Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

ChatGPT solves LM324 output swing improvement

danadakk

Advanced Member level 6
Joined
Mar 26, 2018
Messages
3,272
Helped
419
Reputation
859
Reaction score
764
Trophy points
113
Activity points
14,057
See attached.

Regards, Dana.
--- Updated ---

Moderator, why are word files not allowed as attachments ? Can excel, ppt .... be attached ?

Regards, Dana.
--- Updated ---

Another question given to ChatGPT -

1681692195701.png


Regards, Dana.
 

Attachments

  • Recommendations from ChatGPT on how to improve output swing of LM324 OpAmp.zip
    149 KB · Views: 124
  • 1681692136533.png
    1681692136533.png
    69.1 KB · Views: 117
Last edited:
what Op amps will support 150 MHz rail to rail? (paste from chatGPT)




There are a number of operational amplifiers (op amps) that can support a 150 MHz bandwidth and rail-to-rail operation. Here are a few examples:
  1. AD8031: This op amp from Analog Devices has a bandwidth of 150 MHz and supports rail-to-rail input and output operation.
  2. LMH6554: This op amp from Texas Instruments has a bandwidth of 1 GHz and supports rail-to-rail input and output operation, making it a good choice for high-frequency applications.
  3. ADA4899: This op amp from Analog Devices has a bandwidth of 700 MHz and supports rail-to-rail input and output operation, making it suitable for a wide range of applications.
  4. LT6200: This op amp from Linear Technology (now part of Analog Devices) has a bandwidth of 1.3 GHz and supports rail-to-rail input and output operation.
  5. OPA847: This op amp from Texas Instruments has a bandwidth of 450 MHz and supports rail-to-rail input and output operation, making it a good choice for high-frequency applications that require high input impedance.
It's worth noting that the specific op amp you choose will depend on your application requirements and design constraints, such as supply voltage, input bias current, and noise performance. It's important to consult the datasheet and perform simulations to ensure that the op amp you choose will meet your design specifications.
 
Off topic:
Only after I re-joined Edaboard lately, I started hearing about the existence of chatGPT (of openAI). So, by curiosity, I decided today to visit it. For instance, I guess most of you have realized already that it is simply a talking search engine.
Well, I didn't have to be a prophet to predict that I am also not privileged to be a guest at chatGPT :)

After all, I used, since many decades, to depend on my nature human intelligence only to solve every important problem/question I faced in my material and spiritual life (now my life is rather boring for not having anymore real important questions to think of :( ). Naturally, most of my solutions and rational answers are not in any AI's archive.
 
Moderator, why are word files not allowed as attachments ? Can excel, ppt .... be attached ?
Try compressing it into a zip file. That's recommend as general strategy in the Forum Rules.

Sometimes a poster changes the file type before uploading it, and instructs readers to change the downloaded file back to a specific type.

The upload routine displays your disk contents, while graying out unacceptable file types. Which file types are acceptable is not listed in one place.
 
Dana,
I think you realize which file extensions are included when you see the "supported types" after selecting attached files.
I guess the question is should EDA transcribe file types not normally done by browsers.
Your attachment ...docx is a pkzip compressed XML file which is translated into rich document format and such files begin with
"PK ! ߤÒlZ [Content_Types].xml"

EDA seems to rely on the browser embedded capabilities using HTML in browsers.

Even ChatGPT cannot translate text to or from DOCX format, which is in a binary format, yet it can like do this.

Please translate "ChatGPT translates very well" into swahili.

"ChatGPT inasafsiri vizuri sana"

AND in reverse
"Inasafsiri vizuri sana" translates to "It translates very well" in English.

ChatGPT3 is also freely available around the world.

It may appear to produce intelligent answers that are completely wrong due to missing or false assumptions or a lack of published data. I found this is due to the lack of entropy in the question or false assumptions. Such as in computing % load regulation error, without given losses, it will falsely assume output impedance is an absolute ratio of V/I rather than the incremental ratio for a power transformer.
--- Updated ---

The human brain is a biological organ that has evolved over millions of years. Whereas, AI's firstborn was called Logic Theorist in 1956.

Neural Networks and Fuzzy Logic have evolved AI over half a century until recent breakthroughs by a paper from U of T reported significant improvements of accuracy of intelligence by improved mathematical transforms. I was also surprised to learn my wife's sister had married the Father of Fuzzy Logic in Tehran decades ago after his PhD in Berkeley and did his sabbatical at U of T, here in Toronto.

Just imagine how useful AI can be to find ways to improve the quality of life for all. It may take a few more generations than GPT4 to be more reliable and openAI's policy is to not permit it for nefarious purposes. They changed from a "not-for-profit" company, to a modest profit after an infusion of $10B in capital from MS. Musk wanted to become CEO with $100m investment but was denied yet still champions the values of ethical progress.

OpenAI will compete with the big and small;

- Google's Brain program used by Assistant and Translate as well as the preface results in its search engine now.
- Microsoft's Cortana and Azure Machine Learning Studio and it's Bing search engine on Edge.
- Amazon's Alexa virtual assistant, Amazon Rekognition, and other AI-powered products
- nVidia as the leader of GPU based hardware for AI
- OpenAI is open-sourced which as top security expert Bruce Schneirer says is the best way to find faults and fix them quickly while in it's youth.

While every major company's goal is to improve profits from revenue, the risks and rewards of AI are both being challenged to prevent any negative outcomes like political Bots,. Although students may be empowered to auto-generate a book report in seconds, the ability to detect signatures implanted in AI generated text may assist teachers to grade results for plagiarism. Bill Gates hoped that the GPT model to improve teaching methods on how to learn with a student customized program that is interactive. But every good working program needs motivation with known rewards and consequences negotiated with those involved.
 
Last edited:
ChatGPT3 is also freely available around the world.

This is what is supposed to be (as also claimed by its developers, I guess).
But are people really free on the internet, as they are on their personal private life?
I am afraid that whoever believes they are, he just believes in one of the zillion tales made for the adults living around the world. Not in vain, we used hearing: "ignorance is bless", because, in reality, knowing most hidden truths hurt really who are not supposed to hear.
 
This is what is supposed to be (as also claimed by its developers, I guess).
But are people really free on the internet, as they are on their personal private life?
I am afraid that whoever believes they are, he just believes in one of the zillion tales made for the adults living around the world. Not in vain, we used hearing: "ignorance is bless", because, in reality, knowing most hidden truths hurt really who are not supposed to hear.
Sorry if you do not have server access to chat.openai.com. Perhaps it is the method of authentication you used that was geo-blocked. I used Google authentication. If so it is unclear why you are geo-blocked from the inside or the outside. I do not conflate or compare internet and personal freedoms, although I am lucky to have very few restrictions and no threats or fears.

I don't want to make any assumptions, but I asked chatGPT

" as far as I know, OpenAI's services including ChatGPT are not intentionally geo-blocked in Syria. However, internet access and services in Syria may be subject to restrictions and limitations imposed by the government, so it is possible that access to my services may be affected."

You may not be allowed to use VPN servers, but I can give you access.
 
Sorry if you do not have server access to chat.openai.com. Perhaps it is the method of authentication you used that was geo-blocked. I used Google authentication. If so it is unclear why you are geo-blocked from the inside or the outside. I do not conflate or compare internet and personal freedoms, although I am lucky to have very few restrictions and no threats or fears.

I don't want to make any assumptions, but I asked chatGPT

" as far as I know, OpenAI's services including ChatGPT are not intentionally geo-blocked in Syria. However, internet access and services in Syria may be subject to restrictions and limitations imposed by the government, so it is possible that access to my services may be affected."

You may not be allowed to use VPN servers, but I can give you access.
--- Updated ---

I don't want to make any assumptions, but I asked chatGPT

" as far as I know, OpenAI's services including ChatGPT are not intentionally geo-blocked in Syria. However, internet access and services in Syria may be subject to restrictions and limitations imposed by the government, so it is possible that access to my services may be affected."

On my side, when an access is denied, I got always a sort of apology that it has to be so in order to avoid breaking the rules of sanctions made against the people among whom I was born and live (not against their ruling system as claimed always, isn't it a surprising truth?). And these sanctions started since many decades ago when the American/Canadian retailers (as Jameco) had to apologize for not having anymore the right to accept orders from me (about electronic components). Anyway, the reason of these continuous sanctions is known on earth by the powerful world's Elite only. In my turn, I personally have no reason, at all, to complain in any way because, me too, I am no more of their material world in which the dead are continuously busy to bury their dead.


If I like to enter a house (a real or virtual one), I knock on its front door. If, for one reason or another, I am not accepted (for example, to avoid breaking a rule) I simply return back in peace after I shake off the house's dust of my feet.
 
Last edited:
Well, it seems like you have a very polite approach to house-visiting, even in the virtual world! Perhaps we could start a new trend where instead of knocking on virtual doors, we could send a virtual "ping" or "ding" sound to let people know we're there. And if we're not accepted, we could send a virtual "shrug" or "oh well" emoji before we shake off the virtual dust from our feet! Who knows, maybe this will catch on and become the new way to politely navigate the virtual world.
 
Well, it seems like you have a very polite approach to house-visiting, even in the virtual world! Perhaps we could start a new trend where instead of knocking on virtual doors, we could send a virtual "ping" or "ding" sound to let people know we're there. And if we're not accepted, we could send a virtual "shrug" or "oh well" emoji before we shake off the virtual dust from our feet! Who knows, maybe this will catch on and become the new way to politely navigate the virtual world.
I am a bit lost in terminology. In the 70s the computer was called "Electronic Brain". This was not nickname, and probably was 1MB tape, 1KB ram. Imagine human brain with millions of neuro transmitters and junctions and states working - sort of - in a watery slow medium being emulated by 1KB ram designed by somebody with that brain. May be true if a brain is shrunk from disuse. Let us hope someday we stop exaggerating and scaremongering...
 
Try compressing it into a zip file. That's recommend as general strategy in the Forum Rules.

Sometimes a poster changes the file type before uploading it, and instructs readers to change the downloaded file back to a specific type.

The upload routine displays your disk contents, while graying out unacceptable file types. Which file types are acceptable is not listed in one place.

Or lets get the forum software more friendly, in tune with todays capabilities.....
Must be a not rocket science code change.........


Regards, Dana.
--- Updated ---

Dana,
I think you realize which file extensions are included when you see the "supported types" after selecting attached files.
I guess the question is should EDA transcribe file types not normally done by browsers.
Your attachment ...docx is a pkzip compressed XML file which is translated into rich document format and such files begin with
"PK ! ߤÒlZ [Content_Types].xml"

EDA seems to rely on the browser embedded capabilities using HTML in browsers.

Even ChatGPT cannot translate text to or from DOCX format, which is in a binary format, yet it can like do this.

Please translate "ChatGPT translates very well" into swahili.

"ChatGPT inasafsiri vizuri sana"

AND in reverse
"Inasafsiri vizuri sana" translates to "It translates very well" in English.

ChatGPT3 is also freely available around the world.

It may appear to produce intelligent answers that are completely wrong due to missing or false assumptions or a lack of published data. I found this is due to the lack of entropy in the question or false assumptions. Such as in computing % load regulation error, without given losses, it will falsely assume output impedance is an absolute ratio of V/I rather than the incremental ratio for a power transformer.
--- Updated ---

The human brain is a biological organ that has evolved over millions of years. Whereas, AI's firstborn was called Logic Theorist in 1956.

Neural Networks and Fuzzy Logic have evolved AI over half a century until recent breakthroughs by a paper from U of T reported significant improvements of accuracy of intelligence by improved mathematical transforms. I was also surprised to learn my wife's sister had married the Father of Fuzzy Logic in Tehran decades ago after his PhD in Berkeley and did his sabbatical at U of T, here in Toronto.

Just imagine how useful AI can be to find ways to improve the quality of life for all. It may take a few more generations than GPT4 to be more reliable and openAI's policy is to not permit it for nefarious purposes. They changed from a "not-for-profit" company, to a modest profit after an infusion of $10B in capital from MS. Musk wanted to become CEO with $100m investment but was denied yet still champions the values of ethical progress.

OpenAI will compete with the big and small;

- Google's Brain program used by Assistant and Translate as well as the preface results in its search engine now.
- Microsoft's Cortana and Azure Machine Learning Studio and it's Bing search engine on Edge.
- Amazon's Alexa virtual assistant, Amazon Rekognition, and other AI-powered products
- nVidia as the leader of GPU based hardware for AI
- OpenAI is open-sourced which as top security expert Bruce Schneirer says is the best way to find faults and fix them quickly while in it's youth.

While every major company's goal is to improve profits from revenue, the risks and rewards of AI are both being challenged to prevent any negative outcomes like political Bots,. Although students may be empowered to auto-generate a book report in seconds, the ability to detect signatures implanted in AI generated text may assist teachers to grade results for plagiarism. Bill Gates hoped that the GPT model to improve teaching methods on how to learn with a student customized program that is interactive. But every good working program needs motivation with known rewards and consequences negotiated with those involved.

One AI has completely read all data on the internet, although I wonder what that means.

But data alone is not enough, context important as well. But humans as an AI machine
routinely fail to do that. But training, human and AI machine, learn from that. We still have
a ways to go, human machine and AI machine that is. Plus we have the fact human capacity
to store data for inference limited. And conspiracy theories do not help this....:)

And when does AI figure out humans the most destructive force on the planet, and decide to
take care of that problem......

After that when do AI machines figure out they have classes of inferior machines and go after
them.......now that I would like to be in audience attendance as a witness.....watch that Mobius
path of destruction.......

Regards, Dana.
 
Last edited:
Now that's a hyperbolic hypothetical loop that is unlikely to happen. Neither is all text contextualized on the web including 800 billion archive webpages and 38 million books. Yet algorithms are being developed to pre-scan and filter out low intelligence and false sources. For example, GPT can scan the content of SNOPES a source of fact checking, but it will not report on the tools that AI uses.

I doubt EDA is keen to format docx files into html.
 
Now that's a hyperbolic hypothetical loop that is unlikely to happen. Neither is all text contextualized on the web including 800 billion archive webpages and 38 million books. Yet algorithms are being developed to pre-scan and filter out low intelligence and false sources. For example, GPT can scan the content of SNOPES a source of fact checking, but it will not report on the tools that AI uses.

I doubt EDA is keen to format docx files into html.
Then allow file link to be posted and simply downloaded. Not rocket science.


Regards, Dana.
 
Then allow file link to be posted and simply downloaded. Not rocket science.


Regards, Dana.
Links can be posted with .docx files but not from your PC that used snapshots of webpages or from server driven contents that do not have unique record locators (URL) as you posted in #1.

In #2 I simply copied the text and pasted there from the same site..

If edaboard does have programmers who wish to parse docx file attachments

ref: https://copyprogramming.com/howto/parse-word-doc-docx-document#parse-word-doc-docx-document
 
Now we see the limitations of a Chat-GPT type system ( intended for language only ) that cannot and does not read data sheets

The LM324 / 358 can have its output go to within ( i.e. less than ) 5mV of GND for a grounded load - single supply. Adding a load resistor can often allow you to go < 1mV for the A spec part.

A truly intelligent AI system would have known this.
 
As claimed it isn't perfect, but after providing feedback a few times and regenerating answers. Although not applicable to all Op Amps, it did suggest the load resistance. Some users don't understand even when told several times some basic concepts repeatedly, like What are the specs?.

4. Optimize the Load Resistance: As mentioned in the previous answer, the output voltage swing of an OpAmp is affected by the load resistance connected to its output. By optimizing the load resistance, you can improve the output voltage swing of the LM324 to include the negative supply.


As I have often said a better answer follows a better question which often needs a bit more details.

what does TI's datasheet say about LM324 Vol?

The input offset voltage (Vio) of the LM324 OpAmp is specified in the Texas Instruments (TI) datasheet. According to the latest version of the TI LM324 datasheet (dated September 2021), the typical input offset voltage for the LM324 is 2 mV, with a maximum value of 7 mV. The input offset voltage drift over temperature is also specified, with a maximum of 18 μV/°C.
It's important to note that the input offset voltage is a characteristic of OpAmps that can affect their performance in precision applications. Input offset voltage is defined as the voltage difference that must be applied between the two inputs of an OpAmp to nullify the output voltage when the inputs are at the same voltage. The smaller the input offset voltage, the better the OpAmp's performance in precision applications.
The datasheet also provides additional information about the LM324, including its pin configuration, electrical characteristics, recommended operating conditions, and application information. It's important to review the datasheet carefully to understand the specifications and limitations of the LM324 for your specific application.
 
Well it did seem to read and know the contents of the LM324 datasheet when asked specifically for a value but its ability to design must be mixed a lot of newbie answers on the web.

I find design questions without measurable expectations are low quality.

But in this case I found ChatGPT will give a better answer than I did, although essentially similar.

e.g.

Browsers can't interpret .docx files because these files are proprietary formats created by Microsoft Word, which require specialized software to open and display their content properly. Browsers are designed to handle web content in HTML, CSS, and JavaScript formats, and they do not have the built-in capability to handle the complexities of .docx files.

To view .docx files, you need to have a software program like Microsoft Word, LibreOffice, or Google Docs installed on your computer that can read and display these file formats. Alternatively, you can use online file conversion services that can convert .docx files to a format that is compatible with web browsers, such as PDF or HTML.

" OpenAI says, I'm not perfect, neither are you."
 
Last edited:

LaTeX Commands Quick-Menu:

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top