Forum Discussion
Cyclone IV configuration problem
Hi All,
I have a problem with a Cyclone IV device not working correctly on my own hardware PCB. If I connect to the board using a USB blaster cable, Quartus can see the device correctly. If I configure it, Quartus thinks it configured OK (at least it says 100% success with a green bar). But the hardware is not working correctly. Even with a very simple design just to drive some outputs high or low, even this is not working. After looking around with a scope, all of the power supplies look good (1.2V, 2.5V and VCCIO=3.3V) but the CONF_DONE pin is at about 1 volt, even though it has a 10kohm pullup to VCCIO. Does anyone have any idea what might be going wrong here? What might cause the device not to configure correctly, even though Quartus thinks it is OK? thanks! Waljit9 Replies
- Altera_Forum
Honored Contributor
Check first please if you're using the pin DEV_OE in your design. In the menu bar of Quartus: Assigments->Device->Device and Pin options, etc.
- Altera_Forum
Honored Contributor
Are you using a part with an exposed pad on the bottom? I've seen this type of behavior when the pad is not connected to ground. In particular the CONF_DONE at ~1V is a tip-off.
- Altera_Forum
Honored Contributor
in reply to bertulus> Thanks for the reply. DEV_OE is not enabled (neither is DEV_CLRn). I tried settings for both passive serial configuration (i.e. JTAG over a download cable) and also active serial. The MSEL pins are setup for AS config, although I believe these are ignored when using JTAG anyway? Unused pins are set as inputs tri-stated.
Thanks, Waljit. - Altera_Forum
Honored Contributor
in reply to gj_leeson> Thanks for the reply. Very interesting, as yes I am using a QFP-144 package which has the exposed pad underneath. I am just trying to ascertain from the gerbers if this pad is connected to ground or not, but I have a suspicion it might not be. I'll check it and report back here ..... if this is the problem, don't suppose there is a workaround other than modifying the board?
thanks, Waljit. - Altera_Forum
Honored Contributor
I'm not aware of any other work around.
- Altera_Forum
Honored Contributor
--- Quote Start --- if this is the problem, don't suppose there is a workaround other than modifying the board? --- Quote End --- See http://www.alteraforum.com/forum/showthread.php?t=26883 - Altera_Forum
Honored Contributor
I made a dev board with Cyclone 1 EP1C3T100 ( QFP 100 pins ). The board works perfectly until one of my students used it. After that the FPGA behaved like you said: it can be programmed ok with USB-Blaster+Quartus but the I/O don't work.
With careful and patience I remove the chip and put it a new one. It comes to life again. Can you remove the fpga? - Altera_Forum
Honored Contributor
I am assured by the PCB layout guy that the exposed pad underneath the device IS connected to ground.
Does anyone have any other ideas as to what might cause this problem? Here is what I have: - the standard 10-pin programming header connected to TMS/TCK/TDI/TDO - Pullups (4k7) on TMS and TDI to VCCA which is 2.5V - Pulldown (4k7) on TCK - Pins 2 and 10 of the header are ground - Pins 4 and 6 of the header are VCCA (2.5V) - Pins 7 and 8 are no-connect - MSEL0 is connected to GND through a zero ohm resistor - MSEL1 is connected to VCCA (2.5V) through a zero ohm resistor - MSEL2 is connected to GND through a zero ohm resistor - This selects AS mode configuration (which should not matter when using JTAG mode) - CONF_DONE, nSTATUS and nCONFIG are pulled up to VCCIO (3.3V) through 4k7 resistors. I have also tried 10k resistors - nCE is pulled down to GND through a 4k7 resistor - I also have a config device connected to DATA0/DCLK/ASDO/nCSO - There is a 4k7 pullup to VCCIO on nCSO and a 22 ohms series resistor in the DATA0 line symptons - Quartus seems perfectly happy. No errors during programming. - I can also program the config device through the same JTAG connector using a JIC file and the SFL. Again, no errors from Quartus but - the design does not work. nothing. no matter how simple the design. All I/O pins seems to be at 3.3V - the CONF_DONE pin is low during configuration but only comes up to about 1 volt afterwards thanks Waljit - Altera_Forum
Honored Contributor
The PCB layout guy can't tell you if the part actually got SOLDERED to ground. This happened to me once on a prototype board that was assembled by hand, they neglected to put solder paste on the exposed pad area.