Bravo V2 Measurements
Aug 11, 2015 19:16:34 GMT
Post by maxhawk on Aug 11, 2015 19:16:34 GMT
Hi all. This is my first time posting here. I bought a Bravo V2 recently on a whim just to see what a $70 tube headphone amp could do. At work I have access to an Audio Precision SYS2722 audio analyzer and decided to measure the Bravo like I did with my Bottlehead Crack. All the data was exported to an Excel sheet where these graphs were created. Input level for all measurements was 1Vrms and the output level (1Vrms unless indicated otherwise) was adjusted using the amp's volume knob. Bias for each tube was adjusted to 13.5VDC at the plate. The circuit was allowed to warm up until the output level and distortion stabilized. Graphs are referenced to 1KHz.
My mods are:
Below is the frequency response of the stock amp with the stock Shuguang 12AU7 tube. Load is 300 ohms and output level is 1Vrms.
This is the distortion (measured in percent) of the stock amp. I can't explain the strange behavior at 200mV and 500mV output levels.
The following is the frequency response of the modded V2 at various output levels:
And distortion of modded amp at various output levels:
I tried various input tubes and readjusted bias for each one. I'm really surprised at the difference in low end response of the various tubes.
And distortion of the various input tubes:
And to more easily show the difference between stock and modded, here's the stock Shuguang before and after. The loss of bass is due to reduction of the output cap from 1000uF to 330uF. The crack uses 100uF so I figured 330uF would be fine. However I'm not sure why there's a difference at the high end as well.
The higher distortion at the low end might be due to the rolled off frequency response. Maybe at some point I'll try paralleling a larger cap to see if distortion changes.
My mods are:
- Replace IRF610 with IRL530
- Replace LM317 with LM317A
- Replace 1uF input caps with 1uF non-polarized MUSE caps
- Replace 6800uF/25V bulk filter cap with Nichicon 4700uF/35V KA series 105°C cap
- Replace stock 1000uF output caps with Nichicon 330uF/25V MUSE KZ series cap
- Replace 3K bias resistors with 5K pot
- Add 100uF cap across red LED
- Replace stock pot with Alps RK097 audio taper pot
Below is the frequency response of the stock amp with the stock Shuguang 12AU7 tube. Load is 300 ohms and output level is 1Vrms.
This is the distortion (measured in percent) of the stock amp. I can't explain the strange behavior at 200mV and 500mV output levels.
The following is the frequency response of the modded V2 at various output levels:
And distortion of modded amp at various output levels:
I tried various input tubes and readjusted bias for each one. I'm really surprised at the difference in low end response of the various tubes.
And distortion of the various input tubes:
And to more easily show the difference between stock and modded, here's the stock Shuguang before and after. The loss of bass is due to reduction of the output cap from 1000uF to 330uF. The crack uses 100uF so I figured 330uF would be fine. However I'm not sure why there's a difference at the high end as well.
The higher distortion at the low end might be due to the rolled off frequency response. Maybe at some point I'll try paralleling a larger cap to see if distortion changes.