[Allstar Digital] DV3000 testing results
Steve Zingman
szingman at msgstor.com
Wed Jul 12 15:33:48 EDT 2017
We have spent the last couple of weeks working with the DV3000u and
below are our results.
This what we sent to NorthWest Digital
===================================================================================================
* Hardware under test
o DV3000U set up as direct serial and AMBEServer configurations
(at least 4 devices have been tested)
o X86 based computers (Linux, OSX)
o ARM based computers (Raspian, Comapss, Armbian)
o Power supplies
+ Anker 60 watt multi-port supply
+ Power cube with 1.35A @ 5V output
+ X86 using internal PS
* Test scenarios
o Run a python test application designed to exercise and test
returned results
o Application is run on X86 machines and ARM based machines with
both power supply configurations
o Test loops are run from between 10,000 or greater passes
o Each pass both encodes and decodes AMBE frames
o Tests may be run with the command line
+ python AMBETest4.py -v -e -n -s /dev/ttyUSB0
+ Where
# -v = verbose operation
# -e = stop on errors
# -n = use 460800 baud serial
# -s port = use device “port”
# -I address = use AMBEServer “address” (default port)
+ https://github.com/N4IRS/Analog_Bridge/raw/master/DV3000/AMBEtest4.py
* Test results
o X86
+ All tests perform as expected, both serial and AMBEServer pass
o ARM
+ Using the Anker 60W PS all tests pass
+ Using the Anker PS with a short USB extension, tests fail
+ Using the smaller PS tests fail (expected)
o Observations
+ The failure mode is consistent, the device will reset when
under load
+ We see the PKT_READY frame returned from the DV3000 whenever
the device fails
# The frame 61 00 01 00 39 is documented in the AMBE3000
reference as PKT_READY
# The doc states that this frame is produced when ever the
device undergoes a hardware reset
+ Failure happens on either AMBE -> PCM or PCM -> AMBE
+ We see no indication of any serial data loss
# No data corruption
# No data loss, all data is sent and received
+ The short jumper is of good gauge wire and shows no
observable damage
* Conclusions
o The DV3000u will run fine on x86 and SBC machines provided that
+ USB power is sufficient to keep the dongle running
+ Under load, the power may dip causing an install that seems
good to fail
+ The reset is clean, leading me to think that the TI TPS3805
is doing the reset, not the AMBE3000 itself
+ You are welcome to use the AMBETest4.py app, add to it, etc
as needed.
+ You should make your customers keenly aware that all PS are
not created equal and care must be exercised.
+ We have modified our Analog_Bridge application to sense this
reset and output error messages in the log
Two additional pieces of information.
I have a Pi2 that has the PiDV installed running Raspbian. All test pass
on the PiDV
I added a ThumbDV to the same board and ran the test against the ThumbDV
root at pi-star(ro):DV3000# python AMBEtest4.py -v -e -n -s /dev/ttyUSB0
Setting serial port
Serial port parameters:
Port name: /dev/ttyUSB0
Baudrate: 460800
Byte size: 8
Parity: N
Stop bits: 1
Xon Xoff: False
RTS/CTS: False
DST/DTR: False
*********************
Testing Reset DV3000
Test result: Success (6100010039)
Testing Get Product ID
Test result: Success (61000b0030414d4245333030305200)
Testing Get Version
Test result: Success
(6100310031563132302e453130302e585858582e433130362e473531342e523030392e42303031303431312e433030323032303800)
Testing Set DSTAR Mode
Test result: Success (610002000a00)
Testing Reset DV3000
Test result: Success (6100010039)
Testing Set DMR Mode
Test result: Success (610002000a00)
Testing Decode AMBE
Test result: Success (6100010039)
Error, the DV3000 has unexpectly reset
root at pi-star(ro):DV3000#
I moved the DV3000u to a D-Link powered hub. I ran AMBEtest4.py for
100,000 iterations.
All tests pass.
---
"What are HB_Bridge, IPSC_Bridge and Analog_Bridge?"
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.allstarlink.org/pipermail/allstar-digital/attachments/20170712/91c9e489/attachment.html>
More information about the Allstar-Digital
mailing list