I thought it would be fun to get the waveform. I have some 100 uF samples coming that should have even worse cv curves. Somebody sells a ceramic thing that's specifically intended to be an RF varicap; that could be useful but the tempco os probably rotten.
I can't stand that whiney guy, so I don't watch his slo-mo lectures.
And one of those giant b+w Polaroid scope cameras.
Charging_47u.JPG
But seriously you wouldn't have ended up with that 'impossible' curve if you'd used analogue storage.
Um, can you explain that observation? Do you thing the curve is unrealistic?
The post-it pic is a little fuzzy, but it's a mu, not a p. The jpeg file name is
It's a 47 uF, 6.3 volt ceramic cap. The "ceramic" on the post-it is hard to confuse with "electrolytic" but you can do that, too, if you are really determined. Ask Sloman for help.
It's a very commonly-observed curve as you well know. But it's not right from your screen shot IMV. If you examined the trace with an analogue scope (such as the 466 just to take a random example) you'd see the real, natural beauty of it.
Can you do the same experiment using a good 47 uF cap to validate the setup?
Tom
It's a perfectly good 47u ceramic cap, fresh off the Digikey strip. A square wave, a resistor, and a scope hardly need validating. The waveform is logical.
I thought the waveform was interesting, and all sorts of loonies here are deliberately (or more likely ignorantly) misinterpreting it and getting worked up.
What's not right? Do you think the scope is somehow distorting the actual waveform?
ANALOG scopes had nonlinearity from CRT physics! Trace rotation, too; that had to be adjusted now and then. And vertical and timebase errors, especially scopes with tubes in the signal path.
I think a good digital color scope waveform can be beautiful, better than a fuzzzy green analog scope pic. Wanna post some and compare?
It might have been kinder to use a 47uF 25v capacitor in the test setup though if you are going to apply a 20v square wave to it.
A well-behaved cap would be no fun.
My ancient HP6212 bench supply only goes to 120 volts, but the cap is still alive there. It's leaking about 60 uA. I'll let that soak for a while and see if anything happens. The voltage rating on those caps is maybe based more on c loss than on failure.
Hmmm. When a 24v switching regulator (a wall wart) drives a 5v switching regulator (a negative impedance load) which drives a 1.8 volt switching regulator (another negative impedance load) there will be all sorts of opportunities for instability.
On Wednesday, October 19, 2016 at 11:01:02 AM UTC-7, John Larkin wrote: ...
...
When I was faced with this problem a couple of years ago I found that in ma ny cases increasing the voltage rating on a cap did not reduce the effect o f voltage - it just extended the degradation further. Few manufacturers wo uld document it and they would never give guaranteed limits.
We had to use a cap a with the marked value 10 times what we wanted to get the required amount. We were looking for ~1uF at 20V in an 0402.
I also found that there can be a lot of hysteresis and aging in the caps if they are subjected to voltages over minutes to days - and the changes were always in the wrong direction :-(
kevin
Not sure what you mean by a negative impedance load. But I did a thing with 24V wall wart driving +15V spms which drives a -15V spms. Everyone seemed fine with big caps and/or a small R as load.
George H.
So if I have, say, 10 volts that I want to bypass with a 1206 cap, what would give me the most capacitance? A 10 volt cap or a 20 volt cap with lower nameplate capacitance?
Makes a boy want to buy aluminum polymers just to not have to think too much.
I just got some 56uF 25v polymers that I'll have to blow up, too.
On Wednesday, October 19, 2016 at 11:31:54 AM UTC-7, John Larkin wrote: ...
...
We tried different dielectrics and voltages and there are some manufacturer s doing medium K types to address this problem (I think Samsung was one) bu t found that in general the larger nameplate capacitance won even though it decreased more than the others.
We ended up using tantalums - they were only just becoming available in 060
3 and 0402 sizes in 2012.kevin
He bias's a 6.3v cap to 5Vdc, puts a 1V square wave on top and measures Tc vs. just a 5V square wave and measuring Tc. Plus a second comparison. I think you would enjoy it, just celebrate his different style.
Mikek
Look closer and open your mind to other possibilities. Mikek
When the switcher input voltage goes up, the input current goes down, so the slope is negative. We had one rig where the wart fed a switcher and it oscillated. A big alumimum cap across the wart output fixed that.
47rho?
Have something to add? Share your thoughts — no account required.
Ask the community — no account required