[sdiy] Resistor Networks....

[Pete Hartman]

Another use case which I think better illustrates the routing challenges....

A 560R 4 resistor network (8 pins) used in the stability networks of all
four units in a 2164.  Having them be in the same package forces all the
routing to go to one side or end of the chip, instead of being able to lay
out two individual resistor/capacitor combinations on one side, and the
other two on the other side.  This would be true for the SMD package as
well though the reduced footprint would make it perhaps easier to route
around.

On Sun, Aug 30, 2020 at 12:10 AM Pete Hartman <pete.hartman at gmail.com>
wrote:

> I haven't done an exhaustive search, but a randomly selected 4 resistor,
> not bussed network, by bourns, is about 2x the size of an  0805 resistor.
> Yikes :-D.  I was thinking "hand soldering" here, so ...  We'll see.
>
> On Sat, Aug 29, 2020 at 11:38 PM Pete Hartman <pete.hartman at gmail.com>
> wrote:
>
>> #1: ignorance that such a thing even existed, I will go look at that more
>> seriously, thank you.
>>
>> #2: the main layout issue is not simply the overall size of the networks
>> (they're .1" spacing, so a given resistor doesn't take an awful lot more
>> space than an 0805 resistor) but the forced linear arrangement, and the
>> occasional "extra" resistor that gets used somewhere nearby, but not
>> necessarily easily routable nearby.  Individual resistors could be placed
>> more conveniently to their actual use -- the example of a summer, perhaps
>> not so much, but there are others in this design that aren't quite so
>> localized.
>>
>> I don't know if matching is the reason or not, but Eric's comment about
>> affordable .5% resistors etc is a further argument that matching in
>> particular shouldn't be a criterion of concern.
>>
>> Thanks!
>>
>> Pete
>>
>> On Sat, Aug 29, 2020 at 11:17 PM Brian Willoughby <
>> brianw at audiobanshee.com> wrote:
>>
>>> Why not use SMD networks?
>>>
>>> I've used the following
>>>
>>> https://www.mouser.com/ProductDetail/774-746X101103JP
>>>
>>> https://www.mouser.com/ProductDetail/774-746X101472JP
>>>
>>> … in digital designs without any issues. Granted, these have a common
>>> pin, but if you check the manufacturer (CTS) data sheets and product family
>>> you'll see every variation of parallel and network configuration available.
>>> I've only used them for digital because my analog circuit layout never
>>> seemed to facilitate a network. I'm sure that an SMD network will work just
>>> as well as discrete SMD resistors in an analog circuit.
>>>
>>> Sounds like you're considering mixing the original through-hole resistor
>>> networks with modern SMD, and I wouldn't bother with that. Just use SMD
>>> resistor networks.
>>>
>>> Brian Willoughby
>>>
>>> p.s. Some digital designers don't like the resistor networks because
>>> they find them harder to lay out. I did not have that problem. I believe
>>> it's the folks who actually like using 0102 SMD that find the networks "too
>>> big" - but I have only used 0102 in one design and was told to never do
>>> that again for the sake of sanity (at least during prototype stages).
>>>
>>>
>>> On Aug 29, 2020, at 8:32 PM, Pete Hartman <pete.hartman at gmail.com>
>>> wrote:
>>> > I'm working on laying out a classic through hole module design with
>>> SMD parts.  The original uses a fair number of resistor networks, most of
>>> them just parallel resistors, not common bus versions.  The resistor
>>> networks are, well... huge by comparison.
>>> >
>>> > Unfortunately the original designer isn't around for me to ask for
>>> input.
>>> >
>>> > My *impression* based on how the networks are being used (often times
>>> for things like input resistors and a feedback resistor in a summing op amp
>>> config, just as one example) is that part of their purpose was to get parts
>>> that were well matched.  But that's not anything but a gut feeling.
>>> >
>>> > So I'm looking for seasoned opinions about whether there is any
>>> important reason to keep resistor networks, when I can sub in 1% SMD parts
>>> that will take up less space and be more flexible for placement.  Are the
>>> networks either 1) better matched even than 1% parts, or 2) have some other
>>> characteristic which I'm not thinking of that might be important?
>>> >
>>> > Thanks
>>> >
>>> > Pete
>>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20200830/d9581cc0/attachment.htm>