I don't know if you want answers or guidance. Here is guidance.

The left circuit looks like an attempt to add a buffer stage to Ian's post #19. It is close. Consider a straight voltage follower (no feedback resistance), and do the differentiating at the non-inverting input.

In the right schematic, two things. 1. Standard Schmitt Trigger inverters, such as the CD106, always are inverters. 2. You need only two inverters, one for each signal phase.

Please add reference designators to all components. It will save both of us a lot of time and electrons.

AND, because I'm biologically incapable of not saying this - GND symbols ***always*** point downwards.

NOTE: All of the schematics in 19 and 20 have the same possible issue, a minimum voltage drop across the driver circuit. IOW, with 6 V or +/-6 V rails, a 6 V relay coil might see only 4 V. A way around this is to drive the coil with saturating switches rather than emitter followers. Gotta run, schematics later if you want them.

ak

 

What signal level is actually at the point where the intent is to add some control of the signal level? And what sort of signal is being controlled?
One choice not mentioned is a light dependent resistor and a variable light source. That will avoid the noisy relay contacts and take up MUCH less space. LDRs have been used quite a bit in the past and they work very well.

 

Unfortunately, that's out if it's a commercial product. LDRs contain cadmium, which is banned under RoHS. (They do contain rather less cadmium that the amount of arsenic that's in a red LED, but somehow that's OK)

 

I don't know if you want answers or guidance. Here is guidance.

The left circuit looks like an attempt to add a buffer stage to Ian's post #19. It is close. Consider a straight voltage follower (no feedback resistance), and do the differentiating at the non-inverting input.

In the right schematic, two things. 1. Standard Schmitt Trigger inverters, such as the CD106, always are inverters. 2. You need only two inverters, one for each signal phase.

Please add reference designators to all components. It will save both of us a lot of time and electrons.

AND, because I'm biologically incapable of not saying this - GND symbols ***always*** point downwards.

NOTE: All of the schematics in 19 and 20 have the same possible issue, a minimum voltage drop across the driver circuit. IOW, with 6 V or +/-6 V rails, a 6 V relay coil might see only 4 V. A way around this is to drive the coil with saturating switches rather than emitter followers. Gotta run, schematics later if you want them.

ak

Reference designator? there is the symbol of the schmitt trigger in the shape of Z.
unlike those below the schmitt trigger that you see above is not inverting.
if only two gates are used, it cannot work because at the output there would be pulses with the same sign synchronized and positioned on the same rising or falling edge. I had already taken into account the voltage drop of the push pull, the dual power supply voltages can be increased to have an output voltage suitable for the relays, perhaps a push pull mosfet is better. do you have integrated analog switches in mind?
I did not understand what you wrote on the opamp circuit, can you attach a diagram?.

What signal level is actually at the point where the intent is to add some control of the signal level? And what sort of signal is being controlled?

I did not understand this sentence

One choice not mentioned is a light dependent resistor and a variable light source. That will avoid the noisy relay contacts and take up MUCH less space. LDRs have been used quite a bit in the past and they work very well.

LDRs would be a better choice but it is not a linear component which is why it is not possible to use it with my circuit.

 

Reference designator? there is the symbol of the schmitt trigger in the shape of Z.
unlike those below the schmitt trigger that you see above is not inverting.
if only two gates are used, it cannot work because at the output there would be pulses with the same sign synchronized and positioned on the same rising or falling edge. I had already taken into account the voltage drop of the push pull, the dual power supply voltages can be increased to have an output voltage suitable for the relays, perhaps a push pull mosfet is better. do you have integrated analog switches in mind?
I did not understand what you wrote on the opamp circuit, can you attach a diagram?.


I did not understand this sentence



LDRs would be a better choice but it is not a linear component which is why it is not possible to use it with my circuit.

My question was about the amplitude of the signal being attenuated. Is it a low impedance power signal that is feeding a speaker? or a much lower voltage signal in an amplifier intermediate stage? Is the signal a music signal or an instrumentation signal?

AND the fact is that a Light Dependent Resistor IS a linear device, it is not at all like a photo-transistor or a photodiode, and so it doe not have a current or voltage dependent resistance change. Probably the deired attenuation system would require assembling an arrangement for one light source to affect multiple LDRs, so it would not be a commercially available package.

 

What signal level is actually at the point where the intent is to add some control of the signal level? And what sort of signal is being controlled?

While it never is stated explicitly, this is almost certainly for a line-level audio signal.

ak

 

Reference designator? there is the symbol of the schmitt trigger in the shape of Z.

A reference designator is a notation next to each component on a schematic giving it a unique identifier. For resistors, it would be R1, R2, etc. Cx for capacitors, Qx for transistors, Ux for integrated circuits. It makes life much easier. Rather than say "You need to connect the resistor in the top left corner of the schematic, the one below the capacitor and to the right of the upper transistor ...", you can sat "Connect R17 ...". Without reference designators, discussing the connections, value, or performance of individual components can be very difficult.

ak

 

An interesting thought came up, which is functionally the equivalent of one of the old style knob controlled attenuators that were used in the broadcast industry way back in the long-past times. But instead use a mechanical up/down stepper relay, and build the attenuation network right on the relay. The better relays were very reliable and the ones that had shorting contacts would provide noisless switching, and they consume no power except when changing positions. Those steppers may still be available although they will probably be new-old stock. And it would be more stable than most electronic devices.

 

Large, expensive, both mechanically and electrically noisy.

AND, what about the circuitry needed to turn the output of a ***bidirectional*** binary counter into the correct sequence of step pulses. Even if you eliminate the counter, and have its clock and control lines drive the relay, it has to be bidirectional.

And it would be more stable than most electronic devices.

Maybe in the 1940's and 50's ...

ak

 

Speaking of eliminating the bidirectional counter ... Another approach is to use a digipot, an IC that is an electronic volume control. Maxim is big in these, including ones designed specifically for audio circuits - logarithmically weighted, clickless, etc. with a simple up-down interface; no uC needed.

ak

 

My question was about the amplitude of the signal being attenuated. Is it a low impedance power signal that is feeding a speaker? or a much lower voltage signal in an amplifier intermediate stage? Is the signal a music signal or an instrumentation signal?

it is a musical line attenuator placed before the amplifier.

AND the fact is that a Light Dependent Resistor IS a linear device, it is not at all like a photo-transistor or a photodiode, and so it doe not have a current or voltage dependent resistance change. Probably the deired attenuation system would require assembling an arrangement for one light source to affect multiple LDRs, so it would not be a commercially available package.

I don't know if I understood correctly because I don't know the English language well.
did you claim that LDRs are linear? .
from the current-ohm curve that you see in the attached file it seems to me exactly the opposite.
http://www.farnell.com/datasheets/102616.pdf

 

Speaking of eliminating the bidirectional counter ... Another approach is to use a digipot, an IC that is an electronic volume control. Maxim is big in these, including ones designed specifically for audio circuits - logarithmically weighted, clickless, etc. with a simple up-down interface; no uC needed.

the high-level volume attenuators make use of relays to avoid fouling the signal as it passes through the semiconductors

 

it is a musical line attenuator placed before the amplifier.



I don't know if I understood correctly because I don't know the English language well.
did you claim that LDRs are linear? .
from the current-ohm curve that you see in the attached file it seems to me exactly the opposite.
http://www.farnell.com/datasheets/102616.pdf

It's linear in that it behaves as a resistor which remains resistive as the signal level varies, in other words, it cannot distort the signal.
The graph shows the relationship between light-level and resistance, which is not the same thing.
However, as I said before, it's cadmium content would prevent its use in a commercial product.

 

I believe the TS is referring to a linear relationship between the control voltage (or current) and the series resistance. I do not think this is an issue, because the part would be used as an on-off switch, not a variable attenuator.

A more valid concern is that any attenuator based on an LDR never is completely off. To get an on-off ratio of over 90 dB (a bare minimum requirement IMO) would take an L or T combination, just like the internal circuit of the best analog switches designed for pro audio and/or pro video applications. This is a significant increase in both cost and complexity.

ak

 

I had already taken into account the voltage drop of the push pull, the dual power supply voltages can be increased to have an output voltage suitable for the relays

Good. This makes things easier.

Where are you located?

ak

 

No mention was made of complete shutoff. I addition, I find the whole complaint that semiconductors cause distortion rather strange, since the only ways to amplify signals are to use either semiconductors or vacuum tubes. And really, we have heard not much about the application, except that it is a musical line attenuator. And we have this description of the motivation:" the high-level volume attenuators make use of relays to avoid fouling the signal as it passes through the semiconductors", which is a claim that would really only apply to poorly designed systems. A correctly designed system will not contribute measurable distortion. I am wondering if the purpose is to allow remote control of the volume. AN LDR circuit could certaily do that very well. It would not be able to provide complete muting, but that is an extreme condition.

 

No mention was made of complete shutoff.

I think that a digital output pin from a binary counter controlling a relay is pretty much an on-off kind of thing.

I agree with you that when done carefully, solid state switching devices contribute negligible distortion to an audio signal; measureable, but negligible. I also think the intent of the project is clear.

ak

 

It's important not to lose sight here of how much power is being saved by using a latching relay - probably only 0.1W per relay.
You can soon add enough drive circuitry to use up the 0.1W that is being saved, in which case, the simple, non-latching relay is the better solutioun.

 

since you insist with the LDR I would like to know how you would do the circuit.

 

It's important not to lose sight here of how much power is being saved by using a latching relay - probably only 0.1W per relay.
You can soon add enough drive circuitry to use up the 0.1W that is being saved, in which case, the simple, non-latching relay is the better solutioun.

stion of energy saving.
if I use double the bistable relays, I can use a contact to report the status of the outputs 40193 to its own inputs and thus obtain a volume memory, in addition I can adjust the volume in different quantities between one channel and another.
to obtain a memorization, alternatively I have in mind another thing which would be an elegant solution without any control interface circuit and without moving parts.
I take this opportunity to ask you, maybe it has already been done by someone but I don't know.
it is a question of exploiting the magnetic remanence (Br) of those materials called soft magnetic (a core wrapped in copper coils). They must be combined with a Hall effect magnetic sensor. it has the task of reading the polarity of the residual magnetism and providing a high or low output voltage.
there are sensors that do not behave in a linear way but in a latch way. to drive a memory of this type of memory you can use the same pulse used to drive the relays