post fader amplifier questions

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Now that I know that I could use a single pole switch in series after the fader, I think I will use single DG468 for each channel. It’s super cheap (about 70 cents) and tiny (TSOP package). And it can be powered with +/-15V
Ther are some differences between the DG44X and the ADG40x in terms of distortion and "offness". You may want to evaluate them and not base your selection only on price and bulk.
 
Why would you want a different supply for the AG442? It works perfectly up to +/-22V.
yes, you can, i choose the wrong example, should have used CD405x.

But, it works pretty well.
the main power rail never changes in amplitude, it has more force than the ic's in the circuit.

The CD405x series do the job very good or just as good as the DG series, unless you use a >10k us$ scope.
You can switch it with more than 1 mHz.

eurorack - switch.jpeg
That's the eurorack module, i made using this technique.
It's awesome.
 
It's been awhile since I fiddled with this, but I recall decent "offness" and essentially zero distortion with the switching element inside the feedback loop.

Bri
Yes, it was more or less mandatory with the early CMOS switches, which had relatively high residual resistance, relatively low offness, and particularly couldn't cope with the audio level, being limited to +/-7V (and then with significant distortion).
 
The CD405x series do the job very good or just as good as the DG series, unless you use a >10k us$ scope.
How do you cope with the fact that the CD series can handle only about +/-7V, when the level in a decent mixer can reach twice that?
It can be done only by changing the structure of the circuit, which I don't think is what the OP wants.
 
All I can say about the design I concocted years ago is that it achieved very respectable performance...although required an inverting opamp stage for that result.

Bri
 
All I can say about the design I concocted years ago is that it achieved very respectable performance...a
My comment wasn't meant to diss your circuit. It was just a comment that because of the intrinsic limitations of the CD40XX series, it needed a specific arrangement.
lthough required an inverting opamp stage for that result.

Which is not very convenient for a post-fader amp, where polarity ("phase") preservation is desirable. It takes only an additional inverter.
 
How do you cope with the fact that the CD series can handle only about +/-7V, when the level in a decent mixer can reach twice that?
It can be done only by changing the structure of the circuit, which I don't think is what the OP wants.

i used the opamp circuit shown in a previous post to drive a CD4053 on +10V/-10V.
Works great.

I'm also in the process of designing a DIY mixer, didn't know that audio level was higher than line level inside a mixer.
But I'm thinking of converting audio in 10 V.p.p eurorack level.
I did however hear about a mixer that used 100 V.p.p. or so.

datasheet CD405x:
Wide Range of Digital and Analog Signal Levels
– Digital: 3 V to 20 V
– Analog: ≤ 20 V P-P
Low ON Resistance, 125 Ω (Typical) Over 15 V P-P
Signal Input Range for V DD – V EE = 18 V
High OFF Resistance, Channel Leakage of
±100 pA (Typical) at V DD – V EE = 18 V

applications:
– Analog and Digital Multiplexing and Demultiplexing
– A/D and D/A Conversion
– Signal Gating
– Factory Automation
– Televisions
– Appliances
– Consumer Audio
– Programmable Logic Circuits
– Sensors
 
Ther are some differences between the DG44X and the ADG40x in terms of distortion and "offness". You may want to evaluate them and not base your selection only on price and bulk.

thanks for pointing that out..I didn't know that ADG and DG were different.
comparing the datasheets I noticed a difference of about 20dB in offness (80dB off isolation on ADG419, 60dB on DG468, is this the value I need to check for offness?). But I can't understand how to compare distortion :-(

I only found some info on the ADG1201 datasheet:
Total Harmonic Distortion + Noise (THD + N) = 0.15 % typ
 
Last edited:
Which is not very convenient for a post-fader amp, where polarity ("phase") preservation is desirable. It takes only an additional inverter.

How about a non inverting opamp configuration?
I always found than kinda strange and the the resistors if not matched would or could offset the output.
So i kept using inverting config but requiring a second to set things straight again.
Are non inverting config usable? i saw them used in some datasheets of DAC's.

Op-amp_Non-Inverting_Amplifier-2951746236.png
 
i used the opamp circuit shown in a previous post to drive a CD4053 on +10V/-10V.
Works great.

I'm also in the process of designing a DIY mixer, didn't know that audio level was higher than line level inside a mixer.
The old rule of thumb was to offer at least 20dB headroom between the nominal operating level and the maximum possible level (clipping).
For circuits powered with +/-17V, it means nominal operating level set at about +1.1 dBu.
For convenience, many designers chose to operate at -2dBu, which results in about 23dB headroom, and gives a neat factor 2 for outputting the reference level of +4dBu.
But I'm thinking of converting audio in 10 V.p.p eurorack level.
10Vp-p nominal or max?
I did however hear about a mixer that used 100 V.p.p. or so.
100Vp-p is 35Vrms. It's way too much for any receiver (converter, tape machine, power amp) so it just wastes energy, unless there are step-down transformers. It is a good point for teh mktg dept, though.
datasheet CD405x:
Wide Range of Digital and Analog Signal Levels
– Digital: 3 V to 20 V
– Analog: ≤ 20 V P-P
Low ON Resistance, 125 Ω (Typical) Over 15 V P-P
Signal Input Range for V DD – V EE = 18 V
High OFF Resistance, Channel Leakage of
±100 pA (Typical) at V DD – V EE = 18 V
It's well-known that the gates in the CD40xx series have significant variation of Rds on with voltage, which results in distortion. that's why the series-shunt applications were preferred.
The new ones ADG4xx are way better. More expensive yes, but easier to operate.
 
Last edited:
The old rule of thumb was to offer at least 20dB headroom between the nominal operating level and the maximum possible level (clipping).
For circuits powered with +/-17V, it means nominal operating level set at about +1.1 dBu.
For convenience, many designers chose to operate at -2dBu, which results in about 23dB headroom, and gives a neat factor 2 for outputting the reference level of +4dBu.

I'm not really used to look or watch in dB, log scales, etc, i just look at the absolute voltage in a Synthesizer witch is 10 V.p.p.
the opamps go on +12V/-12V, so i got like close to 7 v headroom for a sound or like 2 V when it's a gate signal.

10Vp-p nominal or max?

100Vp-p is 35Vrms. It's way too much for any receiver (converter, tape machine, power amp) so it just wastes energy, unless there are step-down transformers. It is a good point for teh mktg dept, though.

It's well-known that the gates in the CD40xx series have significant variation of Rds on with voltage, which results in distortion. that's why the series-shunt applications were preferred.
The new ones ADG4xx are way better. More expensive yes, but easier to operate.


I'm also thinking about using these:
74LVC1G3157 (6 Ω (typical) at V CC = 5 V)
74LVC1G66 (6 Ω (typical) at V CC = 5 V)

Tiny, Litle logitc ic's and driving them on a +2.5V/-2.5V psu.
Line level should be like 1 V.p.p

easy to be soldered packages, pretty much the latest on ic biz.
 
I'm not really used to look or watch in dB, log scales, etc, i just look at the absolute voltage in a Synthesizer witch is 10 V.p.p.
Mixers and synths are very different. In a synth, the max level is constrained by the supply rails and the sound generators tend to operate with similar levels, close to the max amplitude possible with these rails.
In a mixer, it's very different; signals are impredictable, with a potentially very large crest factor.
Look at the transient response of a mic'd snare drum; you'll see that the initial transient, which last just a few ms, is about 10 times higher than the rest that follows.
That's why headroom is needed.
Line level should be like 1 V.p.p
Nominal so-called "pro" line level is +4dBu (1.228Vrms or 3.74Vp-p). With typical +/-17V rails, it leaves about 18dB headroom, which is considered a little short according to "pro" standards, but is generally OK. In order to provide more than 20dB headroom, many designers chose to operate at a nominal level of -2dBu, which just required gain of 2 or attenuation of 0.5 to interface with the rest of the world.
Now you may want to operate at the "Tascam" level, which is -10dBv (db referred to 1V), which is 0.316Vrms or 0.89Vp-p. If you want 20dB headroom, you need only +/- 5-6V supply rails.But then, you may have difficulties interfacing with some converters, that operate at higher level.
For example RME converters, on their highest sensitivity, require +13dBu for 0dBfs, that's 4.5Vrms or 12.6Vp-p, which suggests power rails of about +/- 7-8V minimum.

I suggest you read about level diagrams in mixers.
https://mega.nz/file/zPYiQZhS#-Yz3qjowAdogxoQYw5WlZPoFN_MLsAK-Ql7bfVy3d5ghttps://groupdiy.com/threads/steve-dove-designing-a-professional-mixing-console.77353/
 
Mixers and synths are very different. In a synth, the max level is constrained by the supply rails and the sound generators tend to operate with similar levels, close to the max amplitude possible with these rails.
In a mixer, it's very different; signals are impredictable, with a potentially very large crest factor.
Look at the transient response of a mic'd snare drum; you'll see that the initial transient, which last just a few ms, is about 10 times higher than the rest that follows.
That's why headroom is needed.

Nominal so-called "pro" line level is +4dBu (1.228Vrms or 3.74Vp-p). With typical +/-17V rails, it leaves about 18dB headroom, which is considered a little short according to "pro" standards, but is generally OK. In order to provide more than 20dB headroom, many designers chose to operate at a nominal level of -2dBu, which just required gain of 2 or attenuation of 0.5 to interface with the rest of the world.
Now you may want to operate at the "Tascam" level, which is -10dBv (db referred to 1V), which is 0.316Vrms or 0.89Vp-p. If you want 20dB headroom, you need only +/- 5-6V supply rails.But then, you may have difficulties interfacing with some converters, that operate at higher level.
For example RME converters, on their highest sensitivity, require +13dBu for 0dBfs, that's 4.5Vrms or 12.6Vp-p, which suggests power rails of about +/- 7-8V minimum.

I suggest you read about level diagrams in mixers.
https://mega.nz/file/zPYiQZhS#-Yz3qjowAdogxoQYw5WlZPoFN_MLsAK-Ql7bfVy3d5ghttps://groupdiy.com/threads/steve-dove-designing-a-professional-mixing-console.77353/

I can convert or attenuate the amplitude before it enters the mixer and again when it leaves, just another set of opamp and/or switches. I have a RME digi96/8 pad pci card in my g4 powermac.

Cool, HI506 analog multiplexer from the document on page 27 still is being made and even shows the "shunt" mute circuit using a CD4066, fig 28C on page 26.
https://www.renesas.com/us/en/document/dst/hi-506-hi-507-hi-508-hi-509-datasheet
Dual Anti Log pots are not easy to come by on page 42, EQ section.
Page 12, fig 8 and again on page 47, fig 56 why would i use such transistor output amp?

Very cool document although not always clear hat ic are used, like page 62, fig 72.

These
74LVC1G3157 (6 Ω (typical) at V CC = 5 V)
74LVC1G66 (6 Ω (typical) at V CC = 5 V)
are tiny and cheap, used on car, phones and tableds.
 
I can convert or attenuate the amplitude before it enters the mixer and again when it leaves, just another set of opamp and/or switches.
Just consider that the lower the nominal operating level, the higher the relative noise.
Cool, HI506 analog multiplexer from the document on page 27
In the Steve Dove document, the HI506 is used as a demux, routing one signal to one in 16 positions.
He does not suggest using it for the Mute function.
A simple on/off CMOS switch would be enough.
HI506 is a venerable circuit,, having been designed about 25 years ago.
The performance of more recent gates outperforms it.
 
Last edited:
I have mentioned this several times around the forum... CMOS switches / distortion, noise

TG can be made pretty low distortion embedded inside a NF loop.

JR
a CMOS analog switch inside the feedback loop?
You don't happen to have a drawing on hand?


Mute transistor.jpeg
https://www.onsemi.com/download/data-sheet/pdf/dta114e-d.pdf
https://groupdiy.com/threads/cmos-switches-distortion-noise.44722/post-560279
A special transistor package, a PNP with 2 10K resistor inside, SOT package for MUTE circuits.
 

Latest posts

Back
Top