It depends on the input impedance of the equipment that the capacitor is working into and also your own target value for how extended you want the low frequency response to be . The capacitor and the input 'resistance' form a first order high pass filter.
Fig #1 here
https://en.wikipedia.org/wiki/High-pass_filter
If the cap is to small then the low frequencies are attenuated.
Fig #1 here
https://en.wikipedia.org/wiki/High-pass_filter
If the cap is to small then the low frequencies are attenuated.
FWIW ... I have a lot of the way better SSM analog audio switching ICs. I think it are the 2402 and 2404. In all the situations where I see those dreaded HEF405x I replace the whole shebang. Whatever the reasons are, they are simply not suitable for quality audio but they are cheap so one sees them mainly in British audio. Of course now the rare exceptions will be mentioned where brand X that costs 3000 Euro uses them etc. Well, replace them for something better (relays, switches, SSM IC's, all are better) and one will notice. I can not even imagine someone wanting to use these for audio voluntarily.
Last edited:
Sadly unobtanium.
They were ramp specific (ie slow and slower) mute switches developed for broadcast use. They had the series/shunt architecture built in and the rampimg prevented any mute artifact.
The problem with much of the SSM stuff was that there was (generally) no 2nd source. No good for manufacturers, irrespective of the benefits.
Also, none of this is answering the original question!!
As said, I have these in stock. More than I need. Maybe the OP has other parts in abundance?! Imagine trading parts.
AFAIK there was not 1 question but several. Fact is HEF405x are not suitable for quality audio which could be a real showstopper. Any attempt in that direction will be worse than simple switches or relays, SSM IC's etc. Why choose a direction that is worse than the average method? Because of the apparently important cheap factor? To burn time and discover they sound and measure so so? So we are supposed to give advice on non quality audio? To recommend HEF405x because there are second sources? To DIYers that are generally used to use obsolete parts too 🙂
The more recent electronic source switching IC's are also better than HEF405x but most are no competition to classic/simple stuff. If OP wants quality I donate quality relays for shipping costs.
AFAIK there was not 1 question but several. Fact is HEF405x are not suitable for quality audio which could be a real showstopper. Any attempt in that direction will be worse than simple switches or relays, SSM IC's etc. Why choose a direction that is worse than the average method? Because of the apparently important cheap factor? To burn time and discover they sound and measure so so? So we are supposed to give advice on non quality audio? To recommend HEF405x because there are second sources? To DIYers that are generally used to use obsolete parts too 🙂
The more recent electronic source switching IC's are also better than HEF405x but most are no competition to classic/simple stuff. If OP wants quality I donate quality relays for shipping costs.
Last edited:
In stock. The bit I chose not to read!
I wasn't recommending the use of one over another based on 2nd sourcing. But that is the main reason console manufacturers did not use the excellent PMI/SSM stuff.
Back to the original thread:-.
Ulricmuc. You will spend days faffing around trying to get this to work in your original design brief.
You need to look at DG507 and its equivalents (Max307 etc.). These are application specific dual 8-wide TTL compatible analog switches.. Yes they are nearly £20 each, but they do the job much better than your choice. Yes, you can do what you want with 405x chips of various types, but you appear to be on a brickwall learning curve as it is, so make it a bit easier.
I wasn't recommending the use of one over another based on 2nd sourcing. But that is the main reason console manufacturers did not use the excellent PMI/SSM stuff.
Back to the original thread:-.
Ulricmuc. You will spend days faffing around trying to get this to work in your original design brief.
You need to look at DG507 and its equivalents (Max307 etc.). These are application specific dual 8-wide TTL compatible analog switches.. Yes they are nearly £20 each, but they do the job much better than your choice. Yes, you can do what you want with 405x chips of various types, but you appear to be on a brickwall learning curve as it is, so make it a bit easier.
There are some really good analog switch parts coming out, unfortunately they are getting really hard to solder. MAX20336 MAX20337 will allow -5.5v - +5.5v (11v peak-to-peak swings) with an on resistance of 0.19ohm and -110dB THD+N even into a 32ohm load. Unfortunately it's a 6 pin BGA part.
I changed my design to use 6 double-pole relays from Zettler electronics @ €1.49 each (about $1.60). This works fine for me,
but I am not an HiFi expert and my ears are 80 years old.
What could be better if you pay $100 to $200?
A relay is a relay is a relay: a coil and a contact.
Dear @Mooly , I know this post is quite old but I am also looking for a reliable way to switch audio signals. I was wondering, whet´s the purpose of the double JFET? Could it be sufficient using only one of them? Or you get better attenuation with this topology?
@xenon789
The second FET greatly improves the 'off' isolation. Look at this where the FET on resistance is set at a highish 100 ohm and the off resistance at 30,000,000 ohm. The 5pF is an attempt at adding the stray junction capacitance across the FET that is off.
In practice the isolation was so good that even a signal generator delivering a high voltage squarewave was totally inaudible even with the volume on full. The FET's I uses had an on resistance of around 15 ohms from memory (so much better than in this sim). I've set the off resistance to 30meg ohms but at low frequency it will be much higher.
Look at the scale at the left.
The second FET greatly improves the 'off' isolation. Look at this where the FET on resistance is set at a highish 100 ohm and the off resistance at 30,000,000 ohm. The 5pF is an attempt at adding the stray junction capacitance across the FET that is off.
In practice the isolation was so good that even a signal generator delivering a high voltage squarewave was totally inaudible even with the volume on full. The FET's I uses had an on resistance of around 15 ohms from memory (so much better than in this sim). I've set the off resistance to 30meg ohms but at low frequency it will be much higher.
Look at the scale at the left.
Hi @Mooly, thanks, that makes sense. I am working on a dj mixer and there are two sections where I need JFETS: CUE and Filter. For CUE it is acceptable to have a pop noise when a channel is activated, not for the filter though. So far I have been experimenting with one single JFET in current mode but I did not achieve the result I wanted. My simulation is very simple, voltage follower and inverting opamp with the p-channel JFET in the middle. In the plot below, the blue curve is the output and the green is the Vg voltage. The fading in is asymmetrical, when the input voltage goes more negative, the source voltage goes also more negative and the resulting > Vgs bring the JFET to cut off (1.82V for this j176). Now, I dont know how this would sound but I am pretty sure it is not going to be a smooth transition,,,the only way to fade signals with JFET is to use them in their ohmic region. Does this make sense to you?
That all looks and sound very familiar and so I don't think you will achieve a clean fade with that configuration.
My final design for the input switching gives a noise free switch although if there is signal present at the inputs and the volume is turned up you can hear the distortion (the sort of distortion you show in the sim) during that few 10's of milliseconds as the FET gate voltage ramps up. My gate resistors were 1meg from memory.
If you use the FET as a shunt in its linear region you will never achieve a high 'off' value and the essential series FET (the one in your sim) is never going to pass the signal in a linear way during the transition.
My final design for the input switching gives a noise free switch although if there is signal present at the inputs and the volume is turned up you can hear the distortion (the sort of distortion you show in the sim) during that few 10's of milliseconds as the FET gate voltage ramps up. My gate resistors were 1meg from memory.
If you use the FET as a shunt in its linear region you will never achieve a high 'off' value and the essential series FET (the one in your sim) is never going to pass the signal in a linear way during the transition.
what is weird is that until Vgs gets closer to Vgs(off) I do not see any fading ... @Mooly I am thinking about changing completely approach with this. When you start adding JFETs and the electronics around it makes sense to evaluate another option: use VCA such as the SSM2164 as analog switches. Very simple, you create two complementary control voltages, the signal you want to select will get 0V at the VCA control pin, the other 3.3V (100dB attenuation).
Absolutely. The fade requirement is a totally different ball game to just want click free switching.