CD4016 bi lateral switches.

[Rob Flinn]

Hi

I've just been fixing an old Soundcraft desk. This has the CD4016 quad bilateral switches in it.

I was wondering whether they are any good ? I've never come across them before & am trying to get a handle on whether there is any degradation in audio quality once a signal has passed through them.

I found this page which has some application info on these & other bilateral switches. http://members.shaw.ca/roma/switching.html

 

[JohnRoberts]

They are generally not very low on impedance, and that impedance varies with terminal voltage so you need to understand their limitations. There are ways to use them that mitigate their weaknesses.

First look at their data sheets, and then app notes.

JR

 

[clintrubber]

If they're in a circuit where lowering the value of R_ON is relevant you could try replacing with '4066, which is pin compatible.

I'm not aware of any drawbacks by doing so, yet there may be, otherwise I can't imagine why the 4016 lived on after the 4066 appeared.

Regards,

Peter

 

[SSLtech]

...Or design a piggy-back board using a couple of SSM2402's...

It's what I'd be thinking of... IIRC, Soundcraft used the CD4016 for passing the control room monitor signal through at all times and enabling 'dim' and 'mute' functions through a series of them, depending on the model.

Before now I have replaced them with relays, and made the piggy-back board plug in through a 14-pin socket which replaced the 4016... but nowadays I'd probably go the 2402 route instead.

Keith

 

[dale116dot7]

The 4016's do not have buffers on the digital inputs so you can kinda 'soft switch' them a bit. I think I would prefer something else but they seem to work ok in my 2400 so I'm leaving them for now.

-Dale

 

[bcarso]

There's a thread I started in here discussing 4016 parts I believe, where I was mostly concerned about driving an "off" switch with voltages higher in magnitude than Vdd or Vss. But I believe there is some discussion of the distinction between 4016 and 4066.

The reason the parts are still desirable in some applications is the much lower transient induced into the channels of the switch FETs during switching. The linearization parts in the 4066 have a nasty way of coupling something to one rail or the other during switching. It can put a brutal transient in the audio, even when you switch in tens of nanoseconds. If you are using the parts for pwm or switching modulators, etc., it renders them virtually unusable many times.

Otherwise, if you seldom switch, and if the impedances are reasonably low, the "on" switch variation in resistance versus voltage is much smaller in 4066, hence the distortion is also lower, making it the better part.

By the way, the system using the 4016 parts to switch among various EQs for different mounting positions of speakers in a motorized screen assembly, is finally shipping.

 

[bcarso]

[quote author="dale116dot7"]The 4016's do not have buffers on the digital inputs so you can kinda 'soft switch' them a bit. -Dale[/quote]

There is however one inverter driving one of the transmission transistors, so the two paralleled switches will not switch together if the control line is slowed down a lot. If the slow on time of the direct-coupled switch is helping to establish the d.c. potential in a circuit, it's possible that the slow control signal could help a lot anyway.

And I guess, as there is no schmitt-trigger action in that one inverter, you could slow the drive signal down enough so even the inverter-driven device is turning on arbitrarily slowly. To the extent to which sections within a package match, you could contrive a closed-loop driver with a dummy switch and some other stuff.

 

[bitman]

The dynamic range is limited with those buggers as I recall. It's been a while since I did anything with them however.

 

[bcarso]

B series CMOS is good for about 18V p-p swings. The on resistance of the switches gets smaller at the higher voltages and really bad at low ones.

It also tends to peak at about halfway between Vdd and Vss.

The HC versions of the 4066 (no comparable parts for 4016 exist AFAIK) do about 12V p-p in the switch channel, and have both more constant and quite a bit lower Ron. But as mentioned, put nasty spikes of current into the signal path when switching.

 

[JohnRoberts]

The unmeasurable distortion was also my experience back when I used them that way in an old Loft console back in the '70s. While unmeasurable was perhaps an easier threshold back then... :roll:

JR

PS: I think I did a trick where I shorted the unselected input pair to ground to keep crosstalk inside the switch down. Ground layout also makes a difference with that current dumped into it.

 

[bcarso]

[quote author="mediatechnology"]
Brad: I had to use +/-7.5V for the CMOS rails and the off-state shunt does see far more than Vdd and Vss though it's current-limited by a 10K. What did you determine from when you last asked that question?[/quote]

I couldn't take any chances, so I diode-clamped the open switch ends to +/- 7.5V. I think they would have been fine without, but this was Epson :roll:

 

[JohnRoberts]

Better safe than sorry...

Little things like that can bite you...

JR

 

[bcarso]

[quote author="JohnRoberts"]Better safe than sorry...

Little things like that can bite you...

JR[/quote]

This was after trying very hard without "success" to induce the dreaded parasitic SCR action in a number of samples BTW. But one does not have control over what may be installed in China. If it is a counterfeit part that's another story---bad news but no engineer's culpability at least.

That reminds me of another kink in the 4016 versus 4066 story. For a while some not-too-prescient manufacturers were putting 4066 chips inside 4016-marked packages. Since they exceeded the performance of the 4016 in various static parameters they thought it would be o.k. Suddenly, many pieces of equipment had big problems, and people were having to re-qualify various vendors. Definitely a quality assurance parable and object-lesson. For another good read see Tenner's Why Things Bite Back: Technology and the Revenge of Unintended Consequences.

 

[Salvatore]

Here are some schematics I found,,

http://www.doctronics.co.uk/4016.htm

http://www.geocities.com/thetonegod/switches/switches.html

http://www.geocities.com/j4_student/dldbypass.gif

I wonder if they perform well..

 

[bcarso]

[quote author="Salvatore"]Here are some schematics I found,,

http://www.doctronics.co.uk/4016.htm

http://www.geocities.com/thetonegod/switches/switches.html

http://www.geocities.com/j4_student/dldbypass.gif

I wonder if they perform well..[/quote]

The R. G. Keen article linked in the third paragraph of the second reference has some useful stuff.

 

[JohnRoberts]

[quote author="Salvatore"]Here are some schematics I found,,

http://www.doctronics.co.uk/4016.htm

http://www.geocities.com/thetonegod/switches/switches.html

http://www.geocities.com/j4_student/dldbypass.gif

I wonder if they perform well..[/quote]

I see no attempt to mitigate the shortcomings we noted.

Maybe do a search here... we have discussed these before.

JR

 

[gar381]

Can any body tell me the difference between CD4016BC and
CD4016BEE4 ? They are both 14-DIP package so its not a
package difference. The data sheets seem to be no help here.
Can I use them interchangeably ??

Thanks guys

GARY

 

[JohnRoberts]

C is typically commercial grade, EE may higher temp? The data sheet should say.

JR

 

[gar381]

Thanks John and Wayne

I got a few of the BEE4 versions and I am doing a project that
requires some 4016s Sounds like these will work as I can't
find any difference in the data sheets. I've just never heard of that version before. :?

GARY