Hi,
I have come across this link in search of an simple audio delay line.
They claim it works, but in simulations I don't see any pwm-like signal, it just acts like a high-pass filter, just as I suspected the first time I looked at the schematic.
The only thought that comes to my mind is that they use the unstable nature of the comparator to create a pwm signal by having it oscillate.
Does anyone know?
I have come across this link in search of an simple audio delay line.
They claim it works, but in simulations I don't see any pwm-like signal, it just acts like a high-pass filter, just as I suspected the first time I looked at the schematic.
The only thought that comes to my mind is that they use the unstable nature of the comparator to create a pwm signal by having it oscillate.
Does anyone know?
I wanted to experiment using a 20 mhz PIC16F84A and using some sort of sigma-delta conversion to get the data in the pic and storing it for 400 µs before sending it out again to make a digital crossfeed for my headphones.
This seemed an easy way to start since I already have most of the components laying around.
However I can't figure out why the comparator should produce a pwm signal in this configuration.
Wat is a latching comparator anyway? I couldn't find any usefull information about wat is is.
This seemed an easy way to start since I already have most of the components laying around.
However I can't figure out why the comparator should produce a pwm signal in this configuration.
Wat is a latching comparator anyway? I couldn't find any usefull information about wat is is.
The circuit from the link will definitely NOT work.
If one takes the feedback from the output of the first flip-flop then we have a delate-modulator which gives some improvement.
The way the circuit is originally drawn we only end up with the input signal transformed into a rectangular and then being delayed.
The best way would be to use the low pass filter(s) (or even better integrator(s) ) is in the forward path of this modulator circuit and thus building a noise-shaping delta-sigma modulator.
Regards
Charles
If one takes the feedback from the output of the first flip-flop then we have a delate-modulator which gives some improvement.
The way the circuit is originally drawn we only end up with the input signal transformed into a rectangular and then being delayed.
The best way would be to use the low pass filter(s) (or even better integrator(s) ) is in the forward path of this modulator circuit and thus building a noise-shaping delta-sigma modulator.
Regards
Charles
The circuit was described in the link. It is delta-modulator (not delta-sigma). It is not pwm signal btw, although it looks like one.bobo1on1 said:
Latching comparator is basically comparator with d-flop built in, connected like the first dflop right next to comparator in this circuit. I believe it is misprint and it was implied to take feedback from the dflop.
You need output from comparator to the RC have exactly same shape as that of after delay line. Then both RC's behave identically, and as comparator forces left RC to track signal, so will the other.
Why would you like to cripple the sound with such el cheapo when you can have decent crossfeed with analog means?
wimms said:
Because I want to experiment with different delays, changing delays to give a moving source etc.
But this thing just converts the audio waves in to square waves and passes them through a lowpass to get sort of sine waves, is that right?
Well that would be of no good use to me, I guess I have to use a triangle wave generator an fast comparator.
error in schematic
The schematic for the article that you pointed to had an error.
The first RC should have been attached to the output of the flip-flop, not the comparator.
If you fix that, then all should be well. You will have to use a pretty high data rate to get this to work with reasonable fidelity. Overall, the data rate required for a 1-bit delta modulator without any adaptive capability is significantly higher than the data rate from a comparable n-bit A/D. If you add adaptation, then the bit-rate can be very low (~32Kbps for telephone quality).
One-bit AD's such as this are often used high quality audio data streams because anti-aliasing can be avoided during conversion (remember that very high sample data-rate?). The rates can be daunting, however. For example, if you want to have the equivalent of 16 bits with 1 V pp signals, then you are limited to digitizing signals with a slew rate of less than 30uV / dt. A 1Khz 1V pp sign wave has a max slew rate of 1000V/s which covers 30uV in much less than a microsecond. In fact, you need a cool 33M samples per second to meet that spec. Real systems get by with less than this because perfectly full scale sine waves don't happen.
The schematic for the article that you pointed to had an error.
The first RC should have been attached to the output of the flip-flop, not the comparator.
If you fix that, then all should be well. You will have to use a pretty high data rate to get this to work with reasonable fidelity. Overall, the data rate required for a 1-bit delta modulator without any adaptive capability is significantly higher than the data rate from a comparable n-bit A/D. If you add adaptation, then the bit-rate can be very low (~32Kbps for telephone quality).
One-bit AD's such as this are often used high quality audio data streams because anti-aliasing can be avoided during conversion (remember that very high sample data-rate?). The rates can be daunting, however. For example, if you want to have the equivalent of 16 bits with 1 V pp signals, then you are limited to digitizing signals with a slew rate of less than 30uV / dt. A 1Khz 1V pp sign wave has a max slew rate of 1000V/s which covers 30uV in much less than a microsecond. In fact, you need a cool 33M samples per second to meet that spec. Real systems get by with less than this because perfectly full scale sine waves don't happen.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.