TEK MDO4054-6 Hung at Splash Screen

[JohnB]

Several days ago I turned on my MDO4054-6 and the only thing visible was the splash screen, no other buttons were illuminated.  I left it that way for about 10 minutes and nothing happened.  I also tried turning power off (AC still connected) and inserted my current (3.18) flash image USB stick and tried again without success.

I then disassembled the unit so I could measure power supply voltages and check fuses. All the fuses check Ok and the voltages were as follows:

+5V      5.0369
+3.3V   3.3549
+2.5   2.4873
+5VSB   5.0448
+5VA   5.0448
+1.8V   1.8303
+12V   11.905
-5V      -4.8949
+1.5V   0.0 with noise
-2.5V   -2.4954
-12V   -12.012
+1.225V   1.2183

My only concern was the +1.5V which did not appear to be present.  I looked at the test point with my old TEK TDS224 and the signal was at 0 volts with some noise.  I pulled the main board to check for traces on the back but no traces to this test point were visible.  It is possible that it is on an internal layer or not present in my scope model.

The following day I put everything together, connected the AC and turned it on.  Everything worked fine.  Shortly after the splash screen many of the buttons were illuminated followed by a second set of buttons illuminated and finally the scope display replaced the splash screen. (Photos of first and second button illumination attached).

I though everything was fine and shut down the scope.  Latter that day I turned it on again and nothing, only the splash screen.

One of the voltages listed above (+5VSB) is present whenever AC power is connected.  I unplugged the AC and let the scope that way overnight.  The next day I turned on the scope and it powered up just fine.  I did get a message that a flash USB was found but the firmware was not updated as it was the same level as what was currently on the scope.

It seems like when the scope shuts down something does not get cleared and remains as long as +5VSB is present.  Just leaving the AC disconnected for a short while does not fix the problem.  Sometimes the scope will only get to the first set of button illuminations and sometimes it will make it to the second but never display the the scope data.  I am not sure how long AC power must be disconnected but leaving it off overnight seems to make everything work.

While I had the scope apart I did notice a 10 pin connector on the main board (J9201).  It is possible that this connector has serial data signals from the processor, I did not check.  I also have an old TDS7104 that uses this type of connector for the serial data.  I have to use that on the old TDS7104 to restore the PowerPC boot settings.  I have attached a connector diagram for the TDS7401.  If I open my MDO4054-6 again I will try using this connector for the serial port.  That might help, however this is very little information available for the newer Tektronix scopes on line.

If anyone has any additional information on this issue, please post.

John

Update Jan 5, 2020
Thinking about this problem a little more my best guess is that some component operating from the +5VSB supply is not getting or responding to a Power On Reset.  Leaving the AC power disconnected long enough allows the +5VSB to go low enough to force the component to actually reset.

[JohnB]

Scope has been working fine this past week provided that I remove AC power when I am done for the day.

I decided to take the covers off again and add test wires to +5VSB. +1.5V and add a 10 pin ribbon cable and IDC connector to J9201.  I ran the test wires through holes in the metal covers and out one of the narrow cover slits on the case back cover.  I added another 10 pin IDC connector after passing the ribbon cable through the case back cover.

I checked J9201 for ground connections before closing the covers and pins 2 and 10 are connected to ground.

Before connecting the AC power and turning on the scope I connected a logic analyzer (Intronix LogicPort) to the ribbon cable and manually triggered it as I pressed the power on button.  As expected, the scope powered up as AC had been removed for a long period of time.

I was able to verify that even when the scope is completely operational, +1.5V test pin indicates 0V.  Must not be needed with my version of the MD4000 series.

The logic analyzer did not capture anything of value except for occasional narrow pulses on pin 3 of J9201.  I put a meter on that pin and it read about 1V with the logic analyzer connected and 0V with the logic analyzer removed.  All the other signal pins were at 3.3V with or without the logic analyzer connected so they must all be outputs.  There was probably enough noise on the signal to fool the logic analyzer into thinking pulses were present.

I connected 3.3V from a lab power supply through a 10K resistor to J9201 pin 3 thinking that this pin might be used for external connection detection.  I powered on the scope again (AC had not been removed) and I expected the scope to fail.  I triggered the logic analyzer on one of the other J9201 pins that goes to 3.3V at power on.  Much to my surprise, the scope powered up completely.  This time the logic analyzer captured data with pulses detected on  J9201 pins 1, 4, 5, 8 and 9.  None of the pulses looked like a clock.  It is not yet clear to me how these 5 signal lines are being used.

If anyone is interested in seeing the logic analyzer output, send me a PM and I will forward the file.  To view the file you must install the LogicPort application and run it in DEMO mode.  Everything works in this mode with the exception of capturing data since no hardware is present.

Now the stranger thing.  The scope has powered up fine every time I tried if for the last 2 days without removing AC.  There is nothing connected to any of the test wires I added.  It is possible that this is a connector problem (but I don't yet understand why removing AC power fixed things).  There is a logic board on the back cover that contains all of the rear connectors, the clock battery and the two front USB jacks.  The logic board has a card edge connector that plugs into what looks like a PCI socket on the main board.  Possibly the multiple removal and insertions of this connector wiped the pins enough to clean the connections.  This scope is almost 8 years old and last week was the first time I have ever opened the scope.

Update
I found a service manual for the MDO4000 series scopes, J9201 is listed as a JTAG port.

[JohnB]

I think that I understand what the problem might have been.

After reviewing the block diagram in the service manual it is clear that the I/O board which supports all of the rear connectors and the 2 front USB connectors only has 3 connections.  One is to the AC power connector (used for power line sync), the fan control connector and the card edge connector which plugs into, what appears to be, a PCI Express connector on the main acquisition board.

I believe that the Power On Reset (POR) signal is generated on the main acquisition board.  There is also a reset push button on the main acquisition.  If the PCI Express connector is using normal wiring then pin A9 would be the +TRST# signal.  If that pin was not making proper contact with the card edge connector on the I/O board then the logic on the I/O board would not see a proper POR.  When AC power was removed for a long enough period of time (allowing +5VSB to get to 0V) then that logic would be at a power on state and function normally when AC was applied.

The multiple insertions and removal of this connector most likely wiped the contacts enough to allow a proper connection.  (The I/O board is mounted on the rear metal cover and removing the cover pulls the card edge connector out of the PCI Express connector on the main acquisition board).

The scope is still powering on properly and hopefully it will remain that way.

I have owned many Tektronix scopes starting with an old 545 in the mid 60's.  All of the older scopes came with complete schematics (some of which had small cartoons drawn on them) and the scopes were easy to troubleshoot.  Sadly those days are gone.  Very little information is available for the current Tektronix scopes and I suspect all repair are at the board level (very expensive for an MDO scope with much of the scope on a single PCB).

John

[koziymf]

Thanks for sharing troubleshooting steps..
I have a similar problem with mdo4054b-3.
After leaving it powered off overnight, the unit won't boot up, hanging in logo / fan at full speed. Then after 1-2 power cycles it will boot up and work normally.
I tried to clean pcie connectors on both sides - it didn't help. Getting boot log with UART would help but I don't know where those pins are located.
Another thing to try would be to re-flash the FW, but I can't find info on how to force fw update.