Resistor Networks vs Discrete Resistors

[RJDog]

So, I have a design that has 7 220ohm and 6 1kohm resistors all clumped together, so I thought I would be smart and use a resistor network.  Now, the design requires the resistors all be independent (no common pin), but no biggie, independent resistor networks are easy to come by on DigiKey.  But, I'm looking at the price of a 8 x 200ohm chip compared to 8 x individual discrete 220ohm resistors (smt or through-hole, doesn't matter), and the resistor network chip (I am looking at SOIC-16) is significantly more expensive (>25%) than the required number of resistors.

Am I missing something?  Looking an unpopular chip package maybe?  Why would people choose a resistor network if it's more expensive?

[danadak]

Some network chips are laser trimmed for matching, absolute accuracy. Some even temp
compensated. Reason prices are lower on discrete resistors is volumes manufactured,
and to a lesser extent test time. Higher the volume, generally speaking lower the price.


Regards, Dana.

[cowana]

Another point is the assembly cost - assemblers often charge per part, which is reduced significantly when an array is used.

In addition, they're often smaller than individual discrete resistors.

[KL27x]

^ +1.  If resistor costs 1/5th of a cent to begin with, it maybe isn't too hard to make up the $ difference in assembly?

I know if I were hand assembling a board with so many of the same value resistor, I would love using these. Even with a PnP, it is going to go much faster with these.

[AndyC_772]

Pricing is all down to quantity with these parts. Individual resistors are widely used and cost virtually nothing, whereas networks are less common and therefore more costly. It has little to do with the technical qualities of the components.

A resistor network will cost you 1x component cost plus 1x placement cost, whereas the individual parts will be 'n' x each. The extra assembly charges may well exceed the difference in component cost, making the network more attractive.

In theory the network will save PCB space, but that assumes the individual resistors would be located next to each other. With the network, you may end up taking up more space with tracking, which eats into the saving. You'll probably need to swap the connectivity around to optimise the layout too.

[RJDog]

Another point is the assembly cost

Ah, yeah, I think you got it there.  I always forget about the assembly cost.  Makes sense to me now.

As far as space on the board, due to how the resistors are all grouped together anyway, the board layout is nearly identical with discrete or the array, so that's not really a factor (at least for me, for this project, anyway).

[vealmike]

Another point is the assembly cost - assemblers often charge per part, which is reduced significantly when an array is used.

This doesn't take into account the pain in the arse factor caused by arrays. When your component manufacturer can't get the RPAK you've spec'd, you find yourself having to hunt wide and low for an available second source or approving by offspec a slightly different shaped of part bodged in with rework.

Neither does it take into account the extra time needed in layout to route everything to Rpaks, instead of giving the PD guy the flexibility to drop an 0201 exactly where it's needed. Of course this is compounded by any SI issues caused by the extra routing.

There are cases for Rpaks, but generally the cons outweigh the pros in my very 'umble 'pinion guv'.

Resistor packs: putting the "ain" from "pain" into "ainus".

[elecman14]

My other concern with resistor packs is the pin pitch is generally closer than a discrete resistor. Bridging is a higher risk for those parts depending on the package type. If they bridge they will need touch up which add risk to the process. Or the CM may need to re-spin the stencil and that adds lead time.