I just had some interest in them becauses both the
NES and the
Famicom uses this techniques to amplify sound : They use an integer circuit normally reserved for digital applications as some kind of operational amplifier, using a feedback resistor !
I must admit I'm quite amazed this works at all, but apparently it even works very well.
Apparently the NES uses a special 74HCU04 to perform this, but I wonder how different is it to a standard 74HC04 ? According to datasheets found on the net, the HCU version has only one inverter stage (two transistors) while the HC version has more stages. However, more stages means more gain, which normally should make the chip closer to an ideal op-amp, isn't it ?
Has anyone here any experience with this kind of stuff ?
I recall a schematic of a simple radio in an old (early 1990s) radio DIY magazine that was completely made out of logic ICs.
Here is other radio on a single CD4001.
Well I'm not sure what this does but it looks like it's doing some kind of AM demodulation using 4xNAND gates as amplifiers by tying both inputs togethers and using feedback resistors for the frist 2 stages. What bothers me even more is that there is no feedback for the last 2 stages ! Normally the output should be constantly in a "staturated" state..
Why you don't want to use low voltage analogue amplifiers? They are pretty good these days
No it's not I don't want to use them ! It's just for a matter of curiosity, I find the concept to use digital circuits as analog amazing.
You've probably seen this before.
Applying a feedback resistor make the CMOS inverter roughly linear in a limited input range (the resistor is applying negative feedback). It can be indeed a cheap voltage amplifier.
Quote:
You've probably seen this before.
In fact no I haven't seen it anywhere, I've learned to use an Op-amp or possibly a shmitt trigger to do something in the like, but never a "normal" gate.
The CMOS inverter is like an inverted class AB amplifier. The feedback will bias the gate input at the threshold where it would oscillate, but the signal input will hold it steady. A series R before Rf should set the gain (I think 74HC will support ~10). These amplifiers are good for lo-fi audio and oscillators, not much else unfortunately.
I tried to do it with a 4009 and a 4011, and a gain of 3.3 (done with 1kohm and 3.3kohm) got the same result on both : Amplification works but a very nasty high freq oscillation happens on the output. The "unbuffered" chips weren't available in the lab unfortunately.
I think you need much larger resistor values. Try 10k and 33k at least. Do you have access to 4069? I think they are all unbuffered. 74HC04 should work too.
kyuusaku wrote:
I think you need much larger resistor values. Try 10k and 33k at least.
I would say ten times more
Well I guess I currently suck at analogs
I'd have guessed smaller resistors would make a stronger feedback and would cause amplification less likely to fail, however it sounds my guess was wrong.
Negative feedback can lead to instability, if there's too much gain when phase lags by 180˚ (negative feedback becomes positive feedback at this point, and you know it's instable). So either reduce the amount of feedback, or stick in a small cap either at the input or output to further reduce gain at high-frequencies. But the fact that the output is buffered can indeed not help.
OK I now have another question. Apparently if the negative feedback is too strong (the resistor too small) an amplifier (that haven't been designed for small gains) will oscillate.
What happens if you do the other way arround, that is choosing resistors too big ? For example if I use something in the MOhms as a feedback resistor, will anything bad happen ?
Theoretically, the undesirable effects are decrease of linearity, higher output impedance and lower bandwidth, but you get more gain.