Forum Discussion

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

Configuring Altera devices on long-distance cabling

Hello,

We have a need to configure Altera devices (EPC configuration devices and some FPGAs) through a long cable (~ 3 m). For the sake of simplicity let's say that the setup is a PC, 3 m of cable and a card with the device on it.

We've found that USB blaster can't program at such distances, but ByteBlaster can, which can be probably attributed to the JTAG communication speed.

Are there other ways to achieve this ? Using USB blaster, if possible ? Perhaps there's a way to make it run a bit slower which will make longer distances feasible.

Thanks in advance

8 Replies

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

    Although I am not sure if this fits into the constraints at your work-place but you can use Ethernet-Blaster to program FPGA boards located remotely.

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

    I'm mostly using a 5m cable with my USB blaster when working at target hardware. It can be further extended by available USB repeater cables, if necessary. In some special applications with a high noise level (industrial power electronics), I use an isolated JTAG interface, additionally. Alternatively, isolated USB couplers could be used. Other options are a remote JTAG server on a slave computer, could be battery powered and accessed through WLAN in HV applications.

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

    I should probably have said that the long cable is between the blaster and the board, not between the PC and the blaster !

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

    I understood pretty good what's your problem. The JTAG side connection can't be simply extended. As you gave not a single reason, why you didn't consider to extend the USB side, I just reported my solution for a similar problem.

    In other words: When extending the JTAG interface (with acceptable speed), there must be a driver near to the board connector. Why not using the USB Blaster as the driver?
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    17 years ago

    --- Quote Start ---

    I understood pretty good what's your problem. The JTAG side connection can't be simply extended. As you gave not a single reason, why you didn't consider to extend the USB side, I just reported my solution for a similar problem.

    In other words: When extending the JTAG interface (with acceptable speed), there must be a driver near to the board connector. Why not using the USB Blaster as the driver?

    --- Quote End ---

    The board is buried rather deep inside a large case full with other boards. The configuration lines were taken through the motherboard to the connector of the case. Besides, the USB blaster is planned to be made a part of a special test equipment that connects to the case through a proprietary connector, and we want to lead the configuration lines through this connector. The total distance can reach a few meters.
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    17 years ago

    You may have a look at the JTAG signals with the osciloscope, if you can.

    If you see that the problem is coming from superimposed noise, you may simply try to shield your flat cable in some way (aluminum sheet for example).

    If the problem is a strong attenuation of the signal you may use a multi-thread cable with lower attenuation per meter.

    If neither of the two solution works, certainly Frank's suggestion to use a long USB cable remains the only solution.
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    17 years ago

    Thank you for clarifying the motivation of extending the JTAG side. The USB Blaster is using a TCK of 6 MHZ, which isn't that fast, to my opinion. I think the problems are rather related to cable impedance matching, possibly manifesting by a ringing TCK signal.

    I recently did some comparative tests with different USB Blaster versions and different cables (original Altera low-impedance flexprint and IDC cable, as used by Terasic). Generally, a moderate length IDC cable (60 cm) works O.K., but shows already some ringing due to the too low TCL driver impedance. The signal quality can be improved by increasing TCK (and optionally other signals driver series resistance).

    Interestingsly, this doesn't work on some boards. I found that they had interfering signals that coupled to TCK and generated false clock edges. In this cases, a very low TCK impedance as with the original USB Blaster is preferable.

    A 3 m cable has a round-trip delay of about 30 ns, no problem at a 6 MHz TCK. But you should use matched cable impedances and optionally, but most likely not necessary, individual screened signals.

    At least, you should have IDC cables with interleaved GND wires. This ends up in a 50 to 70 ohm imdedance for single ended signals, that has to be matched at the respective driver sides (assuming source termination only). As I mentioned previously, it would be helpful to use buffers for both directions near the board JTAG connector.

    If TCK speed turns out to be a problem (I don't think so), the 24 MHz crystal oscillator at the USB Blaster may be replaced by e. g. a 12 MHz device. I didn't check, but from my understandig of USB Blaster operation, I assume that it should work.