BUF16821-Q1 rewritability of Nonvolatile Memory

Dear Masanori,

The team will have a look at this and answer this after the weekend.

Thanks,

Carolus

Dear Carolus-san and support team,

I got another question about BUF16821-Q1 from the customer.

Once the writing to the non-volatile memory is performed, the device can not read out by I2C access.
Even if BKSEL is switched, the status doesn't return to normal.
And the power supply must be turned on again.

Is such behavior the specifications of the device?

Thanks and best regards,
M. Tachibana

Tachibana-san,

 

Can you please confirm the following condition was not violated while attempting to communicate with the part? If not, what you described is not the correct behavior of the device. We’ll be glad to support but please provide detailed timing information.

 

Regarding the questions in your first post:

There is no write-protect for the NVM data.

There are 16x2 sets of NVM cells for each channel, these are one-time programmable. Each time a write command is issued, the data gets stored in the next blank NVM cell, until all 16 are exhausted in either bank.

 

Since the NVM cells can’t be erased electrically, it is therefore “one-time”, although it can be erased through other means. The data retention time is virtually infinite.

 

Regards,

Guang-Apps CS Products

Dear Guang-san,

Thanks for your reply.

Regarding the rewritability of nonvolatile memory, our customer and I understood.
But the customer said the NVM could be overwritten 20 times.
He confirmed that the data of  the NVM have been rewritten in each overwriting.
Why can the NVM be overwritten more than 16 times (up to 20 times)?
Actually, how many times can the NVM be overwritten?

By the way, I've gotten another question from the customer.
How long the memory retention of the NVM is at 85°C, and at 95°C?
(100 years at 25°C is seen in Electrical Characteristics)

Regarding the behavior of the devece, my explanation was insufficient.
The customer says for example,
1. Write a certain data in OUT1 register of the NVM.
2. Read register, so data of OUT1~OUT16, VCOM1 and VCOM2 is all the same, which is same as OUT1.
3. Switch BKSEL, but the read data doesn't change from above.
4. Re-turn on the power, the read data change to correct one.
Is such behavior normal?

Thanks and best regards,
M. Tachibana

Dear Guang-san,

Thanks for your great support.
I understood your advice and passed it to our customer.

Thanks and best regards,
M. Tachibana

Dear Guang-san,

You informed me that retention time at 85° C is about 2 years.
For example, retention characteristics of popular EEPROM are 100 years at 25° C and 20 years at 85° C.
The temperature coefficient you informed is very different from EEPROM’s one.
Would you please check again?

Best regards,
M. Tachibana

Tachibana-san,

 

The OTP is 100% tested at wafer level and it is also subjected to qualification test during product development.

 

If applying the 100% test condition, you’ll arrive at the roughly two year retention time. While applying the qualification test conditions (higher temperature and longer test time) the number is much higher, typically 10 to 20 years range at high temperatures. I checked the OTP document, it looks like the design and qualification calls for 10 years at 125 degree C. This is in line with industry standard.

 

The industry practice calls for qualification test condition when it comes to calculating the retention time, not the production test condition, as I thought initially, which is intended to catch defective/week bits.

Thanks for poniting this out. 

 

Regards,

Guang-Apps CS Products

Dear Guang-san,

Thanks so much for your information.

The customer’s application is automotive use, not industrial.
Can I think that the retention time 10 years at 125° C even for automotive use?

Is 10 years the value considering the duty cycle?
(For example, working 8 hours per day, etc)

Best regards,
M. Tachibana

Tachibana-san,

 

The retention time is continuous.

 

Regards,

Guang-Apps CS Products

Dear Guang-san,

Thanks for your reply.

Can I ask you one more question to make sure?
I understood that 10 years at 125° C of the retention time you say is under 125° C for 10 years continuously.
Is my recognition correct?

Best regards,
M. Tachibana

Tachibana-san,

 

Yes correct, the 10 year is at 125⁰C and it is continuous, 24 hour a day. 

 

Regards,

Guang-Apps CS Products

Dear Guang-san,

Thanks for your prompt reply.
I'm relieved to hear that.
I'll inform it to our customer soon.

Thanks and best regards,
M. Tachibana