Bypass capacitor question
Quan Tran
I have a pair of NHT speakers that have been modified quite a bit. I
fact I have replaced the tweeters and modified the crossover as
needed.
I used SCR capacitors (French, 5.5uF) on the high pass filter. I am
now considering bypassing these SCRs with Hovland Musicap (0.47uF).
I would appreciate if anyone could advise whether this upgrade worths
the price because the Musicaps are quite expensive.
The cap on the midrange is still the original electrolytic. Its value
is too big to be replaced by PP. Should I bypass this cap too?
Thank you in advance
Quan Tran
P.S. I understand that the frequency response will change with the
addition of the bypass cap and I know how to correct it by changing
the Lpad.
Stephen McLuckie
news:okxhc.183689$JO3.105110@attbi_s04...
> I have a pair of NHT speakers that have been modified quite a bit. I
> fact I have replaced the tweeters and modified the crossover as
> needed.
>
> I used SCR capacitors (French, 5.5uF) on the high pass filter. I am
> now considering bypassing these SCRs with Hovland Musicap (0.47uF).
>
> I would appreciate if anyone could advise whether this upgrade worths
> the price because the Musicaps are quite expensive.
0.47uF is too large for a bypass cap. Use 0.01uF. In my experience the
quality of the bypass cap isn't too critical but MKPs are relatively
expensive in this value.
>
> The cap on the midrange is still the original electrolytic. Its value
> is too big to be replaced by PP. Should I bypass this cap too?
Yes.
Stephen
Stephen McLuckie
> P.S. I understand that the frequency response will change with the
> addition of the bypass cap and I know how to correct it by changing
> the Lpad.
You cannot correct for increasing the value of a crossover cap by using an
L-pad. It's not that simple.
Stephen
Stephen McLuckie
> expensive in this value.
>
Should have been inexpensive, of course.
Stephen
Quan Tran
I thought that when the value of the tweeter cap is increase then the
xover point will be lower. In that case I can change the Lpad in away
that decreases the total impedance of the tweeter network so the xover
point will be put back upward. It is also necessary to keep the ratio
between the value of the serial resitor and the combined impedance of
the parallel tweeter and resitor the same so that the sentitivity
remain unchange.
By doing the above I think the xover point remains the same. The
voltage applied to the tweeter will not change either. Are there any
other problems that I may have?
"Stephen McLuckie" <stephen....@ntlworld.com> wrote in message news:<c66k4...@news4.newsguy.com>...
Quan Tran
Thanks for your reply. Actually my current 5.5uF cappacitors are MKP.
But I heard that film and foil capacitors are even better and want to
use those for bypassing the MKP. The Hovland Musicap 0.1uF costs $9
each.
What brand do you recommend for bypass caps?
"Stephen McLuckie" <stephen....@ntlworld.com> wrote in message news:<c66uf...@news3.newsguy.com>...
Norbert Hahn
>"Quan Tran" <quan...@hotpop.com> wrote in message
>news:okxhc.183689$JO3.105110@attbi_s04...
>
>> I have a pair of NHT speakers that have been modified quite a bit. I
>> fact I have replaced the tweeters and modified the crossover as
>> needed.
>>
>> I used SCR capacitors (French, 5.5uF) on the high pass filter. I am
>> now considering bypassing these SCRs with Hovland Musicap (0.47uF).
That would increase the total capacitance by roughly 9% and hence
lower the filter resonant frequency by about 4%.
>
>> I would appreciate if anyone could advise whether this upgrade worths
>> the price because the Musicaps are quite expensive.
>
>0.47uF is too large for a bypass cap. Use 0.01uF.
Hm, we're talking about tweeters. If the cross over frequency is about
2 kHz with the original 5.5 µF the additional 0.47 µF will have some
effect way above 2 kHz. Thus, 0.01 µF will have no effect at all.
Or did I overlook something?
Norbert
Stephen McLuckie
> Please correct me if I am wrong.
>
> I thought that when the value of the tweeter cap is increase then the
> xover point will be lower. In that case I can change the Lpad in away
> that decreases the total impedance of the tweeter network so the xover
> point will be put back upward. It is also necessary to keep the ratio
> between the value of the serial resitor and the combined impedance of
> the parallel tweeter and resitor the same so that the sentitivity
> remain unchange.
>
> By doing the above I think the xover point remains the same. The
> voltage applied to the tweeter will not change either. Are there any
> other problems that I may have?
Increasing the value of the cap will change the frequency response of the
crossover in the crossover area, which will be difficult to correct by
altering the values of the L-pad, which operates across the entire tweeter
frequency response.
Stephen
Stephen McLuckie
> Steve,
>
> Thanks for your reply. Actually my current 5.5uF cappacitors are MKP.
> But I heard that film and foil capacitors are even better and want to
> use those for bypassing the MKP. The Hovland Musicap 0.1uF costs $9
> each.
>
> What brand do you recommend for bypass caps?
I can only speak from my experience, and that is that the quality of the
bypass cap is not all that important. I normally use the same type as the
main capacitor.
You could always do that first to see what you think.
Stephen
Stephen McLuckie
> Hm, we're talking about tweeters. If the cross over frequency is about
> 2 kHz with the original 5.5 µF the additional 0.47 µF will have some
> effect way above 2 kHz. Thus, 0.01 µF will have no effect at all.
>
> Or did I overlook something?
No, you're right. The bypass cap is too small to affect the frequency
response. I'm sure the audio engineers on here will tell you exactly what it
does in due time. I read up on the theory a long time ago when I saw JBL
using the technique, but the memory of the exact details has faded.
Bypassing the caps with a 0.01 to 0.1 value brings a small improvement in
transparency for very little additional cost. That's my experience, at
least.
Stephen
Quan Tran
My speaker xover point is about 4000hz. The Lpad includes a 15.7ohm
resistor in parallel with the tweeter and then a 3 ohm resistor in
series.
IF xover at around 2000hz the cap would be 12uF with a Lpad of 10ohm
and 2.7ohm respectively. That's what I found in the schematic of the
NHTPro A10 which use the same tweeter.
I guess that the impedence of this tweeter is quite stable from 2 to
4khz (about 6 ohm).
Is it better if I change to a 6uF cap and the same Lpad as the A10,
i.e. 10 and 2.7 ohm? I guess with a smaller parallel resistor in the
Lpad, the total impedence of the tweeter network will be even more
stable. This also seems to be closer to the NHT's design for the A10.
normanstrong
> Steve,
>
> Thanks for your reply. Actually my current 5.5uF cappacitors are
MKP.
> But I heard that film and foil capacitors are even better and want
to
> use those for bypassing the MKP. The Hovland Musicap 0.1uF costs $9
> each.
>
> What brand do you recommend for bypass caps?
There's another possibility. You could leave it alone and enjoy your
speakers. If an added capacitor changes the sound, then you're
redesigning the speaker--probably in the wrong direction. If it
doesn't you will be wasting your time and money.
Norm Strong
Norbert Hahn
>I think I should have given more information.
>
>My speaker xover point is about 4000hz. The Lpad includes a 15.7ohm
>resistor in parallel with the tweeter and then a 3 ohm resistor in
>series.
I assume the those resistors are used to attenuate the output of the
tweeter. So the working range of the tweeter is roughly 3.5 kHz to
more than 20 kHz.
>
>IF xover at around 2000hz the cap would be 12uF with a Lpad of 10ohm
>and 2.7ohm respectively. That's what I found in the schematic of the
>NHTPro A10 which use the same tweeter.
I took 2 kHz as a number without knowing your speaker. Sometimes it is
better to use some numbers as an example rather than words only.
I acutally wanted to point out that a 5.6 湩 condensor works perfectly
well in the audio band. There's no reason for bypassing it with a smaller
condensor as the self inductance of 5.6 湩 at 20 kHz is some magnitude
smaller than the inductance of the voice coil.
>Is it better if I change to a 6uF cap and the same Lpad as the A10,
>i.e. 10 and 2.7 ohm?
What is the problem that you are trying to solve?
Norbert
Quan Tran
original. When I replaced the old tweeter I also redesigned the high
pass filter to suit the new one.
The old high pass filter had a 4.7uF cap and a 4ohm resistor in
series.
I I designed the new HPF without any quipment but I did refer to the
xover of the NHTpro A10 because they have the same tweeters xovered
at around 2000Hz.
I do listening test by comparing the sound of the speaker with my
Senheiser 414 headphone. With the current 5.5uF (3.3+2.2) cap, 3ohm
and 15.7 ohm Lpad, they sound quite close. I want to fine tune the
xover further. But this is difficult with our measurment equipment.
As the the Lpad on the A10 has 2.7 and 10 ohm resistors, is it a good
idea to copy this Lpad and change the capacitor value to arrive at a
4000hz xover point? (assume the same efficiency).
The second thing I want to do is to bypass the cap with a small good
quality cap (film and foil?) because many say it improve the sound.
What is your advice on this?
Norbert Hahn <ha...@hrzpub.tu-darmstadt.de> wrote in message news:<c6bvb...@news3.newsguy.com>...
> quan...@hotpop.com (Quan Tran) wrote:
>
> >I think I should have given more information.
> >
> >My speaker xover point is about 4000hz. The Lpad includes a 15.7ohm
> >resistor in parallel with the tweeter and then a 3 ohm resistor in
> >series.
>
> I assume the those resistors are used to attenuate the output of the
> tweeter. So the working range of the tweeter is roughly 3.5 kHz to
> more than 20 kHz.
> >
> >IF xover at around 2000hz the cap would be 12uF with a Lpad of 10ohm
> >and 2.7ohm respectively. That's what I found in the schematic of the
> >NHTPro A10 which use the same tweeter.
>
> I took 2 kHz as a number without knowing your speaker. Sometimes it is
> better to use some numbers as an example rather than words only.
> I acutally wanted to point out that a 5.6 µF condensor works perfectly
> well in the audio band. There's no reason for bypassing it with a smaller
> condensor as the self inductance of 5.6 µF at 20 kHz is some magnitude
Norbert Hahn
wrote:
>As mentioned in my first message, the current tweeters are not the
>original. When I replaced the old tweeter I also redesigned the high
>pass filter to suit the new one.
>
>The old high pass filter had a 4.7uF cap and a 4ohm resistor in
>series.
>
>I I designed the new HPF without any quipment but I did refer to the
>xover of the NHTpro A10 because they have the same tweeters xovered
>at around 2000Hz.
If I got that right you replaced a tweeter that covered 4 kHz to over
20 kHz with a different one that goes down to 2 kHz is a (slightly)
different box. Fine. The new tweeter will not have any problem with
the reduced bandwidth.
Of course you may alter the xover to make use of the extended
frequency range of the new tweeter. As I don't know your speakers
I cannot give you more specific tips.
>
>I do listening test by comparing the sound of the speaker with my
>Senheiser 414 headphone.
I'm somewhat sceptical in comparing the sound of a speaker under
construction with the sound of a headphone. I would rather modify
on speaker at a time and compare the modified one with the original,
using a mono signal of course.
>With the current 5.5uF (3.3+2.2) cap, 3ohm
>and 15.7 ohm Lpad, they sound quite close. I want to fine tune the
>xover further. But this is difficult with our measurment equipment.
You need a PC with a soundcard, some software as sound generator and
some software to analyze the results along with a good mic and a mic
amplifier. I use a Behringer ECM-8000 mic and the Behringer Eurack
UB 802 mixer as mic amp. Sound source is some pink noise generated
on a PC and for analysis I use Adobe Audition. There are less
expensive programs than that but I need it as audio editor and
multitrack mixer as well.
>As the the Lpad on the A10 has 2.7 and 10 ohm resistors, is it a good
>idea to copy this Lpad and change the capacitor value to arrive at a
>4000hz xover point? (assume the same efficiency).
Changing the capacitor will change the xover frequency. The
main reason for the resistors is the level match (attenuation) of the
tweeter with the midrange speaker.
>
>The second thing I want to do is to bypass the cap with a small good
>quality cap (film and foil?) because many say it improve the sound.
Bypass capacitors are needed for a large frequency range that extends
over several decades. I remember a project where I need a AC impedance
of less than 5 Ohm in the range from 10 Hz through 30 MHz. I used one
bypass cap for every other decade. The largest cap had to cover 10 Hz
to 3 kHz, the next one 3 kHz to 300 kHz etc.
>What is your advice on this?
Get a (5.6 µF) condensor of good quality and you're set.
Norbert