Vref again

[strider02]

hi,
we need to use internal 2.56 Vref on ATmega16....an external super
stable vref as: zenner or other voltage regulators(if they can assure
very stable vref). If you can find any then it will be great.
thanx in advance..

 

[strider02]

hi,
we need to use internal 2.56 Vref on ATmega16....an external super
stable vref as: zenner or other voltage regulators(if they can assure
very stable vref). If you can find any then it will be great.
thanx in advance..

i forget to add that we can't use vcc to avcc cause its not stable
enough

 

[Tom Bruhns]

hi,
we need to use internal 2.56 Vref on ATmega16....an external super
stable vref as: zenner or other voltage regulators(if they can assure
very stable vref). If you can find any then it will be great.
thanx in advance..

'lo.

Why, yes, I can find some, so "it" must be great.

(What exactly do you mean by "super stable"? 0.01ppm/C? 0.01ppm/year?
Or something a bit more practical? Have you looked at the usual
suspects, like National Semiconductor, Linear Technology, Maxim, Analog
Devices, Thaler?)

 

[Genome]

hi,
we need to use internal 2.56 Vref on ATmega16....an external super
stable vref as: zenner or other voltage regulators(if they can assure
very stable vref). If you can find any then it will be great.
thanx in advance..

Sounds like your problem is some other sort of problem. Unless you really
try to explain what you are doing then you will just sound like some
ignorant sort of person.

I'd like to know the reasoning behind producing a 2.56V reference....

DNA

 

[Arlet]

we need to use internal 2.56 Vref on ATmega16....an external super
stable vref as: zenner or other voltage regulators(if they can assure
very stable vref). If you can find any then it will be great.
thanx in advance..

If you need to use the *internal* 2.56V reference, you don't need an
external reference. Just decouple the AVcc, put a cap on Vref pin, and
select 'internal reference' in the applicable ADC register.

 

[Eeyore]

i forget to add that we can't use vcc to avcc cause its not stable
enough

*Stable* ? You mean noisy perhaps ? Why not filter it ?

Graham

 

[Rich Grise]

Sounds like your problem is some other sort of problem. Unless you really
try to explain what you are doing then you will just sound like some
ignorant sort of person.

I'd like to know the reasoning behind producing a 2.56V reference....

Probably because it's 2^8 / 100, and they want 10 mv steps in their
ADC.

Cheers!
Rich

 

[Eeyore]

Probably because it's 2^8 / 100, and they want 10 mv steps in their
ADC.

Except it doesn't give 10mV steps !

Graham

 

[Robert Baer]

Sounds like your problem is some other sort of problem. Unless you really
try to explain what you are doing then you will just sound like some
ignorant sort of person.

I'd like to know the reasoning behind producing a 2.56V reference....

DNA

Why he is using a 256 bit DAC!

 

[Arlet]

Except it doesn't give 10mV steps !

Correct, the ATmega16 has a 10 bit ADC, and according to the datasheet
produces an ADC value equal to:

ADC = ( Vin * 1024 ) / Vref

So, with a 2.56V Vref, it has 2.5mV steps.

 

[Eeyore]

Correct, the ATmega16 has a 10 bit ADC, and according to the datasheet
produces an ADC value equal to:

ADC = ( Vin * 1024 ) / Vref

So, with a 2.56V Vref, it has 2.5mV steps.

2.50244 mV in fact ( 2.56/1023 )

Graham

 

[Arlet]

2.50244 mV in fact ( 2.56/1023 )

Graham

Maybe true for other ADCs, but not for the ATmega16 the OP is using.

See http://www.atmel.com/dyn/resources/prod_documents/doc2466.pdf on
page 217.

It even states:

"0x000 represents ground, and 0x3FF represents the selected reference
voltage minus one LSB."

 

[Eeyore]

Maybe true for other ADCs, but not for the ATmega16 the OP is using.

See http://www.atmel.com/dyn/resources/prod_documents/doc2466.pdf on
page 217.

It even states:

"0x000 represents ground, and 0x3FF represents the selected reference
voltage minus one LSB."

So not actually 2.56 V then !

Graham

 

[Arlet]

So not actually 2.56 V then !

Graham

Correct, the highest ADC output 0x3FF does not correspond to the 2.56V
reference voltage, but 2.5575V, for a step of 2.5575V/1023 = 2.5mV per
LSB.

 

[Spehro Pefhany]

Correct, the highest ADC output 0x3FF does not correspond to the 2.56V
reference voltage, but 2.5575V, for a step of 2.5575V/1023 = 2.5mV per
LSB.

Taking this a bit further, for Microchip's on-chip ADCs.

Code Ideal input voltage Ideal Middle input voltage
0x000 0.. 1 LSB 0.5 LSB
0x001 1.. 2 LSB 1.5 LSB
....
0x3FF 1023LSB..1024 LSB* 1023.5 LSB

Where LSB = Vref/1024

*Actually anything higher than 1023 LSB gives 0x3FF

So, the middle (most likely?) input voltage is
Vref * (code + 0.5)/1024, and a 2.56V reference would (ideally) give
0..2.55875V with steps of 2.500 mV.


Best regards,
Spehro Pefhany

 

[redbelly]

2.50244 mV in fact ( 2.56/1023 )

No. It would have to be 2.56000 V for that to be true.

Mark