I'm probably overthinking this, but whatever - comment welcome.
rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-16 8:48 am #1 2016-12-16 8:48 am #1 I'm probably overthinking this, but whatever - comment welcome. Attachments pre05.png 154.5 KB · Views: 348 pre05.asc 7.1 KB · Views: 46 ••• More options Share A AndrewT R.I.P. Joined 2004 2016-12-17 12:02 pm #2 2016-12-17 12:02 pm #2 I can see NFB around the Current Mirror. But I can't see feedback around more of the stages. Have you missed some part? ••• More options Share rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-17 3:31 pm #3 2016-12-17 3:31 pm #3 No, it runs as is, with the buffer open loop. I tried the sim with both connections, it seems more stable - less bandwidth - as shown. ••• More options Share H Hans Polak Member Joined 2005 2016-12-17 7:04 pm #4 2016-12-17 7:04 pm #4 I have added in LTspice .options numdgt=24 and .options plotwinsize=0 for greater accuracy. And I have given Q2 and Q8 their own bias with R14-R16. ••• More options Share rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-18 8:24 am #5 2016-12-18 8:24 am #5 Thanks for the suggestions. Actually your changes separate the buffer and bring it back to close to what I had originally before I smushed it into the current mirror of the CFA. numdgt=24 and .options plotwinsize=0 for greater accuracy Click to expand... Greater accuracy of the FFT it seems, but what exactly are those parameters you have set there? ••• More options Share sgrossklass Member Joined 2004 stephan.win31.de 2016-12-19 2:15 am #6 2016-12-19 2:15 am #6 The latter turns off lossy compression of transient waveforms. Basically a must if you need to sim low distortion. ••• More options Share rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-19 11:47 am #7 2016-12-19 11:47 am #7 Definitely a useful trick. I may have been tilting at windmills. The simple version of the circuit seems to work just as well as the complication introduced above. Attachments pre06 simple.asc 5.9 KB · Views: 37 pre06.png 76 KB · Views: 187 ••• More options Share Show hidden low quality content Status This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button. Share: Facebook Twitter Reddit Pinterest Tumblr WhatsApp Email Share Link Home Source & Line Analog Line Level Current feedback amplifier circuit (preamp/headphone amp)
A AndrewT R.I.P. Joined 2004 2016-12-17 12:02 pm #2 2016-12-17 12:02 pm #2 I can see NFB around the Current Mirror. But I can't see feedback around more of the stages. Have you missed some part? ••• More options Share rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-17 3:31 pm #3 2016-12-17 3:31 pm #3 No, it runs as is, with the buffer open loop. I tried the sim with both connections, it seems more stable - less bandwidth - as shown. ••• More options Share H Hans Polak Member Joined 2005 2016-12-17 7:04 pm #4 2016-12-17 7:04 pm #4 I have added in LTspice .options numdgt=24 and .options plotwinsize=0 for greater accuracy. And I have given Q2 and Q8 their own bias with R14-R16. ••• More options Share rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-18 8:24 am #5 2016-12-18 8:24 am #5 Thanks for the suggestions. Actually your changes separate the buffer and bring it back to close to what I had originally before I smushed it into the current mirror of the CFA. numdgt=24 and .options plotwinsize=0 for greater accuracy Click to expand... Greater accuracy of the FFT it seems, but what exactly are those parameters you have set there? ••• More options Share sgrossklass Member Joined 2004 stephan.win31.de 2016-12-19 2:15 am #6 2016-12-19 2:15 am #6 The latter turns off lossy compression of transient waveforms. Basically a must if you need to sim low distortion. ••• More options Share rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-19 11:47 am #7 2016-12-19 11:47 am #7 Definitely a useful trick. I may have been tilting at windmills. The simple version of the circuit seems to work just as well as the complication introduced above. Attachments pre06 simple.asc 5.9 KB · Views: 37 pre06.png 76 KB · Views: 187 ••• More options Share Show hidden low quality content Status This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button. Share: Facebook Twitter Reddit Pinterest Tumblr WhatsApp Email Share Link
I can see NFB around the Current Mirror. But I can't see feedback around more of the stages. Have you missed some part?
rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-17 3:31 pm #3 2016-12-17 3:31 pm #3 No, it runs as is, with the buffer open loop. I tried the sim with both connections, it seems more stable - less bandwidth - as shown. ••• More options Share H Hans Polak Member Joined 2005 2016-12-17 7:04 pm #4 2016-12-17 7:04 pm #4 I have added in LTspice .options numdgt=24 and .options plotwinsize=0 for greater accuracy. And I have given Q2 and Q8 their own bias with R14-R16. ••• More options Share rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-18 8:24 am #5 2016-12-18 8:24 am #5 Thanks for the suggestions. Actually your changes separate the buffer and bring it back to close to what I had originally before I smushed it into the current mirror of the CFA. numdgt=24 and .options plotwinsize=0 for greater accuracy Click to expand... Greater accuracy of the FFT it seems, but what exactly are those parameters you have set there? ••• More options Share sgrossklass Member Joined 2004 stephan.win31.de 2016-12-19 2:15 am #6 2016-12-19 2:15 am #6 The latter turns off lossy compression of transient waveforms. Basically a must if you need to sim low distortion. ••• More options Share rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-19 11:47 am #7 2016-12-19 11:47 am #7 Definitely a useful trick. I may have been tilting at windmills. The simple version of the circuit seems to work just as well as the complication introduced above. Attachments pre06 simple.asc 5.9 KB · Views: 37 pre06.png 76 KB · Views: 187 ••• More options Share Show hidden low quality content Status This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button. Share: Facebook Twitter Reddit Pinterest Tumblr WhatsApp Email Share Link
No, it runs as is, with the buffer open loop. I tried the sim with both connections, it seems more stable - less bandwidth - as shown.
H Hans Polak Member Joined 2005 2016-12-17 7:04 pm #4 2016-12-17 7:04 pm #4 I have added in LTspice .options numdgt=24 and .options plotwinsize=0 for greater accuracy. And I have given Q2 and Q8 their own bias with R14-R16. ••• More options Share rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-18 8:24 am #5 2016-12-18 8:24 am #5 Thanks for the suggestions. Actually your changes separate the buffer and bring it back to close to what I had originally before I smushed it into the current mirror of the CFA. numdgt=24 and .options plotwinsize=0 for greater accuracy Click to expand... Greater accuracy of the FFT it seems, but what exactly are those parameters you have set there? ••• More options Share sgrossklass Member Joined 2004 stephan.win31.de 2016-12-19 2:15 am #6 2016-12-19 2:15 am #6 The latter turns off lossy compression of transient waveforms. Basically a must if you need to sim low distortion. ••• More options Share rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-19 11:47 am #7 2016-12-19 11:47 am #7 Definitely a useful trick. I may have been tilting at windmills. The simple version of the circuit seems to work just as well as the complication introduced above. Attachments pre06 simple.asc 5.9 KB · Views: 37 pre06.png 76 KB · Views: 187 ••• More options Share Show hidden low quality content Status This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button. Share: Facebook Twitter Reddit Pinterest Tumblr WhatsApp Email Share Link
I have added in LTspice .options numdgt=24 and .options plotwinsize=0 for greater accuracy. And I have given Q2 and Q8 their own bias with R14-R16.
rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-18 8:24 am #5 2016-12-18 8:24 am #5 Thanks for the suggestions. Actually your changes separate the buffer and bring it back to close to what I had originally before I smushed it into the current mirror of the CFA. numdgt=24 and .options plotwinsize=0 for greater accuracy Click to expand... Greater accuracy of the FFT it seems, but what exactly are those parameters you have set there? ••• More options Share sgrossklass Member Joined 2004 stephan.win31.de 2016-12-19 2:15 am #6 2016-12-19 2:15 am #6 The latter turns off lossy compression of transient waveforms. Basically a must if you need to sim low distortion. ••• More options Share rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-19 11:47 am #7 2016-12-19 11:47 am #7 Definitely a useful trick. I may have been tilting at windmills. The simple version of the circuit seems to work just as well as the complication introduced above. Attachments pre06 simple.asc 5.9 KB · Views: 37 pre06.png 76 KB · Views: 187 ••• More options Share Show hidden low quality content Status This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button. Share: Facebook Twitter Reddit Pinterest Tumblr WhatsApp Email Share Link
Thanks for the suggestions. Actually your changes separate the buffer and bring it back to close to what I had originally before I smushed it into the current mirror of the CFA. numdgt=24 and .options plotwinsize=0 for greater accuracy Click to expand... Greater accuracy of the FFT it seems, but what exactly are those parameters you have set there?
sgrossklass Member Joined 2004 stephan.win31.de 2016-12-19 2:15 am #6 2016-12-19 2:15 am #6 The latter turns off lossy compression of transient waveforms. Basically a must if you need to sim low distortion. ••• More options Share rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-19 11:47 am #7 2016-12-19 11:47 am #7 Definitely a useful trick. I may have been tilting at windmills. The simple version of the circuit seems to work just as well as the complication introduced above. Attachments pre06 simple.asc 5.9 KB · Views: 37 pre06.png 76 KB · Views: 187 ••• More options Share Show hidden low quality content Status This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The latter turns off lossy compression of transient waveforms. Basically a must if you need to sim low distortion.
rjm Member Joined 2004 phonoclone.com Paid Member 2016-12-19 11:47 am #7 2016-12-19 11:47 am #7 Definitely a useful trick. I may have been tilting at windmills. The simple version of the circuit seems to work just as well as the complication introduced above. Attachments pre06 simple.asc 5.9 KB · Views: 37 pre06.png 76 KB · Views: 187 ••• More options Share
Definitely a useful trick. I may have been tilting at windmills. The simple version of the circuit seems to work just as well as the complication introduced above.
