X5R vs X7R MLCC

Ah! same thing! The LM311 is used with PIC 16F84 as oscillator. In my case I use the PIC 16F648, the nice thing about that PIC is that it has a build in hardware comparator that can be routed to an output pin, that comparator is used as oscillator.

mmm he uses a relay to switch the reference cap, I use a transistor...

And hard to replace with surface mount: most polymers won't take a temperature profile suitable for surface-mount soldering. Surface mount plastic film DOES exist, but not in high values at attractive prices.

Kemet makes some, their 'F161' series.

I hate to say recycled parts, and I'm not recommending anyone here do that. For decoupling the odd failed or off spec part is no big.

NT

I got an email from Digi-Key saying that they'd fixed the Samsung characteristics links, and that the fix would roll out next web site update (which should be pretty soon).

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

e:

a

re

stable

the

nput

seeing

ction

.

s any

nsity

ves

d

d

nF

(X7R

ing

.

up:

Thanks, Phil, they have fixed the links. I re-charted my data so it is in percent-delta-C to match their chart. Looks like my measurement very tight ly correlate with theirs:

Samsung 'Characteristics' Datasheet on the left, my data on the right.

Good!

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

e:

ote:

or a

are

re stable

n the

input

om seeing

duction

rd.

in

is any

e

density

urves

a

e

and

ce

e

and

70nF

.

e (X7R

m r

sting

o).

etup:

t

.

n percent-delta-C to match their chart. Looks like my measurement very tig htly correlate with theirs:

Getting back to the subject title: X5R vs X7R MLCC...

I got in the replacement capacitors and ran them through the same test:

These are made by Taiyo Yuden. 35V X5R in 0805

P/N: GMK212BJ474KG-T Source: Digi-Key

Note that the Samsung parts were 50V X7R in the same package. I would expe ct the 50V X7R parts to have less droop over voltage. At 20V the X7R (Sams ung) parts are down 28% while the X5R (Taiyo Yuden) parts are down only 24% . At 30V it is even more dramatic: 59% vs 45%

The Taiyo Yuden caps start a bit lower in capacitance, but at the bias volt ages I use them for, they are higher than the Samsung parts I was previousl y using.

Neither of these have deal breaking deratings for my use. Since I don't ru n these over 24V, and my boards operate from 4C-40C, I think I'm just going to switch over to these Taiyo Yuden parts...unless I run into trouble in m anufacturing or later. They are a tiny bit cheaper currently ($0.05390 vs $0.05417 in reels of 1000). Most importantly, they are IN STOCK!

Samsung has decent characteristic curves published on Digikey.

Yeah, C(V) curves are all over the place.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

These are made by Taiyo Yuden. 35V X5R in 0805

P/N: GMK212BJ474KG-T Source: Digi-Key

Note that the Samsung parts were 50V X7R in the same package. I would expect the 50V X7R parts to have less droop over voltage. At 20V the X7R (Samsung) parts are down 28% while the X5R (Taiyo Yuden) parts are down only

24%. At 30V it is even more dramatic: 59% vs 45%

Why would you expect that?

Type 2 ceramic caps are electric ferrite beads. They have a voltage rating instead of a current rating, and they saturate at an often unspecified fraction of the rating. They're also lossier than their proper cousins, have worse tempco, etc.

Ran across a cap the other day that was -90% at rated voltage. The voltage rating is electrically meaningless. Ignore it completely and shop by C(V) curve instead. No curve? No sale!

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Design 
Website: https://www.seventransistorlabs.com/

nly

ng

ge

)

I'm wondering why the big concern. Decoupling caps are unfussy things, if t hey're ceramic, have enough capacity, which can be under 1/10th rated, are adequately voltage rated & not unsealed MLCCs - once those are satisfied, w hat's the problem? One really could use salvaged caps of any unknown cerami c dielectric & be ok. (I'm not suggesting actually doing so.)

NT

Preach it, brother.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

Decaps are less sensitive to this sort of problem, but not all MLCCs are used as decaps, by a lot. AC coupling, timing, filters, lead-lag networks, composite amplifiers, capacitance multipliers, and many other applications are unlikely to work properly if the capacitance is uncertain by a factor of 10.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

I would expect that because of charts like this:

...found all over capacitor datasheets. These typically show X5R to drop faster than X7R, and an even bigger effect from voltage rating. Notice that the second cap I tested wasn't just X5R, but a lower voltage rating as well.

As Phil was trying to point out, manufacturer makes a much bigger difference. I posted my data as evidence to that fact.

Those plots are illustrative but entirely worthless for design purposes. Capacitor datasheets are about the crappiest technical docs in captivity, with the possible exception of mechanical drawings of connectors.

Even for a single manufacturer they're still all over the place. Samsung is a prime example.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

Nah, laser data sheets are generically worse.

--

John Larkin   Highland Technology, Inc   trk 

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

I hear you, but that's often because laser diodes are all individuals--their thresholds can vary by a factor of 2 unit-to-unit and over temperature.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

If the value can be 1/10th then you're specifying a 10x too large part!

Do a basic PDN analysis. Not hard once you know how to do it (and build your circuit in such a way that the analysis is easy to do).

Which will be much more stable, assuming they fill the same volume of ceramic chip with fewer electrodes, wider spaced -- effectively 1/10th value at a 10x higher voltage rating (even though they might still call it "50V"). Stores the same energy at saturation but doing so at 10x the voltage. Consequently 10x flatter at low voltages.

Or they make it with the same layer thickness but fewer layers because they're cheap, and you get the same size with the same voltage curve and a fraction of the energy storage at saturation. You never know...

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Design 
Website: https://www.seventransistorlabs.com/

t:

X7R

n only

ating

s,

ltage

C(V)

if they're ceramic, have enough capacity, which can be under 1/10th rated, are adequately voltage rated & not unsealed MLCCs - once those are satisfie d, what's the problem? One really could use salvaged caps of any unknown ce ramic dielectric & be ok. (I'm not suggesting actually doing so.)

I very much doubt there's anyone on here to whom that is not obvious.

NT

f

are

,

depends what voltage you run them at versus rating. But that is of course a n issue with low cost ceramics.

lue

").

a

If you can get any cap you want immediately, great, no problem. My point wa s that when you can't, the options are especially wide for decoupling caps. Even more so if you've added a few extra pads so aren't relying on any one cap to provide enough capacity every time. I use/specify salvaged parts ro utinely to keep BOM rock bottom.

NT