I was looking for 5V Reference for ADC and found the attached component (Datasheet) to be best.
But if there are better ones please guide me to one.
From the datasheet I chose AD586MN for minimum error ( I believe) i.e 2mV.

I was looking to buy it and found the numbering to be "odd".

Look at the picture ..Are they really numbered like that..The "MN" part to be specific.



Like I said, Are there any component better than this part

 

The datasheet does not indicate anything about how its marked..
Only Analog would know how it should be marked and if any changes to the marking were done and from what batches those apply..

Could be fake parts..could be good parts and they make have simply changed their process along the way..

 

I see the white square "MN" ones from most of the sellers

 

How good do you need?

 

I see the white square "MN" ones from most of the sellers

You can do a quick marking test: http://www.erai.com/CustomUploads/ca/wp/2012_7_Awareness_of_Marking_Permanency.pdf

 

I need minimum error. It's not like I can buy when need it with in a day.
So If I buy I buy what I can use for everything..Mainly ADC 12 to 16 bit voltage measurement accuracy via a PIC

 

If I were going to make counterfeit chips, I wouldn't bother making a rubber stamp with a 1997 date code!

And, I wouldn't make it incorrectly (vs. original) by creating the MN in negative space (white background/chip color for letter color) - much harder than a plain rubber stamp with a positive space image (with white ink).

If it is a hack/counterfeit, then I am fairly confident that the original also had the MN in negative space.

All that to say, a counterfeiter would not do something more difficult that does not match the original (they would do something less difficult and risk not matching the original).

 

Test them, see what they read.

 

Why not start with a PIC device that includes a built-in FVR (Fixed Voltage Reference) module? This includes all of the "enhanced mid-range" devices which provide a mechanism for using the 1.024, 2.048, or 4.096 volt FVR output as the +Vref for the ADC modules (without using an I/O pin).

 

Why not start with a PIC device that includes a built-in FVR (Fixed Voltage Reference) module? This includes all of the "enhanced mid-range" devices which provide a mechanism for using the 1.024, 2.048, or 4.096 volt FVR output as the +Vref for the ADC modules (without using an I/O pin).

For a high resolution ADC the long term stability and accuracy of the internal FVR as vref is, lets just say, not very good.

 

MN might be a country code, something that rarely is on a datasheet.

For the voltage reference to contribute less than one LSB of ambiguity to a 16-bit A/D output value, it needs to have an absolute accuracy of less than 0.0015%. That ain't gonna happen. One way around this is to use a reverence that is very stable, but not necessarily accurate, and deal with the accuracy in hardware or software knowing that once it is set, it won't move. One option is a super stable reference in a very-low tempco adjustment circuit (and a super accurate voltmeter). Another is a stable reference that is read by the A/D and used for self-calibration. Does your application require absolute accuracy or relative (ratiometric) full scale readings?

Don't know where you've looked, but a while back Linear Tech and Maxim got in a voltage reference accuracy war. They both have parts better than the best grade LM4040.

ak

 

MN might be a country code, something that rarely is on a datasheet.

For the voltage reference to contribute less than one LSB of ambiguity to a 16-bit A/D output value, it needs to have an absolute accuracy of less than 0.0015%. That ain't gonna happen. One way around this is to use a reverence that is very stable, but not necessarily accurate, and deal with the accuracy in hardware or software knowing that once it is set, it won't move. One option is a super stable reference in a very-low tempco adjustment circuit (and a super accurate voltmeter). Another is a stable reference that is read by the A/D and used for self-calibration. Does your application require absolute accuracy or relative (ratiometric) full scale readings?

Don't know where you've looked, but a while back Linear Tech and Maxim got in a voltage reference accuracy war. They both have parts better than the best grade LM4040.

ak

MN is on the datasheet. M is for the 2mV accuracy and N is the N-8 package. That wasn't the concern. The concern was how it is printed on the package - and if it should be printed like that or not. The MN version is listed on the last page of the datasheet.

 

@MMcLaren
When I run out of old PIC's I will buy Enhanced PIC's. For now the chip is cheaper option.
@Lestraveled
I need to buy to test. I cancelled the dispute about the 2N3373 ones as it did not show disappointing results. Only that the labeling is questionable and the seller insisted that they were genuine. Only time will tell.

So we can say the components is Genuine.

 

@MMcLaren
When I run out of old PIC's I will buy Enhanced PIC's. For now the chip is cheaper option.

Just curious... is it really cheaper?

 

Just curious... is it really cheaper?

Good point. I was surprised how affordable some of the new pics are.

 

Hmmm.
You guys make me wonder.
Still...When I buy PIC's I always buy around 10pcs each.
To you guys it may be cheaper but for me, I do not think so.
That is why I want to use what I have and spent 10 bucks for the reference chip to use all the PIC's I have when I need to u know.