IIRC in recent tear down here or somewhere else, at least focal alpha have quite some really crap rated brand electorlytic caps in their amp... really wonders why cheap this out for reliability issue just for what, less than a cent per cap savings?
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Capacitor upgrade in crossover - You CAN'T handle the TRUTH! - Part 3
Bold statement for nothing to back it up. What exactly is "crap rated brand"? Is it because You read the specs of the actual devices, and these didn't comply to it, approved by Your measurement? I'm not asking for Your statistics in examining the output of the vendor You didn't mention by name. What is that?
You told me that >I< was to measure ESR. Obviously You are just speculating about the ESR, maybe alarmed by internet-hogwash. Why do You think I should take the effort? And then You consequently dismiss the very success of the caps in the product, in disregarding the relevance of the final outcome. What is that?
Last edited:
Inside the Focal Alpha 65
www.audiosciencereview.com
it is this one which I remember seen a while ago, Chengx caps don't seems like among the better reliability caps. remember in computers they are pretty prone to swelling and die prematurely. crap rater maybe too far for a grade, but then it won't go into any of the "reliable electrolytic cap brands" either
Inside the Focal Alpha 65
it is this one which I remember seen a while ago, Chengx caps don't seems like among the better reliability caps. remember in computers they are pretty prone to swelling and die prematurely. crap rater maybe too far for a grade, but then it won't go into any of the "reliable electrolytic cap brands" either
You introduced rumours regarding a brand being "crap ...". The final statement in that thread so far came also from You. But without any (!) data to prove or even suggest it. Neither the specs are documented, nor the caps being out of specs, and likewise that the usage in that product would somehow use the caps off limits.
Finally I argue that those caps are used in an active speaker in the power section. Thread title here is: "
Capacitor upgrade in crossover - You CAN'T handle the TRUTH!
What are You after?I am with the dilemma between passive crossover and single amplification, and multiway amplification with digital crossover.
In your opinion there are more risks of failure having a passive crossover or N amplifiers? (considering the capacitors involved in the two solutions I mean).
In my case, I am speaking of 3 way passive crossover (817hz+4750hz) vs 3 Hypex NCxxxMP module.
In your opinion there are more risks of failure having a passive crossover or N amplifiers? (considering the capacitors involved in the two solutions I mean).
In my case, I am speaking of 3 way passive crossover (817hz+4750hz) vs 3 Hypex NCxxxMP module.
In general passive crossovers are going to be more reliable and "safer" i.e. less to go wrong to blow out your drivers but really if you are smart about the design and use of your system it is not that big of a deal. I think the potential SQ improvements are worth the slight risk. One thing to consider is that for tweeters and even a small midrange any Hypex module has way more power than you need and this is where you can run into problems. Any "protection" scheme is flawed and the best protection is to use lower powered amps especially on the tweeters. An "accident" with a 40 watt amp can likely be survived by a tweeter, with a 200 watt amp doubtful.
Yes, for the tweeters I have the 50w NC52MP module planned.
Wait I don’t get where your rage comes from, I was just replying your question when another member said that when crap caps are used generally they die first in a system (I think it’s pretty common sense, in most electronics instead of an IC or driver, the electrolytic caps are the ones to die first, when they are of less durability), so when you ask for examples of bad caps being used, I just pick from memory seeing that thread a while ago, thus reply that it’s actually out there with products using bad brands inside speakers, nothing more, that thread of course have my reply also else I won’t have such clear memory
For caps i don’t think failure rate can be found so easily as they are dirt cheap to be documented by consumer or so so no, I don’t have definitive proof that it’s always fails easily, but then in PC power supply etc. there’s a long regarded reliable brands and unreliable ones, the even lesser known ones generally goes into the unknown bad brands which fails easier under load.
Yes the focal evo is having them in the power amp plate, off topic of the crossover topic but it’s just to show that inside speakers they are used, and if fails it likely affects some functionality if not sound signature as its there to do something.
And I have to be clear that I get into this thread long ago to make it into follow list because I completely agrees that audiophile upgrade of crossover caps are meaningless and don’t change a thing when ratings are the same. For me the crap caps vs good ones aren’t changing the <$1 @ Rubycon to some silver caps costing $100+ each, it’s the same <$1 caps we found in all electronics, but bad being less reliable brands vs similarly cheap but very reliable Japanese ones. That’s it, the expensive silver ones are just snake oils
In 1982 Mark Levinson sales director was asked "do you design by ear on by measurements?"
He answered: we design by measurements, but if our listening experience does not coincide with measurements, we design new measurements"
The guy who started this discussionn is sure, that everything can be measured with a handheld RCL-meter.
Designing new measurements in this field must start with a long meditation on psycho-physiology of hearing.
Yes, the differences in frequency response are the first things in differences bitween speakers that are noticed, but it's also been proven some 45 years ago
that certain things, like aligning radiating centers of speaker elements, will be clearly hearable only when most of things that affect frequency balance of sound that reaches the ear, are flattened out. Etc etc etc .....
SO: we don't know what people, who are making high-end capacitors and cables, do not tell us about the things they measure in these things.
AND: everybody has a bit different hearing mechanism connected to a different psychology. If you like the sound of you hear, it is good FOR YOU.
And Mother Nature has made things the way that yout GOOD not be as good for some other guy.
He answered: we design by measurements, but if our listening experience does not coincide with measurements, we design new measurements"
The guy who started this discussionn is sure, that everything can be measured with a handheld RCL-meter.
Designing new measurements in this field must start with a long meditation on psycho-physiology of hearing.
Yes, the differences in frequency response are the first things in differences bitween speakers that are noticed, but it's also been proven some 45 years ago
that certain things, like aligning radiating centers of speaker elements, will be clearly hearable only when most of things that affect frequency balance of sound that reaches the ear, are flattened out. Etc etc etc .....
SO: we don't know what people, who are making high-end capacitors and cables, do not tell us about the things they measure in these things.
AND: everybody has a bit different hearing mechanism connected to a different psychology. If you like the sound of you hear, it is good FOR YOU.
And Mother Nature has made things the way that yout GOOD not be as good for some other guy.
You think its a secret?
Is that what he said?
Is this your field?
That's a great quote to make people think there is some mystery hidden in there. There isn't.
How about the designer of Quad Amps who takes the opposite approach? No listening needed...
Do you think Benchmark tunes by ear?
Not really sure what that has to do with whether something can or can't even be heard in the first place.
MAB
Major Contributor
The sales director said this? Verbatim??? Were there audio luminaries in the room? Was anybody in the room who even knew how to do a measurement? The sales director, yes...
No, that doesn't appear to be the premise of the original poster. Can you show me where they say that? Can you also show that all of the measurements were done with a handheld meter? Are handheld meters OK?
If the OP was Spock and the meter was a Tricorder would that be OK? But most importantly, can you show us where the OP said this?That's the great thing about capacitors, they actually don't need new measurements. The physics of how they operate is quite well understood, and they lend themselves to precise electrical characterization with respect to the signal that goes through them. And, each of the capacitor types do have different electrical responses with electrically measurable differences, the better the LCR (or Tricorder) the deeper one can resolve those subtle but simple differences. And the results of those differences can even be measured at the speaker, just not by ear. Certainly not by your or my ears since we are human.
Correct, they measure the price, then they measure the cost, then they subtract the two; they never tell you how absurdly high that number is for no benefit to the sound.
But, wait for it, I knew it was coming, the golden-ear argument.
Plenty of stuff have measurable differences that cannot be heard (like capacitors). If in your meditations you come up with something that can cause an audible change but that cannot be measured, please post!
musicisair
Member
- Joined
- Nov 11, 2024
- Messages
- 31
- Likes
- 7
I'm very new to this, but these threads and measurements are fascinating!
I've been using ChapGPT to learn more, and one of its claims is that in bass circuit, several smaller capacitors (like 5 25uf caps) in parallel are better than one larger one (like 125uf). Specifically, it says:
> The parallel arrangement of capacitors ensures the circuit can handle transient peaks better, allowing the woofer to stop more precisely after the signal ends, thus improving decay performance.
Is this totally BS and just not even worth measuring, if not, has it been measured here? There are a lot of pages so maybe I missed something or just didn't understand some if the data that was presented.
I've been using ChapGPT to learn more, and one of its claims is that in bass circuit, several smaller capacitors (like 5 25uf caps) in parallel are better than one larger one (like 125uf). Specifically, it says:
> The parallel arrangement of capacitors ensures the circuit can handle transient peaks better, allowing the woofer to stop more precisely after the signal ends, thus improving decay performance.
Is this totally BS and just not even worth measuring, if not, has it been measured here? There are a lot of pages so maybe I missed something or just didn't understand some if the data that was presented.
MAB
Major Contributor
ChatGPT manages to avoid telling you actually what happens when you parallel several capacitors together, manages to say something that isn’t exactly wrong (the stuff about transients), and get the application to speaker wrong.
Paralleling two identical capacitors doubles the capacitance and halves the Equivalent Series Resistance (ESR). Most capacitors used in crossovers have ESR in the audio band so low it’s a bit difficult to measure without the correct equipment. Much lower than the resistance of the driver circuit’s resistance. The lowered resistance of the parallel caps has negligible effect on the circuit formed with the speaker and amp.
The ESR of the cap can interact with other low impedance components in the crossover. Swapping capacitors with different ESR in an existing crossover can change the filters in the capacitor, sometimes large enough change the filter response. In a well designed crossover, not good. In a poorly designed crossover may help or make it worse. Doesn’t stop people from supposedly ‘upgrading’ their crossovers by swapping caps, adding bypass caps, lauding about the improvement, despite not really knowing what resulted.
A few other things change, similarly unimportant in almost any crossover.
ChatGPT does not understand the Fourier theorem.
musicisair
Member
- Joined
- Nov 11, 2024
- Messages
- 31
- Likes
- 7
It certainly doesn't understand anything, but it can recite more about it than anyone else.ChatGPT does not understand the Fourier theorem.
Thanks for taking me to school!
So knowing the approximate/typical ESR of a capacitor at expected frequencies (like electrolytic vs film) is nearly as important as knowing the value, in very high quality speakers at least?
In other words: the ESR of a capacitor in a well-designed circuit was considered by the designer, and can't just be blindly significantly changed and then replaced by a resistor, because ESR may vary slightly with frequency (higher at high frequencies), whereas resistors (in audible frequencies) do not.
In practice: Swapping a 75uF electrolytic cap for 3 25uf film caps in a band pass filter (like in a mid circuit in a 3-way XO) could audibly alter the FR at the crossover point and higher by reducing the damping at those frequencies due to the significant change in ESR (going from higher and less stable ESR of the electrolytic to significantly lower and more stable ESR of the parallel group of film caps)?
Chat GPT is the next step in dumbing down society. It is terrible to learn something with. I think mankind might be doomed!
Hard disagree. If you work in a technical field and can't/won't use an LLM like chatgpt you will not be able to keep up and will be irrelevant in the next few years. I've seen LLMs solve obscure bugs, in complex mathematical code, in only seconds, that a group of 4 software engineers with combined experience of over 70 years could not over the course of 2 days.
Yes, it still hallucinates and will gleefully write BS fluff pieces, which now pollute the Internet and most books written today, but don't just write it off because of that. It's a tool like any other and tools used properly can be very powerful.
It's of secondary importance. In some cases, it has a small but measurable (and possibly noticeable) effect, in other cases a near-zero effect.
I do work in a technical field. My two best uses for it were writing my AI policy for my course syllabus for Intro Chemistry and for writing my obligatory faculty DEI Statement.
"I've seen LLMs solve obscure bugs, in complex mathematical code, in only seconds, that a group of 4 software engineers with combined experience of over 70 years could not over the course of 2 days."
Thank heaven I was a hardware engineer.
Chat GPT and other AI programs have the same problem as Wikipedia; data and information is not fully vetted so the wrong answers are delivered and blindly accepted by folk who do not know better.
Thank heaven I was a hardware engineer.
Chat GPT and other AI programs have the same problem as Wikipedia; data and information is not fully vetted so the wrong answers are delivered and blindly accepted by folk who do not know better.
Parallel capacitors were touted years ago as being better for wideband signals due to lower ESR and higher self-resonance. Capacitors now (and for the past 10~20 years) generally have lower ESR and much higher self-resonance than those from decades ago. Parallel capacitors take up space and require additional connections (extra resistance etc.) so you may or may not win vs. a good single capacitor. I've presented lectures on power supply decoupling that dives fairly deeply into the type and number of capacitors required; it can get complicated. For example, a low-ESR capacitor may cause high peaking with a low-resistance inductor, leading to poor frequency response and potential instability since the amp sees a highly reactive load. Sometimes a lower ESR capacitor is intentionally chosen for the best overall system performance. The world abounds with DIY folk who do not fully understand the nuances of their design or modification and may make things worse instead of better. Chances are it won't be a huge difference, and somehow the result is always audibly "better".
