Heating problem of the CPLD

Hello,

I am using atmel's ATF1508AS-7QC160. My clock frequency is 20MHz. Now, my cpld some times work and sometimes does not work at that frequency. If I lower the clcok frequency to 12MHz then it always work. bu tin both cases the chip does get hot. I put a heat sink on the chip but if the chip works for long hours then it effects its functionality. Can anybody advice me how to slove this problem?

John

Reply to
john
Loading thread data ...

John,

"Quite hot"

Anyway to be more specific?

If you can hold your finger on it, that tells us one thing.

If you can't even touch it (as it is boiling water) that is another.

If it is somewhere in between, that tells us something else.

Aust> Hello,

Reply to
Austin Lesea

Hello,

Its something in between if the heat sink is on.

John

Reply to
john

John,

Which tells me that with the heatsink, you may be right on the edge of overheating the die (anyway). Without the heatsink, it is hopeless (junction temperature is way out of spec).

Sounds to me like heat is the problem, but without real measurements, I have no way of knowing for sure.

Aust> Hello,

Reply to
Austin Lesea

I'd suggest you look for high-drive I/O or bus contention. Your dynamic power consumption should be minimal so where is the power going? Have you looked at the part in "reduced power mode?" That might give you insight into I/O versus core supply.

If the part draws 200 mA at 5V without I/O, 1W would be used to determine the minimum case temperature from the theta-jc. Do you have thermal data for that package? I only fount the mechanical info in the data sheet.

Reply to
John_H

Any open, unconnected inputs ? Are the inputs on defined valid levels ?

Rene

--
Ing.Buero R.Tschaggelar - http://www.ibrtses.com
& commercial newsgroups - http://www.talkto.net
Reply to
Rene Tschaggelar

First, measure the Icc, and determine if it is close to the data sheet typicals ( see the graphs ). If it is widely higher, then look for shorted pins, floating IPs etc ( enable the PinKeepers )

In the ATF15xx series, you have a Reduced Power Mode (per macrocell) option, that can save power, at the trade off of additional delay. Use that on the MCells that do not have to be fast.

How many levels of logic do you have ? ~20MHz should be ok in a 1508, but if you have a lot of logic levels, you will get the marginal operation you describe - 20Mhz is 50ns full cycle, 25ns half cycle.

Look at the fitter .FIT report file, to see the levels of Logic, and MC options.

-jg

Reply to
Jim Granville

I doubt that the problem has anything to do with 'heat' unless this is a very unusual environment that you're in. That's just a symptom that you're observing. It's most likely a timing problem. Some things to look for...

- Asynchronous input not being handled properly.

- Setup times are being violated.

- Is the clock frequency computed by the tool used to fit your design telling you that it will run at the speeds you're running it at?

The actual timing of a device at a given moment is a function of device temperature. Further, device temperature is a function of clock speed as well.

Another occasional problem is if you're driving a lot of load from this device (either lots of outputs switching simultaneously or just a few that are heavily loaded sometimes). But I'd check into the timing first since that's almost always where the problem is

KJ

Reply to
KJ

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.