Thank you gents, feeling much of a noob here but thinking this is a very worthwhile circuit to play with. A Happy New Year all 😉
So it is posibile to make IV device out from buffer crcuits, Almost all of them...
I tried to sim one with LH0033 module. That is JFET-BJT simple buffer unfortunately out of production... 🙁
Also we have LH0022, technically a sort of diamond buffer...
All these simple devices we can reapeat in discrete mode.
One example is from XEN audio with LH0033. There are some very good photos inside the box...
:
http://xen-audio.com/documents/xen-zgf/ZGF Desktop Description V1.pdf
.
BUF04 is good option because of no huge Io and has offset null pins. (OPA634 is OK too with 16mA Io)
LH0063 is very good but this is very high Io buffer on sims sucking 60mA...
Off course that repeating topology only can make it wit less consumption...
.
I used 2N4401/3 transistors models, tried more but with these models performing better. MMBF4401/3 are SMD versions.
Unusual thing is that Iout grom tha DAC goes to Output of the CFA device...
One more thing when simulated AD811 jung circuit as IV, floor is at exact -200db, and no harmonic content...
But I dont know which type of model is used because i cant read it it is locked in the program.
With almost all of these CFA diskrete circuits floor is typicaly at -130db, but with "real" transistors models...
.
cheers
.
I tried to sim one with LH0033 module. That is JFET-BJT simple buffer unfortunately out of production... 🙁
Also we have LH0022, technically a sort of diamond buffer...
All these simple devices we can reapeat in discrete mode.
One example is from XEN audio with LH0033. There are some very good photos inside the box...
:
http://xen-audio.com/documents/xen-zgf/ZGF Desktop Description V1.pdf
.
BUF04 is good option because of no huge Io and has offset null pins. (OPA634 is OK too with 16mA Io)
LH0063 is very good but this is very high Io buffer on sims sucking 60mA...
Off course that repeating topology only can make it wit less consumption...
.
I used 2N4401/3 transistors models, tried more but with these models performing better. MMBF4401/3 are SMD versions.
Unusual thing is that Iout grom tha DAC goes to Output of the CFA device...
One more thing when simulated AD811 jung circuit as IV, floor is at exact -200db, and no harmonic content...
But I dont know which type of model is used because i cant read it it is locked in the program.
With almost all of these CFA diskrete circuits floor is typicaly at -130db, but with "real" transistors models...
.
cheers
.
Hi,
-200 dB... proves the models given for sim are bads !
About your circuit :
Why the input buffer should have so much mA output ? Current conveyor / Current amp , what is its role here?
LH0033, Input : 1K R .... (not gate input ?)
-200 dB... proves the models given for sim are bads !
About your circuit :
Why the input buffer should have so much mA output ? Current conveyor / Current amp , what is its role here?
LH0033, Input : 1K R .... (not gate input ?)
Last edited:
There are no input R on buffers.
That 1.6K is at first RIV then Rinput to the FET very high impedance gate.
And there is no Feedback as in the standard IV topogies, it is CFA concept.
.
This is a pretty much same as AD844 has
CFA input stage - current mirror out Z compensation point - output buffer.
with a minimum elements and JFET input against BJT input.
.
In discrete version You can determine current in LH0033 By choice of JFET and other values. While the topology remains the same...
Some Io values, with all passive values of elements are same as in the data-sheet SCH, with 2SK30 jfet and most of BJTs are cca. , 2.7mA for jfet branch and same for complementary BJT branch. that is as little bit more than 5mA.
(That value is not one from PDF that stated as 20mA....)
So it is possible to multiply JFETs and BJTs to increase current and decrease output resistance more.
With a noise performance slight improved from multiplying JFETs
.
Anyway the topology is very very good....
.
That 1.6K is at first RIV then Rinput to the FET very high impedance gate.
And there is no Feedback as in the standard IV topogies, it is CFA concept.
.
This is a pretty much same as AD844 has
CFA input stage - current mirror out Z compensation point - output buffer.
with a minimum elements and JFET input against BJT input.
.
In discrete version You can determine current in LH0033 By choice of JFET and other values. While the topology remains the same...
Some Io values, with all passive values of elements are same as in the data-sheet SCH, with 2SK30 jfet and most of BJTs are cca. , 2.7mA for jfet branch and same for complementary BJT branch. that is as little bit more than 5mA.
(That value is not one from PDF that stated as 20mA....)
So it is possible to multiply JFETs and BJTs to increase current and decrease output resistance more.
With a noise performance slight improved from multiplying JFETs
.
Anyway the topology is very very good....
.
Last edited:
@Zoran , why are you developing a circuit with obsolete parts ?
LH0033 is available only through ebay. I would only work with parts that are available today if we want to help the DIY community. As it is, it's just one great story on which everything will remain. I don't believe anyone will solder or test it.
Have you tried using small depletion mosfets instead of jfets in CCSs? You will be surprised by the result.
LH0033 is available only through ebay. I would only work with parts that are available today if we want to help the DIY community. As it is, it's just one great story on which everything will remain. I don't believe anyone will solder or test it.
Have you tried using small depletion mosfets instead of jfets in CCSs? You will be surprised by the result.
At least we can discuss the concept/topology as there are plenty of modern op amp buffer with high mA output.
There was here an incredible composite buffer for headphone iirc made with opa828 input, one of themost powerfull op amp buffer (600 mA) and some solid transistor too. Of course not an I/V but maybe adaptable concept ? Alas can't remember the thread tittle.
Is it a sorta of CFA concept, class AB amp input then buffer but with less current noise input as a bonus vs a classic CFA thhat has no jfet input ?
I amnot sure to understand why CFA should be not feedback concept ? There is still feedback with AD811 ? And also inside LH0033 op amps type ?
So you mean no unity gain external feedback loop ?
May you tell me more please about that buffer input topology and why it matters ?
There was here an incredible composite buffer for headphone iirc made with opa828 input, one of themost powerfull op amp buffer (600 mA) and some solid transistor too. Of course not an I/V but maybe adaptable concept ? Alas can't remember the thread tittle.
Is it a sorta of CFA concept, class AB amp input then buffer but with less current noise input as a bonus vs a classic CFA thhat has no jfet input ?
I amnot sure to understand why CFA should be not feedback concept ? There is still feedback with AD811 ? And also inside LH0033 op amps type ?
So you mean no unity gain external feedback loop ?
May you tell me more please about that buffer input topology and why it matters ?
Maestro, count me in for that task !
I need your precise instructions and the starting file!
No?
It was just hinking of some discrete circuits based on earlier topologies?
They are simple, and cna be reproduced with contemporary availabele parts.
They are CFA type and based on that I think that can fit into the topic?
Heare are inside AD844 BJTs circuit emulated (not sude that ia 100% true)
and OP amp AD844 spice model, different type of spice model.
.
There is a huge difference in results. I set the current to discrete AD844 circuit the same as AD844 spice model giving.
.
The results are similar But with large ditace in noise floor. BJT has -130db and OP model has -200 db.
Same job doing as IV converters. But without standard Feed back topology, because open structure of AD844 makin this possible...
.
Complete sch for BJT discrete AD844 is without single C, and only ine R for current adjust,
Model draws about 7.65mA, not sure that in real is the same? Any eay I set the BJT sch to same Io..
.
Im asking myself abut integrity of other mofels like AD811. But as we dont have internal schematis to emulate it is open Q...
.
and OP amp AD844 spice model, different type of spice model.
.
There is a huge difference in results. I set the current to discrete AD844 circuit the same as AD844 spice model giving.
.
The results are similar But with large ditace in noise floor. BJT has -130db and OP model has -200 db.
Same job doing as IV converters. But without standard Feed back topology, because open structure of AD844 makin this possible...
.
Complete sch for BJT discrete AD844 is without single C, and only ine R for current adjust,
Model draws about 7.65mA, not sure that in real is the same? Any eay I set the BJT sch to same Io..
.
Im asking myself abut integrity of other mofels like AD811. But as we dont have internal schematis to emulate it is open Q...
.
Maybe op660 shematic ?
I stillle don't get the interrest of high impedance output for I/V task in an AC impulsionel environment ? I understand it gives a low input Z though.
Is it to have a stable current bandwidth ?
I stillle don't get the interrest of high impedance output for I/V task in an AC impulsionel environment ? I understand it gives a low input Z though.
Is it to have a stable current bandwidth ?
I was thinking of LH0033 or similar parts .
As for the discrete IV stage, I have nothing against it, only that it is a long way from paper to a finished and most importantly good project. I have been working on an amplifier with laterals in LTspice for a long time and I have no time to solder and at least test it, while in the meantime I have made at least three other amplifiers that I did not have to develop.
If you want to help with the discrete IV, let's try to do it with as few parts as possible, with Rohm or Diode Inc. transistors and depletion mosfets instead of jfets for CCS, in that case I could accompany you with a soldering iron and testing.
few parts as possible ? Why ?
Good project ? Punchline ? What does that means ?
No sense, do you know more than proper engineers at Denon Marantz and so on ?
Does an opa amp has 10 parts inside (only) ? Will we match the Mosfet and Jfet when discrete to acheive good result (I mean in real life and not LT Spice) ? Is it better to do it with less matching, less parts, poor layout, higher noise ?... and most of all shematics that does not work with non ideal pair matched parts ?)
It is really non sense from lthe beginning.... from the high noise AD811 for the I/V task till prophets language non sense, imho !
You just adapted a design to match it to your DAC... and says it is good because it match it in my (so-so) layout, own hifi !
But continue the work, you may find something new. Efforts have sense.... I am not crotisizing this, but just the "sheeps, follow the prophet, verity has to be protected, blah"
Again, it is to balance the anouncments,, all the respect is due to the work.
Good project ? Punchline ? What does that means ?
No sense, do you know more than proper engineers at Denon Marantz and so on ?
Does an opa amp has 10 parts inside (only) ? Will we match the Mosfet and Jfet when discrete to acheive good result (I mean in real life and not LT Spice) ? Is it better to do it with less matching, less parts, poor layout, higher noise ?... and most of all shematics that does not work with non ideal pair matched parts ?)
It is really non sense from lthe beginning.... from the high noise AD811 for the I/V task till prophets language non sense, imho !
You just adapted a design to match it to your DAC... and says it is good because it match it in my (so-so) layout, own hifi !
But continue the work, you may find something new. Efforts have sense.... I am not crotisizing this, but just the "sheeps, follow the prophet, verity has to be protected, blah"
Again, it is to balance the anouncments,, all the respect is due to the work.
Last edited:
It was just thinking on other members questions...
The topologies of that parts are so good even with todays available parts.
CFA type as IC discussed in topic.
And they are so simple. Pretty mush easy to implement.
As LH0033 topology. Some members already did it for other purposes.
If we could discuss the poors and cons for a better understanding, it will be usefull. Typologys are not discussed deeply imho... Or pzrhaps it is clear for you enough...
But if it is only to benchmark by listening , à finished circuity... one can drop here then I ask indeed too much about going further about why.☹
OK guys continue the good work....
But if it is only to benchmark by listening , à finished circuity... one can drop here then I ask indeed too much about going further about why.☹
OK guys continue the good work....
Last edited:
Last edited:
hello all, has there been any development regarding a new pcb for this nice circuit?
I reckon it would be nice to have a small board per channel for the circuit itself, and maybe anyone could add their external psu? seeking positive updates 🙂 thanks!
I reckon it would be nice to have a small board per channel for the circuit itself, and maybe anyone could add their external psu? seeking positive updates 🙂 thanks!