how does this power supply decade resistor network work?

[crazyhog]

Hi all,

I have a Power Designs 2005.  The high precision output of the power supply is selected by a 4 decade resistor network.  I understand the total resistance came from summing different combinations of resistors with 1, 2, 2, and 5 ohms (and their decade multiples).  To get a particular output voltage, different voltage decade is used.  For example, to get a 0.3v output, 300 ohm = 100 ohm+200 ohm.  I've verified with a ohmmeter.  Note that these are precision 0.1% resistors.

However, the schematic is also showing a bunch of extra resistors (4.7, 6.8, 27, 18 ohm and their decade multiples).  See the resistors circle in red.  Furthermore these are 10% resistors!  I am really baffled by the purpose of these very low precision resistors.  Why are they needed in this circuit?

thanks,
James

[bob91343]

That is strange.  The style of drawing makes it a bit ambiguous but it seems those resistors are in parallel with the precision ones, and that makes no sense.

[Ash]

Hi,

You'll probably find it is a variant of a Kelvin-Varley divider.

I have some bigger individual switched dividers like that.

https://en.wikipedia.org/wiki/Kelvin%E2%80%93Varley_divider

Cheers,
Ash.

[Nusa]

Don't make the mistake of thinking 10% marking means the resistor actively changes its value during use. That resistor marked 10% may actually have been selected as actually being very close to the nominal value.

How can this be? One can take a bunch of 10% resistors and go through them with an appropriate measuring apparatus and find ones that are within the required specifications and only use those. It's calling binning. When it's done by the manufacturer, they mark them as precision and sell them for more. But you can get the same end result by doing it yourself.

[grumpydoc]

But you can get the same end result by doing it yourself.

Sometimes - but as you said it is also done by manufacturers - so the 10% tolerance band could be all the resistors that were >5% but <= 10%, the 5% batch all those > 1% and <= 5% and so on.