some guidelines
I am looking at 3 caps of the same dielectric (metallized polypro) and all 3 are 5% tolerance.
Why am I choosing 1 over the other? Is there some criteria I can find in the datasheet that makes the 1 capacitor more desirable/suitable in audio circuits?
(in this specific situation, the capacitor will be in the signal path. But will the choice be different if it's not?)
Thanks
- trying to be objective here. in my financial situation, money don't grow on trees so I want to spend wisely.
- eliminating various criteria like voltage and size and maybe other (let's assume the voltage and size are both suitable to the intended circuit/allocated space)
- the caps have the "same" dielectric
- the caps have the same tolerance
I am looking at 3 caps of the same dielectric (metallized polypro) and all 3 are 5% tolerance.
- Panasonic 4.7uF 450V (RF Microwave Film Capacitors) (datasheet) (cheapest)
- Wima MKP4 4.7uF 250V (general use) (datasheet) (mid-price but around double the cost of Panasonic)
- Vishay Roederstein 4.7uF 160V (AC and Pulse Film Capacitors) (datasheet) (most expensive and 4x the cost of the Wima)
Why am I choosing 1 over the other? Is there some criteria I can find in the datasheet that makes the 1 capacitor more desirable/suitable in audio circuits?
(in this specific situation, the capacitor will be in the signal path. But will the choice be different if it's not?)
Thanks
Form, fit, function, all three function the same so go with form and fit as your main criteria
I've had good success using film capacitors having (i) polypropylene dielectric; and also (ii) highest available dV/dt specification at that value of capacitance.
Place the cap on a bode plot.
For example panasonic ECW’s curve drops like a stone after about 1khz. Good for power and not for full range audio.
For example panasonic ECW’s curve drops like a stone after about 1khz. Good for power and not for full range audio.
I will retry the experiment sometime later this week and post it up. This rolloff is assuming there’s no cable issue.
This was a frequency response measurement of a high pass filter?
These are non-inductive parts. Which exact part number?
These are non-inductive parts. Which exact part number?
I’ll dig that all out. IIRC it was a 600V mouser US part: https://www.mouser.co.uk/ProductDet...u0LICSN1kag==&countrycode=GB¤cycode=GBP
No this was a simple bode plot using a 10Hz-60MHz at 6Vac sine - the signal source cable was a scope cable but i have some 50ohm bnc cables currently sat “delayed” with UPS.
Panasonic's film capacitor catalog includes this graph for their ECWFG product. It's a Bode plot of impedance vs frequency. Looks like a textbook excellent capacitor with very little series resistance, all the way up to the self resonant frequency where lead inductance ("ESL") begins to dominate. No funny business at 2 kHz whatsoever.
_
Attachments
That is what what was odd. I had it in the amp and it sounded dark and veiled, hence changing for wima PPs.
I just tried it out of the blue as i had the kit setup over the weekend.
Maybe that’s a dodgy cap or a dodgy connectiom. Like i said the better leads should arrive and i’ll redo the test.
Panasonic do make decent caps hence the surprise at the plot result.
I just tried it out of the blue as i had the kit setup over the weekend.
Maybe that’s a dodgy cap or a dodgy connectiom. Like i said the better leads should arrive and i’ll redo the test.
Panasonic do make decent caps hence the surprise at the plot result.
Agree that these caps do have some of that sound character, but are still decent enough.
I used them because they were acceptable, and nothing else would fit, other than electrolytics.
I used them because they were acceptable, and nothing else would fit, other than electrolytics.
Last edited:
Ok, so now the new RF cables have arrived, I have a a few connectors try. I also found the banana plug cable I'd been using for grounding had a fuse connector inside but no fuse (it's a green cable.. it just so happens to have an inbuilt fuse holder).
Anyway - the cap seems to be working normally whereas before repeated attempts showed a bend down after 1Khz to the point it would act likes LPF so I'm guessing something else was configured wrong - my fault:
Perhaps the moral of this story is - if it's unexpected check the experiment setup, only then when all options have been exhausted, then call it a result.
From Wima:
Single sided metallized structure(MKP 4) has the smallest possible box size at a given capacity. However, these types have the disadvantage that the capacitor current is only conducted by the thin aluminium metallization. In comparison with the double-sided metallized and film/foil version this construction has the least favourable pulse behaviour. In double-sided metallized (MKP 10) construction, the capacitor electrode is made of film metallized on both sides. The capacitor is produced with 4 film layers; construction has a 5-10 times greater pulse rise time and a much better pulse behaviour. They also have a 30-50% lower dissipation factor than comparable versions metallized on one side only. WIMA film/foil capacitors with polypropylene dielectric(FKP 02 (2.5 mm),FKP 2 (5 mm), FKP 3 (7.5 ,10,15 mm)) represent the ulimate. Due to the very good contact of the capacitor electrodes at the end surfaces to the Leads –“direct bonding”, as well as the large cross-section of the electrode foil, they have outstanding pulse behaviour. In the field of audio application, very precise pulse reproduction is called for, which is assessed in the pulse behaviour criteria. Film/foil constructions cause practically no attenuation of the audio signal and therefore provide absolutely precisely accurate pulse reproduction. FKP 2 versions (33 pF to 0.033 μF only) are also available as a precision capacitor with a tolerance of 2.5%; 1% available on request.
Single sided metallized structure(MKP 4) has the smallest possible box size at a given capacity. However, these types have the disadvantage that the capacitor current is only conducted by the thin aluminium metallization. In comparison with the double-sided metallized and film/foil version this construction has the least favourable pulse behaviour. In double-sided metallized (MKP 10) construction, the capacitor electrode is made of film metallized on both sides. The capacitor is produced with 4 film layers; construction has a 5-10 times greater pulse rise time and a much better pulse behaviour. They also have a 30-50% lower dissipation factor than comparable versions metallized on one side only. WIMA film/foil capacitors with polypropylene dielectric(FKP 02 (2.5 mm),FKP 2 (5 mm), FKP 3 (7.5 ,10,15 mm)) represent the ulimate. Due to the very good contact of the capacitor electrodes at the end surfaces to the Leads –“direct bonding”, as well as the large cross-section of the electrode foil, they have outstanding pulse behaviour. In the field of audio application, very precise pulse reproduction is called for, which is assessed in the pulse behaviour criteria. Film/foil constructions cause practically no attenuation of the audio signal and therefore provide absolutely precisely accurate pulse reproduction. FKP 2 versions (33 pF to 0.033 μF only) are also available as a precision capacitor with a tolerance of 2.5%; 1% available on request.
| Recommendations: | 100 pF to 0.01 μF | FKP 02 (2.5 mm) |
| 33 pF to 0.033μF | FKP 2 (5 mm) | |
| 100 pF to 0.22 μF | FKP 3(7.5 ,10,15 mm) | |
| 0.01 μF to 10 μF | MKP 4 | |
| 1000 pF to 47 μF | MKP 10 |
If you node plot FKP1, MKP10 and MKP4 you will see they’re flat. If you listen the FKP has great claity but due to size constraints is anaemic in bass snd needs support of a larger cap. MKP 10 sounded hideous whilst the MKP4 sounded very good but a notch down in airnesd to the fkp. Just my experience of wimas.
I have always shied away from using MPP caps such as Bennic et-al because they are so damned expensive and not readily available in Australia. It occurred to me that a cheaper alternative is to use MPP motor run/start capacitors because they are readily available in values up to 12uF - and they are inexpensive. They are close tolerance - 5%, and high voltage - typically 500Vac. They are available in wire leaded or Fast-on QC tabs.
https://www.motiondynamics.com.au/cbb61-run-cap-lead-wire/
I can't see why they would exhibit any negative characteristics in speaker x-overs compared to the Bennic units since the polypropylene dielectric film probably comes from a small number of manufacturers and is supplied to all the capacitor manufacturers. My LCR meter says they have negligible dielectric constant leakage.
Has anyone run any tests on them?
https://www.motiondynamics.com.au/cbb61-run-cap-lead-wire/
I can't see why they would exhibit any negative characteristics in speaker x-overs compared to the Bennic units since the polypropylene dielectric film probably comes from a small number of manufacturers and is supplied to all the capacitor manufacturers. My LCR meter says they have negligible dielectric constant leakage.
Has anyone run any tests on them?