Forum Discussion

Altera_Forum's avatar
Altera_Forum
Icon for Honored Contributor rankHonored Contributor
12 years ago

JTAG TCK droop using USB Blaster(Ver B) and Cyclone II

I have a NIOS-II processor in a Cyclone II which I am trying to debug using a USB Blaster (Ver B.) cable and The NIOS-II eclipse tools.

However, the JTAG communications are highly unreliable.

VCCIO and VCCA are 3.3V for all IO banks, VCCINT is 1.2V

JTAG connections are as follows:

TDI and TMS are pulled to 3.3V with 1K resistors

TCK is pulled low with a 1K resistor.

Pin2 is GND

Pin4 is 3.3V

Pin6 is tied to 3.3V via a 1K resistor

I have tried monitoring the JTAG signals with an oscilloscope.

The TDO, TMS and TDI signals all 'seem' to be ok, but the TCK clock looks wrong.

I have attached a photo of the TCK signal on the oscilloscope.

This shows that it is not a nice square waveform.

The frequency is correct (6MHz), but the clock droops from its peak at 3.3V to a level around 2.3V.

When it is not clocking, the 'resting state' of the TCK line is this 2.3V.

If the USB blaster cable is unplugged, the TCK pin on the PCB drops to 0V, so it is not the PCB which is biasing TCK to 2.3V.

Please can you give me some specific hints as to what might be causing this, tests to try, and what I might do to correct it.

Thank you.

PS: This same USB Blaster is always reliable when programming EPCS devices on other boards.

2 Replies

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    12 years ago

    The waveform is within LVTTL voltage limits. I remember to have seen a similar waveform, so I think it's normal operation with Rev. B. TCK is nevertheless prone to pick up interferences, e.g. crosstalk from fast edge signals. Different USB Blaster generations behave slightly different in this regard because the utilize different driver circuits. In some cases, small parallel capacitors near the FPGA TCK input (e.g. 10 pF) helped to supress TCK ringing.

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    12 years ago

    Good call !

    I found that a 22pF capacitor between TCK and ground, right next to the FPGA pin has done the trick.

    It looks like the USB Blaster 'B' has a very weak drive arrangement for TCK. (maybe a 470R series output resistor, and then a 1K pull-down resistor ?)

    Maybe it was picking up noise from somewhere along the 2" of PCB track from the connector to the FPGA, or the 8" of jtag cable attached to the USB Blaster.