I want to build a Hawksford I/V for AD1955, a DAC that puts out 8.4 mA (plus bias). I find power dissipation is simply too high for the usual 2SA970/2SC2240 if I raise voltage enough to get them into a reasonable operating point with that kind of current from the DAC. So I need recommendations for medium power BJTs that are similarly linear (or better if such are around, of course) for audio use, and also have a high fT, like 100 MHz.
Interesting. I'll try them. Not sure what effect the lower hFE will have exactly. There are no noise specs in the datasheet :|
I/V: Fig.4-4 in http://www.essex.ac.uk/dces/researc...Current steering transimpedance amplifier.pdf
I/V: Fig.4-4 in http://www.essex.ac.uk/dces/researc...Current steering transimpedance amplifier.pdf
Hello Abzug,
thank you for posting the AES paper, I was looking for some time for this.
@roender,
I didn't had the chance to congratulate you for your project. It's a great sample of design and enginering; and the thread it's a great learning source!
Am avut un sentiment de mandrie sa vad reactiile celor de pe DIYAudio la un produs romanesc! 🙂
thank you for posting the AES paper, I was looking for some time for this.
@roender,
I didn't had the chance to congratulate you for your project. It's a great sample of design and enginering; and the thread it's a great learning source!
Am avut un sentiment de mandrie sa vad reactiile celor de pe DIYAudio la un produs romanesc! 🙂
With the 2SA970/2SC2240, removing bias from the DAC output and cutting current by half, then doubling degeneration resistors and halving the feedback resistor, I get 1 kHz THD at around 1 ppm (though it goes higher at 20 kHz).
There are various problems with this I/V, including the fact that it defeats steep lowpass filters and limits their attenuation, and is unstable with some filters as well.
I think with medium power transistors and less degeneration it should be possible to get headphone-level voltage without the need for additional voltage gain, just a buffer. Also, improving some of the current sources seems to improve THD in the simulation.
There are various problems with this I/V, including the fact that it defeats steep lowpass filters and limits their attenuation, and is unstable with some filters as well.
I think with medium power transistors and less degeneration it should be possible to get headphone-level voltage without the need for additional voltage gain, just a buffer. Also, improving some of the current sources seems to improve THD in the simulation.
- Status
- Not open for further replies.