The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> High-Q inductor options
https://designers-guide.org/forum/YaBB.pl?num=1323259161

Message started by aaron_do on Dec 7^th, 2011, 3:59am

Title: High-Q inductor options
Post by aaron_do on Dec 7^th, 2011, 3:59am

Hi all,

just wondering if anybody cares to share some high-Q inductor options. For instance,

1) Chip inductors. I've seen Q up to 150 at 1 GHz, but I'm not sure what's typical.
2) Bondwire Q seems to go up to about 25 at 1 GHz.
3) MEMs?
4) IPD?

anything I've missed out?

thanks,
Aaron

Title: Re: High-Q inductor options
Post by raja.cedt on Dec 7^th, 2011, 4:23am

hello,
1. on chip inductor Q won't be 150 i guess, it's much smaller than that, if not all VCO Q at this frequency range (like DCS, GSM) will be dominated by varactor but in reality inductor Q dominates.

2. Bond wire inductor Q will be high.

I would like to add selective substrate etching beneath the inductor technique to increase the Q but with reliability compromise.

what do you mean by IPD?

Thanks,
Raj.

Title: Re: High-Q inductor options
Post by loose-electron on Dec 7^th, 2011, 9:11am

on semiconductor chip or oiff chip?

better performance off chip, see things like Digi-Key and do component searches.

Even off chip though, your network Q is going to be worse than the inductor due to board parasitics.

Higher order networks?
If BPF performance is what you are looking for, then the higher order LC networks may be a possible path.

Title: Re: High-Q inductor options
Post by ACWWong on Dec 7^th, 2011, 9:34am

raja.cedt wrote on Dec 7^th, 2011, 4:23am:

what do you mean by IPD?

Thanks,
Raj.

Integrated passive device.... not to be confused with "chip" inductor which means a surface mount component.

In terms of chip inductors, coilcraft are normally best for performance... but not cost. Q very much depends on technology; multilayer or wirewound.

I'd be interested in gyrator performance too ;) especially at RF freqeuncy using modern sub-micron technologies.

Title: Re: High-Q inductor options
Post by raja.cedt on Dec 7^th, 2011, 10:41am

hello ACWWong,
why u need gyrator, r u talking abut active passives?

Thanks,
raj.

Title: Re: High-Q inductor options
Post by loose-electron on Dec 7^th, 2011, 4:06pm

need to define RF frequencies

up to 200MHz these days you can do with Op-amps

from 100-500MHz gmC will work

200MHz and up traditional "RF" techniques, LC sections and networks.

Some overlap...

"RF" is a pretty meaningless term these days.

Title: Re: High-Q inductor options
Post by aaron_do on Dec 7^th, 2011, 5:41pm

Hi all,

thanks for the replies. Anyway just so that maybe we can stay on topic, as ACWWong pointed out, by chip-inductors I meant surface-mount. I checked up Murata and the highest Q I saw was around 150 at 1 GHz.

Quote:

I would like to add selective substrate etching beneath the inductor technique to increase the Q but with reliability compromise.

does that fall under MEMs? What kind of Q is possible? How about cost-wise? It sounds very expensive.

Quote:

Even off chip though, your network Q is going to be worse than the inductor due to board parasitics.

So for example, if I want to design a VCO with a tank inductance of 5 nH, what is the maximum tank Q I can get, and what kind of technology would I be looking at?

thanks,
Aaron

Title: Re: High-Q inductor options
Post by raja.cedt on Dec 7^th, 2011, 7:17pm

hello,
could you please point me the reference where you have seen Q~150 (i am so curious)

regarding your last question, it's tough to say which tech and it wont change with technology more or less. Inductor Q effected by 1. high from the substrate 2. size of the inductor 3. width of the top metal. So if you have more top metal then it's good for Q, by tweaking the design (green-house formulas) by selecting proper topology (with ground shield, differential coil) you can get better Q.

thanks,
raj.

Title: Re: High-Q inductor options
Post by aaron_do on Dec 7^th, 2011, 9:45pm

Hi Raj,

here's a document from Murata with a lot of RF inductor info. Check out page 155 for a very high Q inductor.

http://www.murata.com/products/catalog/pdf/o05e.pdf

Also, I think you misunderstood my question. When I asked which technology would I be looking at, I was not refering to which process node. I mean what kind of technology such as IPD, surface mount components, MEMs, Bondwire etc. i.e. how can I get the best possible Q, and as a side question, in practice, what options do people normally use. I think bondwire is fairly commonly used as long as the inductance value doesn't need to be too large.

regards,
Aaron

Title: Re: High-Q inductor options
Post by raja.cedt on Dec 8^th, 2011, 3:44am

hello,
yes you are correct, Bond wire Q will be higher more or less. Thanks for the doc.

Thanks,
raj.

Title: Re: High-Q inductor options
Post by loose-electron on Dec 8^th, 2011, 11:17am

Bond wires can be inconsistent on the value due to the variance in the mounting, length and height above package ground.

Title: Re: High-Q inductor options
Post by aaron_do on Dec 14^th, 2011, 7:03pm

Quote:

Bond wires can be inconsistent on the value due to the variance in the mounting, length and height above package ground.

Compared to on-chip inductors I guess that's true, but how much variation is typical?

thanks,
Aaron

Title: Re: High-Q inductor options
Post by raja.cedt on Dec 15^th, 2011, 1:40am

hello arron,
You can refer Wireless cmos frequency synthesizer design book. It has chapter on Bond wire inductance, and unfortunately it depends on many parameters like bonding angle and bond wire length which are very strongly depends on the bonding environment. In this book he shown variation is around 30%.

Thanks,
Raj.

Title: Re: High-Q inductor options
Post by aaron_do on Dec 15^th, 2011, 6:57am

Hi Raj,

thanks for the reference. Are you sure you remembered the number correctly? I checked up the same reference and it mentioned the total variation to expect was on the order of about 6%.

regards,
Aaron

Title: Re: High-Q inductor options
Post by raja.cedt on Dec 15^th, 2011, 7:48am

hello,
let me check, but 6 is very small number seems. Soon i will provide exact number...

Thanks,
Raj.

Title: Re: High-Q inductor options
Post by loose-electron on Dec 15^th, 2011, 7:57pm

aaron_do wrote on Dec 14^th, 2011, 7:03pm:

Quote:

Bond wires can be inconsistent on the value due to the variance in the mounting, length and height above package ground.

Compared to on-chip inductors I guess that's true, but how much variation is typical?

thanks,
Aaron

Very dependent on the mechanics of the situation.
Saw one example using a bondwire iinductance where it all
worked nicely, and then they moved manufacturing to
a different facility. Total fail of the tuned circuit due to new center
point on tuning the device. (had trammed caps)

after talking with the assembly shop they could tweak it
and bring it back to center with adjustment on tension in the
bonding machine.

Sorry, no number available.

Title: Re: High-Q inductor options
Post by aaron_do on Dec 15^th, 2011, 9:22pm

Hi loose_electron,

I'll keep that in mind. Perhaps the variation is not so great, but it may require a little trial and error to get the right value.

In fact the way I normally model the bonding wire like a right-angle triangle with a missing bottom, and a little extension at the top. However, we managed to look at the cross section of the bondwire in question, and found it looked more like a semi-circle. So maybe some iteration is necessary.

cheers,
Aaron