What every Class AB builder needs to be constantly aware of

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Think of an RLC filter, capacitors have resistance and inductance,as they are nothing other than a piece of aluminum foil wrapped around your finger.
Make the capacitor 5 times larger , inductance 5 times smaller, the product LC remains the same.
Resonance occurs when Ft = 1/ 2Pi SQRT(L*C) , if LC remains the same so does the resonance frequency.

Now think of a spring-mass system in place of the RLC.
For a spring-mass system resonance is also determined by the product of the spring stiffness and the mass.
The formula for spring-mass resonance even looks the same.
A combination of a weak spring and high mass has a much higher resonance amplitude compared to a combo of a stiff spring and low mass, despite occuring at the same frequency.

Would you like some complex math ?
 
These are impedance measurements of 2200uF 35V Panasonic cups up to ten in parallel. The resulted impedance is higher than expected due to leads and solder joints resistance. As you can noticed no dips.
 

Attachments

  • capparall.gif
    capparall.gif
    21.7 KB · Views: 338
The graphs are excellent Dimitri.

It seems chemicon have a very non-inductive package if I read those frequencies correctly. And that dip is interesting?


Jacco, If c is x5 and L is /5 , L*C= same so no change in resonance?

Dimitri's measurements actually show the resonance coming down with multiple AND the diminishing benefit of paralleling at HF. However the 40% reduction in impedance at 5-20KHz over the single 22mF could be worthwhile. I note that 4x2m2F is the same at HF as one 22mF.

Are they as close as practicable, Dimitri? Yours are not the Chemi-con?
 
seems chemicon have a very non-inductive package if I read those frequencies correctly
frequency range 10e2-10e7, it appears chemicon inductance two times lower
btw couldn't find any reference to LXF series at their site

this dip between 100k and 1M puzzles me - can be explained only from multisite model

Are they as close as practicable
you mean to each other - exactly
 
B Cullingford said:
?weak + high mass ... ?
yes please, love that complex math and imaginary numbers

RLC/LC filters are often compared to sping/mass systems, afaik.
For spring/mass the resonance factor is SQRT(C*M*G).
C: spring stiffness, M: mass, G: gravitation constant.

With spring/mass systems you see this high excitation behavior for different C-M values, i suspect the same applies for parallel capacitors.

I'd think the shift in resonance freq. is because of the multi-RLC model, as Mr Danyuk suggested for the dips.
The graph for multiple parallel capacitors seems to show that a straight line could be dawn through the resonance frequency of different parallel numbers.

I like to stick to simple comparisons, suggesting complex math is what i refer to as a sense of humor.(i hated complex math at Uni)
 
Re graphs in posts 202&203.
the 10*2.2mF cf 22mf makes sense, the hf performance of 10*2.2mF is superior due to the effect of the smaller caps(less inductance).
The 10*2.2mF cf 1*2.2mF also makes sense. The high frequency performance is dropping off slightly due to the added inductance of the interconnections.
The movement of the minima relative to frequency tell a large part of the story.
It appears from these that many small caps are always better than a small number of large caps provided the interconnections are kept very short.
Has anyone come up with an explanation for the weird behaviour of the earlier paralleled comparison graphs that show gains much greater than 10* and at higher frequencies?
 
Jacco:
At the risk of going to far off topic, what I was questioning was whether you had inadvertently mixed the weak and strong in phrase: "A combination of a weak spring and high mass has a much higher resonance amplitude compared to a combo of a stiff spring and low mass, despite occuring at the same frequency."

However it seems to me that the weaker a spring is the lower the f of oscillation of a spring pendulum, and increasing the mass will also reduce the f, so weak spring + low mass will have same f as strong spring and high mass. High mass and strong springs will have more stored energy for a given amplitude, the amplitude will depend on the energy put into the system (we can also see that the higher the f of oscillation the higher energy since the mass is moving faster)

I too like simple comparisons, though I love math as well and it often gives futher insights into a system.

As to the topic at hand - I have often thought that multiple smaller caps could have advantages but neither properly done the math or tried. I suspect that implementation has alot to do with the results stray inductance and the like.
 
True, reverse.

W= SQRT (K/M) K = spring stiffness. W in radials per second.
this is transferred to C= 1/K
=> W = 1/ SQRT(C*M).
C*M = constant=> same resonance frequency.
Means high spring stiffness and high mass, or weak spring and low mass.
A weak spring and low mass has higher ampitude, but the same resonance frequency as a combination of rigid spring and high mass.
Just as confusing as C/W and W/C.

I am getting old, and shouldnt be posting when i am too busy with work, sorry. :xeye:

I meant to say that as the spring-mass system is often mathematically compared to an RLC system, putting a number of capacitors in parallel may be mathematically comparable with a system of coupled springs and masses.

I mentioned this because i remember from the time that i practiced matrices for multiple spring-mass systems in finite elements programming a long time ago the graphs looked identical.
 
Jacco, - I know that feeling. And I don't want to get into that math right now - studying water rocket math right now water rocket thrust equations

with R being equilavent to damping.
Lots of coupled pendulums can lead to chaotic looking behaviour - maybe an indication that just paralleling lots of caps without due consideration to layout, trace inductance etc is maybe not the best solution. Or to put it another way YMMV, according to your path.

Bill
 
B Cullingford said:
studying water rocket math right now [/URL]

Thanks for posting the link to the water rocket equations, Bill.

I had one of those when i was 10 years old, my uncle brought it as a present from the states.
Unfortunately it broke on the pavement, by the time they sold them overhere i was too old to play with water rockets.
When i moved to the US i only had eyes for other creatures.

I do like math, i will enjoy reading the stuff.
 
amplifierguru said:
Thanks guys.

Now with each output half having to deliver half the waveform, just look at the spray of harmonics that are generated in and around the outputs and power supply lines


Hi Mr.Greg,

Your posting displayed the spray of harmonics in supply lines as well.
My WISE question is this what will happen when a birdged amp sprays the harmonics, Aren't they cancelled[I may be Wrong] due to out of phase waveforms in case of harmonic spray.

Kindly , share your views with regarding us.

Sincere regards,
Kanwar
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.