I have read several post's here about using a 10,000uf capacitor to block DC to a speaker. I cannot find the post where i saw it anymore. someone posted a link to the PASS F4 that uses a 10kuf cap at the output.
as a student i have been experimenting with amplifiers and doing some repairs of vintage gear. I want to use a capacitor to protect my test speakers. I realize there will be some sonic degradation which i don't care so much about, but what is the effect of that capacitor on the amplifier under test? does the amp still see the 8/4ohm speaker load? safe to use with tube amps? any issues with causing oscillations or odd problems?
Thank you for the help
as a student i have been experimenting with amplifiers and doing some repairs of vintage gear. I want to use a capacitor to protect my test speakers. I realize there will be some sonic degradation which i don't care so much about, but what is the effect of that capacitor on the amplifier under test? does the amp still see the 8/4ohm speaker load? safe to use with tube amps? any issues with causing oscillations or odd problems?
Thank you for the help
No problems in general, but with the electrolytic cap you might want to consider two of then in series with some DC bias to protect the caps, this if you are testing higher power levels at lower frequencies.
A series cap used in this way is appearing before a shunt resistance (the speaker or driver), and this forms a first order highpass filter (close enough, since the load is reactive).
It is important to understand that cap is only truly blocking DC under steady state conditions and the important question is more about what happens the instant that the fault occurs (that puts DC across the speaker load)? The situation is like a step response applied to the high-pass filter formed by the capacitor and the load. When you model that, you can see that the instant the fault occurs there is a voltage "spike" almost equal in magnitude to the DC voltage that reaches the speaker terminals for a brief time. How long and how high is this peak? This depends on the RC time constant. You might want to make the capacitor very small to make this time constant also be small, but then the highpass RC filter corner frequency will also be high, e.g. well into the audio band and you will eliminate bass from the loudspeaker. When you make the capacitor large the corner frequency can be made low but then the time constant is long. You are in charge of what the right value should be (but I am sure some wise person can suggest a value that is useful in practice).
After a couple of time constants pass no more current can flow because the cap is fully charged up. Steady state is reached and the voltage across the speaker falls back to zero, e.g. "DC blocked".
It is important to understand that cap is only truly blocking DC under steady state conditions and the important question is more about what happens the instant that the fault occurs (that puts DC across the speaker load)? The situation is like a step response applied to the high-pass filter formed by the capacitor and the load. When you model that, you can see that the instant the fault occurs there is a voltage "spike" almost equal in magnitude to the DC voltage that reaches the speaker terminals for a brief time. How long and how high is this peak? This depends on the RC time constant. You might want to make the capacitor very small to make this time constant also be small, but then the highpass RC filter corner frequency will also be high, e.g. well into the audio band and you will eliminate bass from the loudspeaker. When you make the capacitor large the corner frequency can be made low but then the time constant is long. You are in charge of what the right value should be (but I am sure some wise person can suggest a value that is useful in practice).
After a couple of time constants pass no more current can flow because the cap is fully charged up. Steady state is reached and the voltage across the speaker falls back to zero, e.g. "DC blocked".
Thank you everyone. My intent was to put a capacitor IN my test speakers. I accidentally blew the woofer in my last set of test speakers, slipped with a probe and smoked the woofer with DC. totally my fault in that case. But, I have had some situations where some DC protection would be useful. I had one amp with an intermittent fault. would work perfectly for hours on end then BLAM blow the power fuse. I could never track down the fault as it would pass by the time i could replace the fuse. finally found one transistor had leads bent too tight and had an internal break. at random, it would open and blow the fuse. and each time it would do that. the speakers would just about jump out of the boxes before the fuse could blow.
I asked about tube amps just because I want to leave the cap in the speaker and want to be able to use the speaker with any type of amp.
These would be considered test tools. so I am not the most concerned about sound quality, but more concerned about how the caps would react with the amp. good to know that it will not be an issue. and I could put a fuse after the cap and before the woofer if i really had to. but I think they will survive a short spike.
I have some 10,000uf nichicon caps I can put in series. that should work ok.
I asked about tube amps just because I want to leave the cap in the speaker and want to be able to use the speaker with any type of amp.
These would be considered test tools. so I am not the most concerned about sound quality, but more concerned about how the caps would react with the amp. good to know that it will not be an issue. and I could put a fuse after the cap and before the woofer if i really had to. but I think they will survive a short spike.
I have some 10,000uf nichicon caps I can put in series. that should work ok.
Bear in mind that the series capacitor could see either polarity of DC voltage
so at least connect two equal polarized electrolytic capacitors in "anti-series".
so at least connect two equal polarized electrolytic capacitors in "anti-series".
smaller values (~450uF for "8 ohm driver") are used to "boost" LF in so called 3rd order sealed systems. If you know / can measure the driver parameters and tuning of your test speakers, then you can use hornresp (FREE!!!) to examine frequency response, excursion, impulse response etc.
If anyone wants to calculate/model the effect of the RC highpass formed by the DC blocking cap (assuming the driver acts purely as a resistor) this is a good online tool for that:
http://sim.okawa-denshi.jp/en/CRtool.php
If you enable the step response output you will see the behavior I mentioned as well as the duration of the spike. The highpass corner frequency is also shown.
For example, for an 8 Ohm load 470uF Fc=42Hz, and for C=10000uF (10mF) Fc=2Hz ish.
Using a very large electrolytic cap (or pair, back-to-back) has the advantage that the cap is essentially a short circuit for the audio band and does not contribute significantly to distortion, etc. This is not true for lower values like 470uF because for low bass there will be some voltage drop across the cap. You can find a good review of this behavior in "The Design of Active Crossovers" by D. Self that also cites some old Electronics World articles. Self is writing about DC blocking in active circuits, but the concept is the same for passive, speaker-level DC blocking.
http://sim.okawa-denshi.jp/en/CRtool.php
If you enable the step response output you will see the behavior I mentioned as well as the duration of the spike. The highpass corner frequency is also shown.
For example, for an 8 Ohm load 470uF Fc=42Hz, and for C=10000uF (10mF) Fc=2Hz ish.
Using a very large electrolytic cap (or pair, back-to-back) has the advantage that the cap is essentially a short circuit for the audio band and does not contribute significantly to distortion, etc. This is not true for lower values like 470uF because for low bass there will be some voltage drop across the cap. You can find a good review of this behavior in "The Design of Active Crossovers" by D. Self that also cites some old Electronics World articles. Self is writing about DC blocking in active circuits, but the concept is the same for passive, speaker-level DC blocking.
Last edited: