Author Topic: Dealing with AI posts  (Read 7712 times)

0 Members and 1 Guest are viewing this topic.

Offline floobydustTopic starter

  • Super Contributor
  • ***
  • Posts: 9057
  • Country: ca
Dealing with AI posts
« on: March 30, 2026, 08:20:01 pm »
I "Report to moderator" a 1st post in a necro thread as being AI, and added a reply that it is full of technical rubbish.
I don't want bullshit AI content that is erroneous, ruining threads.
Now see my post is deleted, yet the rubbish post lives on. WTF is going on?

I know this is an old post, but  looked into a resolution for this, seeing it is pertinent to figure something out.

First off, you are looking like you are working on a B&K 501A Semiconductor Curve Tracer ( as  am also researching the same )

Q's 5, 6, 11 and 12  are listed as a SPS2036 NPN, small signal transstor, with a C-B-E ( Base=Center) pin configuration.

In positions Q5~6, they are used as buffer / drivers for the signal needed for IC2, and they isolate the unfiltered DC pulses from the full wave bridge +/- 15V supply to the IC2 Input.
They most likely are capable of at least a 40V Veco ( Voltage between Collecter and Emitter) operation configurtion rating

Q's 11~12 are not really being used as transitors providing gain. They are configured to act as a "DIAC" component , or “Diode for Alternating Current” , which is usually done with a diode of a specific rating that conducts bi-directionally at a specific breakdown voltage in order to protect other components from damage.

(Attachment Link)
Looking at the schematic and particular configuration, they are taking advantage of said transistors "Vebo", or, its reverse biased juncton breakdown voltage between the Emitter and Base, which in
most small signal transtors is about 5 VDC or above, but can be up to 10V when measured. Configured as such ( Back to back and reversed to each other), you would find that the 2 components would conduct from base to base once the 0.7V forward biased B-E junction, and its connected reverse biased E-B junctiion voltage ( about 5V or slightly higher) is met and surpassed. The base to base potential of the AC stairstep signal that the connections are between, is set to 5 volts, so, this DIAC configuration would protect a connected "Devce Under Test" from a potental surge that could damage it.  B&K  warehousing must have had a glut of these particular transistors available, and engineers just decided to use them in a crafty configuration instead of paying more for select DIAC diodes.

 With that being said, most any small signal NPN transistor with a Veco of 40 or more Volts DC, and a Vebo of approxmately 5~6V would suffice as a replacement. The originals are described as having a TO-92 package with a C-B-E pin config ( Base in the center) so any small signal NPN such as a 2N3904 or 2N2222 would most likely be a workable substitute. One could also substitute the 2 transistors with 2 Zener diodes of approx 6 volts breakdown ( creating a circuit with approx 6.7V to start conducting) and configure them back to back ( Either Anodes or Cathodes facing each other) to achieve a similar voltage breakdown circuit with a predictable voltage level instead of "whatever the Vebo voltage is" on 2 chosen transistors.
 

Offline thm_w

  • Super Contributor
  • ***
  • Posts: 9756
  • Country: ca
  • Non-expert
Re: Dealing with AI posts
« Reply #1 on: March 30, 2026, 10:04:19 pm »
I doubt its an AI account, and the text doesn't look AI written.

Plenty of people post misinformation here, not much you can do other than disagree with them, unless they are purposefully trolling then you'd report them.
Profile -> Modify profile -> Look and Layout ->  Don't show users' signatures
 

Offline ebastler

  • Super Contributor
  • ***
  • Posts: 7730
  • Country: de
Re: Dealing with AI posts
« Reply #2 on: March 31, 2026, 06:02:15 am »
It's clearly not an AI post. Looks like someone who is working on the same type of curve tracer, came across that old thread in a search, and wanted to add what he (thinks he) figured out. It may be incorrect, but it's genuine.

I did not see your reply in the thread; maybe the moderator found it too aggressive? If you just re-post a factual technical comment, everything should be alright.
 

Offline floobydustTopic starter

  • Super Contributor
  • ***
  • Posts: 9057
  • Country: ca
Re: Dealing with AI posts
« Reply #3 on: March 31, 2026, 07:24:16 pm »
It sets off my "AI response" flag: 1PBTID, 2 year necro thread plus the reply has a handful of technical errors. It's a 2306 not 2036.
Anyone who knows what a DIAC is compared to the circuit will raise an eyebrow. Plus the AI verbose encyclopedia explanations that are actually incorrect.
I just dislike a thread getting polluted with AI snippets and misinformation, then my post to the bot got deleted.

Yesterday, a thread vanished after another member questioned if it's an AI bot post doing some marketing sounding stuff.
 

Online antenna

  • Frequent Contributor
  • **
  • Posts: 607
  • Country: us
Re: Dealing with AI posts
« Reply #4 on: March 31, 2026, 08:53:02 pm »
It sets off my "AI response" flag
Opinions may differ, but I agree with you.
 

Offline thm_w

  • Super Contributor
  • ***
  • Posts: 9756
  • Country: ca
  • Non-expert
Re: Dealing with AI posts
« Reply #5 on: March 31, 2026, 10:05:21 pm »
It sets off my "AI response" flag: 1PBTID, 2 year necro thread plus the reply has a handful of technical errors. It's a 2306 not 2036.

Put the flag away for now:
- "SPS2036" is not a part as you say, so, why would an LLM generate it
- Double spacing in some places and not others
- Various other typos
- Post was edited after a few minutes
- Screenshot of paint.net with edited in text
- User has an odysee channel with multiple videos
Profile -> Modify profile -> Look and Layout ->  Don't show users' signatures
 
The following users thanked this post: Psi, Siwastaja, ebastler

Offline Psi

  • Super Contributor
  • ***
  • Posts: 12588
  • Country: nz
Re: Dealing with AI posts
« Reply #6 on: March 31, 2026, 10:12:01 pm »
The text style looks very human to me.
Greek letter 'Psi' (not Pounds per Square Inch)
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf