While selecting resistors most often if standard values aren't present , the closest standard values are used which might be a kilo ohm or two away from the design requirements . I want to know how much leeway is there while choosing an inductor or capacitor .I couldn't find a 0.6nH for L1 in the below interfacing circuit and the ones I have ordered are about a week or two away from being delivered and I was sort of wondering if I could use another ballpark value (say 9.9nH) in its place. L1 in combination with C2 is used as some sort of LPF I think to remove ripples from supply voltage to the low noise amplifier SKY65404.
-
2\$\begingroup\$ FWIW, Digikey lists nearly 300 inductors between 0.5nH - 0.7nH, perhaps one of those might suffice? \$\endgroup\$– uint128_tApr 21, 2017 at 16:05
-
\$\begingroup\$ How is 9.9n in the ballpark of 0.6n?????????? \$\endgroup\$– Neil_UKApr 21, 2017 at 16:28
-
\$\begingroup\$ 0.6 nH is like the inductance of a typical bondwire (in a wirebonded chip package). You could probably get this by just cutting out groundplane beneath the track feeding the Vcc pin of the chip. But you probably don't want to change your PCB at this point. \$\endgroup\$– The PhotonApr 21, 2017 at 16:57
-
\$\begingroup\$ @uint128_t, in fact there's 25 different PN's in stock for 0.6 nH specifically, all with SRF > 10 GHz, and on pretty standard footprints. OP, the lesson learned from this should be to check for stock availability of parts before designing them in to your board. \$\endgroup\$– The PhotonApr 21, 2017 at 16:59
2 Answers
There's no general answer to your question. There are lots of factors to consider when selecting an inductor (e.g. ESR, saturation current, interwinding capacitance, lossiness including core and AC losses, etc.).
I wouldn't consider 9.9nH to be "in the ballpark" of 0.6nH as it's more than a decade higher in inductance. In this case, it seems like more inductance would be better at filtering the power supply, but if there are supply transients drawn by the part the higher inductance could lead to higher ringing and overshoot so it would be best to go with the inductance specified in the datasheet.
I'm surprised that they don't give more information about acceptable inductors in the datasheet, because the type of inductor can have a big impact on performance. I would expect that a lossy inductor (maybe a ferrite bead) would be preferred in this application, but again no guidance in the datasheet. I would check to see what they use on the evaluation board for the part and go with that since you know that it has been tested.
-
\$\begingroup\$ "If unsure, check the eval board" is usually sound advice. Also, contacting manufacturer's sales/AFEs might not be the worst thing to do – they'd probably be happy to improve the quality of their docs. \$\endgroup\$ Apr 21, 2017 at 16:15
-
\$\begingroup\$ @MarcusMüller, RE: "they'd probably be happy to improve the quality of their docs" --- don't count on it. I've worked for companies in the same general space as Skyworks and if your project might require 10,000+ units per month we'd send an FAE to sit in your lap, but if you just need a few thousand parts, we'd rather not talk to you -- and AFAICT bad documentation is part of discouraging small customers from bothering with us. \$\endgroup\$ Apr 21, 2017 at 17:27
-
\$\begingroup\$ @ThePhoton meeeeeeeeeeeeeeeeeeh yeah, see your point. \$\endgroup\$ Apr 21, 2017 at 17:39
When selecting alternate values for a component, think about the relative change in value, rather than the absolute change.
If the design calls for a 20K resistor, using 18K or 22K will probably be fine, as that is only a 10% change.
If the design calls for a 5K resistor, using 3K or 7K instead will probably not work - this is the same 2K off the desired value, but is now 40% "wrong".