Hi,
re circuit post116.
the tl072 in the DC servo should be lf411 or similar low offset opamp.
C10 is too small. Try 470nF to 1uF.
Why are you using the second tl072 as a buffer?
The base resistance to ground of the two halves of the LTP are unbalanced and using the servo to cancel out the input offset error.
Try reducing R2 to 1k0 and buffering the input pot. Does R1=100k achieve anything?
No gate to source zeners?
How good is that folded cascode (Q3). It's a topology I like the look of and Borbely, and others, are in love with it.
re circuit post116.
the tl072 in the DC servo should be lf411 or similar low offset opamp.
C10 is too small. Try 470nF to 1uF.
Why are you using the second tl072 as a buffer?
The base resistance to ground of the two halves of the LTP are unbalanced and using the servo to cancel out the input offset error.
Try reducing R2 to 1k0 and buffering the input pot. Does R1=100k achieve anything?
No gate to source zeners?
How good is that folded cascode (Q3). It's a topology I like the look of and Borbely, and others, are in love with it.
AndrewT said:
TL072 is fine here. It has 3mV typ and 10mV max offset voltage.
Why are you using the second tl072 as a buffer?[/B][/QUOTE]
C10 is only needed to filter HF components - to make the integrator life a bit easier. It can be omitted altogether. Second opamp inverts the servo signal and allows to connect it to the inverting input of the amp. I don't like a single opamp non-inverting integrator option.
The important advantage of this circuit is that it doesn't need a preamp or a buffer on the input. Unbalance of resistors is not a problem.
No need - the max Vgs in this circuit is limited to 2xVth anyway.
It is not a folded cascode 🙂 . I have to admit that I took this arrangement from a very old hybrid hi-speed opamp circuit by National. It provides for a local feedback on the pnp invertor and works very well.
Cheers
x-pro
Hi!
Was out of town, at Kolkota. Thanks Kanwar for the design.
Now time for me to do some artwork. BTW can I add another pair of MOSFET on PCB.
I have also noticed this trace broadning effect in Tek LCD scopes at work. I am sure my analog one Tek 2465B at home will give a clearer picture.
Regards
Rahul
Was out of town, at Kolkota. Thanks Kanwar for the design.
Now time for me to do some artwork. BTW can I add another pair of MOSFET on PCB.
I have also noticed this trace broadning effect in Tek LCD scopes at work. I am sure my analog one Tek 2465B at home will give a clearer picture.
Regards
Rahul
Workhorse said:
It is I scheme! It is scheme X-PRO.
In the scheme it is written Alex Nikitin
p.s. X-Pro (Àëåêñåé), excuse for PR of your scheme
Slightly confused Kanwar. The schematics show 11 nos small signal devices. PCB has 13 of them????
Secondly the additional 15V supply line what should be the current.
Are there any mods required or the test results are of different version??
Rahul
Secondly the additional 15V supply line what should be the current.
Are there any mods required or the test results are of different version??
Rahul
x-pro said:
Hi Alex,
So you are an Ex-Creek Chief Engineer, Nice to meet you.....Your circuit is not Rail to Rail swinger....In pro application Paralleling is often the best solution to get high power with reliability, which is a prime importance..You design audio gear for Audiophiles but we design it for for professional people because reliability is prime target...goals are different....thats why the approach is also different.....
Rahul said:
Hi Raaahul,
You can Add upto 10Pairs of IRFP360 with Rails at +-120V very well....
Rahul said:
Yes you are right the missing one are actually cascode loaded trannies at the first differential stage....updated the schematic....have a look...thanks for pointing out the mistake....
The 15V supply is a negative elevation which should supply at least 10mA , you could build it with simple Zener as well...I have also done it with a zener ofcourse..and a filter cap....along with...
The test version is same as that of updated schematic... no mods required.....
Regards,
K a n w a r
Attachments
Thanks Kanwar,
At the input I also see a couple of diodes going to the extended negative line in picture , they seem to be missing in ckt do we need to include them too.
Just want to clear up confusion before feeding everything for a PCB.
Regards
Rahul
At the input I also see a couple of diodes going to the extended negative line in picture , they seem to be missing in ckt do we need to include them too.
Just want to clear up confusion before feeding everything for a PCB.
Regards
Rahul
Hello Kanwar,
You've posted a lot and got lots of response . However you never seem to read between the lines in some of the replies to your posts.
"......So you are an Ex-Creek Chief Engineer, Nice to meet you.....Your circuit is not Rail to Rail swinger....In pro application Paralleling is often the best solution to get high power with reliability, which is a prime importance..You design audio gear for Audiophiles but we design it for for professional people because reliability is prime target...goals are different....thats why the approach is also different..... "
Are you really trying to teach him about something he surely knows more about than you ? Would you imagine that he knows nothing about pro audio design targets ?
There is nothing Audiophile about a circuit . Its the application and performance parameters that "make" it " audiophile " or not.
Get real man. Seasoned designers KNOW what audio design is all about and generally it doesn't end with audio circuits alone which is just one application of electronic circuit design!
Cheers.
You've posted a lot and got lots of response . However you never seem to read between the lines in some of the replies to your posts.
"......So you are an Ex-Creek Chief Engineer, Nice to meet you.....Your circuit is not Rail to Rail swinger....In pro application Paralleling is often the best solution to get high power with reliability, which is a prime importance..You design audio gear for Audiophiles but we design it for for professional people because reliability is prime target...goals are different....thats why the approach is also different..... "
Are you really trying to teach him about something he surely knows more about than you ? Would you imagine that he knows nothing about pro audio design targets ?
There is nothing Audiophile about a circuit . Its the application and performance parameters that "make" it " audiophile " or not.
Get real man. Seasoned designers KNOW what audio design is all about and generally it doesn't end with audio circuits alone which is just one application of electronic circuit design!
Cheers.
Hi Ashok,
I think you have misunderstood my saying....quite well...
I have no intention to teach Mr. Alex anything, I have only said what is seen in that context....
He said paralleling ruins the sound quality, yes its true, but I also said that to get more power one has to do it to achieve reliability...and thats only a matter of conception, noway on teachings...man....no teaching to anyone ....
Secondly He said "The power supply usage is fairly good - about the same as in your circuit with one pair of output devices" which is not true because his circuit isnot a rail to rail swinger amp [because I have already tested that topology]but yes mine is rail to rail....and its true...so how you came upto the conclusion that I am teaching him.....
Yeah sure, the application and parameters such as price consideration, quality performance, choice of components, circuit topologies, layout of pcb's etc are all the approach that classifies the circuit category for an audiophile or pro-gear....
Alex's target is audiophile gear, mine is pro-audio gear....hence no contradiction....both types are different from each other,isn't it...
What lets you to think that i am teaching someone who is Seasoned in that field......
K a n w a r
I think you have misunderstood my saying....quite well...
I have no intention to teach Mr. Alex anything, I have only said what is seen in that context....
He said paralleling ruins the sound quality, yes its true, but I also said that to get more power one has to do it to achieve reliability...and thats only a matter of conception, noway on teachings...man....no teaching to anyone ....
Secondly He said "The power supply usage is fairly good - about the same as in your circuit with one pair of output devices" which is not true because his circuit isnot a rail to rail swinger amp [because I have already tested that topology]but yes mine is rail to rail....and its true...so how you came upto the conclusion that I am teaching him.....
Yeah sure, the application and parameters such as price consideration, quality performance, choice of components, circuit topologies, layout of pcb's etc are all the approach that classifies the circuit category for an audiophile or pro-gear....
Alex's target is audiophile gear, mine is pro-audio gear....hence no contradiction....both types are different from each other,isn't it...
What lets you to think that i am teaching someone who is Seasoned in that field......
K a n w a r
Rahul said:
Hi Rahul,
You dont have to include that diodes, they are just sufficed to get proper voltage drop in series with 10K resistor's which i had at that time because I havent 12K resistors at that time which are currently in schematics...its only a substitute thing....just to accomodate the voltage drop nothing else....just go with the schematic....brother..
K a n w a r
Ashok,
get your sohnenbrillen on 😎 , some people just like to be a swinger and (moon)shine!
Cheers Michael
get your sohnenbrillen on 😎 , some people just like to be a swinger and (moon)shine!
Cheers Michael
Workhorse said:
Kanwar,
if you check the actual losses in you output stage you'll see that it is not quite "rail to rail" for a very simple reason of 0.47 Ohm source resistors in your output stage. Even if the output MOSFET is completely open you'll lose a bit over 1/16th of the supply voltage on these resistors into 8 Ohm and over 1/8th into 4 Ohm. In my design the resistors are the same (thought these are not in the local feedback loop) and so the losses are similar (for a single pair of output MOSFETs), especially if the load impedance is 4 Ohm or lower .
Cheers
x-pro
Ultima Thule said:Ashok,
get your sohnenbrillen on 😎 , some people just like to be a swinger and (moon)shine!
Cheers Michael
Some good people Just like you Michael.....🙂
x-pro said:
Hi Alex,
You are right, if the impedance is somewhat lower, then it tends to drop the voltage across the resistors, but again driving low impedance is generally accomplished by paralleld devices which inturns decreases the voltage drop accross source resistors as they are also acting as parallel paths for current flow, thus output voltage again swings near to its rails.....
Cheers,
K a n w a r
I agree that the power supply usage in x-pro's circuit is very poor. For the positive rail, it clips at (Vrail - three_diode_drops - Vgs(Iout)). For the negative rail, things are worse as it clips at Vrail - Iout*0.47 - Vgs(Iout). For a 4 ohm load this may account for up to 9V rail losses in the negative side and 7V losses in the positive side. That's quite unacceptable for a circuit with +-34V rails (plus sagging!).
The main advantage of the Vas-less and phase-splitter-less circuit is that Vgs is no longer included in the clipping characteristic, leaving Rds-on and Rsource as the only sources of voltage drop. That's what makes the difference between a quasi-complimentary source follower and a rail-to-rail amplifier.
Anyway, both circuits suffer from the same pitfall because they rely blindly into driving the output MOSFET gates with either symmetric voltages or symmetric currents. When driving reactive real-loudspeaker loads, this will yield either crossover distortion or cross-conduction depending on the exact output voltage at which zero current crossing happens, because gate thresold voltages and gate input capacitances are strongly dependent on Vds.
The main advantage of the Vas-less and phase-splitter-less circuit is that Vgs is no longer included in the clipping characteristic, leaving Rds-on and Rsource as the only sources of voltage drop. That's what makes the difference between a quasi-complimentary source follower and a rail-to-rail amplifier.
Anyway, both circuits suffer from the same pitfall because they rely blindly into driving the output MOSFET gates with either symmetric voltages or symmetric currents. When driving reactive real-loudspeaker loads, this will yield either crossover distortion or cross-conduction depending on the exact output voltage at which zero current crossing happens, because gate thresold voltages and gate input capacitances are strongly dependent on Vds.
Scheme X-Pro, does not apply on Rail-to-Rail, is resulted by me as an example of GOOD application N-channel.Eva said:
Eva said:
You are not quite correct with your estimate - devices used in my circuit are "logic level" and have very high transconductance - accordingly Vgs for 4A peak (32V into 8 Ohm) is only about 2.1-2.2V, so the voltage drop on positive side into 8 Ohm is about 4.2-4,3V. When the peak current exceeds 8A (i.e. for the load lower than 4 Ohm) the voltage drop on the 0.47 Ohm resistor will prevail over the voltage drop on the current source and Vgs.
On the negative side this circuit will clip with a voltage drop about 3.8V for 4A - quite close to that on the positive side.
Eva said:
It is my opinion that for that very reason (driving reactive loads) the output stage should have a low open-loop impedance. Gate threshold for the devices used in my circuit does not vary significantly from Vds and this circuit is quite capable of working into any reactive load inside the current/voltage/dissipation limits. The open-loop output impedance is about 0.1-0.2 Ohm and the peak current is near 30A into 0.5 Ohm. If you would like to have a look at the capabilities of this approach, go here:
http://stereophile.com/integratedamps/327/index8.html
It is the "bigger" version of this circuit. Note that the distortion on 1 kHz are about 0.01% even into 1 Ohm. The 4330 (the circuit shown is very close to it) measurements by Stereophile are not on their web site but I could dig these out for you if required.
Cheers
x-pro
Hi Eva,
Yes the gate drives receive symmetrical drive currents but, If you look at the driver structure it very much isolates the varying capacitances of gates of mosfets due to VDS from the drive, as it is buffered by emitter follower, thus again minimizes the artifacts arisen due to reactive loading of output stage.....
K a n w a r
Hi!
Thanks Kanwar about clarifications.
I see lot of gas being released but the fact remains that indeed you have come up with an exceptional design / topology that needs to be given a try. I some how dont digest statements like loss of 1/16. Anyway for me it is good enough. Why even the Citation 12 mosfet mod was good enough for me.
But then I am only a hobbyst.
Regards
Rahul
Thanks Kanwar about clarifications.
I see lot of gas being released but the fact remains that indeed you have come up with an exceptional design / topology that needs to be given a try. I some how dont digest statements like loss of 1/16. Anyway for me it is good enough. Why even the Citation 12 mosfet mod was good enough for me.
But then I am only a hobbyst.
Regards
Rahul
- Status
- Not open for further replies.