Thinking about building another fully discrete phono stage...

[stratokaster83]

Thanks Malcolm, if all is well after testing I will release the Gerber files for non-commercial use

[stratokaster83]

Despite saying that I was going to leave the circuit alone, I couldn't help myself and added capacitor decoupling to the constant current sources 

But now the prototype PCBs are already in production, hopefully I should be able to build a test unit soon.

[damiena]

I'm waiting to hear your verdict 

[stratokaster83]

Also thinking about building a discrete MC frontend, the main issue is that I no longer have an MC cartridge to test it.

This simulates extremely well, and using something like THAT300 would eliminate the painful part (hand-matching transistors by both Vbe and beta).



THAT300 does not have the lowest base spreading resistance but it's still decently low (around 30 Ohms), so paralleling all four transistors in the package together should result in a very low noise amplifier.

The gain is approximately equal to R2/R1, but in reality it also depends on the transistors and the source impedance of the cartridge, 120/10 should be a good enough compromise for the majority of MC cartridges on the market.

[stratokaster83]

 

[vs music]

Also thinking about building a discrete MC frontend, the main issue is that I no longer have an MC cartridge to test it.

This simulates extremely well, and using something like THAT300 would eliminate the painful part (hand-matching transistors by both Vbe and beta).



THAT300 does not have the lowest base spreading resistance but it's still decently low (around 30 Ohms), so paralleling all four transistors in the package together should result in a very low noise amplifier.

The gain is approximately equal to R2/R1, but in reality it also depends on the transistors and the source impedance of the cartridge, 120/10 should be a good enough compromise for the majority of MC cartridges on the market.

Hello !
There is a small error lurking in your scheme. The BC559 transistor ( Q3 )  is a p-n-p.
There should be an n-p-n transistor in this place.
Or there should be another inclusion of this transistor .

[stratokaster83]

Hello !
There is a small error lurking in your scheme. The BC559 transistor ( Q3 )  is a p-n-p.
There should be an n-p-n transistor in this place.
Or there should be another inclusion of this transistor .

Yes, thank you Victor, I've been playing with different transistors in LTspice and forgot to mirror this one. The emitter and collector connections should be reversed.

[stratokaster83]

Another version, using ZTX851 transistors which are famous for their incredibly low Rbb (less than 2Ohm). Should be significantly lower noise than the THAT300 version, but not as foolproof because Q101-104 and Q201-204 should be matched by both Vbe and beta.

[stratokaster83]

Populating one MM board would cost approx. 39 euro at today's prices, or 78 euro for a stereo pair. The largest part of the cost, as usual, is in precision capacitors for the RIAA network.

https://www.mouser.ie/ProjectManager/ProjectDetail.aspx?AccessID=811a902cab