Prasi Jumas recomendation is try to concentrate sensitive GND points to a small area (see post #1). You got a lot of copper and long lines in signal GND (post #168).
According to this recommendation you can reduce the signal gnd area like this.
Or wait for Juma's opinion
According to this recommendation you can reduce the signal gnd area like this.
Or wait for Juma's opinion
Attachments
Last edited:
Prasi Jumas recomendation is try to concentrate sensitive GND points to a small area (see post #1). You got a lot of copper and long lines in signal GND (post #168).
According to this recommendation you can reduce the signal gnd area like this.
Or wait for Juma's opinion
that wont produce good results, believe me.
what Juma says is completely different, he says to concentrate all gnd points at a small location.
placing a long narrow trace is not recommended for ground.
many people recommend a plane for ground as long as you dont produce loops.
many do a gnd plane on one side of pcb while placing all signal traces on other side.
I have done this technique (like the cubie-2 bjt type) for a head amp thats very popular here and many guys have it tested and works well without any hum/hiss of any kind even on sensitive headphones.
Last edited:
Hi Juma,
Since I'm in the process of gathering the components for Cubie2, I have a question regarding the power supply. I'm considering building a double power supply (each channel will have its own PS), thus I think that the PS capacitors value should be halved. Correct?
Also, what is the wattage rating of R1-R2 (1 Ohm) resistors?
Regards,
Vagelis
Since I'm in the process of gathering the components for Cubie2, I have a question regarding the power supply. I'm considering building a double power supply (each channel will have its own PS), thus I think that the PS capacitors value should be halved. Correct?
Also, what is the wattage rating of R1-R2 (1 Ohm) resistors?
Regards,
Vagelis
Giasou Vageli,
they are both very nice but F5 with 2SK2013/2SJ313 has a certain smooth, velvety note to the sound a bit more pronounced. In the long run, however, I preferred Cubie2. It's all nuances that you become aware of with time.
Cubie2 is simpler from construction point of view (less output devices means less metal work) and it has more gain so the preamp is not needed.
Na'sai kala !
they are both very nice but F5 with 2SK2013/2SJ313 has a certain smooth, velvety note to the sound a bit more pronounced. In the long run, however, I preferred Cubie2. It's all nuances that you become aware of with time.
Cubie2 is simpler from construction point of view (less output devices means less metal work) and it has more gain so the preamp is not needed.
Na'sai kala !
Geia sou juma,
Eyxaristo! Thank you for your immediate response (as always).
I will certainly build Cubie2, although I was tempted to try also the F5 with 2SK2013/2SJ313, since I read too many good things about it. But, since they are close in character, I would opt for the Cubie2, both due to its simplicity and because I want to hear how the lateral MOSFETs sound.
Na'sai kai sy kala!
Best regards,
Vagelis
Eyxaristo! Thank you for your immediate response (as always).
I will certainly build Cubie2, although I was tempted to try also the F5 with 2SK2013/2SJ313, since I read too many good things about it. But, since they are close in character, I would opt for the Cubie2, both due to its simplicity and because I want to hear how the lateral MOSFETs sound.
Na'sai kai sy kala!
Best regards,
Vagelis
Geia sou kai pali juma,
In the Cubie3 thread, member Berny aked:
I do like what you did for the offset regulation with P2, can I do this too? (replace R12 =120ohm in Cubie2 by maybe 100ohm + trimmer)
and your answer was the following:
Sure, try 110R + 20R trimmer.
But, in the original Cubie2 circuit, there is already a trimmer between the sources of the input jfets (namely, P1), which adjusts offset.
Could you plese elaborate on this? Are those alternatives, or the R12 + trimmer can be used supplementary to P1?
Vagelis
In the Cubie3 thread, member Berny aked:
I do like what you did for the offset regulation with P2, can I do this too? (replace R12 =120ohm in Cubie2 by maybe 100ohm + trimmer)
and your answer was the following:
Sure, try 110R + 20R trimmer.
But, in the original Cubie2 circuit, there is already a trimmer between the sources of the input jfets (namely, P1), which adjusts offset.
Could you plese elaborate on this? Are those alternatives, or the R12 + trimmer can be used supplementary to P1?
Vagelis
And, I suppose that, in case of R12=110R + 20R trimmer, P1 should be replaced by 2 10R reistors. Right? But, then, how do I connect the feedback resistors?
Also, is there any difference if the DC offset adjustment circuit is in place of R12 or R11?
Taking the idea from Berny's post in the Cubie3 thread, stating that:
as I do not like so much the trimmer in the feedback loop
I thought of following this approach for applying the DC offset adjustment. If that will cause any trouble, I could just follow the original Cubie2 schematic. If, on the other hand, having the feedback loop separately from the DC offset adjustment trimmer would be better, I would like to follow this path.
Thank you again,
Vagelis
Ok, taking a key words out of your post (thought, could, would) the conclusion is simple - no need to change anything, both approaches are good.
And Berny missed to explain what makes him dislike the trimm-pot near the feedback divider.
Designing by unknowledgable assumptions can be good (although rarely) but I suggest building and testing before publishing such ideas. I may not be right but that's my expereince. So, get to work...
And Berny missed to explain what makes him dislike the trimm-pot near the feedback divider.
Designing by unknowledgable assumptions can be good (although rarely) but I suggest building and testing before publishing such ideas. I may not be right but that's my expereince. So, get to work...
So, get to work...
juma,
Thanks for clarifying all that mess inside my head
Yes, I have already got into work. I have the laterals in hand (from a GB here), I have ordered the transformer, I have bought a chassis from ebay and waiting for delivery and need to make an order list from Mouser. Also, working on the PCB.
So, a great thanks again!
hi,
Am i right to say that to be able to reduce the bias of the mosfet for energy efficiency purpose and have similar performance to cubie 3 and simplicity of cubie2, i can simply borrow the mosfet bias and gain methods from cubie 3?
i want to power efficient bookshelves speakers not too loud...the axiom m3.
i.e.: in the schematic in post 1, remove R25 that set the OLG and put the bjt Vbe multiplier from cubie3 instead of Rbias 330 ohms?
regards,
-pierre
Am i right to say that to be able to reduce the bias of the mosfet for energy efficiency purpose and have similar performance to cubie 3 and simplicity of cubie2, i can simply borrow the mosfet bias and gain methods from cubie 3?
i want to power efficient bookshelves speakers not too loud...the axiom m3.
i.e.: in the schematic in post 1, remove R25 that set the OLG and put the bjt Vbe multiplier from cubie3 instead of Rbias 330 ohms?
regards,
-pierre
Last edited:
So i deduce that Cubie3 keeps it's performance because of the excellent linearity of the mosfet devices, despite lower bias current, and higher volume performance being attributed to more voltage headroom.
i thought it was because of the higher OLG caused by removal of R25...
In that case keeping R25 allows indeed to experiment.
i may also try to slap in the simple folded cascode from lsk pre, but with ccs instead of resistor. With headphones or for very low listenning level with speakers, the added simplicity may yield benefits i think...
i also plan to compare laterals with the venerable irfp240/9240 (or 9140)....i should be a good fight.
In any case, this amp looks nice and i can't wait to receive my lateral to begin building.
i thought it was because of the higher OLG caused by removal of R25...
In that case keeping R25 allows indeed to experiment.
i may also try to slap in the simple folded cascode from lsk pre, but with ccs instead of resistor. With headphones or for very low listenning level with speakers, the added simplicity may yield benefits i think...
i also plan to compare laterals with the venerable irfp240/9240 (or 9140)....i should be a good fight.
In any case, this amp looks nice and i can't wait to receive my lateral to begin building.
Last edited:
Yes, the circuit from BAF 2013 is really special and its worth an effort to turn it into power amp with enough gain and Zin high enough so that preamp is not needed any more.
It has that wonderful gentle seductive touch, live and involving character - it's one of the very rare no-feedback-loop circuits that can successfully reproduce much more than just a girl & guitar type of music.
It definitely has a special charm and it only lacks that last step in definition and clarity with complex orchestral music.
Since it stays linear with higher gain settings I tried to overcome that by using different output stage and feedback variations, using Laterals (k1058/j162) and Sony's VFETs (k60/j18) in common source and common drain configurations (with appropriate biasing and feedback settings) and it's not very hard to find a spot where everything sits perfectly.
I listen to it since last autumn and it still thrills me. I prototype and test different stuff all the time, listen to various equipment owned by buddy audiophiles, but this one stays....
Attachments
Thanks for sharing that last schematic...it looks nice (f5 meets lsk pre)....i may build it sometimes....except maybe for the CCS's that i would replace with j310 or something like that.
Concerning the lack of coherence with many instruments playings:
have you considered a series reg, even if it is not in the simplicity spirit, for the output stage at least?
i find most regs leave their trace on sound but i find that the AMB Sigma22 to be very transparent.
it is in fact not very far from my favorite the shunt salas V1.2 in terms of signature, but with much better dynamics , and less heat of course.
Concerning the lack of coherence with many instruments playings:
have you considered a series reg, even if it is not in the simplicity spirit, for the output stage at least?
i find most regs leave their trace on sound but i find that the AMB Sigma22 to be very transparent.
it is in fact not very far from my favorite the shunt salas V1.2 in terms of signature, but with much better dynamics , and less heat of course.