why this many resistors in series for voltage division ?!

[hm_fa_da]

Hello,

This is a schematic from Atmel for sampling 220V input voltage:



The resistors in the picture are in red rectangular, as you see there are 7 x 120K resistors in series !
why didn't they use less resistors ? for example 2 x 360K + 1 x 120K or any other combination with equal resistance but less counts of resistors ?!!!

is there any reason ?!

Regards.

 

[Audioguru]

The peak of a 220V sinewave is 311V and a hiccup will cause it to be higher. Some resistors are tiny and will arc if maybe 50V is exceeded across them. Look at the datasheet for your resistors to see the maximum allowed voltage.

 

[hm_fa_da]

it depends only on size of resistor or a same size may have different maximum allowed voltage ?!
i am using 0805 SMD resistors ! and don't have a special datasheet for it ...

 

[Audioguru]

all half-decent resistor manufacturers have a datasheet. Did you buy cheap no-name-brand ones from ebay?
I looked on Digikey's website where they list many manufacturers and the Panasonic 0805 surface-mount resistors have a maximum voltage rating of 150V.

 

[andre_luis]

It is quite often the cases on what design decisions not only consider technical criteria, even if certain aspects are sacrificed. Although I personally believe in the possibility raised by Audioguru in post 2, there may be situations where it is desired to reduce the diversity of components at the BOM, and so the combination of same values helps to achieve this goal.

 

[schmitt trigger]

In my previous job, we had rigorous safety requirements, and multiple resistors were employed in dividers for the exact reason Audio guru has mentioned.

 

[Audioguru]

Add extra parts because their values are common?
In about 1958 my AM portable radio had a label, "8 Transistors" on it. My friend bought a more expensive radio that said, "14 Transistors" on it. We looked inside to see the difference: his had 6 transistors in a circle connected together but not connected to the radio circuit.:smile:

 

[dick_freebird]

If you want a dead ratiometric division then you
would do best (repeatability wise) with M/N of
same value, same batch. Match of two different
values is more random. Also if there is a significant
TC then different power dissipation will drift output
and vary with ambient conditions. Last, what about
the value of spreading out the ladder thermal
dissipation into more bodies and PCB area?

 

[chuckey]

As the input resistor has at least 230V AC connected to it as well as the resistors being able to cope with the voltage, the PCB track layout must as well. One way would be to put the resistors in line at least 3mm away from any other components. I have recorded transients of over 7KV on the mains supply.
Frank

 

[d123]

I think it's the above mentioned: safe design, and cost (/product availability even) - high wattage resistors cost more than readily available low wattage, and can take up a lot more space, and dividing into smaller equal values reduces PD a lot as it reduces the voltage across each resistor. SMD from brief shopping perusals seem low wattage.
Lots of examples (besides desperate personal experience of how do I pass x current through y rated resistor due to wattage and/or peak voltage = divide by however many smaller resistors necessary, but then take up more pcb space...), one schematic I remember is a DVM which had 7* 1Mohm in series rather than 1 * 7Mohm resistor for higher voltages, needless to say why.