Thomas Lets see, you need 35mA at 2.7V(minimum) for 40mS. This corresponds to 95mW for 40mS or 3.8mJ. Allowing for 1V drop on the storage cap (5V to 4V) - enough for an LDO regulator to easily maintain 2.7V on the output - then E=1/2C(Vh-Vl)^2 or 3.8e-3 = 0.5 x C x (5-4) x (5-4) i.e. C=3.8e-3/(0.5 x 1) = 8mF (8000uF) ! Something of a problem. If you could increase the supply voltage to the regulator to , say 9V, the the cap would only need to be about 3.8e-3/(0.5 x 5 x 5) = 304uF, i.e. your 470uF cap should be OK Or support the Flash from a supercap or in-board battery backup? Richard Richard, I use a simple transistor circuit to detect the voltage drop to shut down to MCU. However, the voltage drop down too fast and in some cases, it screws up the flash if it's in a write operation. I use a separate regulator for the flash. The input of the regulator is connected to 5V and with a 470uF cap to ground. The output of the regulator is connected to 10uF cap. Still, the 470uF cap can't supply enough energy. I checked out the specs, and the flash consumes up to 35mA during the write operation. I need to retain its power for at least 20ms, 40ms is refer... Guess I am stuck and have to live with it :( Thomas >From: Richard Prosser >Reply-To: pic microcontroller discussion list >To: PICLIST@MITVMA.MIT.EDU >Subject: Re: [EE]: Help! Data Flash corruption >Date: Fri, 31 Aug 2001 11:00:34 +1200 > >Thomas >Presumably you are using a brownout detect of some sort to stop the >processor as the volts die. We were having a similar problem (with a >different Atmel chip) but fixed it by enabling the brownout detector & >setting it to 4V. As the volts dropped the program wandered wherever it >liked and often ended up corrupting the EEPROM. >If the problem really is the supply dying during a write, then the only fix >is to make sure you have enough on-board capacity (caps or battery) to hold >the supply up until the EEPROM write is complete. >Since the micro won't be drawing all of the 10mA (I guess) - can you >separate the micro supply from the rest of the circuit using a diode and >use a cap to hold the micro supply up until the write is compete? - You may >need to supply the circuit at 5.6 V and then have 2 diodes splitting the >supplies to 5V for the micro & the "rest". > > >Richard P > > > > >Hello everyone, > >I am using a data flash (ATMEL 45DB041B) to store voice information. I >often get data corruption (I think the table of contents screws up) when I >power up and down the test board (I get data corruption about once every >100 >tries) > >Let say the MCU sends out a "program a page" command, and the data flash >starts to program a page (it takes about 10ms to program one page) and >during the time where the data flash is being programmed , the unit is >powered down. Data corruption occurs!!! > >Is there any way to overcome this problem? Any work-around? I tried to >put >the big cap at the data flash supply to retain its power (for 10ms), but >during the write operation, the data flash consumes lots of current (20 >mA), >so it's too much for the cap. > >Thank you in advance! >Thomas > > >_________________________________________________________________ >Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp > >-- >http://www.piclist.com hint: PICList Posts must start with ONE topic: >[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads > >-- >http://www.piclist.com hint: PICList Posts must start with ONE topic: >[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads > > _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads