Sansui BA-F1 clone

[Morten Gundersen]

Hi!
Together with a friend of mine I am about to build a Sansui BA-F1 clone. So far the PCB(s) have been made and some parts are have been ordered. This project will probably take a while and any constructive advice is welcome!

Sansui is one of the companies that uses special diode parts containing 3 and 4 separate diodes in the package. I have made a shortcut and used an LED instead. Hope this work. Anyway, it is easy to switch to 3x1N4148 if needed. The board we are using is a "driver board", i.e. a separate OPS is needed. This is also finished.

The original uses only 2+2 oputput transistors but we decided on an OPS that could use up to 7+7 output transistors. You may note the extra driver stage on the OPS. This may or may not be used according to the jumper setting. This way, the OPS could be of more universal use in other projects.

Here is what we have so far:

Driver schematic:

Driver PCB:

OPS schematic:

OPS PCB:

PCBs:

morten

 

[Alan]

Appreciate you sharing your project. Seems like a well-thought-out build with some clever flexibility incorporated.

 

[barfle]

This looks like quite a project! I’m curious, though, have you breadboarded the circuit? Circuit boards can be tough to modify if there’s an unforeseen issue. Maybe software simulation? If so, I’d like to try it myself.

 

[Morten Gundersen]

No breadboarding, no simulation. Sorry about that! This is a service manual clone, so it depends on the original circuit to be correct!
I am using KiCAD and the project folder (and anything else I have) is available to anyone interrested.

morten

 

[Alan]

I believe KiCAD is great for sharing, it's so cool that you've made the project available for anyone interested.

 

[Morten Gundersen]

Here are 2 KiCAD folders and 1 misc doc folder to play with.

 

[Alan]

Thanks for sharing those. I really appreciate having the KiCAD folders and the extra documents to dive into.

 

[Morten Gundersen]

Things are slowly starting to materialize!

Today's status:

 

[Alan]

Looking great! That Sansui BA-F1 driver board is really coming together nicely. Layout is neat, soldering looks sharp, and Helion’s 2022 touch adds a modern flair. Definitely turning into a solid build.

 

[Morten Gundersen]

About to finalizing the driver board. First some cooling for the drivers. For initial test, the bias transisitor is placed on the heatsink as well. If necessary, I will move it to the main heatsink later. Next is matching and soldering the small signal transistors.

 

[Alan]

Nice work on the progress! Cooling first is a smart approach. Good luck with the matching and your initial test.

 

[Morten Gundersen]

Only one thing left is trying to find an appropriate LED (D1 & D2) to replace the odd 3-diode part in the original circuit.

 

[Morten Gundersen]

Some progress:

Yesterday we tested the driver board. After some hefty bias current and a little smoke, we discovered we used TO-220 for bias transistor instead of TO-126/TO-92. Changing to the original TO-92, the driver board seemed to function perfectly.

Offset easily adjusted to 0.0 mV (drifted a little during the test but just slightly).
Bias adjusted to minimum gave 20mA through "output stage", which remained ccol during the test. Easyly adjustable to higher current.
The we connected a speaker and an internet radio. First impression was impressive! Sound was clean with no obvious distortion or artefacts.
Suppose the driver board would make a seriously good (and expensive) headphone amplifier on its own!

We tested several LEDs to achieve the desired 2.0-2.1mV to replace the SV-03 tripple diode, at about 8-9mA. Some old orange ones made the trick!
We replace the input FETs (2SK129) with what we had, the 2SK146. Seemed to work fine.

So now the next task is to make and test the ouput boards. Still some components to buy, but we feel optimistic at the moment!

 

[Alan]

That's great progress, good job on the bias transistor swap. The boards look really clean and well-organized, especially the driver section and the silkscreen on the output board. It sounds like you’re almost there I’m looking forward to hearing how the output stage performs.

 

[Morten Gundersen]

5W res and power transistors arrived. Next is building PSU for the input stage. Or maybe an initial test with common supply....