BrianDonegan said:
I had discussed this with Mauro a while back, and it is fraught with intricacies. I do have a layout for the 4780 in parallel already (with offboard power). I am not sure how practical it it is, so I abandoned it. Plus you can already easily drive 4ohm speakers with the LM3886.
Resistor value question
I do not like making changes to a circuit optimized as designed. Due to either existing conditions or available parts I am "dancing" on this edge.
My preamp is transformer coupled on both the input and output. The output iron likes low impedances. To help with this I changed the two input impedance setting resistors to 45k. The highest value ever used before was 25K.
Also in my search for 0.1% matched resistors, I stumbled on some 0.01% Vishay S102 resistors in a local surplus store. The only issue is the closest values are 22.5K ohm and 50K ohm.
To get the input filter frequency to the same as designed I have installed a 2.2 ufd input coupling cap. I am hoping that the 22.5K and the 50K resistors will work fine in circuit.
I guess it will take Mauro to answer this question, but will these changes in resistor value effect the operation of the amplifier? I can always change the input resistors, and hope there are no resonances in the preamp output transformer. The 22K and 47k resistors are a little tougher. I now have some 1/4 watt 1% Dales in the correct values. Just need to match them up.
M equipment is limited to several VOM's and a lcr meter. I cannot actually measure the response of the circuit.
George
I do not like making changes to a circuit optimized as designed. Due to either existing conditions or available parts I am "dancing" on this edge.
My preamp is transformer coupled on both the input and output. The output iron likes low impedances. To help with this I changed the two input impedance setting resistors to 45k. The highest value ever used before was 25K.
Also in my search for 0.1% matched resistors, I stumbled on some 0.01% Vishay S102 resistors in a local surplus store. The only issue is the closest values are 22.5K ohm and 50K ohm.
To get the input filter frequency to the same as designed I have installed a 2.2 ufd input coupling cap. I am hoping that the 22.5K and the 50K resistors will work fine in circuit.
I guess it will take Mauro to answer this question, but will these changes in resistor value effect the operation of the amplifier? I can always change the input resistors, and hope there are no resonances in the preamp output transformer. The 22K and 47k resistors are a little tougher. I now have some 1/4 watt 1% Dales in the correct values. Just need to match them up.
M equipment is limited to several VOM's and a lcr meter. I cannot actually measure the response of the circuit.
George
Hi all.
Any precise statements:
Panelhead, you suggestion to use the res. of the value that serve you directly or on the pre output or on the input connector of the amp, however before the 1uF cap. Everything that is joined in AC doesn't create problems of polarization of LM318. choose a value not unmindful of in conditioning "dynamicses" (AC) will be in parallel to 100K.
LM4780 and 4 ohms:
Russ me has already submitted this problem, and I have answered that serve ( in practice ) it re-planning the circuit. to depart that the parallel of a circuit as the " power current pump" is not has to do as the elementary circuits used in the GC, but 2 complete bridge is had to use in parallel, the principal problem is that be necessary re- calculate the Zout of each bridge to have it was characteristic of MY_REF. In practice, 2 bridge are had to use with a res from 1ohm ( a place of 0.47 ) and is you lose the advantages. In any case, even other techniques ( as the parallel of 2 channels ) go study to found, is not to lose the "audio" quality, is to guarantee the global stability ( not himself has to forget that there are 180 db of gain to check ). The planning and the supervision of this circuit is accessible only to members with a strong technical back ground and with a complete laboratory. If reads this Thread understands that a lot of "experts" developers of audio circuits has the doubts that the circuit is able work well, and this means that the scheme is much gives a complex and non elementary.
Bypass: the EL cap+ low cap work " decently ". If will do you feel better use them you.
MY_REF: I have planed this circuit to listen music in accordance with my parameters. My_REF currently "sound better" ( at home mine ) of:
Borbely 60W mosfet , JLH base version, D.Self " load invariant ", Musical Fidelity A370( modified clone), GC (all the variations ), about + tens of solid state circuits experimented in the years. Is not a toy but my "little gems", elaborated in all the possible variations ( even all the techniques of PS with Reg...).
The official version is the " Rev A", What has the PCB with my layout and it corresponds to the Brian BOM Kit and the PCBs that Russ has distributed. After various test even I, as Russ has chosen this version for my system.
All the other variations are working and it is been able experiment in the case himself you want find an other "sonic equilibrium".
If they build the REV3 suggestion to reducing R41-R42 to 47 ohms or to replace it with a jumper.
I recommend to Russ to foresee on his "mono-block" even the caps among pin 2 and 6 of LM318, to make compatible its "REV3" with the "RevA" because probably will be preferred by many...
The "revisions" are in practice of the variations to the circuit make a will and test.
My interventions will be reduce on account of the work, I hope to read it lends on this THread Your impressions more important: that of the "quality of listening"...
Ciao
Mauro
Any precise statements:
Panelhead, you suggestion to use the res. of the value that serve you directly or on the pre output or on the input connector of the amp, however before the 1uF cap. Everything that is joined in AC doesn't create problems of polarization of LM318. choose a value not unmindful of in conditioning "dynamicses" (AC) will be in parallel to 100K.
LM4780 and 4 ohms:
Russ me has already submitted this problem, and I have answered that serve ( in practice ) it re-planning the circuit. to depart that the parallel of a circuit as the " power current pump" is not has to do as the elementary circuits used in the GC, but 2 complete bridge is had to use in parallel, the principal problem is that be necessary re- calculate the Zout of each bridge to have it was characteristic of MY_REF. In practice, 2 bridge are had to use with a res from 1ohm ( a place of 0.47 ) and is you lose the advantages. In any case, even other techniques ( as the parallel of 2 channels ) go study to found, is not to lose the "audio" quality, is to guarantee the global stability ( not himself has to forget that there are 180 db of gain to check ). The planning and the supervision of this circuit is accessible only to members with a strong technical back ground and with a complete laboratory. If reads this Thread understands that a lot of "experts" developers of audio circuits has the doubts that the circuit is able work well, and this means that the scheme is much gives a complex and non elementary.
Bypass: the EL cap+ low cap work " decently ". If will do you feel better use them you.
MY_REF: I have planed this circuit to listen music in accordance with my parameters. My_REF currently "sound better" ( at home mine ) of:
Borbely 60W mosfet , JLH base version, D.Self " load invariant ", Musical Fidelity A370( modified clone), GC (all the variations ), about + tens of solid state circuits experimented in the years. Is not a toy but my "little gems", elaborated in all the possible variations ( even all the techniques of PS with Reg...).
The official version is the " Rev A", What has the PCB with my layout and it corresponds to the Brian BOM Kit and the PCBs that Russ has distributed. After various test even I, as Russ has chosen this version for my system.
All the other variations are working and it is been able experiment in the case himself you want find an other "sonic equilibrium".
If they build the REV3 suggestion to reducing R41-R42 to 47 ohms or to replace it with a jumper.
I recommend to Russ to foresee on his "mono-block" even the caps among pin 2 and 6 of LM318, to make compatible its "REV3" with the "RevA" because probably will be preferred by many...
The "revisions" are in practice of the variations to the circuit make a will and test.
My interventions will be reduce on account of the work, I hope to read it lends on this THread Your impressions more important: that of the "quality of listening"...
Ciao
Mauro
Mauro,
Thanks you for the input. I do agree that the REV A seems to have a great balance, and it is my personal favorite.
I will do some more work on my PCB layout for the monoblocs to be sure that all three revisions can be built from the same PCB.
So, do I understand you correctly that you think the bypass caps I added are ok? The nice thing is they can easily be omitted on one chooses.
In other respects is the PCB passable? :
Cheers!
Russ
Thanks you for the input. I do agree that the REV A seems to have a great balance, and it is my personal favorite.
I will do some more work on my PCB layout for the monoblocs to be sure that all three revisions can be built from the same PCB.
So, do I understand you correctly that you think the bypass caps I added are ok? The nice thing is they can easily be omitted on one chooses.
In other respects is the PCB passable?
:Cheers!
Russ
Some tweaking
I have been playing with my REV A amp.
I moved things around as Mauro recomended.
I also changed out the resistor for the +/-12V to the LM3128 to 3W IRC thick film types, which seem to work better(get less hot, and now my trafo is drawing fewer milliamps at idle). But it is not a cheap part.
Cheers!
I have been playing with my REV A amp.
I moved things around as Mauro recomended.
I also changed out the resistor for the +/-12V to the LM3128 to 3W IRC thick film types, which seem to work better(get less hot, and now my trafo is drawing fewer milliamps at idle). But it is not a cheap part.
Cheers!
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Member
Joined 2003
The best description I have of the sound of this amp compared to my best GCs (this is not a GC) is that it sounds very much "class A". Its sound compares well to (or better than) a Krell or an Aleph, though some will think me heretical for saying so, but it is true. Mauro engineered the circuit for this very purpose, to use the LM3886 as the vehical to class "A" sound.
Cheers!
Russ
Cheers!
Russ