Hard hack - Embedding a SDR in FT-991A - Need RF designers review

Discussion in 'Radio Circuits, Repair & Performance' started by PY2RAF, Mar 18, 2019.

ad: L-HROutlet
ad: l-rl
ad: L-MFJ
ad: Left-3
ad: Left-2
ad: Subscribe
ad: MessiPaoloni-1
  1. PY2CP

    PY2CP Ham Member QRZ Page

    you made it!! cheers! de Enrique PY2CP
     
    PY2RAF likes this.
  2. IZ0ABD

    IZ0ABD Ham Member QRZ Page

    PY2RAF likes this.
  3. PY2RAF

    PY2RAF XML Subscriber QRZ Page

    Enrique querido, thanks for the kind words!!!
    Miss the friend and look forward to meet you again, rather sooner than later.

    Ciao, Francesco!

    Grazie Mille for the kind words. I appreciate, coming from someone with a proper degree :)

    I saw your board, looks great and nifty!! Wish you success! Got curious here on what’s the U1 IC in your board.

    My board borrows the LPF design from G4HUPs (SK) 70 MHz PAT-70 boards, since this is the IF center of my equipment.

    My board is highly specialized for FT-991/A, as you can deduce from the board layout: OE2DOR designed it to fit perfectly in this equipment. And the USB part is also specific for the 991/A radio.

    In the end, this was more a entertaining and fun enterprise. And lots of learning, of course. I have also added a few extra functionalities in the experimental branch. I have ordered the PCB and components yesterday, which PY2UGO will be kindly freighting to me.

    I think that in 3-4 weeks I’ll have the Rev. M deployed with its whistles and bells and will post the outcome here.

    I’m spending this weekend at the country where I’ll be testing and benchmarking against my 40-20 dipole a brand new b-squared engineering T2FD antenna. Will post (in a different thread) my findings.

    Well that’s it

    Thank you very much again

    73s de PY2RAF

    P.S.: I’m starting to document the whole thing at https://github.com/rfrht/FT991A-PAT/wiki
     
    Last edited: Jun 1, 2019
  4. AD5GG

    AD5GG Ham Member QRZ Page

    PY2RAF likes this.
  5. PY2RAF

    PY2RAF XML Subscriber QRZ Page

    Update.

    [tl;dr]

    Went yesterday to Alpha Telecom office and replaced USB2512B with a USB2514B. Unfortunately, it did not work as planned.

    [long post]
    Yvo replaced USB2512B with USB2514B and it was a breeze. The original Audio and CAT ports were identified and functioned normally.
    After having a hard time figuring out how to tap the signals of interest from USB2514, he managed to solder two thin wires to the terminals, that were used to feed to a USB cable, straight to the SDR.

    Here is the bus enumeration sequence.

    Dectected USB2514 hub
    Jun 15 16:00:04 rf kernel: usb 3-3: new high-speed USB device number 10 using xhci_hcd
    Jun 15 16:00:04 rf kernel: usb 3-3: New USB device found, idVendor=0424, idProduct=2514
    Jun 15 16:00:04 rf kernel: usb 3-3: New USB device strings: Mfr=0, Product=0, SerialNumber=0
    Jun 15 16:00:04 rf kernel: hub 3-3:1.0: USB hub found
    Jun 15 16:00:04 rf kernel: hub 3-3:1.0: 4 ports detected


    Dectected the CAT interface
    Jun 15 16:00:05 rf kernel: usb 3-3.1: new full-speed USB device number 11 using xhci_hcd
    Jun 15 16:00:05 rf kernel: usb 3-3.1: New USB device found, idVendor=10c4, idProduct=ea70
    Jun 15 16:00:05 rf kernel: usb 3-3.1: New USB device strings: Mfr=1, Product=2, SerialNumber=5
    Jun 15 16:00:05 rf kernel: usb 3-3.1: Product: CP2105 Dual USB to UART Bridge Controller
    Jun 15 16:00:05 rf kernel: usb 3-3.1: Manufacturer: Silicon Labs
    Jun 15 16:00:05 rf kernel: usb 3-3.1: SerialNumber: 009FF28C
    Jun 15 16:00:05 rf mtp-probe: checking bus 3, device 11: "/sys/devices/pci0000:00/0000:00:14.0/usb3/3-3/3-3.1"
    Jun 15 16:00:05 rf mtp-probe: bus: 3, device: 11 was not an MTP device
    Jun 15 16:00:05 rf kernel: usbcore: registered new interface driver cp210x
    Jun 15 16:00:05 rf kernel: usbserial: USB Serial support registered for cp210x
    Jun 15 16:00:05 rf kernel: cp210x 3-3.1:1.0: cp210x converter detected
    Jun 15 16:00:05 rf kernel: usb 3-3.1: cp210x converter now attached to ttyUSB0
    Jun 15 16:00:05 rf kernel: cp210x 3-3.1:1.1: cp210x converter detected
    Jun 15 16:00:05 rf kernel: usb 3-3.1: cp210x converter now attached to ttyUSB1


    Detected the audio interface
    Jun 15 16:00:05 rf kernel: usb 3-3.2: new full-speed USB device number 12 using xhci_hcd
    Jun 15 16:00:05 rf kernel: usb 3-3.2: New USB device found, idVendor=08bb, idProduct=29b3
    Jun 15 16:00:05 rf kernel: usb 3-3.2: New USB device strings: Mfr=1, Product=2, SerialNumber=0
    Jun 15 16:00:05 rf kernel: usb 3-3.2: Product: USB Audio CODEC
    Jun 15 16:00:05 rf kernel: usb 3-3.2: Manufacturer: Burr-Brown from TI
    Jun 15 16:00:05 rf kernel: input: Burr-Brown from TI USB Audio CODEC as /devices/pci0000:00/0000:00:14.0/usb3/3-3/3-3.2/3-3.2:1.3/input/input19
    Jun 15 16:00:05 rf kernel: hid-generic 0003:08BB:29B3.0004: input,hidraw3: USB HID v1.00 Device [Burr-Brown from TI USB Audio CODEC ] on usb-0000:00:14.0-3.2/input3
    Jun 15 16:00:05 rf mtp-probe: checking bus 3, device 12: "/sys/devices/pci0000:00/0000:00:14.0/usb3/3-3/3-3.2"
    Jun 15 16:00:05 rf mtp-probe: bus: 3, device: 12 was not an MTP device
    Jun 15 16:00:05 rf kernel: usbcore: registered new interface driver snd-usb-audio


    Detected but failed to enumerate the SDR device
    Jun 15 16:00:05 rf kernel: usb 3-3.3: new full-speed USB device number 13 using xhci_hcd
    Jun 15 16:00:05 rf kernel: usb 3-3.3: device descriptor read/64, error -32
    Jun 15 16:00:05 rf kernel: usb 3-3.3: device descriptor read/64, error -32
    Jun 15 16:00:05 rf kernel: usb 3-3.3: new full-speed USB device number 14 using xhci_hcd
    Jun 15 16:00:05 rf kernel: usb 3-3.3: device descriptor read/64, error -32
    Jun 15 16:00:06 rf kernel: usb 3-3.3: device descriptor read/64, error -32
    Jun 15 16:00:06 rf kernel: usb 3-3.3: new full-speed USB device number 15 using xhci_hcd
    Jun 15 16:00:06 rf kernel: usb 3-3.3: Device not responding to setup address.
    Jun 15 16:00:06 rf kernel: usb 3-3.3: Device not responding to setup address.
    Jun 15 16:00:06 rf kernel: usb 3-3.3: device not accepting address 15, error -71
    Jun 15 16:00:06 rf kernel: usb 3-3.3: new full-speed USB device number 16 using xhci_hcd
    Jun 15 16:00:06 rf kernel: usb 3-3.3: Device not responding to setup address.
    Jun 15 16:00:06 rf kernel: usb 3-3.3: Device not responding to setup address.
    Jun 15 16:00:07 rf kernel: usb 3-3.3: device not accepting address 16, error -71
    Jun 15 16:00:07 rf kernel: usb 3-3-port3: unable to enumerate USB device


    After checking USB2514 design checklist, I have found no special requirement for capacitors or resistors - It should be wired directly to the port, so no breaking news.

    I have also checked for a few possible hypothesis:

    * Bad USB host - Ruled out, have the same behaviour in both Windows and Linux
    * Bad PC negotiation - Ruled out, as the other devices were correctly detected
    * Voltage ripple: Ruled out. There were some voltage variation when plugging in the USB hub. As a counter-measure, I bridged the 13.8V straight to the ENABLE in the power regulator, so it would be on all the time, avoiding the ripples. Issue persisted.
    * Voltage sag: Might be possible. See below thoughts.
    * Design issue: ruled out, after revisiting the USB hub Datasheet and Design checklist, everything seem to be sound and good.
    * Bad USB hub config: Slim chances; There are some "Strap Configs" for this USB hub, but they seem to be more about overcurrent and presenting to the computer as permanently attach or removable devices.
    * Bad wiring: Checked and re-checked and everything seem to be in good order.
    * Bad port: There's a slim chance that Port 3 is just plain bad/damaged. Might be worth a shot trying Port 4.

    Now, a few words about linear power regulators.
    When powering the SDR from the board, the 13.8V -> 9V regulator gets REAL HOT. _Very_ hot. While the 13.8 - 9V regulator datasheet specifies that Tjunction for this regulator is 150 Celsius (and temperature was definitely below that) and a few Ohm law calculation, I'm afraid that while the SDR during operation draws approx. 310 mA in 5 V during peak operation, the efficiency, the PCB substrate and the heat dissipation might put everything at a very tight power / heat dissipation budget. This has to be better evaluated and thought of.

    For what is worth, the circuit was drawing 160 mA from the 13.8V line when powering the SDR.

    As of now, I have reverted to the original setup (just took the opportunity and better wrapped the SDR and added, used a beefier USB cable and added a good grounding to the board, everything got a lot cleaner) and I still have that hateful USB cable coming out of the radio.

    Let's see how that develops.

    My 73s,

    PY2RAF.
     
  6. PY2RAF

    PY2RAF XML Subscriber QRZ Page

    Have not thrown the towel yet.

    While I don't do a final round of tests involving the USB hub, I have released Revision N - In a different github branch, named "NoUSB".

    In this branch, this is a panadapter board without USB stuff, but with a few features, namely:

    1. IF TAP
    RATIONALE: You can select to tap or not tap the signal from your SDR. By default, the Panadapter board will not consume any IF signal if you don't explicitly enable it.
    STATES:
    * Isolate the SDR (default) - No signal is consumed from the radio and the SDR is left grounded
    * Tapping IF signal - Feed the signal to the next stage in the Panadapter board.

    2. BAND-PASS FILTER
    RATIONALE: Ability to choose to either a filtered signal (with insertion loss and reduced bandwidth) or the whole IF signal
    STATES:
    * Pass-Through (default) - The IF signal is forwarded untouched and unaltered ahead to the next stage. There is no insertion loss and the signal features the entire IF bandwidth.
    * Band-Pass - The signal is filtered at the center of the IF frequency (69.450 MHz), giving 3 MHz of bandwidth, filtering strong signals that might generate image artefacts in the SDR. The RTL-SDR optimum bandwidth is 2.4 MHz.

    3. HIGH IMPEDANCE AMPLIFIER
    RATIONALE: Ability to use the high impedance amplifier + low noise amplifier or choose to tap the IF signal unaltered
    STATES:
    * Amplified (default) - The IF signal enters a JFET source follower to present a high impedance to the IF filter, minimising any loading on the main signal path of the radio. This is fed in a 2SC5086 as an emitter follower to amplify the signal.
    * Pass-Through - The signal is tapped from the IF signal and forwarded straight out to the SDR port.

    A few more changes:
    * Got rid of the G4HUP LPF. Reasoning: Found it of no use, since it cuts the signal 24 MHz above the IF center frequency. The selectable BPF suits better (fingers crossed).
    * Moved regulators and signal pads to the upper board side (access & heat dissipation)
    * A slimmer BOM (more than 10 components removed)

    Board ordered, around a month to arrive, will let you all know the outcome.

    73s,

    - Rodrigo.
     
  7. PY2RAF

    PY2RAF XML Subscriber QRZ Page

    IT'S ALIVE!!!!!!!! \o/

    And it WORKS!!!

    Everything coming down through a single USB cable!!!! :D

    Video here.

    It was a missing pair of 22 ohm resistors on the USB line. And that was it!

    The documentation and test results are all in the Project's Wiki.

    Now everything coming down from the Radio's USB port, in tandem; the Audio, CAT and SDR:

    [​IMG]

    Everything working nicely and great! And fortunately this is no longer the case here:

    [​IMG]

    or

    [​IMG]

    MANY THANKS to Everyone that helped this to come true!

    73!!!
     
    KC9ONY likes this.
  8. KC9ONY

    KC9ONY Ham Member QRZ Page

    Congratulations Rodrigo! Persistence pays off!
    Like the documentation as well. :cool:
     
    PY2RAF likes this.
  9. KC3LNW

    KC3LNW Ham Member QRZ Page

    PY2RAF,

    Firstly I would like to thank you very much for the time and effort your have put into this great idea. It's obvious you have a passion for ham radio and solving design problems.

    I have decided to build and install the PAT into my 991A. I've just received my revision-O boards a few days ago and my components from Digi-Key. (Thanks again for those links they really save time).

    So far things have been progressing quite well. I've got my board assembled.

    There was only one minor descrapency that I would like to question. The Revision-O board notes that F1 is 150ma, but the Digi-Key parts list includes only a 50ma fuse. I assume the parts list is accurate and that the board should read 50ma. Is this correct?

    Additionally, I do not have access to any sophisticated board testing equipment other than a power supply and simple volt/amp meter. I've tested for shorts and open connections as best I could with the tools I have, and I've even inspected each component under a microscope to make sure there were no pins shorted, but I would still like to be extremely cautious before installing into the radio.

    Can you explain the step you describe in the WIKI, to power and test board, prior to installing it into the radio? Can I simply supply +13.8v power and measure the current draw?


    I have a few questions about connecting the PAT to the RST-SDR and the radio.

    1) The GetHub instructions note to "Route the GPIO cable and connect the respective GPIO cables to the RX_EN, BPF and BYPASS pads" It appears that the Revision-O board does not have a RX-EN pad. Which pad should I connect RST-SDR GPIO Pin 32 to? LINK
    2) The Installation block diagram indicates that SDR GPIO Pin 29 should be conneted to the OR Gate of the PAT. Could you tell me which PAD on the Revision-O board this corresponds to? LINK


    Finally, I've modified my RTL-SDR. However I'm not ready to replace the USB hub in the radio yet. While my skills with the hot air soldering station are ok, they I'm not yet confident enough to attempt replacing the USB Hub. I noticed that during your design progress, you conducted an intermediate step where you ran the USB cable through the back of the radio. This is my intention as well, until I can find an experienced technician to replace the hub, or learn to do it myself. I may 3D Print a USB Receptacle which I can fasten to the back of the radio in order to provide a cleaner appearance until I can replace the USB.


    So I have the following questions:

    1) Is the functionality of the 991A hampered in any way, if I choose to NOT connect the USB for the RST-SDR?
    2) Would it operate normally if left powered or do I always need to keep the RST-SDR powered when using the radio?
    3) Also can you think of any problem that could occur electronically with the RST-SDR ground connected to both the external USB cable and the internal GPIO Shielded cable?


    I'm really looking forward to finishing this project and using my 991A again. I realize we all have day jobs and our Hobbies come behind work and Family. Thank you in advance for your time and assistance.

    Regards,

    John (KC3LNW)
     
    PY2RAF likes this.
  10. PY2RAF

    PY2RAF XML Subscriber QRZ Page

    Hey there John!
    Answers below, inline.

    > Firstly I would like to thank you very much for the time and effort your have put into this great idea. It's obvious you have a passion for ham radio and solving design problems.
    Thank YOU very much - This is exactly what makes me happy, sharing a project and people finding it of use - It humbles me.

    > I have decided to build and install the PAT into my 991A. I've just received my revision-O boards a few days ago and my components from Digi-Key. (Thanks again for those links they really save time).
    Glad that it was of help!

    > So far things have been progressing quite well. I've got my board assembled.
    Looks GREAT!!!!! Well done!

    > There was only one minor descrapency that I would like to question. The Revision-O board notes that F1 is 150ma, but the Digi-Key parts list includes only a 50ma fuse. I assume the parts list is accurate and that the board should read 50ma. Is this correct?
    That's correct - And thanks for pointing that. Will correct it in the coming days.
    For reference, this is the current draw of the Rev. I board:
    https://rf3.org:8443/q/pat-991a/rev-i-tests/03-current-draw.jpg
    9.5 mA - So 50 mA ought to be more than enough. Sorry about that again.
    For what is worth this is a simpler design than Rev. I - Expect even less current draw.

    > Additionally, I do not have access to any sophisticated board testing equipment other than a power supply and simple volt/amp meter. I've tested for shorts and open connections as best I could with the tools I have, and I've even inspected each component under a microscope to make sure there were no pins shorted, but I would still like to be extremely cautious before installing into the radio.
    Man, I know the feeling - trust me ;-)

    > Can you explain the step you describe in the WIKI, to power and test board, prior to installing it into the radio? Can I simply supply +13.8v power and measure the current draw?
    See previous the image link that I sent you - That's what I've got in my first test. And good call; will add the image and figures in the wiki. Great question you made.

    > I have a few questions about connecting the PAT to the RST-SDR and the radio.
    A foreword: The Wiki was written taking into account the Revision N and there were a few logic and assumptions change. Now let's move forward.

    1) The GetHub instructions note to "Route the GPIO cable and connect the respective GPIO cables to the RX_EN, BPF and BYPASS pads" It appears that the Revision-O board does not have a RX-EN pad. Which pad should I connect RST-SDR GPIO Pin 32 to? LINK
    SO. Previously, the board was **disabled by default** (the first RF switch was putting the IF port to reflect and connecting Ground & SDR in). This now have changed.
    In revision O, the board is ON BY DEFAULT, meaning, the if port conduct to SDR in. Now, it will change to IF Reflect/ Connect Ground & SDR IN in the event of the TX9 signal (notice here, big change!) **OR** a GPIO command.

    Meaning: We will flip the SDR to ground & untap FI on the event of a transmission OR if you decide to do so using the SDR's GPIO signal.

    This is a suggestion on where to tap the TX9 signal:
    https://rf3.org:8443/q/pat-991a/tx9-board.png
    Schematic:
    https://rf3.org:8443/q/pat-991a/tx9-schema.png
    The TX9 signal is created at this point:
    https://rf3.org:8443/q/pat-991a/tx9-source.png

    > 2) The Installation block diagram indicates that SDR GPIO Pin 29 should be conneted to the OR Gate of the PAT. Could you tell me which PAD on the Revision-O board this corresponds to? LINK
    TX9.

    > Finally, I've modified my RTL-SDR. However I'm not ready to replace the USB hub in the radio yet. While my skills with the hot air soldering station are ok, they I'm not yet confident enough to attempt replacing the USB Hub. I noticed that during your design progress, you conducted an intermediate step where you ran the USB cable through the back of the radio. This is my intention as well, until I can find an experienced technician to replace the hub, or learn to do it myself. I may 3D Print a USB Receptacle which I can fasten to the back of the radio in order to provide a cleaner appearance until I can replace the USB.
    Nice, John!! Do **not** forget the ferrite choke in the USB line - Otherwise you will experience SDR hangs during your TX.
    See https://rf3.org:8443/q/pat-991a/rev-i-tests/12-chassis.jpg

    > So I have the following questions:
    > 1) Is the functionality of the 991A hampered in any way, if I choose to NOT connect the USB for the RST-SDR?

    I have found no loss of functionality or sensitivity - FT8 still does happily with almost impossible to hear signals, etc.

    > 2) Would it operate normally if left powered or do I always need to keep the RST-SDR powered when using the radio?
    Mine is exactly in this scenario. I would like to be able to switch it off somehow but I didn't find a simple way to achieve it <shrugs>

    > 3) Also can you think of any problem that could occur electronically with the RST-SDR ground connected to both the external USB cable and the internal GPIO Shielded cable?
    I found this of no significance for ground loops, as I ran for a good amount of months with TWO USB cables, the SDR and the Radio (see previous pic). No enhancement (Besides being tidier of course) or impairment noticed between both scenarios.
    Remeber the ferrite choke!!!!!!!!

    > I'm really looking forward to finishing this project and using my 991A again. I realize we all have day jobs and our Hobbies come behind work and Family. Thank you in advance for your time and assistance.
    Feel free to reach - Again, I'm honoured that you tried it!

    One last thing, I haven't tested the BPF yet - In vitro and In vivo are *very* different things
    The design came from this thread if you haven't seen: https://forums.qrz.com/index.php?threads/filter-design-help.671664/

    This is my last try in a LC filter - If this one fails too, I'll move to a SAW filter in the next iteration.

    73 de PY2RAF o/
     
    KC3LNW likes this.

Share This Page