Basicly this is a good approach to check out the sonic quality. Unfortunately your test point is befor your voltage regulator instead in parallel to the load.
This means misrepresentation due to the influence of the voltage regulator and the interplay of the following caps.
In the attachement you will find the royal way to ckeck the sonic quality of caps - especially interesting for RIAA head amps. The capacitor to be tested is connected parallel to the load (here an OP-Amp).
Isn't there any capacitor before regulators in your solution?
You could not compare the works of capacitors before regulation and after regulation.
- Capacitors before regulation provides 100% of current (DC and AC),
- They are charging and discharging 100/120 time per seconds,
- They needs high current when charging (90% bigger than DC needed)
- Don't provide any current when charging. (Regulator directly fed by rectifiers)
Capacitors after regulators are used for AC transient response and regulator stability. For transient response, a MKP capacitor seems to be a better solution than electrolytic capacitor.
Isn't there any capacitor before regulators in your solution?
You could not compare the works of capacitors before regulation and after regulation.
- Capacitors before regulation provides 100% of current (DC and AC),
- They are charging and discharging 100/120 time per seconds,
- They needs high current when charging (90% bigger than DC needed)
- Don't provide any current when charging. (Regulator directly fed by rectifiers)
Capacitors after regulators are used for AC transient response and regulator stability. For transient response, a MKP capacitor seems to be a better solution than electrolytic capacitor.
I agree, but I don't like serial regulators in general.
Isn't there any capacitor before regulators in your solution?..............yet, but here you must not use the highest quality that you could get, additional by parallel regulation I can choose the capacity in very high values in parallel to the shunt. This I cannot do by serial regulator output without trouble (except I use a shottky diode for regulator protect and an inrush current limiter resistor between regulator output and capacity input/supply rail).
For this reason, in most cases the normally used serial regulators are not the royal way - I think.
Today I received 80 pieces new capacitors. This ordered type and brand was the same than the one I use already since 10 years for upgrade amplifier power supplies.
Capacity is 4700uF and voltage is 100V (105 degrees solder legs "snap in")
The brand is one of the most respected manufacturer for audio and so far I was very satisfied.
I note now by this charge follow differences against to the old devices:
1) a bit different labeling (primarily the marking for the neg. pole)
2) different wight - the most striking difference immediately after opening the package - only 50 gramms instead 80 gramms, the wight of the old ones
3) muddy soundquality
4) ESR/Z
measured values (100Hz/10KHz):
old ones = 30/20 milli ohms new devices: 49/34 milli ohms
values according datasheet (100Hz/10KHz):
old ones = 33/26 milli ohms new devices: same size: 33/26 milli ohms and smaller size: 45/33 milli ohms
To verify all parasitic values I need actually a network analyzer to measure not only the capacitance and ESR.
The capacitor charge next week will be returned for exchange.
Whether the manufacturer brand has really cheap products obtained from India or China and the manufacturer itself only carry out the heat shrink tubing with it's own brand logo (Please note: I am not a manufacturer of aerospace devices)? I can't rule out (have a look to Google: "bad caps").
Now two questions
1) what could be the most likely cause for the significantly lower weight of currently version from the same type in same size?
2) Anyone know a (personal computer based) network analyzer optimized for checking electrolytic capacitors between 0,1mF and 100mF (100uF and 100.000uF)?
Thank you for your advices.
Capacity is 4700uF and voltage is 100V (105 degrees solder legs "snap in")
The brand is one of the most respected manufacturer for audio and so far I was very satisfied.
I note now by this charge follow differences against to the old devices:
1) a bit different labeling (primarily the marking for the neg. pole)
2) different wight - the most striking difference immediately after opening the package - only 50 gramms instead 80 gramms, the wight of the old ones
3) muddy soundquality
4) ESR/Z
measured values (100Hz/10KHz):
old ones = 30/20 milli ohms new devices: 49/34 milli ohms
values according datasheet (100Hz/10KHz):
old ones = 33/26 milli ohms new devices: same size: 33/26 milli ohms and smaller size: 45/33 milli ohms
To verify all parasitic values I need actually a network analyzer to measure not only the capacitance and ESR.
The capacitor charge next week will be returned for exchange.
Whether the manufacturer brand has really cheap products obtained from India or China and the manufacturer itself only carry out the heat shrink tubing with it's own brand logo (Please note: I am not a manufacturer of aerospace devices)? I can't rule out (have a look to Google: "bad caps").
Now two questions
1) what could be the most likely cause for the significantly lower weight of currently version from the same type in same size?
2) Anyone know a (personal computer based) network analyzer optimized for checking electrolytic capacitors between 0,1mF and 100mF (100uF and 100.000uF)?
Thank you for your advices.
Last edited:
I have observed this also with Nichicon Caps. The replacement for the VX series is the VR which is smaller. I think the electrolyte is supposed to be more efficent but the newer smaller ones definately don't have the specs of the older ones. What I have found with this is by ordering a higher voltage such as 50V to replace a 35 volt you get a larger cap with almost identical specs as the old one. Also I would let the caps burn in 100 hours-The nichicon KW series sounded muddy to begin with then after 100 hours it sounds better through it's own output than through a dedicated DA converter. The unit I did was a Krell CD DSP.
I charged capacitor with 5V and with screwdriver make short circuit. Second touch with screwdriver i saw and heard spark. third touch again. Than i charge again capacitor to 5V and repeat the same proces again. After several times when i charge capacitor and touch it with screwdriver, capacitor is discharged complete. How i Know? second touch with metal nothings happend and voltage on it is almost 0V. i am not try in electrici circuit how long is enough to form capacitor. I only try this for testing capacitors.
But in the digital audio i use Nipon Chemicon PSE and this capacitor need about 200-400 hours for burn in. Muse KZ need about 100-200 hours. My friend is burn in capacitors for only three days. After each song he is turn off cd wait for 15 seconds and turn it on again.
But in the digital audio i use Nipon Chemicon PSE and this capacitor need about 200-400 hours for burn in. Muse KZ need about 100-200 hours. My friend is burn in capacitors for only three days. After each song he is turn off cd wait for 15 seconds and turn it on again.
Hi impuls,
You do know you are damaging the capacitor - right?
The current exceeds the rating. This causes the foil to burn away a bit at a time as you short it. Heaven knows what is happening within the foil plates. If you need to "burn in" a capacitor, you have to keep the peak current flow within the ratings of the part. In fact, shorting a charged capacitor with metal is not advised. In fact, it can be dangerous.
-Chris
There is great talk about burning parts in, and the best capacitors brand and model. I think many people have far too much time on their hands. Yes, some are better than others, but the good ones are more similar than you want to believe.
Network analyzers and electrolytic capacitors are perhaps not the way to go here. I have an HP 4263A LCR meter, and a spectrum analyzer. The test you want should really be done with an oscilloscope and signal generator, along with some resistors. That will stand a chance at showing you what could possibly be happening.
You do know you are damaging the capacitor - right?
The current exceeds the rating. This causes the foil to burn away a bit at a time as you short it. Heaven knows what is happening within the foil plates. If you need to "burn in" a capacitor, you have to keep the peak current flow within the ratings of the part. In fact, shorting a charged capacitor with metal is not advised. In fact, it can be dangerous.
-Chris
There is great talk about burning parts in, and the best capacitors brand and model. I think many people have far too much time on their hands. Yes, some are better than others, but the good ones are more similar than you want to believe.
Network analyzers and electrolytic capacitors are perhaps not the way to go here. I have an HP 4263A LCR meter, and a spectrum analyzer. The test you want should really be done with an oscilloscope and signal generator, along with some resistors. That will stand a chance at showing you what could possibly be happening.
Yes i know that is dangerous but it is only for testing. How quickly test capacitor. "That's one small step for [a] capacitor, one giant leap for science". this is the fastest way. 5V is so small volrtage to give high carrent with ESR. I never didi this with higher voltage.
But i need some kind of device for burning capacitors for the short time?
But i need some kind of device for burning capacitors for the short time?
Where is "There"? Here on this forum?
I think this also. Therefore some years ago I have invented a method to make a preselect to get a small selection of interesting brands and types.
I have prepare the TV. The checkpoints are pos. and neg. rail of the vertical flyback power amp (aoound 35VDC). The condition to choice of TV it was that the voltage supply was create about the horizontal line transformer. I remove the internal smoothing capacitor there. So I compare the Vss value of the residual (and there unwanted) sawtooth signals of horizontal and vertical flyback by a wide range of capacitor brands. There was very large differences.
The correspondence of the results with the listening tests are very great (the goal was the lowest possible value within a certain size and certain minimum voltage, not necessarily the same capacity).
I suppose I must now repeat this procedure because many capacitor manufacturers introduced due to cost and weight savings other material and other production processes (and other production places - like Siemens/EPCOS from Germany to India e.g.).
Because today's capacitors are mostly in use by SMPS (SPS) instead traditional power supplies with only a rectifier, the behavior and characteristic now more matched to SMPS - this could also a reason for the bader audio behavior).
Is this maybe on of the reason, why by new amplifiers often is to find bader sound quality than by the old devices (previous models of the same brand in the same price class)?
Last edited:
This is a very good question!
There's who swear about ceramic caps for bypass or as snubbers...
Hi impuls,
I use an HP 4263A LCR meter and look at the dissipation figures at different frequencies. That tells me a great deal about the part.
Perhaps the best way is a jig that someone used (sorry, I forgot who now) to measure THD and non-linearity caused by the capacitors themselves. I think Walter Jung was involved and he may even have that report on his web site. It is a good read and I have confirmed it's findings. Caps with dielectric issues cause trouble when they develop a voltage across them. Like a low frequency across a small capacitor, or when their impedance begins to affect the circuit, like Cdom in an amplifier.
Hi tiefbassuebertr,
Too many different threads to mention over the years. I would start with manufacturer's application notes for the truth. Not audio "white papers" or similar.
Your method of testing does hit the mark, it just doesn't have any numbers attached for comparison. See, you can measure these things! The SMPS capacitors do seem to work better for audio from my experiments. Of course, that isn't a general thing. You will always have to test to make sure the parts are suitable.
Also, using new capacitors with the same case volume (higher voltage) will measure and sound more closely to the older part you are replacing.
Last note. Bypassing an electrolytic cap will only lower the high frequency impedance, it will not correct at all for electrolyte problems like dielectric absorption. That's the thing that distorts a waveform. To fix this, you need to find a capacitor without those problems. A higher voltage part is often the solution, possibly in a different brand. Just understand that there are more than one mechanism for "bad sound", or distortion in capacitors.
-Chris
I use an HP 4263A LCR meter and look at the dissipation figures at different frequencies. That tells me a great deal about the part.
Perhaps the best way is a jig that someone used (sorry, I forgot who now) to measure THD and non-linearity caused by the capacitors themselves. I think Walter Jung was involved and he may even have that report on his web site. It is a good read and I have confirmed it's findings. Caps with dielectric issues cause trouble when they develop a voltage across them. Like a low frequency across a small capacitor, or when their impedance begins to affect the circuit, like Cdom in an amplifier.
Hi tiefbassuebertr,
Too many different threads to mention over the years. I would start with manufacturer's application notes for the truth. Not audio "white papers" or similar.
Your method of testing does hit the mark, it just doesn't have any numbers attached for comparison. See, you can measure these things! The SMPS capacitors do seem to work better for audio from my experiments. Of course, that isn't a general thing. You will always have to test to make sure the parts are suitable.
Also, using new capacitors with the same case volume (higher voltage) will measure and sound more closely to the older part you are replacing.
Last note. Bypassing an electrolytic cap will only lower the high frequency impedance, it will not correct at all for electrolyte problems like dielectric absorption. That's the thing that distorts a waveform. To fix this, you need to find a capacitor without those problems. A higher voltage part is often the solution, possibly in a different brand. Just understand that there are more than one mechanism for "bad sound", or distortion in capacitors.
-Chris
If somebody needs them, i have both Jung articles about capacitors. They are very interesting esp. regarding to the most common materials.
Your observation about using different brands is a smart one
I usually use a signal generator with known accurate resistor close to 50 ohms in series with the DUT electro and sweep to find the minima which using the voltage divider rule can be calculated very close to the ESR for all practical accuracy. keep sweeping upward to find the freq at which Xc starts rising and you've found the self resonance freq.
Hi infinia,
The self resonant frequency is normally more pronounced in ceramic and other smaller capacitors. The internal losses in electrolytic types would normally damp out (= low Q) any resonant peak. Having said that, I will admit to never trying this with electrolytic types. What have you found in the real world?
The effects most important to we audio types would be the non-linear defects that lead to distortion. The very same issues affect signal processing in things like pulse generators and oscilloscopes. I'll bet that the wrong capacitor type in a CD or DAC reconstruction filter might cause a lot of trouble also.
Hi Telstar,
Very good of you to offer our members. I have them somewhere, but others may not.
Highly recommended reading everyone. After you read this stuff, just consider the effects and how to minimize them in a design. There are times when this is unavoidable, but at least you know then how to go about minimizing the effects.
-Chris
The self resonant frequency is normally more pronounced in ceramic and other smaller capacitors. The internal losses in electrolytic types would normally damp out (= low Q) any resonant peak. Having said that, I will admit to never trying this with electrolytic types. What have you found in the real world?
The effects most important to we audio types would be the non-linear defects that lead to distortion. The very same issues affect signal processing in things like pulse generators and oscilloscopes. I'll bet that the wrong capacitor type in a CD or DAC reconstruction filter might cause a lot of trouble also.
Hi Telstar,
Very good of you to offer our members. I have them somewhere, but others may not.
Highly recommended reading everyone. After you read this stuff, just consider the effects and how to minimize them in a design. There are times when this is unavoidable, but at least you know then how to go about minimizing the effects.
-Chris