Interesting link. I didn't realize the PT2399 uses a simple first-order delta modulator as an ADC. Neither does the person who wrote that website, but it is quite clear from the information on the site that that is what it is. I now understand what the comparator is for, it's the quantizer of the delta modulator.
See figure 2 of https://en.m.wikipedia.org/wiki/Delta_modulation The first-order low-pass filters (pin 9-pin 10 and pin 11-pin 12) are used as the integrators.
I built a doorbell with an equally simple first-order delta modulator when I was 17. It sounds awful at too low signal levels; it has no dithering, so you can get all sorts of limit cycles at low levels. At higher levels, it sounds quite reasonable, especially at high bit rates.
Regarding using the internal op-amps for the balanced in- and outputs that hobby_guy wants to have: I don't really see how you could do that, given that the op-amps are already needed for anti-aliasing and reconstruction filtering and that they all have one pin connected to the reference. On top of that, there is almost no information about them in the datasheets.
Regarding the capacitors between pin 10-pin 9 and pin 11-pin 12: as long as you use equal capacitor values for both, they should in principle not have any impact on the frequency response. That's because the first-order filter/integrator of the modulator is in its feedback path, and compensates for the first-order filter/integrator of the demodulator.
There is a compromise to be made, though. The larger the capacitors, the less quantization noise you get, but the easier you can drive the modulator into slew rate limiting with signals with a lot of treble.
There is a compromise to be made, though. The larger the capacitors, the less quantization noise you get, but the easier you can drive the modulator into slew rate limiting with signals with a lot of treble.
Not much; 31 nV/√Hz and only unity gain is supported. It's meant for balanced line level inputs rather than as a microphone preamplifier.
This is great, really enjoying following the technical discussions, as best I can. Thankfully, working with sound for live music is very forgiving, as quite a lot of distortion can be accepted and just labeled "sound coloration", or "effect" ;-)
This is for learning, and for personal use, so I'm really looking forward to getting my 9V prototype finished. I'll probably post an audio clip
This is for learning, and for personal use, so I'm really looking forward to getting my 9V prototype finished. I'll probably post an audio clip
I am working on patching my (home made, CNC-machined) PCB for the input/output stages to the corrections and improvements given in this thread.
My first iteration will be a 9V battery version, where I replace the INA333 with a TLV2462. I will later change this to an OPA, just have to order more components ;-)
One question: The non-inverting input on the differential opamp connected to "AGND" (2.5V) instead of ground (0V). The input signal (from the mic) is centered around 0, but the single power supply (5V) means it needs to have a 2.5V reference -- and not 0. I get that. But doesn't this also mean I'm adding a 2.5V DC offset to the "cold" (pin 3) part of the signal, and not the "hot" (pin 2) part?
Also, I'm using a TLV2462 which has 2 op amps, and only using one. So I need to "disable" the unused op amp, by connecting output to inverting input and non-inverting input to ground. Should this be ground, or "AGND" then?
My first iteration will be a 9V battery version, where I replace the INA333 with a TLV2462. I will later change this to an OPA, just have to order more components ;-)
One question: The non-inverting input on the differential opamp connected to "AGND" (2.5V) instead of ground (0V). The input signal (from the mic) is centered around 0, but the single power supply (5V) means it needs to have a 2.5V reference -- and not 0. I get that. But doesn't this also mean I'm adding a 2.5V DC offset to the "cold" (pin 3) part of the signal, and not the "hot" (pin 2) part?
Also, I'm using a TLV2462 which has 2 op amps, and only using one. So I need to "disable" the unused op amp, by connecting output to inverting input and non-inverting input to ground. Should this be ground, or "AGND" then?
Thanks for responding!
So, just so I understand, the non inverting input must have the same reference as the output signal (which is "lifted" by 2.5V offset).
But there will be a difference on the two input pins, right? The cold will swing around 2.5V whereas the hot will swing around 0. But that's the way this is supposed to (or has to) work.
So, just so I understand, the non inverting input must have the same reference as the output signal (which is "lifted" by 2.5V offset).
But there will be a difference on the two input pins, right? The cold will swing around 2.5V whereas the hot will swing around 0. But that's the way this is supposed to (or has to) work.
An op-amp + and - inputs can operate anywhere over the "common mode range" which is usually within a couple volts from either rail. The output DC will go to where the results is ~zero volts between + and - inputs, provided the feedback is designed properly. A "single supply" op-amp can operate with + and - inputs at the negative rail, and the output at Vcc/2. The output has to be in the middle of the supply if you want +/- voltage swing, ie AC audio, which will be decoupled with a DC blocking capacitor. The easiest way to do this is to operate all op-amp pins at Vcc/2 and decouple/block the DC with capacitors so that the AC ground can be the negative rail. Theoretically all rails are AC ground, but the reality is that power supplies have some noise on them. You have to arrange the feedback to the negative input to cancel any voltage on the positive input. The Voltage between inputs is not actually zero, but it is the output divided by a huge number called the open loop gain, so you design as if it's zero. The job of the negative input is to sense the output, or some fraction of the output, and make the op-amp copy the positive input. If you feedback only half the output, then the output will double the positive input in order to make the neg input equal the positive input. The output "returns" to the power supply after the load, whether that be a center voltage of one of the rails. The inputs only care about the difference between them, within the power supply limits.
For an inverting amplifier, the positive input is set to zero (AC) and resistors from the input and output provide currents which cancels each other for essentially zero volts (AC) on the negative input.
I'm sure there is a better tutorial on the web. I have ignored feedback stability which involves frequency and phase shift. Most common op-amps are built to be stable with 100% feedback, ie unity gain. Some "chip amps" are only stable if you wire then with 10% feedback or less, because the cost of stability is slew rate.
For an inverting amplifier, the positive input is set to zero (AC) and resistors from the input and output provide currents which cancels each other for essentially zero volts (AC) on the negative input.
I'm sure there is a better tutorial on the web. I have ignored feedback stability which involves frequency and phase shift. Most common op-amps are built to be stable with 100% feedback, ie unity gain. Some "chip amps" are only stable if you wire then with 10% feedback or less, because the cost of stability is slew rate.
Thanks for the great explanation!
But then, why is it not a problem that the inverting input is not also "given" a 2.5V (Vcc/2) offset? As far as I can tell, the inverting input sees a signal centered around 0. Sorry if you did actually explain this and I just didn't catch it...