One more thing that needs to be mentioned: the differences between caps are very small. The differences between film caps are exceedingly small. It takes instrumentation beyond a DVM and signal generator to detect these differences- that means impedance bridges and instrumentation amps. You can't measure an amoeba with a yardstick!
The other thing to keep in mind is application; as couplers, capacitors have very little AC voltage across them, so their effects are vanishingly small. In higher current situations, the differences are a bit clearer. DC parameters are important in timing circuits, but not so much in AC circuits. And so on. Don't fall into the common trap of trying to draw conclusions about use in circuit A from some tests under conditions B which aren't relevant to your use.
I echo DF96's comment. You're trying to learn the basics and putting in the effort, which is far more than most people do.
The other thing to keep in mind is application; as couplers, capacitors have very little AC voltage across them, so their effects are vanishingly small. In higher current situations, the differences are a bit clearer. DC parameters are important in timing circuits, but not so much in AC circuits. And so on. Don't fall into the common trap of trying to draw conclusions about use in circuit A from some tests under conditions B which aren't relevant to your use.
I echo DF96's comment. You're trying to learn the basics and putting in the effort, which is far more than most people do.
Forgive me I already haven't read most of Bateman articles yet. I will surely do this later.
I strongly believe in sonic differences of different capacitors, but I've never heard any
differences after 'burn in'. Nevertheless I'd guess lytics may show some slight differences
before/after.
What about getting three different caps of the same value and redoing the measuring?
What about varying the procedure so that a difference is to be seen? What does this
change need to be? Besides voodoo and wishful thinking there are some differences on
these caps depending on construction. At least they have a different ESR. But will they
show a different behaviour in HF waveform?
I strongly believe in sonic differences of different capacitors, but I've never heard any
differences after 'burn in'. Nevertheless I'd guess lytics may show some slight differences
before/after.
What about getting three different caps of the same value and redoing the measuring?
What about varying the procedure so that a difference is to be seen? What does this
change need to be? Besides voodoo and wishful thinking there are some differences on
these caps depending on construction. At least they have a different ESR. But will they
show a different behaviour in HF waveform?
+1 on trying to learn. What Sy said is important. Measuring the differences between various film caps is difficult. That's why I wanted you to start out with some common electrolytics, which can really be any moderate values. I wouldn't worry too much about specific circuits until you understand the fundamentals concerning components. After that the correct interpretation of what's going on with some specific audio circuit will be much clearer.
Thanks DF96...I do want to understand...audiophool that I may be - the theory is important! And thx for your encouragement.
After doing the calcs for all the caps, I came to the stunning conclusion the only way I was going to get more voltage across the caps, without going over about 20mA of current through my precious Green Pre, is by changing the frequency of my input signal (I say stunning tongue in cheek; I know this is old hat for some!). But what's kind of exciting for me is I am now coming at this in an experiential way...I'll start to really understand capacitors at a basic level. Although again this is obvious to many, it's new to me, so I'll post the equations and then discuss how I came to that conclusion, as well as my experiments in that direction.
I'll post more later; thanks again. I'll also post more sillyscope readings.
Thanks,
Mike
After doing the calcs for all the caps, I came to the stunning conclusion the only way I was going to get more voltage across the caps, without going over about 20mA of current through my precious Green Pre, is by changing the frequency of my input signal (I say stunning tongue in cheek; I know this is old hat for some!). But what's kind of exciting for me is I am now coming at this in an experiential way...I'll start to really understand capacitors at a basic level. Although again this is obvious to many, it's new to me, so I'll post the equations and then discuss how I came to that conclusion, as well as my experiments in that direction.
I'll post more later; thanks again. I'll also post more sillyscope readings.
Thanks,
Mike
PS sorry I only just noticed the page had advanced and the nice comments from Sy, Conrad and ticktock..thanks I appreciate it. I hear you about the difficulty in measuring diffs in film caps; makes sense.And I agree with you Conrad that understanding the components first will make the circuits that much easier to grasp.
I'd like to be a contributing member. Muchos gracious.
I'd like to be a contributing member. Muchos gracious.
Wow I just lost a huge post! It gave me an error and told me to hit the back button, bam, a LOT of data lost! frack!
OK...to summarize...I spent the latter part of my evening measuring caps. That's what happens when you don't have a gf on a friday night...sigh...
I changed the frequency of the cap values because I wanted to maximize current and voltage across each of the caps. I'll just give the results of the math; without proper formatting the math looks like heck. I posted the formulas above anyways.
.01uF cap, 10kHz square wave:
Xc = 1592R
Z = 1701R
I = 5.9mA <--5.6mA measured
Vr = Ir = 3.53v <--3.7v measured
Vc = XcI = 9.6v <--9v measured
.1uF cap, 2k square wave
Xc = 796R
Z = 997R
I = 10mA <--10.1mA measured
Vr = 6v <--6.1v measured
Vc = 8v <--7v measured (hey lost a volt!)
1uF cap, 200Hz square wave:
Xc = 796R (same as above - increased cap value by 10; reduced freq by 10 for similar voltage/current values)
Z = 997R
I = 10mA <--10.3mA measured
Vr = 6v <--6.2v measured
Vc = 8v <--6.85v measured (lost 1.15v this time!)
OK so I picked the new frequencies to keep the current and voltages in the same ballpark; changing the resistor values doesn't give as great a change as changing the frequencies. I needed a decent value for Xc, and changing the resistor to say, 100R gave me only 15.9R - not enough V across the cap.
I'll post my sillyscope readings next.
BTW, I was tired, and accidentally connected my ammeter across the output of my mic pre for a few seconds, without a load...couldn't figure out why I was getting 33mA for the cap measurements...wow! Hopefully didn't fry the PSU, but it seems to be fine. It's a Green Pre. Anyway, now I know the unloaded output of a single channel Conrad!
OK...to summarize...I spent the latter part of my evening measuring caps. That's what happens when you don't have a gf on a friday night...sigh...
I changed the frequency of the cap values because I wanted to maximize current and voltage across each of the caps. I'll just give the results of the math; without proper formatting the math looks like heck. I posted the formulas above anyways.
.01uF cap, 10kHz square wave:
Xc = 1592R
Z = 1701R
I = 5.9mA <--5.6mA measured
Vr = Ir = 3.53v <--3.7v measured
Vc = XcI = 9.6v <--9v measured
.1uF cap, 2k square wave
Xc = 796R
Z = 997R
I = 10mA <--10.1mA measured
Vr = 6v <--6.1v measured
Vc = 8v <--7v measured (hey lost a volt!)
1uF cap, 200Hz square wave:
Xc = 796R (same as above - increased cap value by 10; reduced freq by 10 for similar voltage/current values)
Z = 997R
I = 10mA <--10.3mA measured
Vr = 6v <--6.2v measured
Vc = 8v <--6.85v measured (lost 1.15v this time!)
OK so I picked the new frequencies to keep the current and voltages in the same ballpark; changing the resistor values doesn't give as great a change as changing the frequencies. I needed a decent value for Xc, and changing the resistor to say, 100R gave me only 15.9R - not enough V across the cap.
I'll post my sillyscope readings next.
BTW, I was tired, and accidentally connected my ammeter across the output of my mic pre for a few seconds, without a load...couldn't figure out why I was getting 33mA for the cap measurements...wow! Hopefully didn't fry the PSU, but it seems to be fine. It's a Green Pre. Anyway, now I know the unloaded output of a single channel Conrad!
OK now for the revised sillyscope readings.
1uF cap, 200Hz signal:
Now, why is this so much closer to classical expectations for voltage across a cap? Well, as I figure it, with my original input of 10kHz square wave and 100R resistor, the cap time period as given by T = C x R was .000001 x 100 = .0001s. This means the cap had 10,000 periods per second, and the input signal was 10,000 periods per second. In other words, at best the cap had 1 time period per cycle, not enough to fully charge it, so the wave form looked triangular.
With the new values: T = C x R = .000001F x 600R = .0006s, or 1667 periods per seconds. At 200Hz input signal, this gave roughly 8 cap time periods per input signal cycle - more than enough for the cap to fully charge/discharge!
Am I waaay off in this analysis?
OK, the .1uF cap:
With a T = CR of .00006s, or 16,667 time periods per second, and an input signal of 2kHz, this cap had roughly 8 time periods per input cycle; more than enough to fully charge and thus the nice looking wave form.
The .001uF cap:
Yikes! What happened! I was using my DAW to output the square wave, and had re-recorded it. I checked the input signal directly, and this is what I got:
Clearly I need to re-record the square wave again; I believe this accounts for the crappy looking .001uF cap waveform.
The analysis, however, is similar:
T = C x R = .01uF x 600R = .000006s
This gives 166,667 time periods per second for the cap, and with an input signal of 10kHz, this is roughly 16 time periods for the cap per input signal cycle. More than enough that this waveform should look nice! I'm too tired to redo it right now; I'll redo it and post this later.
Thank-you gentlemen and ladies for indulging my foolishness and encouraging me. I'm still an audiophool, but more dangerous, because "I think I know something!"...Hmmm.
1uF cap, 200Hz signal:
Now, why is this so much closer to classical expectations for voltage across a cap? Well, as I figure it, with my original input of 10kHz square wave and 100R resistor, the cap time period as given by T = C x R was .000001 x 100 = .0001s. This means the cap had 10,000 periods per second, and the input signal was 10,000 periods per second. In other words, at best the cap had 1 time period per cycle, not enough to fully charge it, so the wave form looked triangular.
With the new values: T = C x R = .000001F x 600R = .0006s, or 1667 periods per seconds. At 200Hz input signal, this gave roughly 8 cap time periods per input signal cycle - more than enough for the cap to fully charge/discharge!
Am I waaay off in this analysis?
OK, the .1uF cap:
With a T = CR of .00006s, or 16,667 time periods per second, and an input signal of 2kHz, this cap had roughly 8 time periods per input cycle; more than enough to fully charge and thus the nice looking wave form.
The .001uF cap:
Yikes! What happened! I was using my DAW to output the square wave, and had re-recorded it. I checked the input signal directly, and this is what I got:
Clearly I need to re-record the square wave again; I believe this accounts for the crappy looking .001uF cap waveform.
The analysis, however, is similar:
T = C x R = .01uF x 600R = .000006s
This gives 166,667 time periods per second for the cap, and with an input signal of 10kHz, this is roughly 16 time periods for the cap per input signal cycle. More than enough that this waveform should look nice! I'm too tired to redo it right now; I'll redo it and post this later.
Thank-you gentlemen and ladies for indulging my foolishness and encouraging me. I'm still an audiophool, but more dangerous, because "I think I know something!"...Hmmm.
What do you mean by "clean"? What you have there is pretty much what you'd expect.
Are you familiar with Fourier series? Also, do you have a soundcard for you computer? If so, you can have some fun comparing the change in spectra from one signal to the other and see how close your predictions can be. As a side note, I'd recommend putting together something better for a square wave generator- it's incredibly easy to do, just taking an inverter, resistor, and capacitor or (if you want to be slightly fancier) a 555 chip and a few auxiliary parts.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.