My first posting on a forum will be a question:
I built your preamplifier with a PCI made by myself. The result of THD and noise were excellent (like published values), and sounds very good, but for a reason I still do not know, this pre saturates the output voltage by approximately + VCC for no apparent reason (at random times), but I can not detect RF oscillations or performance problems in the circuit (when running)...
I used different transistors in the circuit. Someone commented to me these transistors could be suffering from high shot noise or flicker, or is it not the case?
I appreciate if anyone has any suggestion or suspicion.
It could be a flaky electrolytic that intermittently leaks through. Hard to find this kind of issues. Is it only one channel or both? If both, probably in the power supply.
Jan
Thank you for answer and help!
Indeed, intermittent defects are things that remind me of the time when I worked with electronics repair: I always wanted never get products with intermittent faults ...
The capacitors I used here are all taken from other projects, so there is a great chance to be (solder and desolder and solder...).
I had checked the power supply and it does not change at all during the event (in fact, this power supply is very good, with good low impedance and low noise). This happens in a channel but during an audio test I suspected that the defect appeared on another channel, but may have just been an swapped channel in the RCA connections during the rapid test.
Another likely defect could be the current source (it it stop providing the output current, logically the voltage output will rise dramatically), but a quick test showed that was still properly polarized (I used a two transistors ring bipolar instead of DN2540).
Some resistors are used but in the previous location they were OK (and them survive better than capacitors to this abuse), yet not rule out the possibility of being one of them.
Probably the fastest would change the capacitors/some res, but as I always like to discover the real cause, so be it ...
Out of curiosity, I used a LM725 opamp in the power supply (with typical over-complicated freq compensation of LM725), and the output reg transistor is a bipolar NPN (BUY49)(with an 6V2 zener in series with base, and an 100R from base to emitter), and proved to be much better in performance compared to "traditional" supplies with LM317 etc but I don't now compared with the original.
Indeed, intermittent defects are things that remind me of the time when I worked with electronics repair: I always wanted never get products with intermittent faults ...
The capacitors I used here are all taken from other projects, so there is a great chance to be (solder and desolder and solder...).
I had checked the power supply and it does not change at all during the event (in fact, this power supply is very good, with good low impedance and low noise). This happens in a channel but during an audio test I suspected that the defect appeared on another channel, but may have just been an swapped channel in the RCA connections during the rapid test.
Another likely defect could be the current source (it it stop providing the output current, logically the voltage output will rise dramatically), but a quick test showed that was still properly polarized (I used a two transistors ring bipolar instead of DN2540).
Some resistors are used but in the previous location they were OK (and them survive better than capacitors to this abuse), yet not rule out the possibility of being one of them.
Probably the fastest would change the capacitors/some res, but as I always like to discover the real cause, so be it ...
Out of curiosity, I used a LM725 opamp in the power supply (with typical over-complicated freq compensation of LM725), and the output reg transistor is a bipolar NPN (BUY49)(with an 6V2 zener in series with base, and an 100R from base to emitter), and proved to be much better in performance compared to "traditional" supplies with LM317 etc but I don't now compared with the original.
Solution...
Last Saturday I was researching a failure in a timer that I did recently, and the problem lay in the PCB. I "sanded" tracks around the suspect sector, and that solved the problem. And in the pre phono the problem was the same: some invisible residue left over from the corrosion time (I always clean the solder flow, but such cleaning is not the same as "sanding" the PCB).
Maybe it's time for renew the PCB's "acid" (in fact, ferric perchloride is an salt)...
In short, I could now have some time to hear the pre and really it is a high-performance specimen!
***"sanded" = in reality is an "brute force" with an endmill attached in an minidrill***
In time, I used selected metal cased BCY58VIII in this project, with good results.
Last Saturday I was researching a failure in a timer that I did recently, and the problem lay in the PCB. I "sanded" tracks around the suspect sector, and that solved the problem. And in the pre phono the problem was the same: some invisible residue left over from the corrosion time (I always clean the solder flow, but such cleaning is not the same as "sanding" the PCB).
Maybe it's time for renew the PCB's "acid" (in fact, ferric perchloride is an salt)...
In short, I could now have some time to hear the pre and really it is a high-performance specimen!
***"sanded" = in reality is an "brute force" with an endmill attached in an minidrill***
In time, I used selected metal cased BCY58VIII in this project, with good results.
Yes intermittent faults are a pain.
How did you verify or discover that the performance of the 725 supply was better?
Jan
LM725...
The LM725 is an ampop with extreme DC gain (>2.000.000, almost comparable to some modern OpAmps), and has a noise figure already interesting, but is much more "slow" that modern options (not to be confused with the LM723, dedicated IC made for regulated sources, more noisy than my option). I used an metal can LM725 in place of the opamp proposed in the project, with proper complex compensation required.
When applying a FFT (using the analog input from an Asus Xonar audio board) direct to the power supply, when using an LM317 I noticed significant more modulation ("invasion") of the power supply than the shunt supply with LM725, which means that the LM725-based impedance must be lower than the LM317.
The LM725 noise level I can not detect in my simple tests, however I can measure the residual noise of the LM317, which indicates that the supply with LM725 is less noisy. And far less noisy than an shunt with TL431.
As I have not built the supply with the original THS opamp I can not say whether it's better or worse, but I imagine it must be even better than the LM725 at HF, and maybe somewhat worse at near DC, IF comparing in same board/implementation.
All my personal use shunt supplies now are based on them.
The LM725 is an ampop with extreme DC gain (>2.000.000, almost comparable to some modern OpAmps), and has a noise figure already interesting, but is much more "slow" that modern options (not to be confused with the LM723, dedicated IC made for regulated sources, more noisy than my option). I used an metal can LM725 in place of the opamp proposed in the project, with proper complex compensation required.
When applying a FFT (using the analog input from an Asus Xonar audio board) direct to the power supply, when using an LM317 I noticed significant more modulation ("invasion") of the power supply than the shunt supply with LM725, which means that the LM725-based impedance must be lower than the LM317.
The LM725 noise level I can not detect in my simple tests, however I can measure the residual noise of the LM317, which indicates that the supply with LM725 is less noisy. And far less noisy than an shunt with TL431.
As I have not built the supply with the original THS opamp I can not say whether it's better or worse, but I imagine it must be even better than the LM725 at HF, and maybe somewhat worse at near DC, IF comparing in same board/implementation.
All my personal use shunt supplies now are based on them.
To aid with building the preamp, the Gerber files with BoM and building tips are now also available; you can find them at
Online Articles
Happy building!
Unfortunately the gerber files are not available anymore. Is possible to have it somehow ?
Thanks,
Adrian
Hi everybody,
sorry to be late, as I didn't note at all that this thread came back to life!
Great to hear your problem is fixed! Enjoy this fine little amp!
@Adrian: sorry to hear the Gerbers are no longer online (apparently due to a broken link).
No problem at all, just send me a pm with your email and I mail you the files.
Hannes
sorry to be late, as I didn't note at all that this thread came back to life!
Great to hear your problem is fixed! Enjoy this fine little amp!
@Adrian: sorry to hear the Gerbers are no longer online (apparently due to a broken link).
No problem at all, just send me a pm with your email and I mail you the files.
Hannes
I put the Gerbers back on the download page, here: Downloads | Linear Audio .
Sorry for the mix-up.
Jan
Sorry for the mix-up.
Jan
Boards available!
Hi everybody,
one of those things happened that are only made possible by the Internet - pcbs are now available!
Malcolm from Tasmania had ordered a few boards from a board house, but apparently there was a misunderstanding and he got much more boards than intended for the excellent price.
The surplus boards he kindly send to me and I'll gladly give them away for the cost of shipping only (8 Euro for worldwide shipping)!
Whoever is interested to build the high octane is welcome to contact me for boards!
Many thanks Malcolm!!
Hi everybody,
one of those things happened that are only made possible by the Internet - pcbs are now available!
Malcolm from Tasmania had ordered a few boards from a board house, but apparently there was a misunderstanding and he got much more boards than intended for the excellent price.
The surplus boards he kindly send to me and I'll gladly give them away for the cost of shipping only (8 Euro for worldwide shipping)!
Whoever is interested to build the high octane is welcome to contact me for boards!
Many thanks Malcolm!!
Attachments
Hi everyone!
Many thanks for your messages to all of you - I feel honored by all this interest!
Please give me a few days to respond to each of you, nobody will be forgotten!
A pm helps me a lot to track the interest, please do not rely only on posting here - thank you!
All the best, Hannes
PS: there are 25 boards to be given away, so there are still a few left
Many thanks for your messages to all of you - I feel honored by all this interest!
Please give me a few days to respond to each of you, nobody will be forgotten!
A pm helps me a lot to track the interest, please do not rely only on posting here - thank you!
All the best, Hannes
PS: there are 25 boards to be given away, so there are still a few left