Hi,
has anyone considered TI's (BB's) MPC506/507 (16channel single ended/8 channel symmetric)?
It has "break before make" (click-free) circuitry and runs from +-15V. I guess it's thought as an alternative for the established SSMs.
I'm going to receive some soon, but it'l take a while for me to test them. So this is probably more a request than a suggestion...
Sebastian
PS: Usability of the 4066 in audio also depends on the used voltage levels, thus on the used input sources... Not that I like 4066s with low voltage levels, but I think 4066s are a no-no with higher (dynamic) levels or when you connect unknown sources (aka not build by yourself
) ...
has anyone considered TI's (BB's) MPC506/507 (16channel single ended/8 channel symmetric)?
It has "break before make" (click-free) circuitry and runs from +-15V. I guess it's thought as an alternative for the established SSMs.
I'm going to receive some soon, but it'l take a while for me to test them. So this is probably more a request than a suggestion...
Sebastian
PS: Usability of the 4066 in audio also depends on the used voltage levels, thus on the used input sources... Not that I like 4066s with low voltage levels, but I think 4066s are a no-no with higher (dynamic) levels or when you connect unknown sources (aka not build by yourself
) ...
mux stuff
OK John, I'm with you.
You are right. But as with many things, also here you get conflicting requirements. Making the series resistance too high gets you into noise country. As always, good engineering is the art to choose between several evils, correctly. But you know all this.
Jan Didden
OK John, I'm with you.
You are right. But as with many things, also here you get conflicting requirements. Making the series resistance too high gets you into noise country. As always, good engineering is the art to choose between several evils, correctly. But you know all this.
Jan Didden
Re: mux stuff
I'm still missing it.
Let me redraw as shown below.
The input signal at Vin is the same for both.
With the same input signal to both, which switch has the most current throught it and the most voltage across it (consequently, having its Ron modulated the most)? I understand the switch at virtual ground has one end "fixed" at zero. The other switch has both ends moving, but I think the actual swing "across the switch" itself is less?
Be gentle, it's been a rough week
mlloyd1
I'm still missing it.
Let me redraw as shown below.
The input signal at Vin is the same for both.
With the same input signal to both, which switch has the most current throught it and the most voltage across it (consequently, having its Ron modulated the most)? I understand the switch at virtual ground has one end "fixed" at zero. The other switch has both ends moving, but I think the actual swing "across the switch" itself is less?
Be gentle, it's been a rough week
mlloyd1
janneman said:
Attachments
Yes, nearly Friday....
Hi Mlloyd1,
Although I have stated that I favour circuit 2, I agree that resistance modulation is probably worse. What I've tried to get over, is that it's effects can be made so small by a careful choice of series resistor, and chosing a reasonable switch IC.
My reasons for going the circuit 2 way, are largely based on it's better crosstalk performance.
Cheers,
Hi Mlloyd1,
Although I have stated that I favour circuit 2, I agree that resistance modulation is probably worse. What I've tried to get over, is that it's effects can be made so small by a careful choice of series resistor, and chosing a reasonable switch IC.
My reasons for going the circuit 2 way, are largely based on it's better crosstalk performance.
Cheers,
Re: Re: Yes, nearly Friday....
If we say that most of the crosstalk occurs inside the analogue switch, then in circuit 1, capacitive coupling will be greater, due to the higher impedance.
My measurements have agreed with this, though I'm sure there are other factors that I've not thought of, or cannot recall.
An admission I must make, is that I was also concerned with "mixing effects". In other words, it mattered to me that Input A didn't feed back up into Input B. This was because they also fed other switches. Virtual eath systems are good at this.
Cheers,
This is my theory:mlloyd1 said:
If we say that most of the crosstalk occurs inside the analogue switch, then in circuit 1, capacitive coupling will be greater, due to the higher impedance.
My measurements have agreed with this, though I'm sure there are other factors that I've not thought of, or cannot recall.
An admission I must make, is that I was also concerned with "mixing effects". In other words, it mattered to me that Input A didn't feed back up into Input B. This was because they also fed other switches. Virtual eath systems are good at this.
Cheers,
sek said:
Hi Sebastian,
I missed your post yesterday...
I've used the '506. The performance is good. However, I moved to DG211's because the board layout is easier, and more flexibility in address decoding, for my applications.
With all these devices, I recommend running at maximum supply rails +15 -15v.
Cheers,
My design...
Hi everybody, and thanks a lot for your input!
I've now decided to go for the "virtual ground switching"
solution, since the benefits seems to be much greater that
the disadvantages.
I'll use 10 kOhms input resistors for each input, together
with 2x1N4148 to "cut" the signals that are not selected
(found that in the SSM2404 datasheet). The RDS(on) for
the ADG406 switch seems extremely flat if it's used in the
mV range, and It'll run on +/- 18V to have the lowest
possible RDS(on).
sek: I also have two MPC506 switches to try out, but just
looking at the DS they don't seem to have the same
performance as the ADG406. (The MPC506:es are pin
compatible with the '406:es.)
Now I just have to drill the PCB...
Again, thanks everybody for your input!
/ Per
Hi everybody, and thanks a lot for your input!
I've now decided to go for the "virtual ground switching"
solution, since the benefits seems to be much greater that
the disadvantages.
I'll use 10 kOhms input resistors for each input, together
with 2x1N4148 to "cut" the signals that are not selected
(found that in the SSM2404 datasheet). The RDS(on) for
the ADG406 switch seems extremely flat if it's used in the
mV range, and It'll run on +/- 18V to have the lowest
possible RDS(on).
sek: I also have two MPC506 switches to try out, but just
looking at the DS they don't seem to have the same
performance as the ADG406. (The MPC506:es are pin
compatible with the '406:es.)
Now I just have to drill the PCB...
Again, thanks everybody for your input!
/ Per
Hi Mlloyd,
you are right, #1 has little variation of voltage ACROSS the switch, the residual voltage depending only on the ratio of r_on to 1 Meg. However, there is a huge voltage variation against the switch substrate. This will to some degree influence r_on (which is only relevant in terms of delta r_on / 1 Meg, i.e. pretty irrelevant), but it is bound to influence the capacitance to substrate or ground. This is why this circuit should only used with very low source impedances.
Circuit #2 has virtually no voltage and hence capacitance variation against ground, but the voltage varation will be higher than in #1 due to the higher current.
The best way, IMHO, is what I called #3, i.e. to place the switch between the inverting input and the point where input and feedback resistor meet. This way, there will be no current and no voltage variation. If you are concerned about the residual (static) capacitance on the input, use a 100 R isolation resistor in series.
Regards,
Eric
you are right, #1 has little variation of voltage ACROSS the switch, the residual voltage depending only on the ratio of r_on to 1 Meg. However, there is a huge voltage variation against the switch substrate. This will to some degree influence r_on (which is only relevant in terms of delta r_on / 1 Meg, i.e. pretty irrelevant), but it is bound to influence the capacitance to substrate or ground. This is why this circuit should only used with very low source impedances.
Circuit #2 has virtually no voltage and hence capacitance variation against ground, but the voltage varation will be higher than in #1 due to the higher current.
The best way, IMHO, is what I called #3, i.e. to place the switch between the inverting input and the point where input and feedback resistor meet. This way, there will be no current and no voltage variation. If you are concerned about the residual (static) capacitance on the input, use a 100 R isolation resistor in series.
Regards,
Eric
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.