Hi,
I have been looking at using decoupling caps with the X8R dielectric instead of X7R. According to Kemets datasheet their X8R caps have a lower voltage dependence than X7R. That would be nice for decoupling. But have any of you tried capacitors with X8R or do any of you have comments about what Kemet claims. Unfortunately I haven't seen other manufactures saying the same about X8R and I can't really find any capacitance vs DC bias graphs for the X8R dielectric. We use quite a lot of X8R at work, but that is because of its temperature rating. I could probably run a test to check, but I am not sure how to test that properly.
http://www.kemet.com/Lists/ProductCatalog/Attachments/9/KEM_C1007_X8R_ULTRA_150C_SMD.pdf
I have been looking at using decoupling caps with the X8R dielectric instead of X7R. According to Kemets datasheet their X8R caps have a lower voltage dependence than X7R. That would be nice for decoupling. But have any of you tried capacitors with X8R or do any of you have comments about what Kemet claims. Unfortunately I haven't seen other manufactures saying the same about X8R and I can't really find any capacitance vs DC bias graphs for the X8R dielectric. We use quite a lot of X8R at work, but that is because of its temperature rating. I could probably run a test to check, but I am not sure how to test that properly.
http://www.kemet.com/Lists/ProductCatalog/Attachments/9/KEM_C1007_X8R_ULTRA_150C_SMD.pdf
For most uses of decoupling caps the dielectric nonlinearity does not matter. That is why high-k ceramics can be used.
Do you suspect the voltage coefficient is that different from X7R? I guess you could either use a higher voltage part or one with a higher initial capacitance to compensate for the capacitance reduction at your operating voltage.
Well, I'm not worried about nonlinearity. For decoupling it is more the loss of capacitance that bothers me, I can of course compensate by using a higher C cap or more caps. But then I would have more inductance, lower resonance frequency and worse pulse current capability. As consequence you can have worse EMI tolerance, EMC and signal integrity. So in general it would be nice to use a capacitor that is 100nF independent of voltage.
Nice Torgeir,
this is exactly actually what I have been looking for. Slide/page 10 has a capacitance vs DC bias graph and it says that Kemets Ultra stable X8R has no voltage dependence. The slides also shows that X8R capacitors in general have the same voltage dependence as X7R.
this is exactly actually what I have been looking for. Slide/page 10 has a capacitance vs DC bias graph and it says that Kemets Ultra stable X8R has no voltage dependence. The slides also shows that X8R capacitors in general have the same voltage dependence as X7R.
Last edited:
Bigger footprint, bigger height and much bigger price than cog/npo, and only in some ways as stable/good as cog/npo, but they found customers, what do I miss?
I only compared 100nF you mentioned at Digikey, then Kemet x8r is bigger and 4 to 5 times more expensive than cog/npo other brands same value/voltage, for 100nF it seems like a weird choice to me, inferior, bigger and more expensive.
If the decoupling capacitor is charged up to supply rail voltage when there is no change in current consumed, then the capacitance will have a value dependent on the charged voltage.
If that supply has a changed current demand, the decoupling capacitor will supply some of that changed current demand and it's charge voltage will change.
What happens to the capacitance value during that time period of trying to meet the changed current demand?
If that supply has a changed current demand, the decoupling capacitor will supply some of that changed current demand and it's charge voltage will change.
What happens to the capacitance value during that time period of trying to meet the changed current demand?
I haven't bothered to look up the Kemet X8R series, but I suspect there are some parameters that are not being discussed here that might be relevant to the design engineer who might be inclined to specify this part for a specific application probably having nothing to do with audio. FWIW there is no such thing as an audiophile grade smd ceramic.. (yet 😉 )
It comes in both commercial and automotive grades and has an operating temperature range of -55C to 150C as well as a number of other somewhat tighter than usual specs. The extended temperature range is probably most of the cost driver, note that it is also available in flexible termination which would likely further increase cost if this option was chosen.
http://www.kemet.com/Lists/ProductCatalog/Attachments/9/KEM_C1007_X8R_ULTRA_150C_SMD.pdf
It is definitely overkill for this application unless this project is going to live in an engine compartment...
http://www.kemet.com/Lists/ProductCatalog/Attachments/9/KEM_C1007_X8R_ULTRA_150C_SMD.pdf
It is definitely overkill for this application unless this project is going to live in an engine compartment...
Since "Audiophile Grade" is a marketing term and without formal meaning, it can and is used all over the place as meaninglessly as desired.
I fearlesssly predict that "Audiophile Grade SMD" will be widely used as soon as someone discovers it can be used for appreciable financial gain.
The TDK 0603 CGA3 100nF 25V works out at about 4p each @100
RS and Element14 prices have no relationship to what Bosch would be paying in volume
RS and Element14 prices have no relationship to what Bosch would be paying in volume
You're right. Kemet's "Ultra Stable" X8R is a bit big, but if you look at X8R from other manufactures their size is smaller, but then you will have capacitance voltage dependence. You can get 100 nF X8R in 0805, I haven't found any 0805 100 nF C0G, only 1206.
We use a lot of X8R at work, but that is for downhole instruments were electronics are deisnged for temperatures up to 200C/~400F
We use a lot of X8R at work, but that is for downhole instruments were electronics are deisnged for temperatures up to 200C/~400F
- Status
- Not open for further replies.