Feedback loop length
Dear all,
The saga continues... I have corrected all reported errors in the schematic, and renumbered the components to fit Pass' drawings.
I have added three jumpers to cut power to the MC stage and the Inverter stage, if they are not used. I figured that can help clean up the power supply even more. (Should I also cut the leads to the ground?) The schematic should be flawless now. If anyone feels he has time to spare, anyone can contact me and get the drawings.
Second, I have found and corrected a few reported footprint errors, (Q10 - Q13) and I am searching for component layouts to support different types of caps, pots, etc.
Third, since I had to re-design half of the board, I decided to do an entire makeover. Better component placing, improved trace length, etc. Which brings me to my main question:
The MM stage has a feedback cirquit in it. Is it an important thing to make this loop as tight as possible? I mean, every square mm added to the loop, will decrease the speed and stability of the loop? Would it be audible to make it as short as possible?
Bouke
Dear all,
The saga continues... I have corrected all reported errors in the schematic, and renumbered the components to fit Pass' drawings.
I have added three jumpers to cut power to the MC stage and the Inverter stage, if they are not used. I figured that can help clean up the power supply even more. (Should I also cut the leads to the ground?) The schematic should be flawless now. If anyone feels he has time to spare, anyone can contact me and get the drawings.
Second, I have found and corrected a few reported footprint errors, (Q10 - Q13) and I am searching for component layouts to support different types of caps, pots, etc.
Third, since I had to re-design half of the board, I decided to do an entire makeover. Better component placing, improved trace length, etc. Which brings me to my main question:
The MM stage has a feedback cirquit in it. Is it an important thing to make this loop as tight as possible? I mean, every square mm added to the loop, will decrease the speed and stability of the loop? Would it be audible to make it as short as possible?
Bouke
Congratulations, Bouke!
We're all getting exited here, Thumbs up and thank's a lot for keeping the faith...
We're all getting exited here, Thumbs up and thank's a lot for keeping the faith...
Count me in...
Hi Bouke,
When the definitive layout is ready and you are having PCBs produced, I would very much like to join in ordering PCBs...
Keep up the good work!
Regards,
Lucas.
Hi Bouke,
When the definitive layout is ready and you are having PCBs produced, I would very much like to join in ordering PCBs...
Keep up the good work!
Regards,
Lucas.
I would think that in general, shorter is better, for all signal paths; not just for the feedback loop.
However, to answer your question properly, you must consider what compromise you make to shorten the feedback path. What part of the layout would become less ideal? Then the tradeoff can be made.
-- mirlo
Well, I was placing components on a new version which had the MC and the MM input next to eachother, like the original Pass design. But after I had placed the MC stage quite optimal, there was too little space left to have an optimal placment for the MM stage....
So the tradeoff now is: Improved component placing against shorter connection leads for the MM input. I vote for the improved component placing 😉... Argument: inside the case, the signal is better off inside a well shielded cable, than on a partially unshielded copper trace on the board...
Moreover, I realised I was reinventing my own wheel, and I would probably end up with an almost similar layout like the first prototype, so why not just take that design and work on from that... It saves a lot of time 🙂
Bouke
So the tradeoff now is: Improved component placing against shorter connection leads for the MM input. I vote for the improved component placing 😉... Argument: inside the case, the signal is better off inside a well shielded cable, than on a partially unshielded copper trace on the board...
Moreover, I realised I was reinventing my own wheel, and I would probably end up with an almost similar layout like the first prototype, so why not just take that design and work on from that... It saves a lot of time 🙂
Bouke
I would guess that the lower levels present in the MC stage benefit more from careful and short routing than the higher levels in the MM stage, and that the signal levels are a more important consideration than feedback.
And as you suggest, no need to reinvent the wheel.
And as you suggest, no need to reinvent the wheel.
Exactly... So I think I will just work on from the original design, and improve that... I think I have the MC stage pretty optimised by now... The internal leads for the MM stage will be about 10 cm longer than those from the MC input....
Also, I have added the feature of thermally coupling Q1 and Q2. These two provide the bias for Q5, and I think the differential stage will benefit from the least possible difference between those two transistors. Besides, it's fairly easy to implement... What do you think?
Bouke
Also, I have added the feature of thermally coupling Q1 and Q2. These two provide the bias for Q5, and I think the differential stage will benefit from the least possible difference between those two transistors. Besides, it's fairly easy to implement... What do you think?
Bouke
if it will not cost you too much time i would go this way
but it means drifting apart al little from orginal
hmm i vote for termal coupling
/me - awaiting next pcb realese🙂
but it means drifting apart al little from orginal
hmm i vote for termal coupling
/me - awaiting next pcb realese🙂
No it isn't... There is an error on it, and I advise you not to use it...
I will remove it from my website soon, and replace it with an update
Bouke
I will remove it from my website soon, and replace it with an update
Bouke
hello!
did you seen layout from Edward Robinson : http://www.diyaudio.com/forums/showthread.php?s=&threadid=13201&highlight=ono
maybe it is possible too mix up this and yours layouts? if you want this layout in jpg files to see how it looks mailme
did you seen layout from Edward Robinson : http://www.diyaudio.com/forums/showthread.php?s=&threadid=13201&highlight=ono
maybe it is possible too mix up this and yours layouts? if you want this layout in jpg files to see how it looks mailme
XONO /ONO
In the very first mail of this thread, KDT asked what the differences are between the ONO and the XONO.
Here it is
1) Each channel has now its own set of power supply regulators by means of two times 7815 giving +/- 30Volt instead of one single regulator with MosFets for both channels
2) Power Cap's on the +/- 30Volt lines have been increased from 1000uF to 3300uF//150nF
3) C10 has been changed from 10uF into 220uF//47nF.
4) C1 and C41, both 220uF, have been changed to one single 1000 uF.
5) on the unregulated side the four 1000uf Cap's have been replaced by four 10.000uF Cap's
6) C16 330pF, connected to Q14, has been deleted
That's it
In the very first mail of this thread, KDT asked what the differences are between the ONO and the XONO.
Here it is
1) Each channel has now its own set of power supply regulators by means of two times 7815 giving +/- 30Volt instead of one single regulator with MosFets for both channels
2) Power Cap's on the +/- 30Volt lines have been increased from 1000uF to 3300uF//150nF
3) C10 has been changed from 10uF into 220uF//47nF.
4) C1 and C41, both 220uF, have been changed to one single 1000 uF.
5) on the unregulated side the four 1000uf Cap's have been replaced by four 10.000uF Cap's
6) C16 330pF, connected to Q14, has been deleted
That's it
please clarify!
Sorry, what do you mean with 2x7815 regulators?
I cannot understand your description, and probably a simple sketch/ schematic could clear the doubt.
Thanks
Paolo
Sorry, what do you mean with 2x7815 regulators?
I cannot understand your description, and probably a simple sketch/ schematic could clear the doubt.
Thanks
Paolo
Ono/Xono
The 7815 is a fixed 15Volt/1Amp Voltage regulator, sold by many suppliers as LM7815, or uA7815 etc.
By using a voltage divider at the input consisting of two 1Kohm resistors, the 15 Volt output can be altered to 30 Volt, see the technical documentation of the various manufacturers.
The 7815 is a fixed 15Volt/1Amp Voltage regulator, sold by many suppliers as LM7815, or uA7815 etc.
By using a voltage divider at the input consisting of two 1Kohm resistors, the 15 Volt output can be altered to 30 Volt, see the technical documentation of the various manufacturers.
2SK389 and 2SK170 flavours
Hi folks,
I read somewhere that the 2SK389 was available in the "V" flavour quite cheaply. Would that be satisfactory for an Ono/Xono, or should I go for the 2SK389BL?
Intriguing, too, the flavours of 2SK170 on the Xono/Ono. I thought that BL was the preferred choice, and yet I see others begging to swap their BL's for GR's!
I THOUGHT that Nelson has implied he uses BL's, but now I am not sure exactly what for!
What is the truth of it all?
All help gratefully considered.
Regards,
George.
Hi folks,
I read somewhere that the 2SK389 was available in the "V" flavour quite cheaply. Would that be satisfactory for an Ono/Xono, or should I go for the 2SK389BL?
Intriguing, too, the flavours of 2SK170 on the Xono/Ono. I thought that BL was the preferred choice, and yet I see others begging to swap their BL's for GR's!
I THOUGHT that Nelson has implied he uses BL's, but now I am not sure exactly what for!
What is the truth of it all?
All help gratefully considered.
Regards,
George.
Hi,
I think for the ONO you would need 2SK170 GR´s to get the original gain with the used supply voltage.
The GR´s have the lowest Idss so the voltage drop over the gain resistor will be the lowest for a given source resistor.
If you use BL´s the Idss is higher so you need to lower the value of the drain resistor (and have lower gain) or up the source resistors (and have higher noise).
In my ONO I used BL´s and ommited the highest (76dB) gain setting.
Can´t help with the 2SK389 question as I´m far away from home (Sweden) without any schematics
William
I think for the ONO you would need 2SK170 GR´s to get the original gain with the used supply voltage.
The GR´s have the lowest Idss so the voltage drop over the gain resistor will be the lowest for a given source resistor.
If you use BL´s the Idss is higher so you need to lower the value of the drain resistor (and have lower gain) or up the source resistors (and have higher noise).
In my ONO I used BL´s and ommited the highest (76dB) gain setting.
Can´t help with the 2SK389 question as I´m far away from home (Sweden) without any schematics
William
I was wondering if anyone could help me determine the max voltage that could could arise across the capacitors C1, C27, C10, C19, C37 in the Ono? I don't have a SPICE model and am not positive that my calculations (more like guesses) are close at all 
Thank you so much in advance!
Thank you so much in advance!
Hi,
my guesses.
C1: 6.3V
C27: 1.8V
C19: 16V
C10: 25V
C37: 16V
I think I used 25V for all of them (Panasonic FC)
William
my guesses.
C1: 6.3V
C27: 1.8V
C19: 16V
C10: 25V
C37: 16V
I think I used 25V for all of them (Panasonic FC)
William
hi! thanks for the reply! I was pretty close on all besides C1.. the node above R21 should be close to 0v, right and the other leg of C1 is tied to ground? I guess I'm a bit confused.....
So, I guess I shouldn't try and use a 6.3v Blackgate NX here
?
Also, while I have the Ono experts here, how critical is the value of R41 (listed at 4.02k)? I am trying to use PRP resistors throughout and all I have is 3K9, but I could use a Dale if the value is important... same with R67, is 470R ok (listed value is 475R)? It seems that this might be important to get the voltage drop across R53 and R67 to be around 1v? Finally, how inportant is the value on the output capacitor Cu (it's listed as 680pF)... Digikey is out of this value in the capacitors I'm using, and they have 580pF and 820pF.
Thanks so much!!!
So, I guess I shouldn't try and use a 6.3v Blackgate NX here
?Also, while I have the Ono experts here, how critical is the value of R41 (listed at 4.02k)? I am trying to use PRP resistors throughout and all I have is 3K9, but I could use a Dale if the value is important... same with R67, is 470R ok (listed value is 475R)? It seems that this might be important to get the voltage drop across R53 and R67 to be around 1v? Finally, how inportant is the value on the output capacitor Cu (it's listed as 680pF)... Digikey is out of this value in the capacitors I'm using, and they have 580pF and 820pF.
Thanks so much!!!
- Status
- Not open for further replies.