Because of some project at work unfortunately I am fully booked these weeks. I haven't fully done yet my CFP homework and so far I do not know what questions are really important.
So far, still very happy with your schematic using 317 (1BJT equivalent).
However, concerning the CFP,option coming from the 2BJT world, can you recommend a schematic to beat the THD+N of 317 by 20dB ?
What level is the optimization work involved for stability vs performances? Advanced/experts, or accessible also to guys like me?
We are talking line level, hence keeping the margin to +12dBu is appreciated.
I am considering this CFP option because the 317 might have a signature: as a personal remark I think that with some 15ips hot tapes it does loose a bit of transparency (compared to Fostex R8's line amp, I am sure sure if I understand it correctly but it seems to be the Pin9 of IC2.U2 CX20187).
Explanation could be also this: I have recently upgraded my A77 to 15ips, but still using the previous NAB EQ for 7,5ips (no time and no Ref tape for 15ips). Therefore the program is a bit modified, but despite Tone control adjustments it might also be the reason for the allegedly losing a bit of transparency.
I know, A77 EQ on 15ips is a must for comparing apples with apples. Until this happens, I am trying to prepare a CFP option just in case.
Hi thank you very much again
I will do it for sure because for instance i like also Toshiba bjts a lot Mostly because i have seen them used in very expensive units
I guess that they tend to use the best when they want to make the best quality products. And they are not that expensive as well. Great parts.
If i understand well there is no way to import these models in the LTSpice library ?
Nevertheless to be able to pick them up directly from the library would be very handy
I missed this extremely elegant circuit
Thanks so much for suggesting this. I only have two questions
1) will any LM317 provide good results? (the component used here is quite rare and difficult to get)
2) is the dissipated power given by (Vin -Vou) multiplied by the current Iout? up to what power will there be a need for a heat sink?
Last edited:
With 2Bjt's, the CFP is really the best possible. You can tweak the resistors values, for example to increase the bias current, to make it much larger than the peak current in the load, but this won't fundamentally change the character of the circuit.
The CFP will have a lower noise than the 317, but the THD of the 317 will be lower.....
Hi ! sorry again but i cannot find this file on the pc
I have searched and found nothing This is the biggest issue now
I understand that once the file is found i can just copy and paste the model and save the updated file
But again the file seems missing
Maybe it is hidden ? i am getting mad
The previous version was so much more straightforward I have seen tutorials and the library showed up very easily
They have complicated the sw imho
To add a new model to the library should be the easiest thing to do I guess
Last edited:
Hi thank you so much again The folders were hidden and i could not see them Now i see all the folders/files
I have then added some bjt models taken from the Cordell's file but the program cannot let me to save the mods I mean i edit the standard.bjt file > i save the mods > when i open it again the mods do not show up And when during the sim i pick the bjt the added models of course do not show up in the list
Maybe there is some kind of protection of the original file I will try again soon
The folders were hidden and i could not see them Now i see all the folders/files I have then added some bjt models taken from the Cordell's file but the program cannot let me to save the mods I mean i edit the standard.bjt file > i save the mods > when i open it again the mods do not show up And when during the sim i pick the bjt the added models of course do not show up in the list
Maybe there is some kind of protection of the original file I will try again soon
Hm. This gives me second thoughts. Thanks for aforementioning it.
I'd rather try my luck with fresh new design from scratch using opamps, if improvements are needed. So far not.
Don't know if ANY will do, but my own experience is generally Yes.
Still, better check with LV or Marcel, much more prominent and super-heavy class experts for LM317 among other important related topics too.
There's an easy way to add models without editing the stock file. Copy your file (e.g. Cordell-models.txt) to folder ...\lib\sub, then add to your simulation the directive .lib Cordell-models.txt, that will look for the file in that folder and add all models it contains to your simulation.Hi thank you so much again
I have then added some bjt models taken from the Cordell's file but the program cannot let me to save the mods I mean i edit the standard.bjt file > i save the mods > when i open it again the mods do not show up And when during the sim i pick the bjt the added models of course do not show up in the list
Maybe there is some kind of protection of the original file I will try again soon
Cheers,
Cabirio
Hi thank you very much again I am amazed by how ignorant i am I am sincere
I will try it for sure I have to learn first about LM317 power dissipation to select a proper heatsink
As you said increasing (Vin - Vout) and Iout can improve the performance in terms of THD
Thanks a lot Could it be the case to look for this one in the picture ?
https://m.media-amazon.com/images/I/61KBz0h2hFL._SL1000_.jpg
I actually did such a test, and result was to select a specific load (current) for 'best results' in HF range.
What means 'best results' in this context: noise floor spectrum stable - no jumps/bursts at some particular freqs and harmonics.
For me the sweet spot was starting up of 512Ohm loading, i.e. ca 20mA or less. I have tested down to 220Ohm and up to 680Ohm.
All my attempts to use higher current were dropped because of this. I recommend this additional test.
Funny, loaded with a ferrite coil on Pin2: the bad jumpy behavior happened also with large load resistors, all of them. It is like 371 got upset by reflections. There was a setting with higher current + coil loading which was silent (nothing jumped), but I forgot the numbers (as I did not want to risk using such implementation). Maybe LV and Marcel will explain much more.
I may put here some spectra and recall more details, when I will find my USB thumbdrive.
My Mil parts did not show any different behavior than the plastic ones from two manufacturers, but my parts inclusive plastics are all relatively old and original. It is good to hear inputs that new stock is tested equally good.
Depending on your application, next to noise is the PSRR issue to be optimized. Earlier this thread there were good and clear practical advices shown, again thanks to both.
Thank you very much again I confirm the sweet spot you recommend with 18VDC supply
with 1V input the 2nd order harmonic is at -120dB More than enough
Considering how cheap are these chips is a mystery that they are not more popular as buffers
Personally the best performance in terms of THD i have obtained has come from the circuit attached
It is a CFP taken from the buffer stage of a vintage Kenwood preamp
I have not been able to find a low THD CFP with some gain ... the circuits i have found are all noisier than when used as unity gain buffer
with 1V input the 2nd order harmonic is at -120dB More than enough
Considering how cheap are these chips is a mystery that they are not more popular as buffers
Personally the best performance in terms of THD i have obtained has come from the circuit attached
It is a CFP taken from the buffer stage of a vintage Kenwood preamp
I have not been able to find a low THD CFP with some gain ... the circuits i have found are all noisier than when used as unity gain buffer
Attachments
Last edited: