Here is the basic structure of the input gain cell. It looks very simple on the surface but works better then expected. The way i have drawn it, it is made for battery supply.
There can be many improvements being made, better PSU rejection by CCSs, lower noise etc. but try this before you draw conclusions.
I want to use regulators: I have Salas BiB shunt regs or will be better other kind of regs?
When you use a PSU it may be better to supply the input with CCSs.
I hope i find the time to draw up something for you.
You can use any PSU as long as ripple rejection is good enough so you do not get any hum.
The problem with the Salas is that there are so many versions and some of them can oscillate. That is not a design flaw but because the output impedance up to very high frequencies is very low based on the shunt mechanism. So when your version is stable with my MC stage, fine.
I hope i find the time to draw up something for you.
You can use any PSU as long as ripple rejection is good enough so you do not get any hum.
The problem with the Salas is that there are so many versions and some of them can oscillate. That is not a design flaw but because the output impedance up to very high frequencies is very low based on the shunt mechanism. So when your version is stable with my MC stage, fine.
Curve tracer got today here in WI, hoever I wasn't at home so they will re-deliver tomorrow!!
Question: on my current schematic I haven't included cascode for input differential.
Usually on JFETs I would consider that mandatory..but for BJT I didn't feel the need of it...till a friend of mine brought this up to my attention.
What do you guys think?
Is there any real sonic benefit by cascoding BJTs?
Usually on JFETs I would consider that mandatory..but for BJT I didn't feel the need of it...till a friend of mine brought this up to my attention.
What do you guys think?
Is there any real sonic benefit by cascoding BJTs?
Why do you consider mandatory to cascode jfets ? Miller ?Question: on my current schematic I haven't included cascode for input differential.
Usually on JFETs I would consider that mandatory..but for BJT I didn't feel the need of it...till a friend of mine brought this up to my attention.
What do you guys think?
Is there any real sonic benefit by cascoding BJTs?
Why do you consider mandatory to cascode jfets ? Miller ?
yes!
Here is the basic structure of the input gain cell. It looks very simple on the surface but works better then expected. The way i have drawn it, it is made for battery supply.
There can be many improvements being made, better PSU rejection by CCSs, lower noise etc. but try this before you draw conclusions.
Transistors received today, thank you so much Joachim.
I don't have caps 2200uF on hand I guess I can use two 100uF in parallel?
Best, Felipe
Attachments
I have another question regarding the mirrors.
For the mirror attached here for example, beside Q33 that dissipates a significant amount of power all the other BJTs have very little power.
So my thought was the following:
Would I create any problem on the mirror if I would use dual small power BJTs for Q35/Q36 and use a TO220 device for Q32/33 (with heatsink)/34?
I was thinking that as long as the pair in the top are symmetric and the pair at the bottom with its helper are the same matched device it should be ok, but maybe I am missing something.
The reason for doing that is to save space on layout and also thinking that having a dual BJT on the tracker above would help keeping the current mirror on track better than having two single devices thus improving a bit output DC offset/
Any thoughts on this?
For the mirror attached here for example, beside Q33 that dissipates a significant amount of power all the other BJTs have very little power.
So my thought was the following:
Would I create any problem on the mirror if I would use dual small power BJTs for Q35/Q36 and use a TO220 device for Q32/33 (with heatsink)/34?
I was thinking that as long as the pair in the top are symmetric and the pair at the bottom with its helper are the same matched device it should be ok, but maybe I am missing something.
The reason for doing that is to save space on layout and also thinking that having a dual BJT on the tracker above would help keeping the current mirror on track better than having two single devices thus improving a bit output DC offset/
Any thoughts on this?
Attachments
Transistors received today, thank you so much Joachim.
I don't have caps 2200uF on hand I guess I can use two 100uF in parallel?
Best, Felipe
My mistake not two 100uF I want to use two 1000uF in parallel.
Stefano, if you do not want to parallel this is a way to solve the thermal issue.
I think as long as the chip inside the package is the same and PNP and NPN match in the quasi saturation region that should work and even improve the thermal offset when the heat of each transistor is more similar in the group.
If your chosen transistors work well or better then the BC327/337 used in the Paradise i can not tell. That can only be found out on the curve tracer and i do not have this transistors at hand. You got a curve tracer and could make some measurements.
I have shown on the MPP thread how i do the measurements and what i think is important for good mirrors.
I think as long as the chip inside the package is the same and PNP and NPN match in the quasi saturation region that should work and even improve the thermal offset when the heat of each transistor is more similar in the group.
If your chosen transistors work well or better then the BC327/337 used in the Paradise i can not tell. That can only be found out on the curve tracer and i do not have this transistors at hand. You got a curve tracer and could make some measurements.
I have shown on the MPP thread how i do the measurements and what i think is important for good mirrors.
Joachim,
what if mirror is cascoded using dual small devices for the mirrors so that power dissipation on the upper devices is kept pretty low and let the cascoding device handle the power dissipation?
In this circumstance does the helper have to be the same exact BJT? Because with the helper the number is odd of course.
Me and a friend of mine were looking into the DMMT3904 and its complementary for tiny dual.
The only problem there is that if output current is set to be 30mA this device is not so linear at that current level.
What are your thoughts? Do you have a better dual in mind we can use?
Layout is complete with single devices but I decided to improve it further if there is the possibility and using little duals to achieve better tracking for output DC.
what if mirror is cascoded using dual small devices for the mirrors so that power dissipation on the upper devices is kept pretty low and let the cascoding device handle the power dissipation?
In this circumstance does the helper have to be the same exact BJT? Because with the helper the number is odd of course.
Me and a friend of mine were looking into the DMMT3904 and its complementary for tiny dual.
The only problem there is that if output current is set to be 30mA this device is not so linear at that current level.
What are your thoughts? Do you have a better dual in mind we can use?
Layout is complete with single devices but I decided to improve it further if there is the possibility and using little duals to achieve better tracking for output DC.
Attachments
Some say : do not cascode the cascode.
I do not now. The Paradise mirror works because the output impedance is not TOO high.
Raising the output impedance to an insane high level can compound the thermal drift so you end up with a DC offset that floats around between up and down.
In another mirror design i had the same problem with thermal overload and i simply paralleled two matched BJts. Works just fine.
I do not now. The Paradise mirror works because the output impedance is not TOO high.
Raising the output impedance to an insane high level can compound the thermal drift so you end up with a DC offset that floats around between up and down.
In another mirror design i had the same problem with thermal overload and i simply paralleled two matched BJts. Works just fine.
Stefano,
I have been watching this thread for awhile now and was wondering if you think you are close to a final design or will this be an ongoing upgrade project where you are just changing to learn what you can do? I would love to build a nice phono preamp for my turntable which has been sitting for far to long but am not sure how long to watch the changes? I guess I could go back and read the Paradise design thread of Joachim and I am sure I would be happy with either one of them. I know you said in the beginning you would not be making any boards available but perhaps when you get to the end of the road here someone else could do that for those of us who are not capable of doing that kind of board layout? It looks like you have come a long way and at times it seems you have been very happy with the results. I am not sure if you are in Italy or the USA by your listing here but I do think it has been an interesting thread.
Steven
I have been watching this thread for awhile now and was wondering if you think you are close to a final design or will this be an ongoing upgrade project where you are just changing to learn what you can do? I would love to build a nice phono preamp for my turntable which has been sitting for far to long but am not sure how long to watch the changes? I guess I could go back and read the Paradise design thread of Joachim and I am sure I would be happy with either one of them. I know you said in the beginning you would not be making any boards available but perhaps when you get to the end of the road here someone else could do that for those of us who are not capable of doing that kind of board layout? It looks like you have come a long way and at times it seems you have been very happy with the results. I am not sure if you are in Italy or the USA by your listing here but I do think it has been an interesting thread.
Steven
yes, i would keep the mirror as is and parallel similar BJTs.
i now heard the Pardise with a tube PSU and that did sound awesome so i know that the mirror has a very good potential.
why change a winning team ?
i now heard the Pardise with a tube PSU and that did sound awesome so i know that the mirror has a very good potential.
why change a winning team ?
yes, i would keep the mirror as is and parallel similar BJTs.
i now heard the Pardise with a tube PSU and that did sound awesome so i know that the mirror has a very good potential.
why change a winning team ?
yes I agree with you.
It just happened that once I tripled the mirror it seemd that offset worsed a bit but granted I hadn't provided enough cooling for parts while in this next prototype everything will be arranged properly.
What do you think about input cascode?
Steven,
Build Salas Folded simplistic to compare to either Paradise or Stefanoo's version. Two different approaches. Both excellent in their own respects. Simplistic is quick and easy build.
Build Salas Folded simplistic to compare to either Paradise or Stefanoo's version. Two different approaches. Both excellent in their own respects. Simplistic is quick and easy build.
Stefano,
I have been watching this thread for awhile now and was wondering if you think you are close to a final design or will this be an ongoing upgrade project where you are just changing to learn what you can do? I would love to build a nice phono preamp for my turntable which has been sitting for far to long but am not sure how long to watch the changes? I guess I could go back and read the Paradise design thread of Joachim and I am sure I would be happy with either one of them. I know you said in the beginning you would not be making any boards available but perhaps when you get to the end of the road here someone else could do that for those of us who are not capable of doing that kind of board layout? It looks like you have come a long way and at times it seems you have been very happy with the results. I am not sure if you are in Italy or the USA by your listing here but I do think it has been an interesting thread.
Steven
Steve,
thank you very much for the nice post.
I appreciated you have enjoyed this thread.
I am at the final spint.
In other words I am to the point where I am really pleased with the design and how it sounds.
PCBs? Something I will need to consider later on. It is nice if there is some interest however this design is a bit complex to implement as it is fully balance.
One more detial...this design ONLY works on fully balance system and cannot be ran SE or you will loose 6dB gain.
However for your enjoyment and also other's...here is the current placement for parts.
...this is a truly masterpiece... 😀
😎
Attachments
Buzzfored,
thanks for the direction there, I will take a look at the other thread.
Stefano,
I forgot about this being a balance circuit design and so it needs to be tied to a balance circuit following after it. Reading many threads it is always a question whether to go with a balanced circuit design but it sure seems to be appropriate to solve other issues that SE circuits have to deal with in other ways. You and Joachim sure seem to have put a lot of effort into this design and it is always nice when you finally have a finished product. Thank you for posting this thread, it has been a nice way to learn things even those that are over my head.
Steven
thanks for the direction there, I will take a look at the other thread.
Stefano,
I forgot about this being a balance circuit design and so it needs to be tied to a balance circuit following after it. Reading many threads it is always a question whether to go with a balanced circuit design but it sure seems to be appropriate to solve other issues that SE circuits have to deal with in other ways. You and Joachim sure seem to have put a lot of effort into this design and it is always nice when you finally have a finished product. Thank you for posting this thread, it has been a nice way to learn things even those that are over my head.
Steven
Thanks you are really kind!!
Balance brought significa sonic improvement over SE at least in this instance.
Simplicistic seems to be a nice simple project to build..easy to implement and also nice sounding.
Masterpiece is more of a "ultimate" balance design and I would be sad to see it running on nothing other than top quality audio gear surrounding it.
Like buzzford suggested maybe starting with simplicistic then move to paradise and then to masterpiece is a nice path.
Thanks again for following the thread.
If you have liked it please give it a rate!
😎
Balance brought significa sonic improvement over SE at least in this instance.
Simplicistic seems to be a nice simple project to build..easy to implement and also nice sounding.
Masterpiece is more of a "ultimate" balance design and I would be sad to see it running on nothing other than top quality audio gear surrounding it.
Like buzzford suggested maybe starting with simplicistic then move to paradise and then to masterpiece is a nice path.
Thanks again for following the thread.
If you have liked it please give it a rate!
😎
- Status
- Not open for further replies.
- Home
- Source & Line
- Analogue Source
- Masterpiece