Replacing 4053 With…

[Newmarket]

Any thoughts on the switch debouncer mentioned earlier? Forget the diode, a 5V regulator would probably work better and switch common would be routed to the same ground as the fader.

That device appears to be a switch debounce device to debounce a switch that comes before it. And it's a logic level device rather than an analogue switch ? And it has 20ms delay.
Debouncing is not your problem here. The relay looks like your best and no compromise solution as I see it.
But if you can I'd still look at increasing the supply voltage on the 4053.

 

[JohnRoberts]

What problem are we trying to fix... Debouncing the logic signal can possibly reduce clicking but will not affect switch linearity.

JR

 

[Slime_Lord]

I need to not read data sheets when I first wake up . Realizing now the switch debouncer isn’t so much a switch as it is a filter/transient suppressor.

At the end of the day, I can edit out a pop or click in what’s recorded, My end goal is a more transparent audio path.

I’ll continue down the relay path and also check the DG411 series Wayne mentioned.

Again, appreciate all the insight into this!

 

[Newmarket]

I need to not read data sheets when I first wake up . Realizing now the switch debouncer isn’t so much a switch as it is a filter/transient suppressor.

At the end of the day, I can edit out a pop or click in what’s recorded, My end goal is a more transparent audio path.

I’ll continue down the relay path and also check the DG411 series Wayne mentioned.

Again, appreciate all the insight into this!

tbh I think you've already hit on a solution if you can fit a relay in there.
Depending on how much you're willing to mod / design in you could go for "traditional" jfet switching. But you need quite a large negative voltage to turn them off to "mixing desk" standard. But you can build in a ramp to minimise transients (pops) in the signal.

 

[Matt Syson]

I have a feeling that the 4053 effectively has schmitt trigger control port so slowing it down doesn't actually vary the switchng time but the resistor capacitor arrangement might actually be to either stop some 'bounce' when fed from a real switch or reduce capacitively coupled click because a fast edge control trace is near a audio virtual earth point. series switch and shunt to ground is very conveniently done with a 4053 and the DG 246? which was used by Audix in the 1980's . The DG chip being a dual channel single pole changeover Using 'usual' plus and minus 5 Volt rails like the DG411 or other devices being particularly handy. Fast switching is a mixed belessing anyway as audio switching really wants a 'fast ramp' because sharp edged cut of a signal below a round 10KHz tends to give a psychoacoustic 'click' as determined by the BBC and I presume many others years ago. And that assumes you haven't got any millivolts of DC offset on either side of the switch.

 

[rogs]

I remember reading the attached section of a page from the 1980 National Semiconductor Audio Data Book, which suggested that the structure of CMOS TGs - like the 4051/2/3 analogue switch - did not allow for totally silent switching (if that is important?)
I do remember finding the 4051/2/3 analogue switches very useful though, but I found it wise to use the split supply between Vdd and Vee, and switching between Vdd and Vss for maximum headroom.
As I recall, it was not a good idea to AC couple the signals in and out. They allowed audio signal bias to 'drift' away from the centre reference point, depending on the characteristic of the coupling caps.
Also if you do not use a split supply - and use VSS for both signal and switching ground - then you need to DC bias the signal to half rail, to get best results.
I always found the devices pretty clean, audio wise, if biased correctly..

 

[Slime_Lord]

tbh I think you've already hit on a solution if you can fit a relay in there.
Depending on how much you're willing to mod / design in you could go for "traditional" jfet switching. But you need quite a large negative voltage to turn them off to "mixing desk" standard. But you can build in a ramp to minimise transients (pops) in the signal.

More than likely grabbing a few relays, a DG411, and some prototype board to see what works best.

There’s a -17.5V rail close to that area of the board and winters in chicago last forever so, not totally opposed to exploring a JFET mute if I’m not satisfied with either choice!

 

[Brian Roth]

Look at the Panasonic T2Q series of relays. They are in a DIP-sized package. I first encountered those in an Amek 9098 desk ("by Rupert Neve the Designer") where they used them all over the place. Only issue is properly dealing with the current for the coil and adding snubber diodes, etc across the coil.

Bri

 

[Brian Roth]

Back in the early 80s I switched multiple channels of a console with a single logic signal using TG. The trick that I used to deliver very low distortion was to keep the TG inside the negative feedback loop (visualize the NF resistor connected back to the input side of the TG). I was unable to measure distorion using the bench technology of the day (Sound Tech).

JR

PS: You don't need 3 relays, the TG has connected three internal switches in parallel for lower on resistance.

I also had that same idea back in the 1980's! Not bragging....just an idea that struck me one day. Distortion measured with my Amber test set was essentially the same with or without the 4053 in the signal path. I only used it a few times for one-off/custom switchers for selecting multiple monitor speaker paths out of a desk.

Bri

 

[Newmarket]

More than likely grabbing a few relays, a DG411, and some prototype board to see what works best.

There’s a -17.5V rail close to that area of the board and winters in chicago last forever so, not totally opposed to exploring a JFET mute if I’m not satisfied with either choice!

If you can implement a relay then you don't need to look at alternatives imo. Obvs you might want to out of technical interest.

 

[Newmarket]

I also had that same idea back in the 1980's! Not bragging....just an idea that struck me one day. Distortion measured with my Amber test set was essentially the same with or without the 4053 in the signal path. I only used it a few times for one-off/custom switchers for selecting multiple monitor speaker paths out of a desk.

Bri

Can either of you post a schematic or sketch of this configuration. I'm having trouble visualising it. Thanks.

 

[Brian Roth]

Can either of you post a schematic or sketch of this configuration. I'm having trouble visualising it. Thanks.

A quick pencil scribble from 35+ year memory, but as far as the R values are concerned, it still makes sense to me. R2 and R3 in parallel is 10K. Hence unity gain (but inverted polarity) from in to out.

Bri

 

[Brian Roth]

Followup....in a later version, I used something like a 74HC4053 which somehow ended up slightly better in some form or fashion. LONG time ago....lol.

Bri

 

[Newmarket]

Followup....in a later version, I used something like a 74HC4053 which somehow ended up slightly better in some form or fashion. LONG time ago....lol.

Bri

Thanks. I was missing R2 in my understanding. I get it now.

 

[Brian Roth]

R2 is there to keep the opamp from going totally open loop/without NFB for the "nanosecond" when the TG is switching. My choices for R2 and R3 were what I had available on the bench when designing/testing. But since I was a 553x opamp fanboy back then, 110K for R2 worked out well.

R3 is the magic mojo since it wraps the TG inside the NFB of the opamp.

I first began fiddling with the 4016 (SPST) chip, but "off" attenuation wasn't great. I DID discover how to minimize distortion with the wrap-around NFB. The 4053 (SPDT) made the "off" almost solely dependent on things like circuit board layout.

Wiring all three sections of a 4053 in parallel never crossed my mind for extra "off" attenuation.

Bri

 

[JohnRoberts]

A quick pencil scribble from 35+ year memory, but as far as the R values are concerned, it still makes sense to me. R2 and R3 in parallel is 10K. Hence unity gain (but inverted polarity) from in to out.

Bri

I don't have a schematic either, but that is the general concept. Used inside the NF loop of a simple unity gain inverter stage the TG is passing minuscule current/voltage.

IIRC I used 4066, the lower on resistance version of the 4016. I was switching an entire console monitor system between tracking and playback using a single logic signal. I wrapped two pairs of input/feedback resistors around a single op amp in each channel. The TG connected the op amp's minus input to one of the two feedback networks. I probably used around 20k input/feedback resistors and a 1M or several hundred k "always there" feedback resistor with a small C across the op amp (I would have likely used a TL07x op amps). I probably ran the CMOS TGs from a simple +/- 7.5V rail.

To improve switch kill and suppress crosstalk, I used a second TG to alternately ground the feedback network node that was not active. Used this way, the TG distortion contribution was unmeasurable with bench equipment available to me 40+ years ago.

JR

PS: I have probably posted about this before.

 

[Doc Morgan]

I used the 405x data selectors in several designs by putting them in series with the inverting inputs of op amp stages. The somewhat sketchy Ron of the CMOS parts was swamped out by the input resistor of the inverting amp. They worked quite well. If any of you are familiar with the THD+N specs of the Valley gear, you can appreciate that most of them used that configuration in the audio chain without adding any significant distortion.

 

[thor.zmt]

Is it possible to switch the 4053 out for 3 small relays or is there a better solution/replacement for this chip?

First, used as shown, distortion will be low.

Second, with 5.6V PSU, check if the IC is in fact a 74HC4053 OR a HEF4053. If so, a 74HC4053 will have lower resistance and distortion. There may be even lower resistance and distortion versions limited to 6V.

Third, you could make a small PCB and replace 4053 with a pair of J-Fet's (e.g. MMBF4393).

Designing in a small relay is also possible, but be aware that relays have a relatively long on/off timing, which may be a problem. Also relay coils need a fairly high drive power, compared to MOS IC's. So you need power and a driver transistor.

Thor

 

[Slime_Lord]

First, used as shown, distortion will be low.

Second, with 5.6V PSU, check if the IC is in fact a 74HC4053 OR a HEF4053. If so, a 74HC4053 will have lower resistance and distortion. There may be even lower resistance and distortion versions limited to 6V.

Third, you could make a small PCB and replace 4053 with a pair of J-Fet's (e.g. MMBF4393).

Designing in a small relay is also possible, but be aware that relays have a relatively long on/off timing, which may be a problem. Also relay coils need a fairly high drive power, compared to MOS IC's. So you need power and a driver transistor.

Thor

They most definitely are HEF4053B’s. Appreciate the suggestion for the other chip but I’m hesitant to try another CMOS switch in their place.

I’ve removed/shorted the 4053 on the first 6 channels as well as one of the group outputs and have been a/bing all week between the channels/outputs with and without. The mono input channels have more headroom and the EQ isn’t muddying up the signal when the low/high shelf bands are boosted.

Summing has improved when going through the shorted group outputs as well. There’s a significant improvement with low/low-mid response, signal is no longer fighting for space in this frequency range and I can hear more subtle detail in the mix.

The improvements are a combination of multiple changes though, I do want to point that out. All electrolytics have been replaced, the ceramic caps in the eq section (including the 2 SMT 3.3nF on the sub-module) replaced with higher quality film, etc.

I ordered 4 relays to test with and they showed up yesterday. Going to take a break from working on the console this weekend but will see where I land this week with the relays.

 

[thor.zmt]

They most definitely are HEF4053B’s.

These have quite high resistance, especially at 5.6V.

I ordered 4 relays to test with and they showed up yesterday. Going to take a break from working on the console this weekend but will see where I land this week with the relays.

If you use relays, you can remove the entire circuit starting at C29 and go directly to the output (labelled "pre" in your schematic).

All of it becomes superfluous.

Thor