likely improve results a lot. And extra finally: there is a typo in the printed manual, hotspot the PI is looking for is NOT krakensdr but KrakenAndroid.

@thekraken2086

Жыл бұрын

Thanks for the feedback, happy to hear you've had good results with it. Yes the new software should definitely improve results with the Kerberos too, but do remember to change to the KerberosSDR preconfig, and manually disconnect antennas during intial calibration. I'll think about how to implement quick heading method and freq switches. I assume that you would have some compass or other device that would tell you your heading when changing to manual heading?

@bodstrup

Жыл бұрын

@@thekraken2086 Sure, using compass on my watch/phone. Best when outside vehicle. More feedback: With default settings, KrakenSDR samples directions, even when there are no signal. Often, ambient noise gives a more or less constant bearing in some random direction. Once signal arrives, I get a clear 'pointed' lobe and the correct bearing, however averaging can skew the result towards the non-signal sample. Setting minimum power and confidence is difficult, very small changes can make a difference between no sample at all - or sample when signal only (desired for me). As there are no feedback visible on confidence - and signal strength on the spectrum do not appear to relate to the '-200 to 0' power level in settings - it is hard to tune the system to avoid sampling when there are no signal. A simple 'squelch' type slider on the Spectrum, setting a minimum dBi level for sampling could be a huge benefit, allowing us to quickly tune sensitivity. I often have noise floor between -70 and -80 dBi, signal at perhaps -35 dBI, so there is a clear difference. Spent hours in the weekend driving around experimenting with settings, but these, minimum power and confidence I have not yet managed to set reliably.

Impressive range for 12 dBm (16mW) - without any special modulation e.g. WSPR, etc. Were the antennas high enough to have direct line of sight coverage most of the time.

@thekraken2086

2 жыл бұрын

Antennas were high on a house roof with LOS to the car for most of the drive. If you're blocked by a mountain or buildings etc, results will be poor, but if you can receive the signal you can probably still get a very rough direction depending on where the signal refracts or reflects on.

Hey, great result! How high is your antenna? You have a beautiful place. Have you tried to work on the radio stations of supermarket employees, truck drivers? Translated by translate google.

Is there someone I can contact to about doing a presentation about this to my radio club?

Would knowing the local elevation (via DEM) make some of the calculations more exact?

@thekraken2086

Жыл бұрын

It's possible, but the algorithms and implementation to take that into account don't exist yet. Definitely a research project for the future.

1.) is it possible to use the individual radio(s) in like SDR# software, like a regular SDR hardware radio? 2.) Can I combine all 5 radios to act like a single SDR receiver to get an increased gain/detail of signals?

@thekraken2086

Жыл бұрын

1) Yes, if you don't use our software the device will simply show up as five separate RTL-SDRs which you can use. 2) Yes, but it will require you to implement your own DSP algorithm to do the combination. It's not just as simple as adding the 5 signals together. A little down the line we may look at providing a GNU Radio flowgraph for this.

@Barc0d3

Жыл бұрын

@@thekraken2086 Thank you very much, you are the best!

can the program save the last averaged result on the card (for example, for 100 ms)? We take a picture of the bearing from one place, move to another place and get a different bearing. Ideal bearings are 90 degrees. Translated by translate google.

@thekraken2086

2 жыл бұрын

The app isn't designed for that, but it should be possible. Just press the stop button in the app, then when your in the new location press it again briefly to log only a few points. Overall, it's much better if you just leave it collecting all data and let it average on the grid by itself.

@bodstrup

Жыл бұрын

@@thekraken2086 Partly disagree, if passing high rise buildings, you will add a lot of false bearings, have seen the estimated location jump 180 degrees briefly while in an area with multipath. A big Stop/Start button on the map IF would be a benefit. As would be a co-driver judging terrain turning on/off collection and planning route.

@thekraken2086

Жыл бұрын

​@@bodstrup Multipath readings will usually diverge as long as the multipath isn't dominant over time, and the vehicle is actually moving and gets good readings at some point in the past or future. So those divergent readings will eventually have a negligible effect on the grid space as only convergent readings matter. But yes if you're starting in an area with only multipath readings it could be confusing because you have no good data at first. That's definitely where some experience and manual stop/stop would come in handy.

@bodstrup

Жыл бұрын

@@thekraken2086 One session I started at the top, open floor of a parking garage - that gave me a DOA that was nearly perfect. Two of these and I can plan. On the humorous side: anyone else think the lobes ocasionally look like a path a naughty pilot would draw in the sky .. ;-)