RCA BA-31B Inspired DI

[BigDad]

Hi there,

I am designing a Direct Box inspired by the RCA BA-31B. To be more accurate, it uses the output stage of the BA-31B, while the input follows the topology from Rod Elliott's website (Figure 7.5):

https://sound-au.com/articles/jfet-design.htm

Here is the schematic of the full DI:



I have built a prototype, and it seems to work fine at first glance, but there are some aspects that disturb me.

Basically, the DI sounds really good, but there's a kind of upper-frequency fry or fuzzy sound that superimposes on the signal when I play loud with a guitar or a bass. It’s really subtle, but it makes the sound not totally clean. I'm not sure if this is coming from the input or the output stage. I plan to connect the input stage with an output capacitor directly to the transformer to be sure. But I want to know if BA-31 users have experienced this with the original design?

Also, this class A output design is clearly not the most efficient, and I don’t really understand how to set the R7 value. I’ve tried adjusting it by using a trimmer, but it doesn't seem to change anything except increasing or decreasing the power dissipated by Q3 and Q4.

I'm also unsure about the best way to calculate the values for R4 and R5. I left them at 2.7k because it seems to work with the 2N4416, and Q2 seems to auto-bias the whole thing, but maybe there's a better tuning?

I will post harmonic and frequency response analysis results, but I didn't see anything wrong during the first tests.

I would be honored to share build files with the community if I can figure out how to fix this problem.

 

[BigDad]

Here is what I’ve measured at 1Vrms on the input. Everything looks okay in this configuration.






I’ve also recorded and isolated the fuzz sound I was talking about.
I used an Apollo X and a Precision Bass with both full tone (A) and tone roll-off (B). The gain was leveled to +50dB in the DAW.

I also tested the first stage alone, and the issue seems to come from here.

 

[MidnightArrakis]

[and the issue seems to come from here] -- Then what happened???

/

 

[BigDad]

I thought the problem was coming from the first stage because it seemed more obvious when using the first stage only. But I think it was actually coming from the mic input of the interface. Using an external mic pre into the UA Line input gives a much cleaner sound.

Also, there were output transistors that were overheating due to the lack of a resistor between the -15V and the output stage V- (like in the original RCA design). This affected the Q4 voltages and caused the entire output stage to consume more current than needed.

I also tried replacing Q2 with a 2N404 to get a higher DC offset at the output stage input, aiming to get closer to the original bias and design spirit. It works great in this design as well. The sound is a bit more compressed with more harmonics, and I personally preferred the result with the 2N4037.

I’ve made a new PCB design and am now waiting for delivery to try a cleaner prototype assembly.