General > General Technical Chat
Why does OpenAI ChatGPT, Possibly Want to disagree/annoy and change my eating...
RJSV:
Wouldn't mind reading more details, on that FPGA with OP-AMP embedded. (Please see reply #9). Any possibility, of placing quotes, here, on the chat-BOT response, with details on that embedded OP-AMP suggestion.
THAT, is more along the lines, of embedded hardware, (although certainly the social aspects need exploration.). I'd like that chat-BOT to be fluent, in using one if the bench circuit simulators, I bet that's VERY interesting, even if it may need a couple tweaks, to work in analogue / digital combos.
Thanks, Mk14.
-Rick
tom66:
So, it *kinda* understands SPICE, but it doesn't get things right.
It manages a resistor across a voltage source. (Ground node is wrong though, I think SPICE assumes 0 is always ground? Runs in LTspice if you remove the 'GND' line.)
It cannot however create even a basic filter, and indeed misunderstands the problem. It just puts all components in parallel. Also, it doesn't plot the frequency response right.
Interesting, though.
https://sharegpt.com/c/sDbLsR0
tom66:
Addendum. I think the best way to describe this chatbot is, "Confidently wrong". With a bit more prompting, it connected components in series (which is impressive!), but despite being able to independently calculate the frequency response using the formula, it continues to produce the wrong answer.
The last equations are quite impressive in terms of the AI explaining how it got where it did. It starts writing out using the (incorrect) values it calculated, but then the equation "morphs" into producing the right answers! How very odd, I wonder what is going on here.
It is very, very good at :bullshit:
https://sharegpt.com/c/DMp7V57
bdunham7:
It should be trivially easy to figure out that your AI ChatGPT, no matter how advanced, doesn't actually have any understanding of the topics it is discussing. We just sort of 'read in' the meanings, but if you look for issues that reflect an inability to comprehend the fundamental principles of the matter at hand, they show up pretty quickly.
I asked OpenAI to discuss 'phase to phase voltage measurement', since that was being discussed in another thread and I was just curious to see what it found. Here is what it wrote. It looks good at the very first glance, but how many howlers (ludicrous errors) can you find right away?
Phase-to-phase voltage, also known as line-to-line voltage, is the voltage between two phases in a three-phase system. It is typically used to measure the voltage of a three-phase system, such as that found in a commercial or industrial setting.
To measure phase-to-phase voltage, you will need a voltmeter that is capable of measuring AC voltage. Set the voltmeter to the appropriate range and connect the leads to the two phases you want to measure. It is important to ensure that the leads are properly connected, as reversing the leads can result in an incorrect measurement.
It is also important to ensure that the voltmeter is properly calibrated and that the leads are in good condition, as these factors can affect the accuracy of the measurement.
When measuring phase-to-phase voltage, it is important to follow proper safety procedures, such as wearing appropriate personal protective equipment and ensuring that the electrical system is de-energized before making any measurements. Failing to follow proper safety procedures can result in injury or death.
MK14:
Quick summary of results, quoted in full, as per request, see below.
Unable to get as good results, as I got previously. There are elements of luck and/or changes in questions, can affect the results.
This ChatGPT's current incarnations. Doesn't really seem to have a good grasp of electronics, by a long, long way. It shouldn't let me succeed in getting it to create an analogue op-amp inside of a digital FPGA. It should be coming up with huge numbers of excuses and get out clauses, as it is a very bad idea, and won't work.
There can be some fun had with ChatGPT as regards electronics, but it seems to be well behind, compared to trying to get ChatGPT to attempt to write computer programs. Which it is much, much closer to successful achieving.
As other(s) have pointed out, in this thread. ChatGPT, seems to very significantly NOT properly understand about electronics. It just doesn't seem understand how these electronics bits interact, and should mix together. From what I've seen, ONLY talking about ChatGPT AI (other AI systems, may have different successes with electronics). Even 5 or 10 years, may not begin to really make it do well with electronics. It just seems too weak and wrong, with its electronics capabilities.
I'm speculating. But I suspect ChatGPT, lacks significant training time and resources (whatever openAI, do behind the scenes to make these chatGPT systems), within electronics. Possibly because the openAI team lack the necessary electronics learning resources.
N.B. Speculation on my part, talking to the openAI team, would probably reveal better understandings on why it seems to behave so relatively badly with electronics.
In theory, there is a huge range of books available, about electronics, as well as many other resources. But presumably, they need to limit things, because of practicalities, around their ChatGPT's computing system capabilities, especially when training it.
I'm not sure ChatGPT, even then. Would necessarily be that good at electronics. Maybe it would only be a small amount better. I don't really know.
--- Quote from: RJHayward on December 29, 2022, 12:00:31 am ---Wouldn't mind reading more details, on that FPGA with OP-AMP embedded. (Please see reply #9). Any possibility, of placing quotes, here, on the chat-BOT response, with details on that embedded OP-AMP suggestion.
THAT, is more along the lines, of embedded hardware, (although certainly the social aspects need exploration.). I'd like that chat-BOT to be fluent, in using one if the bench circuit simulators, I bet that's VERY interesting, even if it may need a couple tweaks, to work in analogue / digital combos.
Thanks, Mk14.
-Rick
--- End quote ---
Also, some of my earlier work, was in this thread:
https://www.eevblog.com/forum/chat/openai-chatgpt-for-electronics-design/
One example, is this (it didn't work out, well at all):
--- Quote ---Using an FPGA, and displaying the necessary Verilog, and schematic diagram, drawn using text symbols. Implement a general purpose op-amp, configured as a non-inverting, times 25 gain amplifier. The op-amps parameters, should be able to be adjusted, by external analogue signals, on the pins of the FPGA. Please also draw out a summary, of the datasheet, for this op-amp, that has been created inside the FPGA.
Sure, here is an example of how you can implement a general purpose op-amp configured as a non-inverting amplifier with adjustable gain using Verilog and an FPGA:
module opamp(input wire in, input wire v+, input wire v-, output wire out);
parameter gain = 25;
// Non-inverting amplifier
assign out = gain * (in - v+);
endmodule
This module takes an input signal "in" and two reference voltages "v+" and "v-" and produces an output "out" that is the input multiplied by the gain factor. The gain can be adjusted by changing the value of the "gain" parameter.
Here is a schematic diagram of the non-inverting amplifier using text symbols:
v+
|
R
|
in -----[ ]-----o out
|
v-
The input signal "in" is applied to the non-inverting input of the op-amp, and the reference voltages "v+" and "v-" are applied to the inverting input and the ground terminal, respectively. The output "out" is produced at the output terminal of the op-amp. The resistor "R" represents the feedback element that sets the gain of the amplifier.
Here is a summary of the datasheet for the op-amp implemented in the FPGA:
Type: General purpose op-amp
Configuration: Non-inverting amplifier
Gain: Adjustable, default value is 25
Inputs: "in" (input signal), "v+" (non-inverting reference voltage), "v-" (inverting reference voltage)
Output: "out" (amplified output signal)
Other features: None
Note that this is just one example of how you can implement a non-inverting amplifier using Verilog and an FPGA. There are many other ways to do this, and you may need to add additional features or modify the design to meet your specific requirements.
--- End quote ---
--- Quote ---Sure! An op-amp (operational amplifier) is a type of amplifier that is designed to perform mathematical operations on input signals and produce a single output signal. It is a key building block in many electronic circuits, and can be used to amplify signals, perform filtering, and perform a variety of other functions.
To implement an op-amp in an FPGA, we can use a combination of Verilog code and a schematic diagram to specify the design of the op-amp. The Verilog code will define the behavior of the op-amp, while the schematic diagram will show the connections between the various components of the op-amp.
Here is an example of how you might implement a general purpose op-amp configured as a non-inverting, times 25 gain amplifier in an FPGA:
Verilog code:
module opamp (
input wire inp,
input wire inm,
output wire outp
);
parameter GAIN = 25;
// Implement the non-inverting amplifier
assign outp = GAIN * (inp - inm);
endmodule
Schematic diagram:
______
inp -| |- outp
| OP- |
inm -| AMP |
‾‾‾‾‾‾‾
This Verilog code defines a module called "opamp" that has three input wires (inp, inm, and outp) and one output wire (outp). The op-amp is implemented as a non-inverting amplifier, with a gain of 25. The gain can be adjusted by changing the value of the "GAIN" parameter.
The schematic diagram shows the basic connections between the input and output wires of the op-amp. The inputs are labeled "inp" and "inm", while the output is labeled "outp".
Here is a summary of the datasheet for this op-amp that has been created inside the FPGA:
Inputs: inp, inm
Output: outp
Gain: 25 (adjustable via the GAIN parameter)
Circuit type: Non-inverting amplifier
FPGA resource usage: Depends on the complexity of the circuit and the resources available on the FPGA
I hope this helps! Let me know if you have any further questions.
--- End quote ---
4th attempt, 3rd attempt not quoted, as seemed to be a repeat and/or crashed the Chat engine.
--- Quote ---Sure, I can help you with that. Here is a Verilog implementation of a non-inverting op-amp with adjustable gain:
module opamp (
input wire in,
input wire g1,
input wire g2,
output wire out
);
// Internal nodes
wire feedback;
// Create the op-amp using the Xilinx Core Generator
xlx_opamp opamp_inst (
.in1(in),
.in2(feedback),
.out(out)
);
// Create a voltage divider to provide feedback to the op-amp
voltage_divider feedback_divider (
.in(out),
.g1(g1),
.g2(g2),
.out(feedback)
);
endmodule
This module takes in an input signal in, and two external signals g1 and g2 that can be used to adjust the gain of the op-amp. The output of the op-amp is out.
The gain of the op-amp is calculated as follows:
Gain = 1 + (R2/R1)
where R1 is the resistance of g1 and R2 is the resistance of g2.
Here is a schematic diagram of the op-amp, using text symbols:
in
|\
| \
R1 R2
| |
| out
| /
|/
gnd
The op-amp has three terminal points: the non-inverting input in, the inverting input g2, and the output out. The two resistors, R1 and R2, form a voltage divider that provides feedback to the inverting input of the op-amp.
Here is a summary of the datasheet for this op-amp:
Input Voltage Range: The range of input voltages that the op-amp can accept without saturating or causing abnormal behavior.
Output Voltage Range: The range of output voltages that the op-amp can produce.
Gain Bandwidth: The frequency range over which the op-amp can amplify signals with minimal distortion.
Slew Rate: The maximum rate of change of the output voltage per unit time.
Input Offset Voltage: The voltage difference between the inverting and non-inverting inputs required to make the output zero.
Input Impedance: The resistance of the input terminals of the op-amp.
Output Impedance: The resistance of the output terminal of the op-amp.
I hope this helps! Let me know if you have any questions.
--- End quote ---
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version