Max 2 IC occasionally gets destroyed

Recheck all max 2 pins the device on the pcb as it is realy connected with the datasheet

if you use a new unprogrammed device will it be destroyed as well ?

if no check wheter you have two outputs connected.

check the labeling on your device with the datesheet if you have the correct device mounted.

Thank you for your answer, MSchmitt.

--- Quote Start ---

check the labeling on your device with the datesheet if you have the correct device mounted.

--- Quote End ---

It is the correct device. The label on the device package says "EPM570T100C4N". This is equal to the device setting in the corresponding Quartus project.

--- Quote Start ---

if you use a new unprogrammed device will it be destroyed as well ?

--- Quote End ---

No, it is not destroyed, but even when it is programmed, it may work as expected. There are both cases (destroyed and not destroyed) when programmed, and there may be both cases when not programmed - but I can't verify that.

--- Quote Start ---

check wheter you have two outputs connected.

--- Quote End ---

No outputs are connected with each other.

--- Quote Start ---

Recheck all max 2 pins the device on the pcb as it is realy connected with the datasheet

--- Quote End ---

Well, the datasheet (device handbook) does not contain the pinout, does it? Would you recommend to take the ".pin" file of my Quartus project and to compare it with the PCB schematics?

--- Quote Start ---

The label on the device package says "EPM570T100C4N". This is equal to the device setting in the corresponding Quartus project.

Well, the datasheet (device handbook) does not contain the pinout, does it? Would you recommend to take the ".pin" file of my Quartus project and to compare it with the PCB schematics?

--- Quote End ---

You can use this file

http://www.altera.com/literature/dp/max2/epm570.pdf

and check ALL pins between the pdf and your schematics & layout

also use Quartus -> Assignments -> Pin Planer

and check all entries.

Next step, check your project settings for this device how it should handle unused pins.

If there is no mistake between altera documentation and your design, then check for ESD problems in your enviroment. i remember that we had a chair that was not ESD save and depending on the person some pcb were damaged due to ESD. replacing the chair with one that was ESD safe solved this problem.

Also you power suppply and or your design at all can have problems. voltage over and undershots on input pins, gnd bouncing (probably wont destroy the device)

as your problem happens on more than one pcb then it must be a design weakness, quartus settings or schematics or enviroment.

--- Quote Start ---

check ALL pins between the pdf and your schematics & layout

--- Quote End ---

VCCINT and VCCIO are connected to 3.3 V; GNDINT and GNDIO are connected to 0 V; TDI, TCK, TMS and TDO are connected to pull-down resistors and to the Altera Byte Blaster; GCLK and DEV_OE and DEV_CLRn pins are used as "User I/O".

--- Quote Start ---

also use Quartus -> Assignments -> Pin Planer and check all entries.

--- Quote End ---

All "User I/O" outputs pins have a high impedance load; all "User I/O" input pins have an input between 0 V and 3.3 V. All unused "User I/O" pins have either an input between 0 V and 3.3 V or are unconnected.

--- Quote Start ---

check your project settings for this device how it should handle unused pins.

--- Quote End ---

They are reserved "as input tri-stated".

On the conditions above, can any cause of the device destruction be excluded except the following?

--- Quote Start ---

ESD problems in your enviroment [...] voltage over and undershots on input pins, gnd bouncing

--- Quote End ---

regarding power connection, it seems to be okay from schematic view, your pcb is another part how it is layouted from an SI oder EMI view.

regarding the io pins, they are fine from my side.

and for the last point, do you have some kind of GND Loop ? PC -> Blaster -> JTAG -> PCB -> any-Interface -> PC

or via some kind of measurement device like scope ? the GND potential of your scope is normaly PE what is GND on your PC (Blaster)

--- Quote Start ---

and for the last point, do you have some kind of GND Loop ? PC -> Blaster -> JTAG -> PCB -> any-Interface -> PC

or via some kind of measurement device like scope ?

--- Quote End ---

The Byte Blaster is not connected most of the time, but there may be a ground loop sometimes, e.g. because of an emulator cable. Should I disconnect the emulator cable or the scope when the Byte Blaster is connected? I would find it rather awkward. Is there a more convenient method to avoid ground loops?

if there is a loop break it or make it as smal as possible.

if your designs get unstable with such loops then look at it again.

whenever possible just connect everything to one pc and connect all devices on your setup to only one ac source. avoid a setup where pc with blaster is on ac phase 1 and scope on ac phase 2. i have seen currents on such gnd loops. this should not happen as you live in zürich and your equiment needs to meet certification procedures.

what is unclear now is, when will your maxII be destroyed.

it can be destroyed during or after programming.

do you have boards that still survive ?

i knew that cyclone III are more sensitive to the powersupply than cyclone II was.

you say your maxII occasionally gets destroid. is it realy dead ?

a had been seeking very long for the reason why some maxII epm570 devices do not work properly out of a production pcb set and with mysupport at the end it was a quartus setting and now ALL boards work fine.

Thank you for your advice about ground loops.

--- Quote Start ---

do you have boards that still survive ?

--- Quote End ---

Yes, I am working currently with a board which has obviously survived.

--- Quote Start ---

what is unclear now is, when will your maxII be destroyed.

it can be destroyed during or after programming.

--- Quote End ---

I don't know exactly when it is destroyed. There are only few examples of this failure. Maybe it happend always when the boards were not in use for some time. Maybe the storage place for the boards is not quite ESD proof. I will have to keep observing this issue.

--- Quote Start ---

you say your maxII occasionally gets destroid. is it realy dead ?

--- Quote End ---

The current which flows through the destroyed board is more than 500 mA (even more than 2 A in the first moment after power-up). The voltage drops to 0.43 V within a few seconds. (The power supply can't provide enough power.) The Quartus programmer tells me: "Error: Can't access JTAG chain" Therefore I would say it's dead.

Oskar,

You say the MAXII device dies when the power is applied, do you connect 0V or 3.3V first when you apply the power. What kind of voltage regulator do you have on the board? What is the source of all the inputs to the board?

Thanks

--- Quote Start ---

do you connect 0V or 3.3V first when you apply the power.

--- Quote End ---

I think that the potentials 0 V and 3.3 V only have a meaning if there is a voltage from one to another (or to a third potential in the device). So I don't understand the question and its aim.

--- Quote Start ---

What kind of voltage regulator do you have on the board?

--- Quote End ---

It is a low-dropout regulator: ADP3338. It is on a different board.

--- Quote Start ---

What is the source of all the inputs to the board?

--- Quote End ---

At present, there are only a bus (adresses, data, control signals - from a bus transceiver IC), 5 V (from a mains adapter) and 3.3 V (from the power regulator).