The henrik x-bsoz needs 60volt rails so your transformers seems to do the work ok.
If you want you can run the circuit at the same current with half the voltage rails simply dividing the resistors values par 2.
I've try it and it works fine , and it has even lower output impedance.
The resistor are R 101 and 102;
R 103 104 105 and 106
If you want you can run the circuit at the same current with half the voltage rails simply dividing the resistors values par 2.
I've try it and it works fine , and it has even lower output impedance.
The resistor are R 101 and 102;
R 103 104 105 and 106
Vit
I went for the +-60V rails because Nelson regards this as optimum, and you can make it with 2*60V~ transformer for both channels and with separate rectifiers and regulators for each channel. Your transformer at 120VA is more than enough.
I havn´t heard x-bsoz with +-30V rails, but if I should make it, I would do it like Stefano just suggested.
I went for the +-60V rails because Nelson regards this as optimum, and you can make it with 2*60V~ transformer for both channels and with separate rectifiers and regulators for each channel. Your transformer at 120VA is more than enough.
I havn´t heard x-bsoz with +-30V rails, but if I should make it, I would do it like Stefano just suggested.
Re: quiescent current
The higher quiescent current is the better, as the gain is higher and the linearity is better. There's a practical limit, but dissipation and other constraints will get you first, so just design for higher current within those limits, and it'll work out.
Peter Huang said:
The higher quiescent current is the better, as the gain is higher and the linearity is better. There's a practical limit, but dissipation and other constraints will get you first, so just design for higher current within those limits, and it'll work out.
Grataku
Easy done
http://www.diyaudio.com/forums/showthread.php?threadid=3949&pagenumber=17
and download of scematic
http://www.diyaudio.com/forums/attachment.php?postid=67322
Easy done
http://www.diyaudio.com/forums/showthread.php?threadid=3949&pagenumber=17
and download of scematic
http://www.diyaudio.com/forums/attachment.php?postid=67322
So....I did Henrik mods to my BOSOZ turning it into an X-bosoz I guess. The schematic I used is the link a couple of posts ago.
It involves taking a regular BOSOZ with passive CS, shorting the gain resistor/pot combo and a couple more resistors to get max open loop gain of 30 dB, and then chopping off 20 dB via X-feedback. Henrik has simulated...have you actually built it?... this X-ed up version to get a 10 dB gain in the end.
Results? It works, and -seems- to sound rather different. I maybe dreaming, but I don't think so, but I hear better mid/lows (by mid I mean cello frequencies), better imaging, and just a smitherine crispier highs. By crispy I mean just that, not bright!
Bottom line, I like it alot.
It would be nice to do a rigorous A_B between a reg. BOSOZ and the X-version.
Here it is in a great looking Schroff chassis I've had around for 15 years, I wonder if they still make those, an EAO switch, another artifact from ancient history. I used black beauty pots (10k on the outputs), Neutrik XLR, vampire RCA etc, etc, etc.
I forgot to take the pic of the inside...it is as neat as the outside. You are just going to have to take my word for it. 😉
It involves taking a regular BOSOZ with passive CS, shorting the gain resistor/pot combo and a couple more resistors to get max open loop gain of 30 dB, and then chopping off 20 dB via X-feedback. Henrik has simulated...have you actually built it?... this X-ed up version to get a 10 dB gain in the end.
Results? It works, and -seems- to sound rather different. I maybe dreaming, but I don't think so, but I hear better mid/lows (by mid I mean cello frequencies), better imaging, and just a smitherine crispier highs. By crispy I mean just that, not bright!
Bottom line, I like it alot.
It would be nice to do a rigorous A_B between a reg. BOSOZ and the X-version.
Here it is in a great looking Schroff chassis I've had around for 15 years, I wonder if they still make those, an EAO switch, another artifact from ancient history. I used black beauty pots (10k on the outputs), Neutrik XLR, vampire RCA etc, etc, etc.
I forgot to take the pic of the inside...it is as neat as the outside. You are just going to have to take my word for it. 😉
Attachments
Grataku
What a nice peace, congratulation!
I posted my job on XBOSOZ under the X SOZ thred since I made them at the same time and optimized the interface between the two amps.
I also did some tests on CCS and XBOSOZ here on the post "Some tests on CCS´s. ": http://www.diyaudio.com/forums/showthread.php?postid=75141#post75141
Happy listening!
What a nice peace, congratulation!
I posted my job on XBOSOZ under the X SOZ thred since I made them at the same time and optimized the interface between the two amps.
I have, you can se it here http://www.diyaudio.com/forums/showthread.php?postid=102063#post102063
I would like so too, my problem was, that I did the XSOZ first and then the XBOSOZ, so A-B test would not be true to the orig. BOSOZ since it what so ever is capable to drive a singleended load at 120 Ohm, only the X´ed version liks this.
I also did some tests on CCS and XBOSOZ here on the post "Some tests on CCS´s. ": http://www.diyaudio.com/forums/showthread.php?postid=75141#post75141
Happy listening!
X-BSOZ sound
I just converted my BSOZ to Henrik's X-BSOZ design using the passive current source resisitors, but I kept the gain resistor at 680R.
I put the 39K feed back resisitors on a switch so that I could easily flip back and forth from BSOZ to X.
The X version with 680R gain resistor has a very very slightly more focused image. I could not detect any other changes. The output voltage at either switch setting was identical.
I am now going to short out the gain resistor next and see what that does.
I just converted my BSOZ to Henrik's X-BSOZ design using the passive current source resisitors, but I kept the gain resistor at 680R.
I put the 39K feed back resisitors on a switch so that I could easily flip back and forth from BSOZ to X.
The X version with 680R gain resistor has a very very slightly more focused image. I could not detect any other changes. The output voltage at either switch setting was identical.
I am now going to short out the gain resistor next and see what that does.
I just shorted out the 680R gain resisitor.
With the gain resistor shorted out I get way too much gain.
I am at the lowest position on my antenuator and the sound volume is as loud as I can stand.
The sound quality is very interesting. Crispy but not edgy as used by Grataku captures my initial impressions.
It is not as warm as the BSOZ but it is not icey cold either. The image is amazing.
I use an attenuator before the output caps which puts a resistor bewteen + and -.
I need to drastically decrease the gain.
Will I still get a benefit of the "X effect" if I go back to using a gain resistor around 100 to 300R instead of a piece of wire or should I decrease the value of R122/R123 from 39K to around 20K.
ALso I could bleed off more input voltage via a voltage divider?
With the gain resistor shorted out I get way too much gain.
I am at the lowest position on my antenuator and the sound volume is as loud as I can stand.
The sound quality is very interesting. Crispy but not edgy as used by Grataku captures my initial impressions.
It is not as warm as the BSOZ but it is not icey cold either. The image is amazing.
I use an attenuator before the output caps which puts a resistor bewteen + and -.
I need to drastically decrease the gain.
Will I still get a benefit of the "X effect" if I go back to using a gain resistor around 100 to 300R instead of a piece of wire or should I decrease the value of R122/R123 from 39K to around 20K.
ALso I could bleed off more input voltage via a voltage divider?
Re: X-BSOZ sound
ignatz
great!
Yes, driving it with balanced the high gain becomes somewhat of a nuisance (volume at 1/4 of the way for me), with se it's ok. However, it's sounds so much better if driven balanced. I will never go back to se as far as I am concerned.
Maybe I am saved by using a low gain in the AX with 10K as input resistors.
With regard to the gain I would think that everything could be scaled back if maintainig the same ratio of x-feedback. So if one lowers the gain it should increase the value of x-fdb resistor. I think Henrik would be much more efficient than me in simulating the circuit. I wonder if it will change the frequency response...
ignatz
great!
Yes, driving it with balanced the high gain becomes somewhat of a nuisance (volume at 1/4 of the way for me), with se it's ok. However, it's sounds so much better if driven balanced. I will never go back to se as far as I am concerned.
Maybe I am saved by using a low gain in the AX with 10K as input resistors.
With regard to the gain I would think that everything could be scaled back if maintainig the same ratio of x-feedback. So if one lowers the gain it should increase the value of x-fdb resistor. I think Henrik would be much more efficient than me in simulating the circuit. I wonder if it will change the frequency response...
Hi Ignatz and Grataku!
So nice to have someone to share the XBOSOZ with.
Your problem is elsewhere. As you write, you have an attenuator between the drains of the fets. If that is going to work, you would want a <1K log pot, and even then I believe you would have some serious linearity problems. Remember that XBOSOZ has an extreme low outputimpedance at about 10 Ohm singleended, so it need some real low resistive values to be attenuated.
If you want to adjust the gain, don´t use resistors between the sources of the fets, then you will loose some of the X-comunication, that is why it didn´t sound much different when you had the original 680 Ohm gainresistor.
If you uses both singleneded an balanced mode, you could have separate inputresistors for each mode, 10K for singleended and 22K for balanced operation, thus you will obtain the same volumelevel in both modes, something like this is used in the Aleph´s.
If you want to have adjustable gain you should have say 50K lin pots in series with some 15K resistors as replacemens for R122/123.
The methods mentiond will not harm the frequency responce, at least my simulator tells me.
When it comes to the sound of the XBOSOZ as you both describes it, I really reconize my own experience with my XBOSOZ and XSOZ, the imaging is som much better with the X, all contrasts between soft and crispy is just great.
At the moment I use a singleended CD, but I am about to modyfi it to descrete balanced mode.
🙂
So nice to have someone to share the XBOSOZ with.
May be you have too much gain, but I doubt, you should always have a little extra gain. The gain of the circiut is about 11db with singleended loads from 100 Ohm to 100K load, by comparation Aleph 1.7 has ajustable gain between 10 and 30 db as I recall, so 11 db isn´t too much.
Your problem is elsewhere. As you write, you have an attenuator between the drains of the fets. If that is going to work, you would want a <1K log pot, and even then I believe you would have some serious linearity problems. Remember that XBOSOZ has an extreme low outputimpedance at about 10 Ohm singleended, so it need some real low resistive values to be attenuated.
If you want to adjust the gain, don´t use resistors between the sources of the fets, then you will loose some of the X-comunication, that is why it didn´t sound much different when you had the original 680 Ohm gainresistor.
If you uses both singleneded an balanced mode, you could have separate inputresistors for each mode, 10K for singleended and 22K for balanced operation, thus you will obtain the same volumelevel in both modes, something like this is used in the Aleph´s.
If you want to have adjustable gain you should have say 50K lin pots in series with some 15K resistors as replacemens for R122/123.
Don´t do that, change the gain as described.
The methods mentiond will not harm the frequency responce, at least my simulator tells me.
With 680 gainresistor and R122/123 (39K) you shoud have 7db gain, and without R122/123 12db gain. So if you have the same voltage level at the output something is wrong, are you shure you are using inputresistors at 10K?
When it comes to the sound of the XBOSOZ as you both describes it, I really reconize my own experience with my XBOSOZ and XSOZ, the imaging is som much better with the X, all contrasts between soft and crispy is just great.
At the moment I use a singleended CD, but I am about to modyfi it to descrete balanced mode.
🙂
Grataku and Ignatz, just a hint.
Grataku you are driving an AlephX with your XBOSOZ, it´s input impedance is 10K, wich is much higher than nessesary regarding the low outputimpedance of the XBOSOZ. LCAudio suggests a resistor between signal and ground at the recieving end of the interconnects like in the attached picture. I wonder if this goes for balanced operation too. Since I have a XSOZ with an singleended inputimpedance at 120 Ohm I woun´t benefit from such a resistor.
Regards Henrik
Grataku you are driving an AlephX with your XBOSOZ, it´s input impedance is 10K, wich is much higher than nessesary regarding the low outputimpedance of the XBOSOZ. LCAudio suggests a resistor between signal and ground at the recieving end of the interconnects like in the attached picture. I wonder if this goes for balanced operation too. Since I have a XSOZ with an singleended inputimpedance at 120 Ohm I woun´t benefit from such a resistor.
Regards Henrik
Attachments
Grataku and Ignatz, just a hint.
Grataku, I forgot that you have 10K pot at the output, so the trick with the 100 Ohm at the recieving end at the interconnects may give you some unpleasent unlinearity at your volume regulation caused by the increased outputimpedance as you attenuate the output.
Grataku, I forgot that you have 10K pot at the output, so the trick with the 100 Ohm at the recieving end at the interconnects may give you some unpleasent unlinearity at your volume regulation caused by the increased outputimpedance as you attenuate the output.
No input resistors
Henrik/Grataku:
Thanks for your great information. I do not have either 10K on my single ended sources or 22K resistors on my balanced inputs. I will make that change today.
I will also try the 100R resistors to ground at he inputs of my Bryston 3BST. I think in balanced mode its impedance is 10K.
Although I have the Aleph-X boards, I have not yet made that amp.
Currently I am running a 168R gain resistor and 39K feed back resistors.
The sound quality is substantially improved over the standard version with my current set up.
Instruments and voices seem more dense in a good way and richer as a whole. I find it difficult to accurately describe the difference in sound other than to say it is immediately apparent that it is substantially better.
Henrik/Grataku:
Thanks for your great information. I do not have either 10K on my single ended sources or 22K resistors on my balanced inputs. I will make that change today.
I will also try the 100R resistors to ground at he inputs of my Bryston 3BST. I think in balanced mode its impedance is 10K.
Although I have the Aleph-X boards, I have not yet made that amp.
Currently I am running a 168R gain resistor and 39K feed back resistors.
The sound quality is substantially improved over the standard version with my current set up.
Instruments and voices seem more dense in a good way and richer as a whole. I find it difficult to accurately describe the difference in sound other than to say it is immediately apparent that it is substantially better.
- Status
- Not open for further replies.