Which HackRF Firmware?

[robotastic]

Which version of the HackRF firmware does this support? The original version for the jawbreaker had unsigned output and the newer version that supports the HackRF One has signed output.

I have the Jawbreaker and haven't upgraded yet. I can't seem to get Kalibrate to work, so I am wondering if it is an issue with my firmware version.

[ckuethe]

I haven't made any changes to support recent HackRF firmware, so it probably assumes old format data. Thanks for pointing this out; I'll have a look.

[rxseger]

This may be the same problem encountered when porting dump1090 from rtlsdr to hackrf (in what is now dump1090_sdrplus):

http://hackrf-dev.greatscottgadgets.narkive.com/4mLdz4BJ/adding-hackrf-support-to-dump1090

Donald Pupecki 12 months ago
I don't believe hackrf_transfer and rtl_sdr output in the same format. They
are both 8 bit but the hackrf output is signed whereas the rtl output is
unsigned.

If you get this working a hackrf please post the port on github or
somewhere as that would be sweet.

Ilker Temir 12 months ago
Thanks! signed to unsigned conversion was the culprit. I first tested it
with cox by converting the file and then tweaked my code to do that
within the receive path.

itemir/dump1090_sdrplus@f02a8cf#diff-4844415e789781b82d79d4819e97d461R478

[rxseger]

A stab in the dark, but if I make this change:

diff --git a/src/usrp_source.cc b/src/usrp_source.cc
index 4a80657..75965a3 100644
--- a/src/usrp_source.cc
+++ b/src/usrp_source.cc
@@ -295,6 +295,11 @@ hackrf_rx_callback (hackrf_transfer * transfer)
   //  cout << transfer->valid_length  << " " << hackrf_rx_count << " " << bytes_to_write << "\n";
   if (bytes_to_write != 0)
     {
+      /* HackRF One returns signed IQ values, convert them to unsigned */
+      for (uint32_t i = 0; i < bytes_to_write;  i++) {
+        transfer->buffer[i] ^= (uint8_t)0x80;
+      }
+
       memcpy (u->ubuf + u->hackrf_rx_count, transfer->buffer, bytes_to_write);
       //    for (size_t i=0; i<bytes_to_write; i++) {
       //      hackrf_rx_buf[hackrf_rx_count+i] = transfer->buffer[i];

it has no discernible effect (still get about +19 ppm with channel 995, -35 ppm with 993). Same results with 2014.08.1 and 2015.07.2 firmware.

@robotastic Could you elaborate (sorry its been a while, if you remember), how did kalibrate-hackrf not work for you? Were you seeing inconsistent offsets on different channels as I am, or was it something else?