AudioNervosa
Specialists => Audiologists => Topic started by: S Clark on September 16, 2019, 04:58:55 PM
-
Caps in a crossover are something I'm familiar and comfortable with... but amps??? Filter caps, coupling caps, power caps... what exactly is the difference, where are they in the circuit, and where does quality come into play?
There are several of you guys that are amp guys. A little help please.
-
This is a subject that is simple in concept, but a great big ball of fuzz in reality.
The nuts and bolts are these:
Filter caps/power supply caps provide ripple control after the rectification stage in an application. This can be initial power filtration after the power transformer and rectifier or downstream for local ripple and noise rejection in whatever piece of gear. Coupling caps do exactly tha. They couple one stage in a piece of gear to another. This is done for several reasons. Most common is to block any DC component from effecting the music signal (an AC signal) or to provide isolation between stages that operate at different electrical potentials.
That is the very bare bones reply. I’m sure Jeremy (Folsom) and Steve will have a lot more to add.
It’s late and I didn’t want you to wake up in the morning with nothing to read.
Next I’ll dive into -
DA - dielectric absorption
DF - dissipation Factor
ESR - equivalent series resistance (a bit of a misnomer)
Materials
Lots more to come from people here!
-
Thanks for a good start, Dave. I think I understood, mostly, some, maybe. So the initial power (in the US, 120+V) comes into the power transformer, gets bumped up to ???V, contains miscellaneous crap, and goes through a filter/power cap. So are these usually the bigger value caps that so commonly dominate the landscape on the circuit board?
-
All I know is when I upgraded to Duelund CAST output coupling caps in my Consonance Mono blocks my world changed. Then power caps oh my goodness.
charles
-
... when I upgraded to Duelund CAST output coupling caps in my Consonance Mono blocks my world changed.
charles
I bet it did. Your billfold got a lot lighter! :rofl:
But this is one reason for the initial question. I'm trying to have a better understanding of what the caps are doing, so that I can upgrade more intellegently.
-
Yes, the power supply filter caps are usually the large can ones in the chassis. Usually physically bigger in SS vs tube gear.
Dave gave a great intro and cap materials for construction vary greatly from film to electrolytic styles. Large values 25uF+ are usually electrolytics. The ESR, DF and DA are very different for different styles.
There was a very good set of 3 articles in Audio magazine in Feb,Mar, Apr 1980 done by Jung and Marsh that describes caps well and what the types are for. Been a long time, but a great intro to understanding.
Cover of the first article in Audio:
https://americanradiohistory.com/Archive-Audio/80s/Audio-1980-02.pdf
-
To call PSU caps "filters" is kind of silly. The basic meaning of "filter" is that it "filters" 60hz out. It has nothing to do with all the other harmonic crap that they "filter" to some degree or another.
(https://diyaudioheaven.files.wordpress.com/2015/01/rectbr2.png)
The rectifiers give you the newish type waveform, and the black lines from peak to almost peak are the capacitors discharging (the device uses that power being discharged). The capacitors charge to essentially the peak voltage output of the rectifiers (which is whatever your transformers outputs). Between the peaks the capacitors then release their charge, as they're trying to maintain the voltage that they charged to. Because you aren't seeing voltage drops at a given cycle, you're effectively seeing DC - hence "60hz" has been filtered out.
Today I feel no need to call them filter capacitors. For me they are power capacitors because they maintain power. When they discharge they let out voltage and current. The bigger they are, the slower their voltage drops (the slower they discharge relative to the total amount of energy they hold) and the more current they can provide (the more used the faster the voltage drops). One giant one is usually not possible. The more capacitors you have to get the desired amount so your voltage doesn't plummet, but multiplying capacitors also increases some high frequencies bypassing being attenuated by the capacitors.
Proximity matters, too, because the ESR/ESL changes. So sometimes you have parts that make noise in a circuit and you have to put capacitors near them to help filter that. The farther away the more ESL (inductance) and ESR (resistance) so the less effective they will be. You can buy capacitors that are low ESR, effective around the audio band for example, or low ESL that is more effective in the RF range. Proximity is more important for inductance, as the resistance/inductance is still typically low in the lower frequencies than RF.
When it comes to quality there's subjective and objective differences. Technically many attributes Dave mentioned are desirable but sometimes have consequences. For example ringing can happen when a cap is very stiff in measurements (low everything but capacity), or rather it may not be any good to dampen existing ringing. Capacitors are subject to vibrations, and voltages above 12v may excite them to vibrate - the measured result isn't typically easy to see in audio gear but in other equipment sometimes is very obvious and some manufacturers sell low vibration caps.
Choosing the right capacitor is related to what you need it to do. For example before an active filter like in my 7297, I use low impedance (ESL) capacitors because then the filter is more successful in eliminating HF garbage. But I don't necessarily like the way they sound when directly feeding the amplifier circuit . They aren't doing any frequency passing duties so if they absorb some (DA) it really isn't critical. The problem with a lot of discussions on capacitors is people will say something like "DA is what matters" but that's because they expect electrolytics (which have more DA than film) in the signal path. Basically any wise audio manufacturer has mostly eliminated lytics in the small level signal path (stuff your IC's carry). And we don't have to worry about how much total capacitance we need these days, it's cheap and small till you start getting into 100w classA type stuff.
I expect everyone to have slightly different preferences on what their gear sounds like with different capacitors, all of good quality. Unless we're talking about Chinese off-brands (not all Chinese) lytic capacitors, most are of good quality. But some manufacturers have their own secrets for making audio specific capacitors that measure like, well, pretty much all capacitors. When it comes to film they're almost all really high quality and exceed electrolytics in all ways except capacity to size ratio. But sometimes film capacitor sizes mixed with lytics don't do so well and cause ringing - luckily basically no one does that anymore.
When it comes to carrying signal (coupling) you can measure distortion of a film capacitor but it's so low that there isn't really any indicator for why they sound so different. It's just kind of up to you to decide. But I can say my favorite measures the lowest... but it may not sound right in some stereos that I haven't totally gone through and modified everything :rofl: . Certain capacitors measure higher 3rd harmonics, but what's funny is it still shouldn't be audible. However, despite that these measurements are barely exist... I can tell you polyster (measures highest in 3rd order harmonics of film caps) masks a lot of information if it's in my phono preamp. They can however sound exciting sometimes, too. If your stereo is sounding kinda dull the extra "too low to hear" distortion can bump up some excitement. And if your stereo isn't really revealing, you may not notice a non-loss of information. Different people listen for different things... And high resolution is a double edged sword because if it doesn't sound just right it's distracting so getting it isn't always what you want unless everything is kinda perfect.
I know a lot of audiophiles are obsessed with higher quality capacitors for speakers, but I'd say in general a lot of the cheaper capacitors that aren't repackaged lytics, will measure about the same as most expensive ones. That or some expensive ones might measure bad. You're getting *different* sound not necessarily better. For example I think the Jantzen's sound pretty good in some stereos because they make a lot of information readily available. But in my stereo they sound wrong. I think metalized have their own character that tends to do what I described, and it's useful in many stereos; but if you put it in one that has much high levels of resolution then it sounds weird. It's really all personal preference, and you have to pay more for different sound since most manufacturers of general electronic capacitors make pretty much the same capacitors for general use. So I'll say when you open a speaker up and look at a crossover and see some inexpensive capacitors I don't immediately think it's a bad thing, I think either they wanted to save money by not marketing expensive parts that may not be an improvement or it was the best voiced capacitor to the designer. Now what gear they used to design the speaker with will probably reflect that choice. Here's an example, Magico uses a lot of Mundorfs cheaper capacitors despite the fact that they don't have budget constraints - they are the sound they want.
It's hard to have right and wrong answers with capacitors when they aren't too low in capacitance for their use, when you're talking about something that is subjectively appreciated. Even if they can make better measurements, the reality is unless it stops some kind of error in digital equipment or run-away oscillations in analog gear, and appear to be working then, right & wrong is based on what you want to hear.
-
Interesting and differing points of view as to description and beyond my ability to get into such specifics :roll:
I think over the years since I got into audio...late 80s. probably the most expensive caps I’ve had are the Black Gates which were an upgrade to my custom Audience preamp. That’s my guess and I’m also curious what the most expensive caps are nowadays. Dueland comes to mind.
Nick
-
Thanks for that write up. Very informative, and only one question. You mention DA.... dielectric absorption I'm guessing.
-
Ok, let's see if we can go through a simple tube amp schematic. Cleverly, I typed in "Simple tube amp schematic" into Google and picked something that might work. I'm seeing 3 caps. I assume the 3300 uF is the power cap (filter cap). What about the other two? If there's a better schematic, post it and I'll delete this one. Well, crap. It's bigger than I thought. Can we shrink it?
(https://static.hackaday.io/images/4949751414932586198.jpg)
-
Thanks for that write up. Very informative, and only one question. You mention DA.... dielectric absorption I'm guessing.
Correct.
-
Ok, let's see if we can go through a simple tube amp schematic. Cleverly, I typed in "Simple tube amp schematic" into Google and picked something that might work. I'm seeing 3 caps. I assume the 3300 uF is the power cap (filter cap). What about the other two? If there's a better schematic, post it and I'll delete this one. Well, crap. It's bigger than I thought. Can we shrink it?
You're correct about C3. Look at what C1 is connected to, the center pin of the RCA. What's on the center pin?
C2 I'm not completely sure on because I don't know that tube. Maybe I'll look it up in a minute if someone else doesn't chime in.
-
You're correct about C3. Look at what C1 is connected to, the center pin of the RCA. What's on the center pin?
C2 I'm not completely sure on because I don't know that tube. Maybe I'll look it up in a minute if someone else doesn't chime in.
Ok, C1 is connected to the incoming + signal... but what is there to clean up at that point? I guess V2 is a rectifier tube, making V1 the power (amplifying) tube.
The ECL82 (V1) is some kind of triode
(https://www.radiomuseum.org/images/tubeenvdiag_klein/ecl82.png)
-
C1 doesn't clean anything. It's the only path for signal. What it can do is block DC from other gear. The size matters because if it's too small then bass will be reduced. The combination of the impedance of the amp and the size of the capacitor determine what we call the corner frequency. For calculations and a more comprehensive explanation this site is good (http://www.sengpielaudio.com/calculator-RCpad.htm). For a power amplifier you want a -3db of 4hz or lower (some say 7hz, but that's proven to be too little IME).
Since the voltage is only a little bit (often under 2v at max) the capacitor voltage really doesn't matter, particularly with film capacitors because they're usually no less than 60v, if not hundreds.
C3 "cleans" when the frequencies on say the V+ rail side are attracting electrons, by getting them from ground before the amplifier's circuit (bypassing if you will). It prevents these frequencies being pulled through the amplifier.
Because the amplifier is in the path to ground from the signal (C1), the frequencies pulling electrons from the source pull them through the amplifier, hence it isn't "cleaning" them.
-
So is there usually incoming DC from the signal input that C1 needs to remove?
-
Not necessarily. It's common for digital devices to put out DC though.
If you're 100% sure you have no DC coming from anything then you can remove input capacitors on an amplifier. But sometimes the sound may not be what you want as capacitors seem to have an effect that stops fatiguing sound, too.
C1 doesn't remove DC, it simply won't pass it.
-
Scott,
The C1 coupling cap as stated removes any DC offset from the source device, but also stops the DC bias from the tube input from going back to the source. The grid leakage current on the input resistance will cause a DC voltage as well. Small, but that may cause some DC coupled source output stages problems.
If the source is AC coupled (has an output capacitor), then you can bypass the input C1 by shorting it with a jumper. Easily reversed if needed.
-
So is there usually incoming DC from the signal input that C1 needs to remove?
Scott,
I would suggest the easiest method of understanding is to obtain a copy of the Radio Amateurs Handook,
say 1955, or 64, if into ss, then 69 issue, or another source of basic electronics such as a book from Parts Express. They are relatively inexpensive and go into the audio basics. I will use Scott's schematic amplifier in my below example.
(https://static.hackaday.io/images/4949751414932586198.jpg)
The second image shows, in red lines, for future reference.
C3 is the power supply filter, for both minimizing 120hz signal from rectification and, from the tube's side, see my second schematic and red lines, "seeing" the musical signal and "handling" the musical signal current from V1p to ground.
C2 is the coupling capacitor between the two tube stages. C2, along with R2, forms a high pass 1st order filter, whose slope is 6db/octave. C1/R2 should be sufficient to cover the audio band.
Hal did a wonderful job describing C1's functions. I would just like to add that C1, along with volume control total resistance, also forms a 6db/octave high pass filter.
One important point is that the -db of both 1st order filters add. As a simple example, suppose each filter has a
-0,1db point at 20hz. Together, the response is -0,2db at 20hz. When we add the -db points from the source, preamp, and speaker, the response is much more than -0,2db. We should consider the whole system's response and therefore capacitor's values.
cheers
steve
-
Ok, I've got a 1961 copy on the way.
-
Good point about biasing up the line. This is a mixed bag because if your preamp has output capacitors it's not a problem. But some things are DC coupled so it is a problem :rofl: . Hence why most put the capacitors in.
Electrolytics also can get biased and sound different than film in the position of input capacitor. It's really weird, but if they're being biased you can put a film inline with the lytic and it'll sound way better :shock:
-
Ok, I've got a 1961 copy on the way.
I hope you enjoy the book. One chapter addresses power supply introduction, different rectifier types, types of filters, capacitive input, inductor input, regulators, voltage regulation etc.
Another chapter addresses speech ampllifiers, transformer coupled stages, capacitor coupled, etc. High fidelity stages differ by using larger and higher quality coupling capacitors, wider bandwidth transformers for wider bandwidth. I think it is a good overall book and hope it helps you. Parts express might have a good book, but it costs $25.
Capacitor quality is addressed in previous posts, ESR, DA etc, the article "picking capacitors" by Walter Jung and Richard Marsh. I might add that materials used/quality are also important.
cheers
steve
-
The problem with all the types of articles on "picking capacitors" is that there isn't a useful conclusion, not really.
Experience with sound and why the sound is the way it is, is what really matters.
My favorite smaller capacitor doesn't sound right in everything. But with the right control over certain things it excels.
I wish it were as simple as always choosing the "best" but when you go down that road it get complicated because the closer you get, the more everything that isn't stands out.
-
The problem with all the types of articles on "picking capacitors" is that there isn't a useful conclusion, not really.
Experience with sound and why the sound is the way it is, is what really matters.
My favorite smaller capacitor doesn't sound right in everything. But with the right control over certain things it excels.
I wish it were as simple as always choosing the "best" but when you go down that road it get complicated because the closer you get, the more everything that isn't stands out.
:thumb:
-
Hi Folsom and all.
I am not an expert in audio capacitors,
however, I know a thing or two on photographic flash capacitors,
with capacities that varies from 100uF to 6400uf at 350 to 600 volts.
Just want to say.
I also have a basic course in electronics at Teccart Institute in Montreal,
I have to admit that was a very long time ago (1967)
when we study manly tubes and very little transistors.
Guy 13
-
The problem with all the types of articles on "picking capacitors" is that there isn't a useful conclusion, not really.
Experience with sound and why the sound is the way it is, is what really matters.
My favorite smaller capacitor doesn't sound right in everything. But with the right control over certain things it excels.
I wish it were as simple as always choosing the "best" but when you go down that road it get complicated because the closer you get, the more everything that isn't stands out.
Which is why designing high end audio equipment is at times as much of an art as it is a science.
Heck if everything was done "by the book" nobody would be listening to tube amps at all because they cannot match the distortion specs of their solid state counterparts.
The stuff they taught us in engineering school and that is in the books is just a start. A computer can be programed to blindly apply rules and produce a workable product. But the real engineering starts when the engineer is cut loose to step outside of the box and creatively solve problems using his/her intuition and gut instinct. And that is when you separate the true engineers from those that are simply CAD operators. It is also where you get innovation and advancement in the state of the art. Because if you program a computer to design things using yesterday's rules you will never do anything but replicate yesterday's design's.
-
The problem with all the types of articles on "picking capacitors" is that there isn't a useful conclusion, not really.
Experience with sound and why the sound is the way it is, is what really matters.
My favorite smaller capacitor doesn't sound right in everything. But with the right control over certain things it excels.
I wish it were as simple as always choosing the "best" but when you go down that road it get complicated because the closer you get, the more everything that isn't stands out.
Which is why designing high end audio equipment is at times as much of an art as it is a science.
Heck if everything was done "by the book" nobody would be listening to tube amps at all because they cannot match the distortion specs of their solid state counterparts.
The stuff they taught us in engineering school and that is in the books is just a start. A computer can be programed to blindly apply rules and produce a workable product. But the real engineering starts when the engineer is cut loose to step outside of the box and creatively solve problems using his/her intuition and gut instinct. And that is when you separate the true engineers from those that are simply CAD operators. It is also where you get innovation and advancement in the state of the art. Because if you program a computer to design things using yesterday's rules you will never do anything but replicate yesterday's design's.
You guys killed it! Good job.
There is no "one size fits all" device in audio. I like a particular op amp for use in some line driver applications. It is a good piece in almost any application when properly implemented and a GREAT op amp when used in 'some' applications. Again, it has to be treated properly and with respect to get "there".
Same goes for any other audio component. Pick the right one for the right application and you'll have a winner. You can't just willy-nilly take a bunch of devices that may excel in certain applications, string them all together and be guaranteed a winner.
-
I guess I take a different view point, to some extent on this string. I have seen so much confusion and mis information on the internet. How does confusion create order. I will address "synergy" further down.
My reply would be there are better capacitors than others, including accurate ones, but how is one suppose know what to believe? And a truly accurate component will reproduce all kinds of music accurately/naturally. If the recording is bad, well...
A very simple example of mis information. I have seen capacitor testing using a single piece of gear, one size capacitor (the wrong value ufd) and then the author rates the capacitors. Of course the most accurate capacitors are never rated well because the wrong value ufd is used, and the quality of the rest of the circuit is questionable. So the accurate/good capacitor becomes extinct while the inaccurate/bad capacitors flourish.
One cannot make a silk purse from a pig's ear.
Say one uses a 1.0uf capacitor with a 100k ohm load, often called the grid resistor. The reactance of 1.0uf at 20hz is ~8,000 ohms. The frequency response has just dropped off 7.5% at 20hz. At 80hz 1.87%, at 160hz,
~1%, at 320hz 0,5%, and at 640hz the frequency response has dropped off 0,25%. And that is just one high pass filter, one coupling stage. There are many more in an audio system.
Let's take a look at the opposite. If all the prolyprope film capacitors were designed correctly, for accuracy, all the designers would produce the same high quality products. However, different capacitors do affect the sound differently, due to DA, ESR, materials used, thickness of insulation, termination techniques, physical size (L vs D), shape etc. So obviously not all designers/manufacturers understand how to properly design. In fact, very few.
One cannot make a silk purse from a sow's ear.
Resistors the same story as capacitors, they vary the "sound" due to materials used, termination techniques, internal inductance due to physical size, shape etc, maybe even internal capacitance. Crappy parts result in poor sonic quality. Can't make a silk purse from a pig's ear.
What about circuit design. Are all circuits accurate? Nope. Are just the ones with low harmonic distortion accurate. Nope. There are many different kinds of distortion according to the RCA Radiotron Designers Handbook, written by 26 EEs, 1960.
1. Non linear distortion and harmonics. Harmonic Distortion, HD.
2. Intermodulation, IMD, related by not harmonic but sums and differences. It is very generally 3 times that of HD, but sometimes near zero if extremely high figures of negative feedback are used.
3. Frequency distortion, frequency response weaknesses
4. Phase distortion
5. Transient distortion.
6. Dynamic range and its limitation
7. Scale distortion
8. Frequency modulation distortion
9. Variation of frequency response with output level
How about smearing caused by a part as number 10.
So going by the lowest harmonic distortion figures is only one criteria.
Probably the worst flaw of an inaccurate part is that there is NOT just one flaw, but multiple. Synergy will never be optimized using the flawed concept of synergy. Optimum Synergy is only attained when all the parts, components, conditions are optimized. This includes room conditions. One cannot make a silk purse from a sow's ear.
The toughest component to design is an accurate speaker/tranducer. However, my contention is that another huge bottleneck is still the poorly designed recording equipment that is almost always used. We have all seen the long, long consoles with all those sliders, switches. Think of a table radio quality parts or a $100 stereo. The musical signal can "travel" through as many as 100 transistors, and nearly as many stages. It amazes me how the sound quality is as good as it is. Limited amount can be done with original recordings.
Lastly, budget is something we all contend with. We can only purchase what we can afford. But I think knowledge is power, to understand allows one to help his fellow audiophile/music lover to avoid the pitfalls.
Cheers, and great weekend to all.
steve
-
Hi Steve.
Thanks for this very informative write up.
By what you wrote, I can see that you know what your are talking about.
Guy 13
-
Having taught for decades you emphasize that there are no "stupid" questions (usually true). So bear with me please, if C1 and C2 are "coupling" caps ---- I see where C1 could "couple" the input to the power tube. What is the "coupling" that happens with C2 which appears to bypass the power tube? I get that every cap acts as a high frequency filter, depending on value... but why add unamplified high frequency back to the circuit?? Or am I looking at it backwards? What's coupled?
-
The tube is a triode and pentode in one. The C2 couples the output of the triode plate to the grid of the pentode for output.
-
The tube is a triode and pentode in one. The C2 couples the output of the triode plate to the grid of the pentode for output.
I think for this to mean something, he'd need to know the significance of that. (just a guess that it may not be clear)
-
I'll do some quick research on triode/pentode. You're right. I'm aware of the terms but not really the relevance.
Ok, single heater, single cathode, two sets of (differing) grids, two anodes.
The tube is a triode and pentode in one. The C2 couples the output of the triode plate to the grid of the pentode for output.
AHA!! I can see that! Thanks, Rich!
So, is coupling a term for caps that connect portions of the circuit where something is actively altering things... signal coming in, signal being amplified... but not simply filtering DC or frequency?
-
The coupling is of the music signal from the preamp triode section to the output stage pentode driver section, but not the DC bias from the input to output. Since the output is a gain stage, trying to amplify the DC output voltage on the triode would saturate the output tube, so the coupling cap keeps the DC out of the grid of the output pentode. This lets its cathode bias circuit work correctly and get the music signal into the output stage.