How to select input capacitor value and type for audio amplifier projects
Вставка
- Опубліковано 6 вер 2024
- Select the proper capacitance value for your amplifier project so frequency response is not adversely affected.
Link to the Okawa Electric Design site:
sim.okawa-densh...
Since you didn't say it(at least I didn't hear you say it), the reason you want input coupling instead of just hooking it right up is, you want to make sure very little DC bias is on the signal. Not only will that cause transistors to overheat, it could damage speakers.
Hey, John! Love all your posts, but this one in particular is especially exciting, as I have NEVER seen another video address this topic & you made it all so simple & clear! Well done, Sir!!!
John, first I wanted to say great video on this but I wanted to mention one issue I have with your statement of not using any ceramics for coupling purposes in audio. Not all ceramics are the same and they can not be grouped together. I would agree that type 2 dielectric not be used. However type 1 dielectric if available in the values that you need can be phenomenal and just as good as polypropylene caps. There dissipation factor can be much less than 0.1 with calcium zirconate(C0G, NP0, U2J) and you can sometimes get the value that you need with them, otherwise can stack a few of them. It is only the ones made with barium titanite that are no good for this purpose(X7r, X5r, Y5v).
I'm just finding my feet in Audio and you 'style' of teaching is superb...cheers.
I'm glad you're dispelling some audio myths. Good stuff as usual. 👍👍
Good one John. Wanted to know more about input caps. My design specifies a 10uf BP electrolytic, so I'm swapping that for a 6.8uf film cap. Now thanks to you I know what I can expect to see, or hear with this change. I rather like paper in oil as coupling caps but I don't have anywhere near the correct value. I'm in no position to second guess the designer, but if I want fiddle with values, that's up to me I suppose...
at 2:25...
70.7% has nothing to do with pi.
it is 100% * 1/2 sqrt(2)
The point where the signal energy is halved.
...and besides, I've seen many equations but not one had pies in it. I would have noticed, because I like pies.
Really really helpful video! Thank you. When it comes to choosing a power decoupling capacitor, I assume there are no special rules for audio chips. Any ol' guide for choosnig decoupling capacitors for ICs will be helpful, right?
The main issue with electrolytics for input coupling capacitors for me is their leakage current in the case of operating the amplifier input at very high input impedances (> 100k) and very close to REAL mass potential. it could happen in this case that the polarity of the voltage across the capacitor is opposite to the electrolytics polarization. Especially in the case of higher input voltages (> 1V) the capacitor will degrade with time und it will get more and more leakage current. The leakage current itself will compromize the balance of the input differential amplifier and it will cause dc offset at its output. So, especially for dc coupled output amplifiers it is not convenient to work with electrolytics as input coupling capacitors.
Hey John. Would also love to see something like this video on negative feedback loops on the output of chip op-amps if you get a chance or are so inclined. Thanks again for the videos.
Thanks for this video. Great info and a few things I have been curious about as I read more about chip amps and circuits.
Great video. Thank you for the explanation and demonstration. It would have been great to see other value capacitors in the circuit to see how it affects the audio signal. Maybe you could do an addendum.
If I were to use a .1uf cap in the video example, then the pole is moved to 72Hz (too high for a "hifi" amp). I could have demonstrated that but the video was getting long for the subject as it is.
I didn't think the video was too long. In fact i was just getting into it. But thank you for taking the time to explain these things too us. It's just many of us don't have a sillyscope so cant do these experiments ourselves.
Jesus said, if you do not have an oscilloscope sell your cloak and buy one. Well, He would have, if He knew what an oscilloscope was.
If you want to do your own calculations, then a 1μF capacitor will have an impedance of 8j Kohms at 20Hz.
The voltage divider given by ( 22 / 8j + 22 ) works out as 0.88 - 0.32j. That's a complex divider effectively with 88% of the signal in-phase with the input and 32% at 90° out-of-phase with the input.
That is √( 0.88² + 0.32² ) = 0.94 of the input voltage at a phase shift of arctan(0.32 / 0.88) = 20°.
Of course, 94% of the input signal is not discernible from 100% on a logarithmic graph.
Beside the leakage current when reverse biased, the other reason not to use aluminium electrolytics in the signal path is that they are commonly specified as ±20% tolerance - and I've seen -50%/+100% on larger ones! The capacitance itself is rarely stable and drifts with temperature and over time, so it becomes pointless trying to design a repeatable cut-off with them.
Of course, you can simply over-specify the value "to make sure", but then you have watch out for the capacitor charging at switch on if any DC bias actually exists. That can be quite a thump to deliver to loudspeakers unless you take precautions. In other words, setting the cut-off too low has its drawbacks as well.
Very nice presentation of input capacitors.
Cool, I have the same calculator. In the UK, is called a Casio fx-451
Hi John, when I hook up an electrolytic capacitor backwards to my bench power supply and shove 32 volts at 5A through it, is it normal to feel enjoyment when it explodes?
Tommy, For maximum enjoyment, put on a pair of safety glasses and place your face about 8 inches above the electrolitic before you apply the power.
Tantalum caps can be very exciting when they blow up. They shatter and generate plenty of smoke. Older LEDs have been interesting too with explosive shattering
Have noticed on some amps board they critic them to have brum noises from the transformer. That is boards where the rectifiers are ON the boards.
Those boards that have that full solution where you just connect the transformer directly to it.
You think that will solve the noise if that noise if that rectifiers are taken off the board to a separate board?
So the AC power does not affect the amp board
Great video, as always.
Loved this one. Very nicely explained, well done!
Hi John, Interesting video thanks! I'm over here in the UK tinkering with audio filters, in particular Baxandall tone control followed by some gain circuitry. Are Mylar/Polyester caps the best for quality in this use as well? Thanks in advance.
You should make more videos on troubleshooting bad amps/and/or analyzing schematic circuits :)
Alwage nice information
Hi John
Thank you for that brilliant vid, tried the site, put some values in
it works a treat !!
Please carry on with the great work
Thanks again
This clarified one of my questions. Now if I could just completely understand impedance. I know what it does and how it can store current and release it however why things need matching and how to calculate it. Nope.
impedance is simple. It's the resistive equivalent value of a source or load.
Reactance is individual resistive equivalents of capacitance or inductance whose value is dependent upon frequency.
If capacitive and inductive reactance are equal they become a resistive impedance.
Impedance is really a broader term.
Impedance itself doesn't store energy, but reactance does (ie inductors and capacitors store energy). Resistors don't store energy either but also have an impedance. See what I mean in that it's a broader term?
Great content !!
Some day in the future maybe you could revisit the Input Resistance topic?
I hear the term used all the time and don't quite get it.
I have a Dual CV 6065 RC which is class-A 85wt/ch amplifier there is 6800uf 63wt (2nos) and 1000uf 35wt (2nos). i have to replace those caps can i increase the capacitance and power of capacitors ? i have very little knowledge of audio electronics! thank you!
Could you tell what you suggest to use as decoupling transformer for audio signal (line level) - how manny turns of wire, are ferite ring transformer cores from burnt CFL lights/atx PSUs ok, will it act as too much bandpass filter...
For when you expect MASSIVE DC and ocassional 50Hz AC offset (just for seconds - during genset restart usually) occuring - like for connecting power amplifiers, mixer board and computers, cd/vinyl players on multiple generators (or ups or inverter+battery+charger for computer/player if you know the big diesel generator goes out every time they use too much lightshow for over 20 seconds ;-) at events...
should they be non polar ? the input cap... nichicon ues for example ?
Interesting video. Can you tell more about the polarity of an electrolytic in the signal/audio path?
Thnx, for the input your reply is clear to me. For the output, do you mean that the positive side of the cap is to be pointed to the amp and the negative to the speakers? And, what is the reasoning behind this rule? The signal is AC and either side of the caps may be positive with regard to the other?
If the particular circuit has some DC bias, you'd want to orient the cap + side toward the side with the more positive voltage. If the signal is purely AC, then you can orient it any way you like. The polarized electrolytic capacitor will not be harmed by the small signal AC across it nor will it harm the sound. In the passband, the voltage differential across the plates will be very small (few millivolts). If the cap value is 1uf or less, I'd use the film types as there would be one less electrolytic in the circuit to dry out.
@@JohnAudioTech I've seen X-overs in Aussie vintage Philips consoles and speaker systems using 10uF polar caps, very sleazy. I've changed them to non electro types and use 2.2 for 8 ohm tweeter, the 10 was too midrangey
Hi John, this is great for audio amps but what about high frequency amps in the megahertz range? How would I go about choosing my caps for that? I know that I have to use ceramic caps
yeah it dosent effect the sound but it distorts low frequency square waves. i preferr flat tops rather than lopsided tops (and bottoms)
i must say though it makes the startup time of the amp longer
John, I enjoyed your video and would like your thoughts on something i have noticed a number of times. Even though the pole is usually around 6 Hz or so, just as you mentioned in you video, there are cases where low frequencies are intentionally are rolled off to quite an extreme extent. A good example is the Brook 12A amplifier. It is an old design that uses 2A3 in the output and I built one for a lady a few years back and made it true to the schematic. Scanning it with the HP 8903 did not look good at low frequencies. I looked for a clear schematic to attach but could not find a good one but attached what I could find below.
elektrotanya.com/brook_12a_sch.pdf/download.html
Rolling off LF in these older amplifiers is not that uncommon and some, like modifications to some of the Dynaco amplifiers are pretty common right at the front end. I have always wondered why the designer/builder wants to limit LF response. I have wondered if it has something to do with output transformer saturation at LF or what... I don't know. Know that I am not talking about the little RC circuits we sometimes see off the grids of a tube to snub AM radio stations or a 70 KHz instability not uncommon in the older Williamson designs. You see these RC circuits in Heathkit and UTC designs.
Back to the point - just wondered if you had any thoughts on the reason LF response is severely limited in some amplifiers and I am using the Brook 12A as an example. In the Brooks 12A schematic I am talking about the 24K and pair of capacitors coupling the first to second stage. I have a LTSpice model of the Brook 12A if you do Spice and will be happy to send it to you. Thanks...
tnx for sharing sir John..
As always, very interesting, thanks for sharing.
Greatings from Denmark :-)
hey you cant say not to use ceramics for input coupling/filtering because there are class 1 and class 2 dielectrics and C0G/NP0, U2J are extremely stable and actually show distortion measurements just as low as polyproylene according to cyril batemans measurements tests. Yes class 2 dielectric ceramics are no good for this but they are great for decoupling........Just please dont throw all ceramics into the same class.
Nonsense. Ceramics have a grainy sound.
@@pangolinamplifier422 bullshit
What about tantalum drop capacitors? I tend to go for those over electrolytic capacitors in audio circuits where it's not practical to use MKT or other film caps
hi, what is the difference between those epoxy polyester caps (you have) and those box type film caps ? and which is better for what ?
This should explain much of it:
www.wimausa.com/DE/dipped.htm
interesting, thanks for the link. why it did not show up in google though.
I have modded an old radio with AUX input but i am getting around 2 volts there And i dont want them in my phone. Will it work this way ?
Is R in the same as Z in I'm building a class A common emitter amplifier and trying to figure this out
I have made a wien bridge oscillator with an op amp. The op amp output is a 2khz sine wave that is the 1st stage. The second stage consist of 2 BJT's first C945 and second 882 both NPN a 120 ohm resistor on the collector of the 945, the emitter connected to the Base of the 882, a car alarm siren connected to the collector of the 882 and the emitter to ground or 0V. What should be the coupling capacitor resitor between those 2 stages? I am using a 220microfarad electrolytic capacitor in series with a 1kohm resitor to connect the output of the Op-amp to the Base of the C945 but the siren sounds only while the cap is discharging and the intensity of the sound gets weaker until there is no sound at all. What am I doing wrong? Shout I connect a 22kohm resitor connected to ground between the cap and the 1kohm resitor to maintain a stable charge and discharge of the cap so the siren will have a stable 2khz sound?
Are electrolytic not the worst kind in the audio path?
They are fine in the passband but can have higher distortion around the cutoff frequency where voltage change across the cap will be greater.
When going electrolytic do you need to use a bipolar capacitor?
When there is no DC bias and when dealing with small signals, you can use a standard polar cap.
Thanks, but you mean non-bipolar?
Yes, corrected.
JohnAudioTech I have an digital v/u meter specially made for car audio. When connected the v/u meter will not drop backdown to the -00 mark. I have been informed a non polorized film cap of 10 uf 200V was good to block DC offset emoted by my JL Amplifier. What say you?
thank you, Very helpful video.
all you had to do was mention pole frequency to get my attention, I read about it in Doug Selfs book, all research leads me to painful math
Are they non polar ? The electrolytic caps.
I have a question about audiophool vs audiophile.
If you are to different live concerts and are listening, different pianists who flowles plays the same piece of Mozart will properly still sound different, some would call it "with more soul".
Can that be messered or is that due to humans "believe" or ?
Different players playing on the same instrument at the same venue, for example, will have very slightly different playing style, timing and attack on the notes and so on, which can sound different. If recorded on the same equipment can show the difference.
Yes, you are right and that could easily be measured in time, voltage and so on! Thanks:-)
820nF is closer approach than 1uF.
And what about tantalum capacitor? De facto, tantalum capacitor much better than regular electrolytic capacitor. But I don't know about its noise characteristics.
Cheers from Indonesia
Tantualm caps are very very non linear and horrible for audio think have u ever seen a tantalum cap ever in a signal path the answer is no tantualm are good for bypass and especially timing but absolutely no signal
Maybe if your look for a certain non linear respones
Excellent!!
.707 = 1/√2
nothing to do with pi
In radians sin pi/4 or cos pi/4 = 1/√2. Do you think this is coincidental?
Capacitor = distortion component, and transformer = even greater distortion component. Only perfect reproduction :)
John, why are you using polyesters? You are wrong, they have significant distortion, especially compared to Polyprop.
The only reason to use the PE type is if you want some distortion in the signal path, like in a guitar amp. Never ever use electrolytics in the audio path, never.
I'm aware that electrolytic, ceramic and to a much less extent, PET caps have nonlinearities. But it is at the cross over point. The distortion is low in the passband as the voltage change across the cap is very small. Polypropylene caps are ideal but they get large and expensive. Very easy to design a low distortion amplifier using PET caps in the signal path.
In your calculation for input resistor you entered 2200 and called it 22k. 22k would be 22000.
No, he did it right.
What the hell is a frequeNcy John???!! Last I checked its frequency!
Когато входния и не само кондензатори са изградени от няколко се получават добри резултати.
Audiophoolz 😂 🤣 Very true! Those fools who still use geranium diode, insist only germanium diode reign supreme, and willing to pay huge amount of money to buy obsolete germanium diode.
There are fools pretending to be real experts in every field.