/ News & Press / Video / SDG #061 Ceramic Capacitors will Blow Up your PCB
SDG #061 Ceramic Capacitors will Blow Up your PCB
WEBVTT Kind: captions Language: en
00:00:00.709 so ceramic capacitors killing your 00:00:03.35000:00:03.360 electronics what on earth am i talking 00:00:05.12000:00:05.130 about well I thought this was quite an 00:00:06.86000:00:06.870 interesting topic because it's something 00:00:08.36000:00:08.370 that new engineers may not have even 00:00:11.00000:00:11.010 considered but a situation that occurred 00:00:13.75900:00:13.769 at work recently on a project that I was 00:00:15.91900:00:15.929 working on at the situation where there 00:00:18.23000:00:18.240 was a ceramic capacitor on the input 00:00:20.33000:00:20.340 circuitry to the electronics there was a 00:00:22.70000:00:22.710 switch mode power supply and then the 00:00:24.01900:00:24.029 rest of the device and when we plugged 00:00:26.06000:00:26.070 in the power to the device under certain 00:00:28.07000:00:28.080 situations that DC to DC converter no 00:00:31.46000:00:31.470 longer worked from that point onwards so 00:00:34.01000:00:34.020 I thought I'd demonstrate what was going 00:00:35.42000:00:35.430 on here so I've got a little then I 00:00:38.33000:00:38.340 think this is a 20 2 micro farad ceramic 00:00:41.42000:00:41.430 capacitor and I've got my bench power 00:00:44.75000:00:44.760 supply set up so I've got the other lead 00:00:47.75000:00:47.760 here I think this is set to around 12 00:00:50.72000:00:50.730 volts or something yeah 12 volts and all 00:00:53.86900:00:53.879 we're going to do is measure the voltage 00:00:55.45900:00:55.469 across the ceramic capacitor as we 00:00:57.47000:00:57.480 connect up the power so here we go and 00:01:02.54000:01:02.550 if you have a look at the oscilloscope 00:01:03.59000:01:03.600 you can see actually we've got a massive 00:01:05.69000:01:05.700 overshoot a nominal voltage here at 12 00:01:08.92900:01:08.939 volts but the scope is saying that the 00:01:11.63000:01:11.640 maximum here is actually 20 volts and 00:01:13.82000:01:13.830 the situation that I saw it was actually 00:01:16.76000:01:16.770 a lot higher than that and it was 00:01:17.96000:01:17.970 causing the the input to the DC to DC 00:01:20.66000:01:20.670 converter to no longer work because it 00:01:22.91000:01:22.920 was seeing such a high transient and we 00:01:25.07000:01:25.080 can demonstrate this with a smaller 00:01:26.92900:01:26.939 capacitor as well so it's not just 00:01:28.39900:01:28.409 limited to something like one of these 00:01:30.20000:01:30.210 22 microfarad capacitors this one is a 00:01:34.17900:01:34.189 22 nano farad capacitor and if we hook 00:01:37.99900:01:38.009 up the scope again to this and connect 00:01:41.21000:01:41.220 up the power BAM there we go again so 00:01:44.69000:01:44.700 this time we're reading a slightly 00:01:46.28000:01:46.290 higher peak voltage so 22.6 but the 00:01:49.73000:01:49.740 duration is much much shorter here so 00:01:52.63900:01:52.649 what on earth is going on so all we've 00:01:54.62000:01:54.630 got is our 12 volt power supply with our 00:01:56.38900:01:56.399 ceramic capacitor at the other end of it 00:01:57.95000:01:57.960 with a scope connected across the 00:01:59.99000:02:00.000 capacitor and our education and books 00:02:02.24000:02:02.250 should all tell us that the waveform 00:02:04.45900:02:04.469 that we should see is this capacitor 00:02:06.64900:02:06.659 charging waveform that asymptotes 00:02:08.90000:02:08.910 towards 12 volts so what is giving us 00:02:11.78000:02:11.790 the big spike 00:02:13.66000:02:13.670 well basically we've got some non-ideal 00:02:15.46000:02:15.470 characteristics between the power supply 00:02:17.26000:02:17.270 and the ceramic capacitor so in my case 00:02:19.48000:02:19.490 on the bench here I've got some 1 meter 00:02:21.19000:02:21.200 cables so this distance is just 1 meter 00:02:25.30000:02:25.310 but what were actually got here is we've 00:02:28.18000:02:28.190 got some inductance we've got some 00:02:31.08000:02:31.090 resistance and there would be some 00:02:33.34000:02:33.350 parallel capacitance along here just to 00:02:36.04000:02:36.050 buted all the way along but that's 00:02:37.60000:02:37.610 minimal because we've not got wires that 00:02:40.36000:02:40.370 are joined together along the whole 00:02:41.62000:02:41.630 length but what's critical is that this 00:02:43.93000:02:43.940 inductance here is causing us basically 00:02:46.75000:02:46.760 to have a big inductive spike so what's 00:02:50.41000:02:50.420 going on here essentially is that we've 00:02:52.00000:02:52.010 got this LC circuit and a ceramic 00:02:55.24000:02:55.250 capacitors are extremely low ESR devices 00:02:57.82000:02:57.830 which means that they'll charge very 00:02:59.44000:02:59.450 very quickly so what happens is we start 00:03:02.19900:03:02.209 charging the capacitor and we know our 00:03:04.78000:03:04.790 inductors resist a change in current so 00:03:08.14000:03:08.150 what happens is the current slowly ramps 00:03:10.21000:03:10.220 up and then the capacitor charges up and 00:03:13.03000:03:13.040 the inductor still wants that current to 00:03:14.59000:03:14.600 continue flowing so it does whatever it 00:03:16.18000:03:16.190 can do to try and keep that current 00:03:17.62000:03:17.630 flowing and what that happens to be is 00:03:20.19900:03:20.209 by raising the voltage here as high as 00:03:22.33000:03:22.340 it can until all of the energy and the 00:03:24.97000:03:24.980 coil collapses so what we're actually 00:03:27.67000:03:27.680 seeing is basically an RC tank circuit 00:03:31.18000:03:31.190 which is why we're also seeing a bit of 00:03:32.77000:03:32.780 ringing as well so this is a particular 00:03:34.81000:03:34.820 problem in certain applications because 00:03:37.03000:03:37.040 for example where you've got a plug top 00:03:38.94900:03:38.959 power supply if you have this plugged in 00:03:43.21000:03:43.220 and then connect up your DC connector 00:03:46.18000:03:46.190 this is all charged up ready at 12 volts 00:03:48.78900:03:48.799 so you just shove it in 12 volts 00:03:51.03900:03:51.049 directly to this capacitor which wants 00:03:52.69000:03:52.700 to create a really huge current and then 00:03:54.94000:03:54.950 we see a big current and then the 00:03:56.56000:03:56.570 inductor really wants to get that high 00:03:57.94000:03:57.950 current flowing and you see a really 00:03:59.83000:03:59.840 high voltage in the situation where your 00:04:01.96000:04:01.970 power supply is permanently connected 00:04:03.67000:04:03.680 and you turn on the mains instead you 00:04:06.03900:04:06.049 won't see this because the the voltage 00:04:08.41000:04:08.420 on the output of your power supply will 00:04:11.25900:04:11.269 slowly increase and you won't see such a 00:04:13.78000:04:13.790 high current being drawn by your 00:04:16.09000:04:16.100 capacitors so I thought it might be 00:04:18.78900:04:18.799 interesting just to see what other 00:04:19.90000:04:19.910 capacitors are affected by this it's 00:04:22.02900:04:22.039 typically going to be those with a 00:04:23.50000:04:23.510 particularly low ESR so in this case 00:04:26.26000:04:26.270 we've got a winner 00:04:27.12900:04:27.139 film capacitor so let's do the test on 00:04:30.04000:04:30.050 this one and yet there we go so we see 00:04:33.82000:04:33.830 in the same sort of twenty two point six 00:04:35.67900:04:35.689 peak voltage then we've got another type 00:04:39.04000:04:39.050 of film capacitor and there we go very 00:04:43.02900:04:43.039 similar we've got a lot of ringing on 00:04:44.26000:04:44.270 this one I think this is a particularly 00:04:45.79000:04:45.800 low impedance caster and I've got an 00:04:48.58000:04:48.590 electrolytic capacitor so here we go 00:04:51.11900:04:51.129 they were going a very small amount of 00:04:53.58900:04:53.599 overshoot here so only thirteen point 00:04:55.51000:04:55.520 four dropping down to twelve volts so 00:04:58.60000:04:58.610 you might say to yourself well have 00:04:59.74000:04:59.750 normally got a load on the output here 00:05:01.30000:05:01.310 we haven't just got a capacitor with 00:05:03.07000:05:03.080 nothing there 00:05:03.90900:05:03.919 so what I've done now is I've just 00:05:05.55900:05:05.569 connected a resistor across here so this 00:05:08.80000:05:08.810 is a 22 ohm resistor which should 00:05:10.68900:05:10.699 present quite a load so let's connect up 00:05:14.05000:05:14.060 the power supply again and you can see 00:05:17.58900:05:17.599 look we're still getting a very similar 00:05:18.93900:05:18.949 reading nineteen point nine volts at the 00:05:21.64000:05:21.650 peak here settling down to twelve volts 00:05:24.30900:05:24.319 so even with a relatively large load 00:05:26.49900:05:26.509 like this it's having virtually no 00:05:29.37900:05:29.389 effect on that initial spike which is 00:05:31.42000:05:31.430 just passed straight through to our 00:05:32.74000:05:32.750 electronics in our DC to DC and causing 00:05:35.55900:05:35.569 the input circuitry to blow up so to try 00:05:38.37900:05:38.389 and combat this what we actually need to 00:05:39.76000:05:39.770 do is to limit the current through the 00:05:43.20900:05:43.219 capacitor so that we don't see such a 00:05:45.18900:05:45.199 high peak current and thus we don't see 00:05:46.95900:05:46.969 such a high peak voltage when the 00:05:49.30000:05:49.310 capacitor is charged up so basically the 00:05:52.57000:05:52.580 simplest way that we can do this is by 00:05:54.04000:05:54.050 adding some resistance in here 00:05:55.92900:05:55.939 as much as we can get away with without 00:05:58.17900:05:58.189 affecting our electronics but enough to 00:06:01.14900:06:01.159 dampen the effect of the increased 00:06:04.26900:06:04.279 current right so here we are with our 00:06:06.07000:06:06.080 capacitor and our resist a combination 00:06:08.07900:06:08.089 the scope is across the capacitor and 00:06:09.67000:06:09.680 the resistor and if we repeat the 00:06:11.58900:06:11.599 experiment there you go we can see we 00:06:14.58900:06:14.599 get basically no overshoot here it's 00:06:16.86900:06:16.879 just gone over by about one volt at this 00:06:18.99900:06:19.009 point and what we actually do is have 00:06:21.51900:06:21.529 this resistor in series with our 00:06:23.23000:06:23.240 capacitor but our load would still be 00:06:25.62900:06:25.639 across and the overall combination of 00:06:27.90900:06:27.919 the R and the C so what we're actually 00:06:30.21900:06:30.229 doing here is effectively adding some 00:06:32.23000:06:32.240 impedance to our capacitor so we're 00:06:33.96900:06:33.979 increasing the ESR we've still got our 00:06:36.12900:06:36.139 load connected across the combination of 00:06:38.46900:06:38.479 the R on the C so we don't connect our 00:06:40.11900:06:40.129 low 00:06:40.46000:06:40.470 across the sea because that will affect 00:06:42.85000:06:42.860 maximum current that we can draw from 00:06:44.78000:06:44.790 our power supply so I'm just simulating 00:06:47.33000:06:47.340 that here with our resistor across the 00:06:49.70000:06:49.710 combination of the two and if we apply 00:06:51.56000:06:51.570 power again there we go we get basically 00:06:54.92000:06:54.930 no overshoot on the oscilloscope so 00:06:57.83000:06:57.840 essentially we want to get this 00:06:59.48000:06:59.490 resistance as low as possible so that 00:07:01.91000:07:01.920 our circuit still functions as best as 00:07:03.71000:07:03.720 possible because DC to DC converters and 00:07:05.90000:07:05.910 that kind of thing 00:07:06.62000:07:06.630 want extremely high transient currents 00:07:09.14000:07:09.150 so if we start adding big resistances 00:07:11.48000:07:11.490 here you start getting poor behavior of 00:07:13.97000:07:13.980 your DC to DC converter so what you want 00:07:17.15000:07:17.160 to do is get this as low as possible and 00:07:19.01000:07:19.020 basically what all we're doing is 00:07:20.06000:07:20.070 increasing the ESR of this capacitor by 00:07:22.07000:07:22.080 having this series resistance what may 00:07:24.47000:07:24.480 work better for you depending on your 00:07:26.03000:07:26.040 application is to have cascaded stages 00:07:28.30000:07:28.310 so you might want a COC type filter or 00:07:31.64000:07:31.650 something like that so that you still 00:07:32.84000:07:32.850 get a high transient current capability 00:07:36.29000:07:36.300 if your DC to DC converter but you're 00:07:38.60000:07:38.610 just limiting that inrush current the 00:07:40.76000:07:40.770 other thing that you can do is you can 00:07:42.35000:07:42.360 design something like a soft start 00:07:43.76000:07:43.770 circuit and there's lots of ICS that do 00:07:45.59000:07:45.600 this for you and basically during 00:07:48.14000:07:48.150 startup you've got a resistance in 00:07:50.87000:07:50.880 series with all of your electronics and 00:07:52.64000:07:52.650 then a predefined time later it shorts 00:07:55.13000:07:55.140 out that resistor and then from then 00:07:57.08000:07:57.090 onwards your electronics can draw as 00:07:59.18000:07:59.190 much power as it likes so there we go 00:08:01.07000:08:01.080 hopefully found that video useful and 00:08:02.60000:08:02.610 particularly if you're designing 00:08:03.50000:08:03.510 electronics if you come across this 00:08:05.06000:08:05.070 problem in the future you might know 00:08:06.92000:08:06.930 what to start looking for but until next 00:08:09.32000:08:09.330 time thanks for watching 00:08:12.81000:08:12.820 you 00:08:23.10000:08:23.110
Office location
Engineering company LOTUS®
Russia, Ekaterinburg, Lunacharskogo street, 240/12