Nichicon KG capacitors: type I, II and III

[onanysunday]

Let's say KG 4700uf and KG 6800uf are both suitable. Which would be a better capicator, 4700uf type 2 or 6800uf type 1?

Besides the fact they're different capacitance and the effect that would have on frequency response when interchanged in the same capacitor position on the board, I wouldn't expect the difference between I and II to be very noticeable, if at all.

To see how capacitors compare to one another, go to DigiKey and use the search. There, you'll be able to compare the current of different caps at both low frequency and high frequency so you can see where the emphasis will be; either strong bass, balanced, treble emphasis, or over-emphasis on treble and bass.

There are many other specs to consider when choosing electrolytics which effect performance, but this one generally determines the actual sound the most from what I've gathered. Most datasheets and sellers only give you one average ripple value per cap which isn't nearly as useful in determining their sound characteristics at both LF and HF.

 

[onanysunday]

Current at LF is always lower than current at HF in an electrolytic. So competitively listed current at LF but not HF indicates a more balanced sound, with the exception of most audio-rated caps generally sounding better for audio, sometimes in spite of their current ratings when compared to other general-purpose types.

If you look at ripple current at LF and HF for the audio-rated caps, you'll notice the general trend of competitive current at LF but not listed for HF. Elna Silmic is a perfect example; never edgy, bright, clinical or fatiguing because of the silk fibers and less available current at high frequencies while having competitive current at LF.

 

[Oilmaster]

Let's say KG 4700uf and KG 6800uf are both suitable. Which would be a better capicator, 4700uf type 2 or 6800uf type 1?

1) I usually start with the available diameter space, and then increase the rated voltage to the maximum in order to lower the phase shift (as I mentioned in post #16).

2) Between types and brands, the ripple current differs, so also looking at best available ripple current (not an absolute criteria, but it weighs in). You will also notice that the ripple current is a function of rated voltage and capacitance (3 dimensional factor, so pick your poison )

3) Changing the capacitance value is yet another story... as it will change the corner frequency (a.k.a. cut-off frequency). Increasing the value will lower the corner frequency, i.e. more low frequencies are send to the speakers.

Read about it here, and you can calculate yourself "from where / to where" you will shift the corner frequency (and in fact there is no problem to do that)
https://www.electronics-tutorials.ws/filter/filter_3.html

If somebody asks.... why didn't the manufacuturer put in a large value cap to start with: well they had to make a commercial choice between the 5 factors that I identified by now.
And most folks would have low budget speakers connected anyway that wouldn't go to below 40Hz.

Another topic to keep in mind: DC leakage. The output capacitor is primarily there as a DC blocking device. Period.
By design, higher rated voltage caps have lower DC leakage, but leakage increases when capacitance increases (larger film surface).
So, factor #6 in your decision making

 

[Oilmaster]

In other words: picking a new capacitor with space constraints is a 4 to 6 dimensional compromise.
Single dimension questions like capacitance or type are thus not possible.

What are the available diameter and height space for your caps again (in mm) ?

 

[onanysunday]

UKL is the top pick wherever larger capacitance and low-leakage is concerned and one of the only electrolytics available in 10% tolerance. I like them most in power supplies and rectification stages in the larger values, not so much in the smaller signal path values. For tone, there are other suitable choices. However, they are a suitable general-purpose cap for audio. They're clean, crisp, cool and have the most detail I've heard in an electrolytic that borders on sterile neutrality depending where and how many are placed within a piece equipment. Some would prefer this sound, in which case they can be used exclusively in a recap.

For anyone wanting more oomph from their power filter caps, I always recommend increasing voltage instead of capacitance, as the circuit was designed and optimized with the original capacitance values in mind. As mentioned earlier, a higher voltage rated cap is physically larger and can therefore store more (available) current, which gives more liveliness, energy and detail to the sound without messing with frequency response and voicing of the amp.

Sometimes people think a max specified ripple current is just an averaged, theoretical rating limit which is true, but it's also an increase in working current. Physically bigger, higher voltage caps of the same capacitance as the originals produce more current while maintaining capacitance at the specified value.

 

[HPMguy]

The specified ripple current rating is an indication of how much ripple current the capacitor can handle before overheating, it's not an indication of signal quality. Current draw capacity is indicated by impedance rating, so a lower impedance cap is going to deliver more instantaneous current than a higher impedance one, and heat up less at the same time.

Remember that the cap is charging at 120Hz, transient current draw is indeed transient, not sustained. Larger capacitance will give more current draw capacity, that is pretty much the definition of capacitance as I understand it. A higher voltage rating makes for a larger capacitor, but the capacitance has a lot more to do with charge and discharge than the voltage rating.

If you think you are getting audio effects from main filter caps I'd recommend a higher wattage amp (or turn it down) as I cannot imagine how they would affect the audio path short of rail voltage or current supply limitations, neither of which can be fixed by a bigger cap with the same power supply transformer.

I don't use UKL Nichicons in power supplies, that's what UPW caps are for (and a few others). Low impedance caps for power supplies, low noise caps for audio path where low impedance isn't the goal. UKLs are my choice for coupling caps in small values (I don't have any capacitor coupled output amps) when I cannot use film caps. Film caps are better, they introduce far less harmonic distortion, and I want the source amplified without any changes, just the way I am.

Changing the value of output coupling caps should be done with care, as changing the frequency response of the output can have unintended consequences. If the size was chosen to suppress low frequency noise, changing it may let you hear the things I do on my Kenny KA-900 and HPM speakers -- turntable rumble, AC vents blowing on microphones on classical recordings, and other such things. It will also permit high amplitude non-audible "sound" waves below 10Hz to walk the woofers in and out, which is not a good thing. If you extend the upper frequencies too far you may allow super-sonic oscillations that would otherwise be damped -- the latter may not be an issue, I'm not that familiar with filter effects and capacitance.

I seriously doubt any amplifier was tossed together with random parts, someone did the engineering to determine what should be used and how. You might question their choices, but they were not random.

 

[onanysunday]

The low impedance of UPW is nice. They're indeed designed for power supplies- probably lighting ballasts. The problem I have with them is their high ripple at HF makes them edgy sounding and fatiguing at volume- for audio. I like UKL in all but the biggest, main power supply caps because they seem to provide the cleanest (quietest) power for the rest of the system. They're not also as good as tone capacitors in a direct signal path in my opinion, unless you like a double-dose of a more sterile sound, which is perfectly fine. Some do, as it can be very detailed and on the cooler side of neutral. The higher the voltage rating in a given capacitance for a cap, the physically larger storage space it will have and provide less impedance and more current. Ripple current is not just a theoretical heat rating. The larger the can, the more increased capacity it will have to physically store and release more energy. Yes, the characteristics of the power filter cap can make their way through to the audio signal. To test this, you only need to install two different series of power filter capacitors with very different ripple current ratings at LF as well as HF to notice each channel's different output level of emphasis and impact. I have done this experiment many times before with predictably reliable outcomes. Current does not lie. You feel it more than some of other parameters such as leakage, ESR or impedance, as it has impact. "DC High-Current" designs were a successful marketing slogan for this very reason. You felt it.

Here is a comparison to consider on a provided link. Two different capacitors. Both rated at 22,000uF 100V. Should sound the same, right? After all, they have the same capacitance. One has 17A (competitive) current at LF, but is not competitive at HF. The other, not competitive at LF, but has 19.24A (competitive) current at HF. Notice the low ESR and impedance are almost matched perfectly between the two caps which should make them more similar than different, but there's a 0% chance both channels will hit with the same impact at both the low and high frequencies which is what you'd feel and notice more than anything.

https://www.digikey.com/en/products/compare?s=N4IgzCBcDaIExwAxzATkSANOArAFgHY9U8QBdbUyEAegFMA7GgBwCcB7AEwFcBjAFwDONAGYBLADb86rGgEMJ3ALZiGygLR0JdARwkBPfmN7rec5nN5j+7VsJwAOAPyCAvADk8ASQDmAKwBhACFedgBRAHcAQQiggE0A3gA1AEUAcQBZdgARDIC0qIAjKIAvHwBpOKjELyjUCSj3dh8SgA9W6LCo1nTfAIA2ZijsnwCcFJ8gpKiACTiggBkZnxmJDKiAdQBVBYAtByik2IB5AAsugDFUAGUD44jmdi9sgCkgiLDrnHLuGfZ1uSnOA4VC+RAAJWOSn0F0MchSAUQGSSMwAjj5UBcHHgHl45JMNtkGFsUpwStlOCIthkxFsxNlUVE0nEUtkABoBXzZXYANzCKSiABURBJ-twLtcUmyGHgonFBS8cAsJK0AjNjmFWtdwQBrHAvcFJa7LIIABQWpy2BGOKVaSSNURAAF8gA

 

[onanysunday]

To be clear, I'm not saying one cap delivers its full 17A @ LF all the time and the other delivers is full 19.24A @ HF all the time; just that the higher that "theoretical" ripple rating is, the more actual current you will observe. How much more? I don't know exactly, except to say that it can be used as a broad indicator of LF/HF emphasis in general terms.

 

[Oilmaster]

The specified ripple current rating is an indication of how much ripple current the capacitor can handle before overheating, it's not an indication of signal quality. Current draw capacity is indicated by impedance rating, so a lower impedance cap is going to deliver more instantaneous current than a higher impedance one, and heat up less at the same time.

Remember that the cap is charging at 120Hz, transient current draw is indeed transient, not sustained. Larger capacitance will give more current draw capacity, that is pretty much the definition of capacitance as I understand it. A higher voltage rating makes for a larger capacitor, but the capacitance has a lot more to do with charge and discharge than the voltage rating.

If you think you are getting audio effects from main filter caps I'd recommend a higher wattage amp (or turn it down) as I cannot imagine how they would affect the audio path short of rail voltage or current supply limitations, neither of which can be fixed by a bigger cap with the same power supply transformer.

I don't use UKL Nichicons in power supplies, that's what UPW caps are for (and a few others). Low impedance caps for power supplies, low noise caps for audio path where low impedance isn't the goal. UKLs are my choice for coupling caps in small values (I don't have any capacitor coupled output amps) when I cannot use film caps. Film caps are better, they introduce far less harmonic distortion, and I want the source amplified without any changes, just the way I am.

Changing the value of output coupling caps should be done with care, as changing the frequency response of the output can have unintended consequences. If the size was chosen to suppress low frequency noise, changing it may let you hear the things I do on my Kenny KA-900 and HPM speakers -- turntable rumble, AC vents blowing on microphones on classical recordings, and other such things. It will also permit high amplitude non-audible "sound" waves below 10Hz to walk the woofers in and out, which is not a good thing. If you extend the upper frequencies too far you may allow super-sonic oscillations that would otherwise be damped -- the latter may not be an issue, I'm not that familiar with filter effects and capacitance.

I seriously doubt any amplifier was tossed together with random parts, someone did the engineering to determine what should be used and how. You might question their choices, but they were not random.

Remind that this threat is about amplifier output coupling capacitor, and not about PSU caps.
Very different topic.

 

[HPMguy]

True, I got side-tracked -- this particular discussion is identical to ones about filter caps in other places.

Film caps would be a far better choice for coupling caps if you can find some that fit without long leads. All electrolytic caps have harmonic distortion, it's part of their construction and unavoidable, but low leakage types seem to have the least. Don't know if UKL caps are available in the required rating, but if they were that's what I would use. I would expect low impedance caps to have more distortion being designed for filtering, not coupling.

 

[margus_]

Remind that this threat is about amplifier output coupling capacitor, and not about PSU caps.
Very different topic.

Basically the question was whether there is a significant difference between nichicon kg types (especially between types 1 and 2)

 

[Milen]

Hi

Basically the question was whether there is a significant difference between nichicon kg types (especially between types 1 and 2)

Hi, i not have opinion with type 1, but replaced Epcos power caps in my Rod Elliott P101 power amps with type 2, 4700, 80V. Result is silence between instruments is far deeper. And this change everything. In my opinion is beter to use type 2 - price is not so big. Type 2 caps are bigger and this is the key...

 

[rjb5191]

1) I usually start with the available diameter space, and then increase the rated voltage to the maximum in order to lower the phase shift (as I mentioned in post #16).

2) Between types and brands, the ripple current differs, so also looking at best available ripple current (not an absolute criteria, but it weighs in). You will also notice that the ripple current is a function of rated voltage and capacitance (3 dimensional factor, so pick your poison )

3) Changing the capacitance value is yet another story... as it will change the corner frequency (a.k.a. cut-off frequency). Increasing the value will lower the corner frequency, i.e. more low frequencies are send to the speakers.

Read about it here, and you can calculate yourself "from where / to where" you will shift the corner frequency (and in fact there is no problem to do that)
https://www.electronics-tutorials.ws/filter/filter_3.html

If somebody asks.... why didn't the manufacuturer put in a large value cap to start with: well they had to make a commercial choice between the 5 factors that I identified by now.
And most folks would have low budget speakers connected anyway that wouldn't go to below 40Hz.

Another topic to keep in mind: DC leakage. The output capacitor is primarily there as a DC blocking device. Period.
By design, higher rated voltage caps have lower DC leakage, but leakage increases when capacitance increases (larger film surface).
So, factor #6 in your decision making

The only thing I dispute here is that I believe that leakage typically increases with rated voltage as well.

 

[Oilmaster]

The only thing I dispute here is that I believe that leakage typically increases with rated voltage as well.

Leakage increases with the applied voltage across the capacitor, compare it with pressure across an orifice.

From my readings on manufacturer's (white)papers, in general, a higher voltage rating actually means a thicker insulating film between the conductive layers, and as such, in general, means less leakage when (for example) a 1Vpp signal is applied across a 100V rated cap compared to a 16V rated cap from the same series.

But leakage rate also depends on the quality of insulating film used, and this is yet another performance difference between certain capacitor series and capacitor brands.
So many variables....

Signal coupling caps (amplifier input, pre-amp input/output) in the better audio gear have therefore often 100V rated electrolytics, despite a maximum 1Vpp signal.

 

[soundguy9]

I used the type III ‘super trough’ with my Marantz 2220B large coupled capacitors and it pleasantly improved the sound. I was able to confirm this by swapping them in and out with the original capacitors.

Yes, I need them for output coupling. Capicator must be 6800 uf / 63uf / d=30mm. Only KG type I is available. I read that nichicon KS is higher grade than KG type I, but also not available d=30mm. None of them is expensive, I just want a satisfactory result

I used the type III ‘super through’ with my Marantz 2220B large filter capacitors and it pleasantly improved the sound. I was able to confirm this by swapping them in and out with the original capacitors.