AI Designed Phono Pre-Amp?

[Bonsai]

This thread belongs over in the analog source forum. This forum is for power amplifiers. Maybe one of the mods can move it.

😊

 

[MarcelvdG]

Not in the Software Tools forum?

 

[Bernhard]

How about moving C3 and C6 to the collectors? At least C2 and C5 could make sense as ac bypass for 1k emitter degeneration and both stages had gain.

 

[anatech]

Hi Bonsai, Bernhard,
Either one. But Bonsai posted first so there it goes. 🙂

 

[cabirio]

How about moving C3 and C6 to the collectors? At least C2 and C5 could make sense as ac bypass for 1k emitter degeneration and both stages had gain.

The bases are still unbiased, and I don't see anything resembling a RIAA network. You can add all that, plus an output EF buffer, readjust component values as required... In short, scrap it and start from scratch. I suggest googling or searching the forum for "3-transistor RIAA preamp" or something like that.

 

[Bernhard]

As near as possible to the original schematic, gain and riaa maybe not ideal.

 

[rayma]

There is no 75uS time constant provided.
The source impedance seen by the RIAA network has not been taken into account.

 

[Bernhard]

I'm not into phono, so only copied a network from google 😎

 

[as_audio]

In post 29 we read that the AI can not do drawings, this means that all this discussion is fruitless, because the drawing in post 1 may not represent the AI output result. As far as I can see we do not know the original question and the answer of the program used.

 

[cabirio]

Thread 'Simple Transistor based RIAA'

2023-07-14 8:55 am

So the good news. I got my Vintage Woodstock Amp up and running.
The bad news it Ceramic and Magnetic Head Phono Sections seem to be simple amplifiers.
There is no RIAA equalisation.
After spending a month studying the various possible solutions to this I have short listed two.
The requirements are simple. No op amp must be transistor based.
The Akai I have owned back in 79. I know it sounded and worked well. Even though its so basic.
The Self based design I have not experienced.
I guess the long and short of it is that I need to enter up both in a simulator and see how they...

• chinoy
• Replies: 60
• Forum: Analog Line Level

 

[rayma]

I'm not into phono, so only copied a network from google 😎

Apparently Google isn't into phono either.

 

[Bernhard]

riaa better ?

 

[Bernhard]

Apparently Google isn't into phono either.

Just forgot the cap

 

[bucks bunny]

At a first glance I see poor noise performance as well as poor power supply rejection. Why promote a poor design? Better approaches are well known for decades.

 

[Bernhard]

because this is what the op put here for discussion

 

[bucks bunny]

The TO's starting design is a bad joke, not worth considering to "improve" it.

 

[rayma]

riaa better ?

Only if the first stage has zero output impedance.
You have to include the effect of the source impedance due to the first stage, which is a little under 10k.

As you can see from the KAB calculator, the effect is quite substantial.
The values become 18k, 2.617k, 0.121uF, 0.0417uF.

 

[anatech]

I think the point has been made. A.I. can't design circuits without guidance (ie, you know what you want to start with). Unless you know what to expect, you can't trust it either.

The same holds true for anything else A.I. generated. Search results, answers to questions. Go to the source or ask someone who actually does know.

 

[diyrayk]

Let's engage the AI to conduct a virtual listening test of the circuit and write a review. Who knows, it may glean golden ear phrases from other high end reviews on the internet and praise the inky black backgrounds, etc. It may also mention that it has contacted the designer, who is investigating what's causing the low gain.

As a litmus test, try asking Google AI a complex question that you yourself expertly know the answer to and see what you get. It won`t instill a good feeling of confidence in AI.

Ray K