Hi,
One of our rfx2400 seems not work. I’ve tested for a while and found:
the board can be identified by motherboard
when I plugged the db to motherboard and run some scripts, I found
there aren’t any samples come to PC through USB cable. (no wave but a
blank
on usrp_oscope.py)
the TX doesn’t work either
I doubt there is something wrong with the controlling circuit of this db
and
the RF analog part is OK. I’m really a layman in hardware, can you give
me
some idea about which part is possibly broken, so I can exchange them to
repair.
I’m the colleague of Hanwen. Today I checked the broken board. I need
your
help especial Ettus’ help.
We can assure that the mother board is OK, because when we plug other
rfx2400, it works well. For this broken rfx2400:
The switch control signal, U204 7404
A1 B1 A2 B2 : 0 1 0 1
It seems no problem. And 3.3V voltage is normal.
The TX side: For the power supply circuit, U111 ADP3336 has input 6V
and
output 5V and U110 has 5V input and 3.3V output. For the AD8349, VCC=5V,
and
ADF4360 has VCC=3.3V. They all seem normal. But when we run the transmit
program, the graph always stops immediatly and there is no any error
report
on screen. I’m very confused with this.
The RX side: the power circuit U6 has 6V input and 5V output, but U5
cannot output 3.3V. So U3 ADF4360 has no power supply. If U5 is the
reason
for board broken, I can try to replace this chip.
But I’m not sure the broken chip is U5. I remember Ettus said that the
TX
and RX path is completely indenpendent. Why our TX path doesn’t work
either?
Untill now we have already broke THREE rfx2400s, all with the same
problem.
I worry about other boards will be broken too.
Can anybody give some advice? Thank you very much!
What are you doing to the boards?
We are using the rfx2400s to make a 2*2 MIMO-OFDM platform. We also run
some
other programs like usrp-fft, spectrum sensing etc.
What are you connecting them to?
2 common 2.4G antennas, not the PCB antennas from Ettus.
What level signals are you connecting to the inputs?
RF signal from air. According to the received data in PC, I don’t think
they
are too high. And we have used them for several weeks with the current
connection.
What are you using for the power supply? The one that came from Ettus
Research, or something else?
They are exactly the ones from Ettus R…
The story is like this: We use the platform everyday. Then in someday
morning or afternoon, we started the USRP and found they had no
responses?!
Nobody touched the boards during the period. And when we leave the lab,
we
always shut off the power of the boards. The enclosures are usually open
because we often plug in and out the daughterboard. At that time we are
always very careful.
On Fri, May 25, 2007 at 10:15:25AM +0800, Lin HUANG wrote:
What are you doing to the boards?
We are using the rfx2400s to make a 2*2 MIMO-OFDM platform. We also run
some
other programs like usrp-fft, spectrum sensing etc.
to != with
What are you connecting them to?
2 common 2.4G antennas, not the PCB antennas from Ettus.
The story is like this: We use the platform everyday. Then in someday
morning or afternoon, we started the USRP and found they had no
responses?!
Nobody touched the boards during the period. And when we leave the lab,
we
always shut off the power of the boards. The enclosures are usually open
because we often plug in and out the daughterboard. At that time we are
always very careful.
So we are very confused. What’s the reason?
Changing daughter boards while the power is applied could
cause a problem.
Not having the daughterboards seated properly on the stand offs could
cause a problem.
Dropping wires, connectors, tools or any other conductors onto the
USRP could cause a problem.
Writing your own verilog code could kill the main and/or
daughterboard.
Calling the u._write_oe method could kill the main and/or
daughterboard.
As far as I know, you are the only people who have been killing RFX-*
boards, and you’ve managed to kill a lot of them. I suspect that
someone in your lab is mistreating the boards somehow.
The TX side: For the power supply circuit, U111 ADP3336 has input 6V
and output 5V and U110 has 5V input and 3.3V output. For the
AD8349, VCC=5V, and ADF4360 has VCC=3.3V. They all seem normal. But when
we run the transmit program, the graph always stops immediatly and there is
no any error report on screen. I’m very confused with this.
You said you were using your own antennas. What is the input impedance
of your Tx antennas?
I continued checking these broken boards recent days. I found for the TX
side, the graph stops as soon as it starts. It seems the daughter board
gives some signal to stop the motherboard to load data from PC. I read the
.sch file in the subversion repository and the defination of the interface
between db and mb.
IOUTN_B IOUTP_B IOUTP_A IOUTN_A ---- the analog IF signal mb–>db
clock ---- the db is synchronous with mb mb–>db
I2C_A0 I2C_A1 ---- TX/RX RX1/RX switch control signal mb–>db
SCLK SDA ---- R/W interface to E2PROM mb<–>db
IO0~IO15 ---- for debug usage. mb<–>db
Which pin will give a wrong signal to motherboard and disable the board?
IO?
We never used u_write_oe. So it may be as a input?
Thanks for Eric and Nikhil’s help before. Hope you are continously
interested in my bug investigation. Thank you.
Just update the status of our broken boards. Now they are fixed up.
We
replaced the amplifier chip, U4 MGA82563, with a new one. Then
everything
works now.
I’d like to report to all of you our investigation results. At least,
the TX
side works now. But for the RX side, I think it is absolutely broken and
I
give up to repair it.
I changed the resistors R40, R42, R46, R130, R132, R136 to be 1k ohms.
This
prevents the RX side influencing the TX side. And I connected the 6V
power
of TX side with the power pin of the control part (U203), because the
power
of RX is not normal. Then the TX side can work.
