Working with a single MAX2771?

[jmfriedt]

Since the PocketSDR cannot be readily found assembled, I am trying to see if I can get the proposed code to work with a single MAX2771 as found on the evaluation board https://www.analog.com/en/resources/evaluation-hardware-and-software/evaluation-boards-kits/max2771evkit.html which is fitted with a USB->SPI converter but cannot stream IQ data. I have hence additionally acquired an EZ-USB FX2P board https://fr.aliexpress.com/item/1005006134347046.html which I have successfully flashed with the firmware proposed in this repository (PID/VID correct). I have routed the SPI pins (PD0=CS#, PD2=SCLK, PD3=DATA, PD4=lock) to the associated pins of the unpowered microcontroller of the eval board, and routed the IQ stream from the eval board I0,I1,Q0,Q1 to PB0-4 as well as CLK_OUT to IFCLK. Finally SHDN was pulled up to 3V3.

I think I can configure the board correctly since sudo ./app/pocket_conf/pocket_conf returns a reasonable answer starting with

#  Pocket SDR device settings (MAX2771)
#
#  [CH1] F_LO = 1575.420 MHz, F_ADC =  4.000 MHz (IQ), F_FILT =  0.0 MHz, BW_FILT =  2.5 MHz
#  [CH2] F_LO =     -nan MHz, F_ADC = 48.000 MHz (I ), F_FILT =  0.0 MHz, BW_FILT =  2.5 MHz

after I sudo ./app/pocket_conf/pocket_conf conf/pocket_L1L2_4MHz.conf.

However

sudo ./app/pocket_dump/pocket_dump 
  TIME(s)    T   CH1(Bytes)   T   CH2(Bytes)   RATE(Ks/s)
         0.8   IQ            0   I            0          0.0

keeps on returning empty files and 0 Bytes: is there a requirement I am missing for this software to work with a single channel? Am I missing an option that would allow me to collect these data? I have checked that the CLK_OUT->IFCLK is correctly clocked and that the IQ states are changing.

I know this is not strictly PocketSDR related but any hint would be welcome to expand the users of this software (and possibly later assemble a dedicated board).

Thanks.

[jmfriedt]

One more issue identified: PocketSDR is clocked at 24 MHz, MAX2771EvalBoard is 16.368 MHz. Disconnecting the on-board local oscillator of MAX2771EvalBoard and feeding Ref_In with 24 MHz allows to get the right Clk_Out frequency and some communication over USB, yet still some

libusb bulk transfer error (2)
libusb bulk transfer error (2)
        3.6   IQ     12582912   I      6291456       1734.1

error in the transfer (status 2 is LIBUSB_TRANSFER_TIMED_OUT).

Clocking the MAX2771 on the same 24 MHz source than the EZ-USB (XtalOut output pin from EZ-USB to MAX2771 Ref_In) does not help in preventing the libusb bulk transfer error, so the issue is not with the 24 MHz frequency differences.

[jmfriedt]

The USB configuration sounds good with lsusb -v displaying all bulk transfer endpoints

Bus 003 Device 094: ID 04b4:1004 Cypress Semiconductor Corp. EZ-USB
Device Descriptor:
  bLength                18
  bDescriptorType         1
  bcdUSB               2.00
  bDeviceClass            0 [unknown]
  bDeviceSubClass         0 [unknown]
  bDeviceProtocol         0 
  bMaxPacketSize0        64
  idVendor           0x04b4 Cypress Semiconductor Corp.
  idProduct          0x1004 EZ-USB
  bcdDevice            0.00
  iManufacturer           1 Cypress
  iProduct                2 EZ-USB
  iSerial                 0 
  bNumConfigurations      1
  Configuration Descriptor:
    bLength                 9
    bDescriptorType         2
    wTotalLength       0x002e
    bNumInterfaces          1
    bConfigurationValue     1
    iConfiguration          0 
    bmAttributes         0x80
      (Bus Powered)
    MaxPower              100mA
    Interface Descriptor:
      bLength                 9
      bDescriptorType         4
      bInterfaceNumber        0
      bAlternateSetting       0
      bNumEndpoints           4
      bInterfaceClass       255 Vendor Specific Class
      bInterfaceSubClass      0 [unknown]
      bInterfaceProtocol      0 
      iInterface              0 
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x02  EP 2 OUT
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0200  1x 512 bytes
        bInterval               0
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x04  EP 4 OUT
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0200  1x 512 bytes
        bInterval               0
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x86  EP 6 IN
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0200  1x 512 bytes
        bInterval               0
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x88  EP 8 IN
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0200  1x 512 bytes
        bInterval               0
Device Qualifier (for other device speed):
  bLength                10
  bDescriptorType         6
  bcdUSB               2.00
  bDeviceClass            0 [unknown]
  bDeviceSubClass         0 [unknown]
  bDeviceProtocol         0 
  bMaxPacketSize0        64
  bNumConfigurations      1
Device Status:     0x0002
  (Bus Powered)
  Remote Wakeup Enabled

and the IQ stream is fluctuating over time

but the binary files after complaining about timeout only display 01 values so the USB transfer with sudo app/pocket_dump/pocket_dump is failing, not sure why yet. I filled the (missing) second MAX2771 reg values in src/sdr_dev.c with the values of the only MAX2771 on the eval board but same result.

[tomojitakasu]

To enable bulk-transfer to host by slave-FIFO with external clock, you have to configure the pins of EZ-USB FX2LP:

• RDY0/SLRD pull-up (disabled)
• RDY1/SLWR pull-down (enabled)
• IFCLK/PE0 connected to external clock line.
• PA4/FIFOADR0 pull-down (0)
• PA5/FIFOADR1 pull-up (1)

Please check them.

pocket_conf assumes the TCXO frequency 24 MHz . So wrong frequencies are shown with other TCXO freqencies. To correct them, modify the constant F_TCXO in FE_2CH/FW/v2.1/pocket_fw.c and rebuild the F/W for EZ-USB FX2LP.

[jmfriedt]

Thank you for pointing these out. I had missed RDY0 and RDY1 pull up/down indeed, these were missing and I erroneously thought that seeing the bulk transfer endpoints with lsusb -v was enough to assess proper configuration.

Clock to the MAX2771 is derived from the FX2LP 24 MHz oscillator, CLK_OUT is indeed 4 MHz square wave (as configured with conf/pocket_L1L2_4MHz.conf), I and Q outputs are toggling state, PA4 to GND and PA5 to Vcc, RDY1 to GND and RDY0 to Vcc, CLK_OUT to IFCLK but still

libusb bulk transfer error (2)

after the 3 second timeout, no error message in the kernel logs.

[jmfriedt]

Is there a way to debug the firmware executed in the FX2LP? My concern is that the firmware is not executing properly, but doesn't the fact that I can read correctly the MAX2771 status after configuring and that the USB interface is configured with the right PID/VID prove that the FX2LP is running correctly? I tested with https://gist.github.com/siddharthdeore/b1aac400ec94ca53730291a92279ea0e and this example is running correctly with bulk endpoint 6 communicating correctly even though I cannot find its FX2LP source code: my concern is that the PocketSDR firmware address range lies between 0x099C and 0x09F8 when the functional example cited above is in the 0000 to 0x0235 so maybe I am doing something wrong with executing the firmware in the 8051 micrcontroller. I flashed using

sudo ./fxload -D /dev/bus/usb/003/087 -I PocketSDR/FE_2CH/FW/v2.1/pocket_fw.hex -c 0xc2 -s Vend_Ax.hex -t fx2

after unloading the usbtest kernel module.

Another concern might be that the board I am using is fitted with CY7C68013A-56PVXC whereas PocketSDR is using CY7C68013A-56LTXCT so different pin format: can this make any difference?

[tomojitakasu]

Some chinese FX2LP EVK boards have wrong silk prints for the header pins. Refer the following blog with Google translate. I also fell into the pitfall and wasted time.

https://blog.goo.ne.jp/osqzss/e/d86df04de96123fd5c73bbb6db6e8bc5

[jmfriedt]

Yes this was it !

TIME(s)    T   CH1(Bytes)   T   CH2(Bytes)   RATE(Ks/s)
      3.2   IQ     25690112  IQ     25690112       4025.4

Thank you so much for your help. Now I can get to some real work.

[jmfriedt]

For the record, it all ends up working with a single MAX2771 when

sudo ./app/pocket_conf/pocket_conf conf/pocket_L1L6_12MHz.conf 
sudo ./app/pocket_conf/pocket_conf 
sudo ./app/pocket_dump/pocket_dump -r -t 1 ch1.bin                             # -r to NOT interleave
python3 ./python/pocket_acq.py ch1.bin -f 12 -fi 3 -sig L1CA -prn 1-32 

Explanation:

1. while the pocket_L1L2_4MHz.conf allows for detecting a single tone from a synthesizer (tested at -90 dBm) fed before or after the high-band LNA with a frequency offset of 100 to 500 kHz from L1 carrier, this circuit is far too noisy around baseband to detect GPS. Hence the need to offset the IF by 3 MHz and sample at 12 MS/s
2. verification of the configuration (validates SPI and config)
3. Although I did change in src/sdr_conf.c (maybe a bad idea)

   static int max_ch(int type)
   {
   switch (type) {
       case TYPE_POCKET_2CH: return 1;

   by replacing return 2 with return 1, I still had two files created by pocket_dump => using the -r switch led to the accumulation of all IQ samples in a single file representing the single MAX2771 I am using

4. Finally acquisition of all the signals as shown below, matching the constellation seen by my mobile phone.

thank you for all the amazing tools !