Does anyone know a way to take an audio signal and output it multiple times with different amplitude and phases tia sal2
Greetings All
I have one source for an audio signal and would like to send it to 8 different outputs simultaneously, but would like to be able to change the phase and amplitude of each of the
8 outputs separately.
I was thinking of using 8 small amplifiers (which isn?t cheap) to increase the amplitude of the 8 output signals but I?m not sure how I can change the phase of each signal separately. Is there a product on the market that is inexpensive or can someone recommend a circuit I could build?
I would use an oscilloscope to make sure the phase and amplitude of each of the
8 outputs is where I would want it.
Example: Source is a sine wave generated by a laptop
Output 1) starts at 0 degrees Output 2) starts at 10 degrees out of phase increase amplitude by 1 volt Output 3) starts at 15 degrees out of phase Output 4) starts at 40 degrees out of phase Output 5) starts at 60 degrees out of phase increase amplitude by 3 volts Output 6) starts at 100 degrees out of phase Output 7) starts at 120 degrees out of phase Output 8) starts at 126 degrees out of phase
You should probably think in terms of little amplifier blocks. There's a lot of introductory material on the web if you search for 'op amp'.
formatting link
If you're building a basic circuit for audio frequencies, these are dirt cheap. You can get 4 or more on an inexpensive chip. =A0
If you're using a sound card output, can you output the signal on the left and right channels, 90 degrees out of phase where it starts? That will simplify things greatly!
If not, you'll need to generate a 90 degree phase shifted signal from the source. The simple circuits I'm familiar with are highly frequency dependent, but some of the wizards on this group may know some tricks.
After you have 0 and 90 degree signals (call them s0 and s90) at equal amplitudes, you just need seven 2-channel mixers and some trigonometry. Assuming output amplitutde of 1 V.
You've got that already.
Mostly s0, with the appropriate amount of s90 to shift phase. Increase gain on both s0 and s90 to get increase the overall amplitude.
s0 gain =3D 2*cos(10 deg) s90 gain =3D 2*sin(10 deg)
s0*cos(15 deg) + s90*sin(15 deg)
s0*cos(40 deg) + s90*sin(40 deg)
s0*4*cos(60 deg) + s90*4*sin(60 deg)
For phase shifts >90, you calculate it the same, however, be aware that one or both gains (from sin or cos) are going to be negative. Depending on how you're programming your mixer/amplifiers, this may work automatically. Worst case, you just send your two input signals through unity gain inverting amplifiers (a simple op amp circuit) to get 180 and 270 degree signals. This makes a total of four inputs to each of your seven mixers, though no more than two will be nonzero at any given time.
If this only has to operate at a single fixed frequency, this is fairly straightforward. You can use an "all-pass" filter that shifts phase without changing level. (Add gain or attenuation as needed.) The basic circuit uses an op-amp in what looks sorta like a classic difference amp configuration, but with the + and - input resistors tied together. That alone would be a "do nothing" circuit because the inverting gain and the non-inverting gain would cancel each other out at all input frequencies.
However, you make the non-inverting input phase sensitive by replacing one of the resistors with a capacitor.
The two non-inverting resistors normally make a
2:1 voltage divider, so if you replace the ground resistor with a cap you have a low-pass filter going to the + op-amp input. At low frequencies the overall gain is unity (+2 from this input, -1 from the inverting input connected to the same signal). At high frequencies the + input is grounded so you have a simple inverter with a gain of -1. At intermediate frequencies you get a phase shift. You can adjust the phase by adjusting the RC on the + input. (There is no need for the R to match the inverting Rs.)
Note that this approach can do what you request, for a single input frequency, but all the RCs need to be re-adjusted if the frequency changes.
If you want a scheme that will work for any single frequency, you can digitally generate the signals with different phases. The problem is that sound cards typically have only stereo outputs. (You can use the signal generator in Daqarta if you want to experiment with 2 channels... it's free. Daqarta can actually generate 8 independent data streams, but they still have to be combined down to 2 output channels: 4 on the right and 4 on the left.)
There are sound cards that have multiple outputs (5.1, 7.1, etc) but typically these only have two independent channels that you can generate arbitrary signals with... they produce the other outputs by processing and combining those two signals. High-end cards may have independent channels, but it's hard to tell that from the docs. (Daqarta currently only handles 2 signal outputs regardless, so you'd have to write your own generating code for more than 2.)
If you are handy with microprocessors, the general approach is to create a circular buffer that holds one cycle of a sine wave. To generate a frequency, you output samples from that table. (To generate twice that frequency, skip every other sample, etc. ) To generate different phases, just change the starting point in the table. You need a microprocessor with 8 DACs for this project.
If you need to use arbitrary input frequencies from a single source, you might want to look into phase locked loops. If you can adjust the lock phase in a frequency-independent manner (don't know, haven't tried this), you can have a separate PLL for each output. But a better approach might be to use a single PLL to generate a high multiple of the input frequency that is fed to various counters, whose outputs are combined via selected resistors to create a stepped approximation of a sine wave (which can be easily filtered). To get the different phases, you would use different counter arrangements. LOTS of parts to this approach, however.
If you tell us what your overall application is, we can probably come up with something simpler.
Best regards,
Bob Masta DAQARTA v4.51 Data AcQuisition And Real-Time Analysis
formatting link
Scope, Spectrum, Spectrogram, Sound Level Meter FREE Signal Generator Science with your sound card!
This is excellent advice, and is the simplest solution yet! Since it only requires 2 outputs from the sound card, the (free) Daqarta signal generator create the 0 and 90 degree signals with ease. The beauty of your approach is that once the various mixer levels are set, it will work with any arbitrary generated frequency as long as the phases are maintained at 90 degree separation. You can even generate a frequency sweep, or use FM, etc and everything will track.
To the OP: Let me know if you need any advice on how to set up the signals with Daqarta.
Best regards,
Bob Masta DAQARTA v4.51 Data AcQuisition And Real-Time Analysis
formatting link
Scope, Spectrum, Spectrogram, Sound Level Meter FREE Signal Generator Science with your sound card!
Glad to see someone suggested the All-pass filter. This is a simple way to go if you are only working with one frequency at a time... Of course if you have more than one frequency you can't really define a unique phase shift.
It is by far the nicest IMHO since it also has a flat frequency response IIRC. I've used one in a spiral scan radar display where the deflection coils were driven by sine and cosine waveforms and the ampltude was ramped so as to provide a natural r,theta display.
Using the digital system which I described earlier but with a high-frequency clock substituted for the audio input signal: CLK>---[COUNT]--+-[LUT1]--[DAC]--[FILTER]--[OPAMP]--->OUT1 | +-[LUT2]--[DAC]--[FILTER]--[OPAMP]--->OUT2 | +-[LUT3]--[DAC]--[FILTER]--[OPAMP]--->OUT3 | +-[LUT4]--[DAC]--[FILTER]--[OPAMP]--->OUT4 | +-[LUT5]--[DAC]--[FILTER]--[OPAMP]--->OUT5 | +-[LUT6]--[DAC]--[FILTER]--[OPAMP]--->OUT6 | +-[LUT7]--[DAC]--[FILTER]--[OPAMP]--->OUT7 | +-[LUT8]--[DAC]--[FILTER]--[OPAMP]--->OUT8
will ensure that the phase relationships between _all_ of the output signals will remain constant regardless of the frequency of the clock.
If he had to use his analog output as an input to the system I think this would work to keep the phase relationships of all of the outputs constant regardless of the long-term frequency of ACIN:
How many 2kW 40kHz ceramic beamforming transducers have you designed?
BTW, while that AX3500 might be a nice amp, it's not rated for 3500 watts, it's rated for 2600 watts, bridged, into a 4 ohm load.
Also, its THD is 0.02% with only a single channel driven to rated power at 1kHz, hardly the 0.0008% you seemed to be implying it could get in an earlier post.
That power would be 355 watts into 8 ohms, 550 watts into 4 ohms or 725 watts into 2 ohms, so it seems there's a little specsmanship going on there by not rating it with both channels running at the same time and bridged.
It was totally state of the art then. It proved invaluable on another project too. That shows how far ahead of the pack I was. Of course I have a windows version now.
I prefer to clear it with the client first, not least because I believe the info SHOULD be freely available to all who need it like their customers wishing to service their kit, not just those on Usenet.
And a.b.s.e is no longer available to many US users of Usenet too. It hardly addresses the issue adequately.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.