SDRtrunk ramblings with Hackrf - mostly good
954 views
Skip to first unread message
Douglas Welch
Mar 12, 2020, 8:39:45 PM3/12/20
to sdrtrunk
is it normal with Hackrf (15MHZ sample rate - 850Mhz range) to use Heterodyne to keep from buffer overuns? Using the Polyphase i get overruns within 10-15 seconds of startup.
I also have been able to run 20Mhz sample rate using Hetrodyne. Not more than 8 Mhz with polyphase. Hackrf runs about 0.2ppm (pretty good- i think) off after warmup
just using a mediocre laptop
System Type x64-based PC
System SKU LENOVO_MT_20C5_BU_Think_FM_ThinkPad Edge E440
Processor Intel(R) Core(TM) i3-4000M CPU @ 2.40GHz, 2394 Mhz, 2 Core(s), 4 Logical Processor(s)
I now pause or disable the waterfall display which helps alot. Also moved down traffic pool down to 4.
Runs really nice now with a single Hackrf, my other laptop has 3 Rtl-Sdr (1ppm) dongles and does about the same with lower noice levels and better looking graphics. Only draw back ive found so far is high noise levels i have to rub at. I spent hours test every combo from 12-14 to 40-60 with and without amp. the amp doesnt seem to work well at lna:18 & vga:16-18. The spectrum and Freq. display jump bad, so i dont know if its a software or firmware problem.With the hackrf spectrum screen is so full of noise at all working settings you cant see the the wave forms hardly. (big blob of red). unlike the Rtl-Sdr system
I dont seem to see that particular problem running the portapack by itself.
have a couple LNA amps on order from Ali and will be testing much more in few weeks. I need to get a good Discone mounted outside then settings wont be so important. Just using small but perfectly tuned (laugh) small White router antennas. Dont laught they work better than any others tested.for indoor us. I think they are the tplink router 3.25" wirteless antennas
If im not mistaken i found that you can set the center freq using the record freq setting without recording. Is that a Feature, function, or bug? I thought it dynamically set the center freq. (dc-offset)?
RTL-SDR Screen shot
Dean Sauer
Mar 14, 2020, 7:22:52 PM3/14/20
to sdrtrunk
On Thursday, March 12, 2020 at 8:39:45 PM UTC-4, Douglas Welch wrote:
is it normal with Hackrf (15MHZ sample rate - 850Mhz range) to use Heterodyne to keep from buffer overuns? Using the Polyphase i get overruns within 10-15 seconds of startup.
You gave me an idea I had forgot about... try the hetryodyne for the 4 tuners... to ge the BW I need... welp unfotunately with 4 tuners it still USB ERRORS out... I might retry once I can split some up more on a dedicated USB PCI card.. but with the ports on the box I have to test no go...
Thanks for the reminder on this option....may come in hadny with another test I am going to try....
Runs really nice now with a single Hackrf, my other laptop has 3 Rtl-Sdr (1ppm) dongles and does about the same with lower noice levels and better looking graphics. Only draw back ive found so far is high noise levels i have to rub
I in regards to that waterfall and spectrum... I think you have some issues...
That is way way way to much red and yellow in that waterfall... thats noise and that you get anything and decodes with out it BER' out is amazing... compare that to my screen shot attached generic R820Tx units...
You want more blue except where the signals are.. then it should be red and yellow...
I can tell from the shot is there peaks to that stuff not shown??? Otherwise it looks like most of its buried in the noise floor...
I would not add an amp of any sort till you test out some things..
Since you've played with the VGA etc.. Do a factory reset.. if there is one for that unit ( i don't have one of them to say) get it back to factory point....
Are you getting peaks like I show in my shot? Peaks WAY ABOVE the noise floor?????
Change the GAIN.. not the VGA to various levels till you have a blue waterfall background and not the sea of red and yellow...
What is it connected to for antennas now??? Any thing else??? coax, Amps????? Length of coax, type???? HAAT? HABG??????
Now as for discone, my *** Personal OPINION ** of discones ABSOLUTE JUNK! Get a dedicated LMR antanna for the band/bands you need. If you need 850Mhz LMR then get a decent 850Mhz Comtelco or Browing aka Tram antenna for 850Mhz.... Depending on your RF environment getting a 3db gain antenna is probably the max.. more and you may make things worse.... if you need more gain and want to look at a specific tower then look at a yagi.. again that will depend on your setup for system ie: is it simulcast, are you in a spot between tower sites aka the mush zone? Do the groups you want only appear on one of the sites you get ???
Coax??? Length matters at 850Mhz! Any more than 45 feet you REQUIRE LMR400! Period. Otherwise you are wasting your time and $$$. Upto 45feet RG6 will work... RG11 can be used if you have the means to run it.. its not as flexible as RG6, obviously.. neither is LMR 400.. and note LMR400 also requires special connectors and tools to use those connectors... so if you can run it with connectors on get it pre made that way...
Before getting or putting any LNA on that setup I would want to get the noise conditions under control... as it looks like you got a lot of noise there....
Douglas Welch
Mar 20, 2020, 7:36:58 PM3/20/20
to sdrtrunk
Gotcha on antennas, i had one years ago when they got popular. just hand build a small dipole for 850mhz and working better than the wifi router antenna by alittle and lot more directional sitting inside.
will a 1/2 wave or other give me more sensitivity? Never built or desisgned an antenna before.
its bizzare the hackrf can have that much noise. supprising enoug if i turn on built-in lna amp noise levels drop slighly on many freq. but signal isnt really helped much either. probably poor design or components from china.
yes, water fall is full of red but sdrtrunk easily pulls signals out of it. with other programs like sdrconsole or sdrsharp i just lower the waterfall settings to remove red and it leaves signals. We dont have that option with sdrtrunk (idea? maybe).
if anyone else has a hackrf with same problems please chime in and let me know.
hard to beleive a $13 chinese rtl-sdr v2 can out perform the hackrf but totally possible.
im stuck with the hetrodyne option only on this machine at 10mhz BW. but it works pretty damn good.
ill try the handmade antenna with a couple of them to see what i find and report.
i love the sdrtrunk software more than anything available. and ive tried everyone in last 60 days. op25, unitrunker, mostly under linux. and didnt like quality of voice or software usability.
ill be happy to help with the development as soon as i get up to speed with Java. (lol...)
if there is anything simple or more generic i can do to help now, the developer can let me know if i can help.
Dean Sauer
Mar 20, 2020, 8:50:11 PM3/20/20
to sdrtrunk
On Friday, March 20, 2020 at 7:36:58 PM UTC-4, Douglas Welch wrote:
Gotcha on antennas, i had one years ago when they got popular. just hand build a small dipole for 850mhz and working better than the wifi router antenna by alittle and lot more directional sitting inside.will a 1/2 wave or other give me more sensitivity? Never built or desisgned an antenna before.
1/2 wave are 0dbd or 2.14dbi which is db gain v. isotropic radiator, which is a NON existent item.. its theoretical only.
Its much better to just purchase them. Nothing wrong with learning about the design, operation etc.. but I neither have the time nor patience for building.
The ARRL Handbook in the past had a great deal of fantastic info on designing building etc. antennas.. Like I said.. I don't have the time or patience for that...
Theres a lot of theory to antennas as you increase the gain and what that does to the patterns, both reception wise and transmission wise..
Thats why I said you likely only need a 3dbd antenna.. In a lot of installs only a UNITY or 0dbd antenna is put in at various control points.. or a yagi.. I myself tend to prefer a yagi to the tower for FX1 station but thats me... in a simulcast setup if you use the omni should the tower you point at go down you can use the next one.. as most are overlapped pretty well and a 35-45W base (mostly mobiles on a power supply) will connect just fine to the next site.
alot of this stuff is no longer done that way.. its IP connected to the system direct. This has benefits.. like dispatcher can talk on the group over the a mobile keyed up...Can't do that with an FX1 station.
Things start to change if you need to get a site outside your normal operations.. ie: a neighboring system for example.
I basically have 3 systems in my area the 2 for the area I am in, and a neighboring system the overlap from each is limited. So you might need higher gain to get that one and the normal one.
I chose to do a yagi aimed at the neighbor site for better performance, ie: rain fade, and normal propagation margins.
its bizzare the hackrf can have that much noise. supprising enoug if i turn on built-in lna amp noise levels drop slighly on many freq. but signal isnt really helped much either. probably poor design or components from china.
I am not familiar with the device.. but honestly if it has the much noise.. SOMETHING IS WRONG SOMEWHERE either the unit. outside interference from USB, ethernet, other electronics or something...as well as if you have fiddled alot with the settings for the gain and VGA.. then you want to reset them.
If inside you can get a lot of RF from all sorts of things..
Can you take this outside and run it on a laptop etc???? Same noise present?????
yes, water fall is full of red but sdrtrunk easily pulls signals out of it. with other programs like sdrconsole or sdrsharp i just lower the waterfall settings to remove red and it leaves signals. We dont have that option with sdrtrunk (idea? maybe).
Not faimilar with those software so likely they don't exist on Linux.. but from the description I think they are lying to you on the noise with that setting by removing it via some setting.. as the only way that would be possible is if there was something that changed the spacing of the noise ie: 1db noise = 100 px on the screen... Which leads to another question do you by chance have the SDRT water fall zoomed in??????
As for adding the feature you describe with out more info, I would say no.. Again I think this other software is faking something and thus lying about the noise.
> im stuck with the hetrodyne option only on this machine at 10mhz BW. but it works pretty damn good.
I am looking at options to come up with a WIDE BW to cover two 850 MHz systems... Which basically need 4 2.4Mhz BW but with the issue with somethig in the underlying USB libs that doesn't seem to work right now.
So I am hoping that maybe something than can do the 6-8-10 MHz and use the hetrodyne mode to use on lower loevel multi core setups and then ONE box can do both of these... Right now with only 3 SDR's the blocking rate, ie: no tuner available situations is far FAR too great for my uses... Even one missed called is really not workable in my setup.
Part of my issue is that I find some of these units are in my opinion excessively priced.. although I may relent and see if this airspy mini will work for $99 while I still personally think its overpriced.. I might try it...
sdrtrunk
Mar 21, 2020, 5:55:50 AM3/21/20
to sdrtrunk
The sdrtrunk spectral display needs a complete redesign and new implementation. The current code does automatic selection of gain/color based on the average content of the signal spectrum. This makes it bad for HackRF as you've seen. I've wanted to add the manual adjustment of gain/color, but am holding off investing any time until I can rewrite the whole thing in JavaFX.
Denny
Dean Sauer
Mar 21, 2020, 6:29:38 AM3/21/20
to sdrtrunk
On Saturday, March 21, 2020 at 5:55:50 AM UTC-4, sdrtrunk wrote:
The sdrtrunk spectral display needs a complete redesign and new implementation. The current code does automatic selection of gain/color based on the average content of the signal spectrum. This makes it bad for HackRF as you've seen. I've wanted to add the manual adjustment of gain/color, but am holding off investing any time until I can rewrite the whole thing in JavaFX.
Hmmm.. is the average off because of the HIGH BW? And thus if reduced to the 2.4MS/s (2.4Mhz) of the more generic units the issue would go away? and Be more accurate?????
I have both the airspy and airspy mini and both are really good devices, but I prefer the airspy as my main sdr.
I probably going to try the airspy mini and see if it couple with a generic SDR at 2.4 will work to solve the issues, or with 2 and it all still play nice... It would be nice to have the 2 850Mhz systems on one box and the harris 700 on another... we'll see...
Imago Trigger
Mar 21, 2020, 7:25:20 AM3/21/20
to Dean Sauer, sdrtrunk
the airspy mini and 3 rtl nooelec dongles works great. alao instead of a $15 fm notch filter try that 20$ crystek 700mhz highpass filter.
--
You received this message because you are subscribed to the Google Groups "sdrtrunk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sdrtrunk+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sdrtrunk/3d4fe0a0-b0d0-4c28-ac88-c07829185edd%40googlegroups.com.
Douglas Welch
Mar 21, 2020, 7:36:52 PM3/21/20
to sdrtrunk
Some Important new Hackrf observations below.
ill probably start with notch or high/low band filters and see where that gets me. Ill figure out how to check the noise levels on hackrf and report back when i figure it out.
after a little research on FX1 Stations (didnt have a clue) i see what youre saying. I also thought a yagi would be too directional to be effectice unless antenna stations were physically in same directions from me. something i can test with a quick hand built design just for fun. I have time not money so it feasible.
here in our metro area in Kansas City, we have KCMO MARRS (Huge area - largest) and JoCo MARRS (where i live) is transmitting on same control freq 24/7 never loose it, but KCMO similcast is off and on constantally and cant find it on other published freqs. even searched to identify KCMOs other freqs by hand and no luck yet. Ive check all details of the sdrtrunk site detail info also. im missing somthing there i think.sdrtrunk is set to search all 4-5 freqs they have published or i found in detail pane.
Yes i would like to have neighbor sites capabilities, so more research is needed.ill have to get a map of site locations all in one map for KC.
some results i found with hackrf today:
1. looks like i have to over drive the LNA and VGA to get signal lock and keep errors down
(probably do to inside antenna-although rtl-sdr works fine)
2. i can see signal with clean blue backgound at lower settings but cant get signal to sync at all until heavier yellow/red background appears.
3. trying to use 2 settings combo that might be just perfect causes severe jumping of waterfall and signal windows
(this doesnt happen with other software) this range is where i think it would start to receive.
LNA=8/VGA=20-24
LNA=16/VGA=20-22
LNA=24/VGA=14-18
LNA=32/VGA=10-16
LNA=40/VGA=4-6
in other software the signal looks pretty strong at these settings
Ill post in the Hackrf forum and see if they know anything from a design standpoint. maybe the incorrect BW detection?
4. setting bandwith for hackrf is inaccurate:
2.016Mhz is 3.50Mhz
4.464Mhz is 6.00Mhz
5.376Mhz is 7.00Mhz
7.488Mhz is 9.00Mhz
10.80Mhz is 12.00Mhz
12.00Mhz is14.00Mhz
13.440Mhz is 15.00Mhz
14.976Mhz is 20.00Mhz
19.968Mhz is 24.00Mhz
where does sdrtrunk get bw settings numbers? (programmed or usb info?) are thes rtl-sdr predefined settings? do they have any impact on amp settings ( i wouldnt think so.)
i use the china hackrf, portapack,10mhz 1ppm osc. with internal shielding installed and full aluminum case. tested with multiple usb cables (but will try others again)
still shots from screen shots wont show the jumping but it is a constant about 1/4 of the display hieght, and smeared waterfall display (you can see).
On Thursday, March 12, 2020 at 7:39:45 PM UTC-5, Douglas Welch wrote:
is it normal with Hackrf (15MHZ sample rate - 850Mhz range) to use Heterodyne to keep from buffer overuns? Using the Polyphase i get overruns within 10-15 seconds of startup.
lets see if i can upload mp4 video...
looks like it uploaded. it is a video of sdrconsole running at the same freqs. and all different AMP settings that seem to jump in sdrtrunk.
here is screen shot of hackrf running at low LNA/VGA settings -- looks pretty but no usable signal for sdrtrunk
I also have been able to run 20Mhz sample rate using Hetrodyne. Not more than 8 Mhz with polyphase. Hackrf runs about 0.2ppm (pretty good- i think) off after warmupjust using a mediocre laptopSystem Type x64-based PCSystem SKU LENOVO_MT_20C5_BU_Think_FM_ThinkPad Edge E440Processor Intel(R) Core(TM) i3-4000M CPU @ 2.40GHz, 2394 Mhz, 2 Core(s), 4 Logical Processor(s)I now pause or disable the waterfall display which helps alot. Also moved down traffic pool down to 4.Runs really nice now with a single Hackrf, my other laptop has 3 Rtl-Sdr (1ppm) dongles and does about the same with lower noice levels and better looking graphics. Only draw back ive found so far is high noise levels i have to rub at. I spent hours test every combo from 12-14 to 40-60 with and without amp. the amp doesnt seem to work well at lna:18 & vga:16-18. The spectrum and Freq. display jump bad, so i dont know if its a software or firmware problem.With the hackrf spectrum screen is so full of noise at all working settings you cant see the the wave forms hardly. (big blob of red). unlike the Rtl-Sdr systemI dont seem to see that particular problem running the portapack by itself.have a couple LNA amps on order from Ali and will be testing much more in few weeks. I need to get a good Discone mounted outside then settings wont be so important. Just using small but perfectly tuned (laugh) small White router antennas. Dont laught they work better than any others tested.for indoor us. I think they are the tplink router 3.25" wirteless antennasIf im not mistaken i found that you can set the center freq using the record freq setting without recording. Is that a Feature, function, or bug? I thought it dynamically set the center freq. (dc-offset)?
RTL-SDR Screen shot
sdrtrunk
Mar 22, 2020, 5:39:39 AM3/22/20
to sdrtrunk
Douglas,
The buffer overruns are caused by the CPU not keeping up with the sample rate. The Heterdyne channelizer is less CPU intensive than the polyphase. Another option may be to use a smaller sample rate with the polyphase.
I updated the HackRF tuner controller code with a new set of sample rates. After reading your posts, I realized that the previous set of sample rates were all based on some previous constraints that longer exist in the code. The new set is matched to the baseband filter settings that are available from the tuner.
These changes will be in the next release, or they're available on master branch now.
Denny
Dean Sauer
Mar 22, 2020, 8:55:08 AM3/22/20
to sdrtrunk
On Saturday, March 21, 2020 at 7:36:52 PM UTC-4, Douglas Welch wrote:
Some Important new Hackrf observations below.ill probably start with notch or high/low band filters and see where that gets me. Ill figure out how to check the noise levels on hackrf and report back when i figure it out.
Filters such as an FM filter if you don't have nearby FMR (88-108Mhz) towers are not really needed... if you have them within 2-3 miles and/or in the path of a yagi pointed towards a tower within 4-5 miles then maybe maybe you need them... It won't hurt anything other than increase the IL or insertion loss overall of the system. most filters are about 1db give or take IL overall.... Some could and are worse! Especially cheap ones. and especially worse the higher the frequency ie: 850Mhz!
As for filters this stuff.. let me just tell you upfront that getting say an 851-861Mhz bandpass filter... is not cheap. THese are not routinely made or are but require purchase of 100+ units etc...
after a little research on FX1 Stations (didnt have a clue) i see what youre saying. I also thought a yagi would be too directional to be effectice unless antenna stations were physically in same directions from me. something i can test with a quick hand built design just for fun. I have time not money so it feasible.
Sorry, my brain works in one way only... and I used to using that term to describe a particular setup....
I use an omni for the LOCAL systems and the NSPAC MA channels..
The yagi is used for one system as the omni is not up to the tasks due to terrain, ground clutter ie: trees etc...Maybe if I could put it up 50 more feet it wouldn't matter...So the yagi is used to ensure that I get a strong signal 100%.
here in our metro area in Kansas City, we have KCMO MARRS (Huge area - largest) and JoCo MARRS (where i live) is transmitting on same control freq 24/7 never loose it, but KCMO similcast is off and on constantally and cant find it on other published freqs. even searched to identify KCMOs other freqs by hand and no luck yet. Ive check all details of the sdrtrunk site detail info also. im missing somthing there i think.sdrtrunk is set to search all 4-5 freqs they have published or i found in detail pane.
Run the KC system. Look under the DETAILS TAB for the Radio Frequency SubSystem or RFSS info it will list the CONTROL CHANNELS for the site you are locked onto.. Write those down. Now depending on some things that is the channels that the radios know to tune to get a lock on to a system.
SDRT will ONLY USE MULTIPLE CC if IF YOU set it up that via the CHANNEL - MULTIPLE FREQUENCIES setup. Which if you are seeing things like the CC disappearing then you need to set this up.. Most systems don't rotate the CC unless the base goes down, maintenance etc... Most MOTO systems just bang a Quantar base at 100% 24/7/365... Other systems do things different Harris LOVES LOVES to rotate CC's! LOVES IT.. WHy? EDACS. EDACS can use ANY SYSTEM in the setup for a CC.. The radios just scan to find a CC out of the programmed LCN's...Harris just sets the controller to do it say every 2 days or could be every 2 hours... just depends...Moto has the same setting.. its just SOP not to program that way... Sometimes when you do changes in programming it requires that the controller reset and that could cause a change in the CC depending on setup.
Yes i would like to have neighbor sites capabilities, so more research is needed.ill have to get a map of site locations all in one map for KC.
FCC ULS will produce a nice map of where sites are.
Search
https://wireless2.fcc.gov/UlsApp/UlsSearch/searchAdvanced.jsp
License:
https://wireless2.fcc.gov/UlsApp/UlsSearch/license.jsp?licKey=1302192
Map
https://wireless2.fcc.gov/UlsApp/UlsSearch/licenseSiteMap.jsp?licKey=1302192
See screen cap
some results i found with hackrf today:1. looks like i have to over drive the LNA and VGA to get signal lock and keep errors down(probably do to inside antenna-although rtl-sdr works fine)2. i can see signal with clean blue backgound at lower settings but cant get signal to sync at all until heavier yellow/red background appears.
I think alot of this has to do with using antennas inside. I can do that one system, and only, reliably. As I honestly literally can look out a window and see all 1300 feet of my tower... :) And my home is an Faraday cage... concrete block filled with rebar and e/uv film on windows..metal roof.
I can use some things inside if I find the RF Hot/Sweet spot for that including the the Harris 700Mhz system their current 850MHz EDACS and another Moto 850Mhz.. if I find the hot spot I can use my 5dbd 850Mhz antennas..
*** TO - >>> ME <<<< **** I would want to get known good antennas ie: Commercial LMR stuff and put to the test on that thing.. I don't have one.. but these wild fluctuations are a good sign of issues with inside antenna use.
Example:
https://www.rfwiz.com/comtelco-products/comtelco-bs698xl3-base-antenna?page=2&limit=100
This is similar to what I have.. they make the one I have:
https://www.rfwiz.com/comtelco-s-series-base/comtelco-bs800sn-c-base-antenna
I know it says "indoor" I've had mine OUTSIDE UP in the air in SEVERE WEATHER, and I do mean SEVERE FOR YEARS! Its fine! There were never marketed by Comtelco as "indoors" in the years ago That I got them...
I personally have dealt with RFWhiz for years as well. A+ company.
This is why buildings are tested during construction to determine signal levels, and if BDA's are needed.
I basically could use a wet noodle as antenna.. oh wait I have one of those.. its a stubby antenna and it works fine to get the one system...the other not so well...again I am getting baked by RF from one even in my house.. the others are not.
Setting a HUGE GAIN level to get a signal is another sign... You want to find the gain sweet spot.. enough to decode, and fade margin for rain, propagation chanes due to hot or cold weather and terrain changes ie: trees swaying, leaf growth, pine needles! YES PINE NEEDLES are just the right length for 850Mhz! URRRGH! So you set it up just a little more.. and also you don't want to have too much or you get spurs from strong signals.
I think the way SDRT derives the noise level for the HackRF is a glitch per the post here...so maybe its not as accurate??? v. say the more generic SDR's????? Not sure.. I'll let the dev speak to that.. not going to go beyond the spitball I did in the one post.. I don't do FFT software stuff.
I would want to get a good known signal level to work with and then go from there... If you have an OP25 setup on something that has a GREAT GREAT signal display via the 1 key that will give power levels in db..then you can follow around the house and find the RF sweet spot and then run SDRT.... Theres a couple other ways to do it if you have some of the GRE or Whistler scanners as well. OP25 is better because after you find a nice RF hot spot you can check the constellation 2 key .. and see if you have a nice pattern or not.. You can have a high RF signal and it be a crap signal with a HIGH BER or Bit Error Rate.. which won't make for a happy experience.
I would start with finding a good RF spot then go from there..
Douglas Welch
Mar 22, 2020, 5:54:02 PM3/22/20
to sdrtrunk
sweet... Ill compile and test now. thanks for all the hard work your doing great job too.
if only more people knew about sdrtrunk
Reply all
Reply to author
Forward
0 new messages