EMRFD Message Archive 13903

Message Date From Subject
13903 2017-05-25 10:03:56 Ashhar Farhan high q smd capacitors

i wanted to know what are the options for higher Q capacitors in smd. 

how are the MLCC capacitors? are they better or worse that the ceramic discs?
how  are the 'chip capacitors' different?
does size matter? is a 1206 capacitor better than a 0805 capacitor in terms of the Q?

- f
13904 2017-05-25 13:24:53 Graham / KE9H Re: high q smd capacitors

I am not sure "Q" is all that you need to worry about with surface mount capacitors.
They are all typically higher than leaded components.

You need to consider the dielectric.

Usually given as C0G/NPO, X7R, X5R, Y5V and some others.

C0G/NPO are considered to be suitable for RF and signal circuits, and do not change capacitance value much with either temperature or Voltage changes.

All the other "high k" dielectrics, such as X7R, X5R and Y5V exhibit capacitance value change with applied Voltage, and also some piezo-electric effects.

They can be microphonic, and can also "sing."  These are suitable for bypass capacitors and power filtering, but should not be used in signal paths, unless you understand how to use them.

In general, the smaller the capacitor is for the same capacitance value, the worse these effects.

The main benefit of SMD capacitors is that they don't have wire leads, so no parasitic inductance, so can work at much higher frequencies without self resonance.
They are also physically smaller, so you can built smaller, higher frequency circuits.
It is the only way to go, above HF.

--- Graham / KE9H


13908 2017-05-25 18:59:49 Ashhar Farhan Re: high q smd capacitors

about the rest, i know. it was specifically the q that i am confused. it happened like this : for the spectrum analyzer, i used smd caps (i forgot where i bought them from), and they provided pretty solid band pass in the filter. after a while, i used smd in the bitx40, there were losses. however, at 40 meter, a bandpass filter with a 4 db loss doesn't matter, esp if it is followed by an RF amp. 

now, for newer projects, i am still seeing losses in even the low pass filter, where the drop off is not as sharp as it is if used poly caps or disc ceramic NP0. i was wondering which SMDs to pick.

There are MLCCs that are regularly available form mail order shops. Then, are chip capacitors that are sold by some RF shops like this http://www.ebay.com/itm/100x-27pF-5-50V-0805-NP0-SMD-Chip-Ceramic-Capacitors-Ceramic-Capacitors-RoHS-/282236436390?hash=item41b69a3fa6

i am a little confused about which ones i must try.

- f

13910 2017-05-25 19:33:01 Graham / KE9H Re: high q smd capacitors
Well, the Q of the capacitors, of any type, is normally so high (1000 to 2000) as compared to the coils in any filter (typical range of 30 to 300), that the losses in the filter are set by the coils, not the capacitors.  If you think the SMD capacitors are causing your filter loss, then something is very wrong.

Items purchased on eBay can be anything from legitimate parts, to counterfeit items, to defective items rejected by someone else for quality defects. Beware.

I have not had any "Q" problems in RF filters with the NPO/C0G capacitors. I typically use size 0805 in receiver stages, and larger packages with higher Voltage ratings for transmit filters.  They all behave like the simulations, with the capacitor Q set to 1000 plus, and the trick is to get a good estimate of the Q of the inductors.

I have seen cracking of the ceramic capacitors, from thermal shock when soldering, or physical abuse such as bending the PC board after soldering. When you have cracked caps, anything can happen to the filter response.

13911 2017-05-25 19:35:34 Phil Sittner Re: high q smd capacitors


I built an triplexer for HF transceivers based upon an ARRL antenna book article. The author, W2VJN, noted that MLCC type capacitors had better Q than both DUR micas and KGE type devices. The MLCC types also list the devices for RF high current applications and have them available with up to 500 volt ratings. Also, I recently built a microstrip RF directional coupler and used a single pole low pass filter design to contour the the coupler frequency response. The first try resulted in an unexpected resonance at around 70Mhz when using a Murata ceramic chip capacitor. After changing it to a Samsung MLCC-type the resonance problem disappeared. You're on to something here regarding the performance difference between various SMD types.

Phil, KD6RM

13912 2017-05-25 19:50:42 Ashhar Farhan Re: high q smd capacitors

It is pretty much confirmed that the capacitors that i have are of sufficiently low Q to matter. i have a board with squares cut, across which i soldered inductors wound on T30-6. i swap the smd caps with the poly caps that i picked up at the FDIM vendor night and run the scans. the difference is quite noticeable. the corner is sharp at 34 Mhz just as predicted when using the poly. with the smds, 
with the smds, the roll off looks more like a down slope right from 22 MHz and the losses at 30 Mhz are over 3 db.
then, i tried with regular disc ceramics and the roll off was better than SMD but not on par with the poly caps.

- f

13916 2017-05-27 09:49:52 Mvs Sarma Re: high q smd capacitors
Farhan jee
 can you please glance these pages at this link

13918 2017-05-27 10:21:53 Ashhar Farhan Re: high q smd capacitors
thanks, i am looking at them. i just ordered a bunch from mouser. if i had known about these, i would have added them to the bag as well

- f

13920 2017-05-27 10:56:47 Graham / KE9H Re: high q smd capacitors

Do you have example or representative part numbers on "low-Q" SMT capacitors.
I would like to research what the manufacturer's specs say about those parts.

13921 2017-05-27 11:04:00 Ashhar Farhan Re: high q smd capacitors
one reel read 1206Y101M500, the rest were bought as tapped up small runs off the actual reel, so the designations are lost.

- f

13922 2017-05-27 11:33:16 Ashhar Farhan Re: high q smd capacitors

i made some simple measurements with what  i had in the junk box. i posted some of those results in a separate thread

- f

13924 2017-05-27 14:08:23 Graham / KE9H Re: high q smd capacitors
I can't find any manufacturer's reference for that part number.
I would personally consider any capacitor used for RF which had a Q of much less than 1000 to be defective.
Those caps you tested must be made with a dielectric that is lossy at RF.

I think you may have discovered another reason to stay with manufacturers you recognize and buy them through distributors you trust.

Thank you for writing up your experiment results.

13925 2017-05-27 16:00:17 Phil Sittner Re: high q smd capacitors


You have piqued my curiosity but I can't find your separate thread, even in the EMRFD reflector. Can you tell me where to find it or repost? The other thing of note is when expected information is not listed in the manufacturer's data sheets it's time to suspect performance issues.

Phil KD6RM

13926 2017-05-27 22:02:59 victorkoren Re: high q smd capacitors
Ashar, the "M" code means that it is a 20% tolerance capacitor. It hints that it is not a COG capacitor.
Always buy RF capacitors (1000pF and down) with tolerance of 5% (Code J) and better. They must be COG capacitors.
The 20% tolerance of your capacitor hints that it is probably made with a high dielectric constant material, maybe "Y" material as it is written in its name. That's a very bad capacitor for RF. 
For good decoupling capacitors (larger than 1000 pF) buy 10% tolerance capacitors (code "K") that most probably are made with X7R material which is good for decoupling. Never buy ceramic capacitors with tolerance worse than 10%.
Victor - 4Z4ME
13927 2017-05-27 22:48:31 Russell Shaw Re: high q smd capacitors
13928 2017-05-27 22:48:46 Ashhar Farhan Re: high q smd capacitors
Phil, it is a thread with subject line 'Q of capacitors, SMD et al', i posted it yesterday.

- f

13929 2017-05-28 05:58:32 Phil Sittner Re: high q smd capacitors

Thanks Ashar


13934 2017-05-30 01:39:48 biastee Re: high q smd capacitors
Other temperature characteristics aside from C0G/NPO may be required in different RF circuits.

For example, a varicap diode tuned parallel LC circuit has a positive temperature coefficient, and so may use a N750 cap to achieve temperature compensation.

For DC blocking in a broadband amp, an X7R may be better than a NPO cap because the former's in-band series resonances are dampened by the higher parasitic series resistance.

73, Chin-leong, 9W2LC

---In emrfd@yahoogroups.com, wrote :

C0G/NPO are considered to be suitable for RF and signal circuits, and do not change capacitance value much with either temperature or Voltage changes. (snipped)