Built In Test (BIT) Capability

  • 5
  • Idea
  • Updated 3 years ago

The concept is not new as some military transceivers currently make these tests at power up, periodically during transceiver operation and manually upon operator command.

The FlexRadio SDRs contain both a broadband exciter and one or two broadband receivers.  Thus I maintain that if this capability does not presently exist, it could be quickly be implemented and become an invaluable tool to both monitor and maintain the radio’s performance.

The basic concept would be to generate a test signal with the Exciter, pass it out the XCVR port and back into the antenna port with a coaxial jumper, through the full receiver chain to the SCU.

The first receiver frequency sweep would bypass the ATU, preamp and filter bank to create a base line.  If this produced an expected signal level over the swept frequency range, continue to the next test else declare a fault.  Provide an option to save the results.

Increase the signal level and make another sweep.  Repeat N times to check the ADC's dynamic range and the SCU.  If this produced the expected signal levels over the swept frequency range, continue to the next test else declare a fault.  Provide an option to save the results.

Next, with a high signal level, exercise all the path switches, one at a time, to confirm their low insertion loss and high open circuit isolation.  If acceptable, pass else declare a fault.  Provide an option to save the results.

Next, with a high signal level, sweep the ham band filters, one at a time, showing insertion loss, band shape and stop band characteristics.  If acceptable, pass else declare a fault.  Provide an option to save the results.

Next, with a low signal level, check the preamp gains over frequency.  If acceptable, pass else declare a fault.  Provide an option to save the results.

Repeat the above for the second SCU, if present.

Other tests could be generated to test the transmitter parameters using a dummy load.

The benefit to the customer is a quick, simple hardware validation, and fault detection of subtle failures.

The benefit to FlexRadio is to showcase their SDR capability, assure customers of product performance, early detection of trends and improve product capability.

Photo of Larry - W9FLA

Larry - W9FLA

  • 23 Posts
  • 2 Reply Likes

Posted 3 years ago

  • 5
Photo of Alex - KM5YT

Alex - KM5YT

  • 68 Posts
  • 10 Reply Likes
All I can say is YESSSS - to all of this! I second and third this proposal - and it is congruent to what I have wanted Flex to do since the F5000:  go to a cage/VME-style architecture that allows ready testing and substitution of modules w/o a trip to the shop.  

Alex/Km5YT
Photo of Martin Ewing AA6E

Martin Ewing AA6E

  • 325 Posts
  • 77 Reply Likes
I had assumed that something of this sort (self test) is happening at every boot of the 6500.  Else what is this "Calibration" that takes so much time? :)

The power on tests only seem to deliver an "OK" or "Not OK" result.  If you get the latter, you need to ship the unit to the factory.  (No user serviceable parts inside!)

73 Martin AA6E
Photo of W5XZ - dan

W5XZ - dan

  • 571 Posts
  • 86 Reply Likes
Photo of Ross - K9COX

Ross - K9COX

  • 352 Posts
  • 108 Reply Likes
Sorry can't resist...BITE me (did not think you would be offended judging by your meme or whatever it's called)
Photo of ka7gzr

ka7gzr

  • 221 Posts
  • 36 Reply Likes

A few thoughts:

Built In Test software is real costly as it sometimes the size of the operational software. The usual government requirement is to detect 95% of detectable faults. It often requires Hardware when done properly. Than you get into whether you want BIT to run at startup or continuously or both.

If the ultimate goal is detect faults to the piece part level with little ambiguity- bring a lot of money. Most consumer electronics assume factory repair so extensive BIT is not provided. Also most BIT goes to the board level as the repair of the board often is more costly then replacing it.

Myself I would like an indication that the SDR is not performing correctly and a general sense of where the fault is at the block diagram level. Especially if an outside source could be at fault. Ambiguity could be high but at least you could look at the outside sources before mailing the SDR back to the factory.

Jim

ka7gzr 

Photo of lyndy brannen

lyndy brannen

  • 42 Posts
  • 5 Reply Likes
I see this idea as a useless expenditure of development time.