EEVblog #626 - Ceramic Capacitor Voltage Dependency

hi welcome to fundamentals Friday today we're going to take a look at a rather unusual and often little knowing aspect of ceramic capacitors in fact it's a really undesirable aspect and a real trap for young and old players alike if you're not aware of it so let's have a look at some basics first and then we'll very quickly then we'll jump over to the bench and our shows I won't tell you what it is yet I'll let you figure it out on your own but you know a capacitor right simple RC charge circuit like this very common application for capacitors RC time constant charges up gets to a certain threshold and you can detect that and you can use them as time is very typical application not that accurate because as we're aware capacitors are not particularly high tolerance devices and if we take a look at some typical ceramic capacitors we're only talking about ceramic capacitors here then they're basically divided into several classes class 1 capacitors are your NPO and your Co G type and a couple of more obscure ones that you've probably heard of and as you might know they are actually very stable capacitances with temperature okay so once you've actually measured them then they stay pretty stable they basically do not change with temperature so they're typically your low values you know your one nano farad and under you can get in these MPO Co G tights but a larger capacitance values you know your hundreds of nano farad's your micro farad's all your modern ones you know you can get like hundreds of micro farad's now in little surface mounts around packages the technology is absolutely incredible anyway oh I going to a huge amount of detail but basically you've heard of these codes before things like x7r for example what does that mean well there's a three-letter code like this as part of the EIA code system and X for example the first letter represents the low temperature so minus 55 and the next digit here represents what the high temperature limit is in the case of x7r it would be minus 55 to plus 125 degree is Celsius and the are part is the tolerance of the capacitor this case x7r plus minus 15% pretty basic stuff or an x5 R for example is basically the same capacitor but with a slightly lower higher temperature range and you're familiar with these things and as you might know if you're an experienced designer so picking something like an x 7r is a relatively stable high value capacitor to use a relatively stable dielectric so you might use say an x 7r + x fiver in an RC time constant like this but aha wait until we get to the bench I'll show you a trap it's not just about the tolerance here now if we look at the RC circuit which are using the example on the bench in a minute you're familiar with the charging graph charging response of a capacitor and time constants one time constant or tor as it's known is the simple formula R times C and you know a basic building block circuit basic building block formula you should be familiar with and one time constant is the point here it is one time time constant is on the x-axis here and the percent of charge is on the y-axis here and if we put a step voltage in like this then our capacitor will charge up with a exponential formula the main formula is actually down here but you can simplify it we won't go into details on that we'll just use our simple R times C formula today and we'll be measuring this one time constant period one time constant is where the voltage charges up to roughly sixty three point two percent of the final voltage why that particular number well it has to do the fact that when you plug one e to the power of minus 1 up here that's the number that actually spits out of the formula and it's useful and basically a capacitor is essentially fully charged you could say after roughly 5 time constants or five times RC a capacitor is fully charged so if you know your resistance you know your capacitance you can figure out how long it takes to fully charge that cap or or get to one time constant pretty simple pretty basic and of course there's more to this as well there's a class three down here using yet more exotic materials and yet more unstable materials again for ceramic capacitors so pretty much at class one is the most stable class to there okay but as we'll see there's some few traps in there angle class three is even worse so let's not go there what we'll take a look at today is an x5 r or x7 our capacitor on the bench and we won't be taking frequency into account as well it'll be nothing to do with that it'll just be the capacitance this tolerance figure here so if we have a look at say the x5 our capacitor we're going to go look at on the bench now it should measure plus minus 15% does it well yes and maybe no come with me okay the first example we're going to look at is a typical Oh 8 o 5 Class two ceramic capacitor from Mirada one of the top manufacturers and here's the full part number got this from far nails and it's a multi-layer ceramic capacitor X 5 are basically the same as X 7 are it's just slightly lower temperature nominal 10 microfarads and 6.3 volt rated so one of these you know low voltage rating ones you typically only use them on 5 volts or 3.3 volt rails or something like that now it's a normal 10 microfarads with a tolerance defined by this letter here in this case it is K and I'll show you the datasheet in a second that is plus minus 10% now that's not the x7 are the x7 R value of R plus minus or x5 R value of plus minus 15% we saw before that's its temperature or change the temperature coefficient over that full temperature range so but this is the initial tolerance as you'll measure it and buy it from Murata straight out of the packet here so there you go you have to decode these rather complicated part numbers for these things and you'll notice that the 102 is the value there 10 microfarads and we've got K number eight so we have to drop down to the number eight and see what that is and here you go here's the full table available from Murata and kay there it is plus minus 10% so assuming that we measure it under the defined measurement conditions for this capacitor then we should measure 10 microphones plus minus 10% okay now what frequency always supposed to measure this at what do they the specs define it as well here it is capacitance if it's greater than 10 microfarads there it is at naught point 1 kilohertz 100 Hertz there you go at a 1 volt test voltage plus minus 0.2 volts rms so well let's use LCR meter here it is I've a surface mount soldered one of the capacitors onto this little adapter board so we can plug it into our breadboard know that it's completely cooled down from the temperature change it's at room temperature where like 23 degrees C here in the lab so the temperature is not going to change during this measurement and hundred Hertz here we go look we're practically bang on to our 10 microfarads value okay let's round that to 10 microfarads okay this is spot-on okay so let's mount that on our breadboard here we got our bang on 10 micro farad capacitor there we've got a 1k resistor in series with that so we've got that RC circuit I showed you on the board before so we should be able to see the are charge of the capacitor there and I've got that hooked up to my function Jen here and I've got it outputting the square wave so we get a nice big step change from 0 to 5 volts remember this is a 6.3 rated cap so we're safely under that 5 Hertz repetition rate with the 20% a 5 Hertz frequency with a 20% duty cycles that will allow us to see the charge and discharge cycle of the capacitor and here we go if we look our scope here at single-shot capture that bingo there is the charging of our capacitor now there's our green step change on the input there and we're charging up charging up until we get to and upset our we basically then dropping it back down and we don't care about the discharge period here we're only going to be looking at the charging period so it's got enough time to charge ups so there it is there's our beautiful charging waveform now from this we should be able to turn on our cursors measure one time constant here and from that one time constant equals R times C we should be able to verify that that capacitance is correct so if we go into our select our cursors here let's go the Y right down the bottom like that why - let's take that what we want because we've got a 5 volt signal here we basically want that to be our 63.2% of 5 volts so that is three point one six volts so we set our cursor where our Delta Y to three point one six volts will get close enough okay just in the border three point one six two there we go so that's the Delta difference between there so that height is sixty three point two percent or one time constant now we go into our X here and we want our X right at the start period there and X two cursor right where that intercepts the waveform that's pretty close to spot on there and we're getting ten point six milliseconds as the difference there uh-huh let's plug that into the formula alright so what are we have here we measured precisely basically we rounded it to precisely 10 microfarads tor equals R times C we measured at that time constant to be ten point six milliseconds equals one K times C we're trying to calculate C rearrange the formula two point six milliseconds over one K equals 10 point 6 micro farad hey that's not bad that's pretty close really six percent out there still with inside that 10 percent value that we you know the datasheet value yeah it's a little bit offer this six percent to what the our really accurate a Geling LCR meter told us but hey you know you could put that down to the error in the curses and you know things like that eyeballing this sort of stuff so we're going to call that near enough not a problem okay do it one more time for 0 to 6 volts and I've done the cursors here and we're basically getting that 10.1 milliseconds or translates directly into 10.1 micro farad's everything's hunky-dory but what happens if we change this voltage and we add a DC bias voltage in here so our low-level is is our waveform is going from 0 to 6 volts as you can see here's our ground reference level here at DC coupled inputs and 1 2 3 4 5 6 so we're getting 6 divisions there what happens if we actually set this thing so our low level here is say 5 volts so it's going between 5 & 6 volts so the capacitor always has a constant bias voltage of 5 volts on it let's have a look what happens so if we do our single shot capture there there we go there's our ground level down there still exactly the same as what was before 1 2 3 4 5 volts but now now for an our waveform is now superimposed biased 5 volts up but hey that should not affect our formula at all the time constant doesn't matter about what the initial value is it's all about that step change it makes no difference trust me the formula ain't wrong it's been around forever and it you can rely upon it it's 100% correct so let's go in and measure that shall we hmm so let's take that up a little bit more shall we and let's bring that right down to the here say bring that waveform down to here and then we can capture that again and bingo we can go in there and measure that I'll even bring out the horizontal a bit so we can get a reasonable amount of accuracy in there once again we go in there and we use our curses now we've got to set up let's set our y1 minimum our baseline here right down the bottom no problem y2 let's go up because it's a 1 volt change we're looking for 63 point to Delta here difference sixty three point two five there we go excellent so that's fine and we're going to x one over here x ones already set it's that cursor over there so we won't muck around with that X to bring that over to the point where that waveform crosses there what do we get a delta a time constant time it's times difference there once again we're still get into that 63.2% one time period but look at this four point eight two milliseconds you work that back into the formula you get four point eight two micro farad's our capacitance has halved so if we go back to our full formula here for the charging curve of a capacitor will actually just change it slightly v-0 here is actually the origin it's the start of where the voltage starts out okay so it doesn't necessarily have to be zero like we saw in this case it's starting from five volts and going up to six volts the formula still holds I mean you know you read that in any textbook and it is not wrong okay it is absolutely spot-on and of course our tour is our C and well so what is varying okay I'll start voltage is staying the same we've measured it with our oscilloscope there it is we're going up to the six volts and where we were measuring the time constant while measuring the time period correctly about sixty three point five percent of the total change now our resistor in here are well a resistor is probably the world's most basic component these things don't change okay so it's it's a fixed no.11 k so what is left what must be changing because we've measured all the voltages we've conferred on with the oscilloscope well you guessed at C the capacitance is changing it has changed so it's dropped from that value we had before of ten point six microfarads it's dropped down to four point eight two micro farad's with that one volt range with that five volt offset and it's got nothing to do with the one volt range instead of five volts we can confirm that again let's go back remove the DC bias and check that one volt range it's not the range of the voltage it is the DC bias which is causing this capacitance to drop and here we go just to confirm I've set it back to zero to one volt there and there it is there's our zero to 1 volt waveform and what do we get so it's exactly the same apertures before and we get look bang-on 10 milliseconds Wow bet you didn't know that these class 2 ceramic capacitors change their capacitance based on the applied DC bias level unbelievable who knew I tell you that in the data sheets the bastards but not only that the capacitance also changes with the applied DC voltage level as well not just a DC bias offset so the AC level as well in fact the capacitance can go up not decrease as we've seen with an applied AC voltage depending on the level depending on the construction technology and the dielectric used in a particular type of capacitance and it can even change fairly drastically between the same family with different sized capacitors unbelievable so the same family the same type the same xr-7 rating or whatever it is it can change man hate capacitors so are you shocked well you should be now let's go on to a different type here this is a pretty horrible y5v ceramic cap minus 20% plus 80% initial tolerance 10 microfarad 16 volts and let's measure it and see what we get so I measured a value on the LCR meter of 8.6 5 micro farad so let's see what we get on the scope I won't bore you with the operational details so here you go seven-point 8 micro farad 7.8 milliseconds exactly the same resistor everything else so it's a reading quite significantly lower than that 8.6 we were getting before and well that's not a mistake yeah it was we've seen capacitors can vary with bias voltage and also other things like the just the basic applied voltage now let's check the same y5v capacitor from 0 to 10 volts instead of 0 to 1 volts once again there's no DC bias here at all what do we get well look at that it looks substantially different and there we go look at that 4.6 milliseconds goodness how capacitance once again has like halved pretty horrid but look at this the wave shape the charging waveform has actually changed it is not that sort of you know rapid rise and then the decay like that it sort of goes up like this and then decay is totally different to what we saw before ah so these y5e caps are absolutely horrible and there's some weird you know physics going on here based on the dielectric and the construction of these are y5e capacitors they're nothing like those x7 ours that we got before or even like this same y5v but down at 1 volts all we've done is now change it to 10 volts there we go a 2 volts per division and like the characteristic of the charging of this thing has changed ah man you definitely don't want to use these for any sort of timing application and what happens if we add a bias level we'll add a 9 volt DC bias level remember this is a 16 volt cap this capital only go up to our 10 volts maximum output voltage on a square wave so we're going from 9 to 10 so we're going back to that 1 volt difference that 1 volt change but let's have a look at the waveform and this is what we get now yes we're back to that characteristic shape that we saw at 1 volt that proper you know the curve you actually expect for the charging of a capacitor but look what we're at 1 point 5 milliseconds 1.5 micro farad's it's dropped but you know not quite an order of magnitude but jeez it's getting there and this is for a 16 volt rated you know y5v cap and that we measured the value measured the capacitance of with our good LCR meter hopeless now the real problem with all this is that this is rarely mentioned in data sheets there are exceptions to this some manufacturers do actually remind you of it every now and then but high sometimes it's almost next to impossible to find for your particular manufacturer now this particular Mirada capacitor we're actually using here are we use first of all the O 805 here the 6.3 volts 10 microfarads there it is there and if you jump on over to the product page here by the way it's not in the datasheet you have to actually go to the manufacturers product page for that specific capacitor and here look at this look at this graph right here maybe I could zoom into that perhaps not that great but here it is the capacitance note here it is the DC bias characteristics of the capacitor look zero is like this is the change in capacitance ie the drop in capacitance with the DC bias level and that's exactly what we saw here and look at that graph it can drop as much as 60% by the time it gets to 6.3 volts so it's 6.3 volt rating it's dropped by a massive 60 the capacitance has dropped by 60% it's unbelievable and look over on this side here here are the AC voltage characteristics I was telling you about I don't have time to set up experiments to verify this one today I want to keep it a bit short short as possible probably gone long enough already anyway the capacitance change with AC voltage RMS look quite significant up to like plus 15% minus 30% right down at low signal levels unbelievable who knew this stuff right and AVX are quite decent they actually remind you of it here it is capacitance change versus bias voltage look at the Jurassic drop off in this particular one I mean this is just awful look at this this is for one of their general purpose up multi-layer ceramic I suppose so the and actually goes up a bit as I said it can actually go up at small DC bias levels and then it drastically drops down to you know once again these are high voltage caps but he can drop way way off eighty even ninety percent or more of your nominal rated capacitance just by adding DC bias or by changing that AC voltage in any way you can get some interesting a little obscure articles here which tell you all about the physics and stuff like that behind all how all this sort of stuff works and how the various crystalline structures work and look did you see know that the crystalline structure changes with temperature look at that this is for certain type of dielectric and our construction which is found in typical class 2 and class 3 multi-layer ceramic capacitors so depending on the temperature it actually changes the crystalline structure and then the DC bias it goes in to try and explain the physics of how that actually works and how the DC bias affects the actual capacitance so anyway I'll link in this stuff down below check it out and there's a lot more to what some manufacturers don't even mention it but Wow it so can be a real trap capacitance ain't capacitance and here's an example of a VX actually showing you the effects of voltage in this case it's the AC voltage capacitance change versus AC voltage as well you can see how the capacitance can actually increase drastically you know 50 odd percent or so depending on the AC voltage applied to it for these multi-layer ceramic capacitors so yeah really tricky business now a couple of the manufacturers namely our Murata and a VX these are the two of the better manufacturers in this field for analyzing and trying to correct you know they're always improving their manufacturing processes and stuff like that for to you know try and eliminate this kind of effect although it's next to impossible with the class ii capacitors but at least they're aware of it and they do allow you our tools to actually get simulated as well the Narada website has tool called Atmos in Matt surfing and it allows you to actually plot this stuff with our bias values and all that sort of stuff and AVR as well they have spike app software which will do a similar thing as well so you can play around with these simulation tools but there's nothing better than actually whacking it on your bench and seeing it for real like we did today and I won't bore you with the details I'll leave it up to you to experiment with this but if you drop that DC bias level like say one volt at a time so if you went from nine to ten and then eight to nine and seven to eight and six to seven and so forth and dropped it down or increase it then if you plotted that and measured your capacitance at each value you would get that similar characteristic huge big characteristic drop in capacitance versus your bias voltage exactly like the manufacturers tell you if you can find their data so there you go that's an interesting fact that not a lot of even experienced design engineers know about because well they just throw their capacitor in they assume it works here they know about all sorts of other characteristics of the ceramic caps and all your temperature coefficients and mainly all they care about is temperature but what a lot of people don't think about is that capacitance can change with voltage as well not only DC bias voltage that can have a drastic effect but also the applied AC voltage as well so it's not just like RC time constant like this of course if you're using the same circuit for you know you're doing filtering and things like that it can really matter based on the signal level it can be really quite critical so you have to be very careful with how you use these modern ceramic capacitors they're fantastic with all these wonderful materials technology that goes in and give you incredibly high capacitance in incredibly small volumes for SMD and stuff like that but yeah there's a few downsides and it's not just temperature voltage as well trap for young and old players alike certainly now this only applies to class 2 and above ceramic capacitors it does not apply to electrolytic capacitors talams and class 1 npos and things like that so yeah and there's of course I've done videos on other our traps with these ceramic capacitors are as well piezoelectric effect of course well I'll link that in down below if you haven't seen that video so there's lots of stuff to think about here man so many traps got to be careful electronics design ain't as easy as it seems on the surface you dig deeper and deeper and deeper and well if you're in a critical application this can be a really big deal so there you go hope you enjoy in that video there's so much I can do on this subject I can measure you know hey let's not even get into Cheney a frequency that changes and all sorts of other stuff thrown into the mix our goodness could do hours and hours of videos experimenting with this sort of stuff but you should get to the bench and have a play around with it yourself it can be rather fascinating so there you go um hope you enjoyed that one if you want to discuss it jump on over to the EEV blog forum link is down below if you like fundamentals Friday please give it a big thumbs up even though well it's actually Saturday shooting in song yeah I'm a bit late anyway catch you next time you you

Channel: EEVblog

Views: 78,716

Keywords: Capacitor, Voltage, time constant, tau, measurement, how to, tutorial, oscilloscope, capacitor charging, charging, formula, dc bias, Ceramic Capacitor, mlcc, murata, avx, datasheet, curve, frequency, cursor measurement, charge time

Id: 2MQyQUkwmMk

Length: 27min 57sec (1677 seconds)

Published: Sat Jun 07 2014