I can't figure out why changing the feedback network impedance only around the opamp makes any impact on the drive ability. After all, the opamp's not doing the heavy lifting here.
Are you sure you only changed two resistors in the feedback network?
Are you sure you only changed two resistors in the feedback network?
Increasing the impedance of the feedback network will not increase dynamics, it is going the opposite way. 47k gives about 10dB higher Johnson noise than 4k7, 1M about 23dB. Therefore dynamics will decrease. One way to keep noise low ( and dynamic high ) is to keep the impedance of the feedback network as low as possible ( according to drive capability of op amp.) . Driving the original values ( 4k7 .... ) is not a challenge for any audio op amp I know.
You can enhance the dynamic headroom by increasing the supply voltage ( respect the maximum rating of your op amp.), for example from 15V to 17V will give you 1dB ( but it´s not that impressive improvement).
You can enhance the dynamic headroom by increasing the supply voltage ( respect the maximum rating of your op amp.), for example from 15V to 17V will give you 1dB ( but it´s not that impressive improvement).
I am sure about what I changed. And I am not happy with what I am hearing. It sounds fine on the active loudspeakers. Nothing special but on the headphones it's flat and dull. Cannot rule any differently.
I will do some more testing today with different combinations of parts and equipment to figure some more out.
I will do some more testing today with different combinations of parts and equipment to figure some more out.
Ok, I have to share some information here regarding the resistor experiment. I did brief testing and everything sounded really bad. I had to turn on all the gain switches to even get volume through my headphones. I checked the board again and I couldn't find anything alarming. DC offset measured perfectly etc etc.
I just put on my headphones (Mad Dog) and fiddled a bith with the gain switches and tried some different tracks. After a minutes everything started getting louder and opened up. I could turn the second gain switch off again and had proper volume on the headphones.
I can recall I had a sort like experience earlier with the Lehmann. I think since it had been of the power it needed some time to warm up and for the mundorfs to charge properly maybe because the volume is fine now and the sound is much less dull and flat with the new resistors than before.
Right now I'm using 49.9K+33pf NP0 over it, and 22.1K and 6.8K. so approximately factor 10.
What I notice is a more spacious sound. It's not dull or flat anymore. It's a smoother sound I guess and in the beginning I thought it might be a bit too foggy but after dropping in my OPA111VMs this totally settled and I can say it is not the most detailed 'audiophile' sound but it's very pleasant to listen to and not fatiguing at all. I am going to try 22pf for the 33pf caps to see if this helps the HF coming forward a bit more, because it does feel a bit smoothed out.
If you like a very smooth and open sound this is definitely worth a try. I have recall my earlier statement because the reason of the dull and flat sound is not clear to me.
I just put on my headphones (Mad Dog) and fiddled a bith with the gain switches and tried some different tracks. After a minutes everything started getting louder and opened up. I could turn the second gain switch off again and had proper volume on the headphones.
I can recall I had a sort like experience earlier with the Lehmann. I think since it had been of the power it needed some time to warm up and for the mundorfs to charge properly maybe because the volume is fine now and the sound is much less dull and flat with the new resistors than before.
Right now I'm using 49.9K+33pf NP0 over it, and 22.1K and 6.8K. so approximately factor 10.
What I notice is a more spacious sound. It's not dull or flat anymore. It's a smoother sound I guess and in the beginning I thought it might be a bit too foggy but after dropping in my OPA111VMs this totally settled and I can say it is not the most detailed 'audiophile' sound but it's very pleasant to listen to and not fatiguing at all. I am going to try 22pf for the 33pf caps to see if this helps the HF coming forward a bit more, because it does feel a bit smoothed out.
If you like a very smooth and open sound this is definitely worth a try. I have recall my earlier statement because the reason of the dull and flat sound is not clear to me.
By 'dynamics' I didn't mean 'dynamic range' as measured. I agree that increasing the resistor value increases the Johnson noise - however its still going to be lower than the noise on recordings from RBCD with 47k.
And I also agree there - in my experience one does not need to 'challenge' an opamp to get less than optimal sound.
Sorry, it wasn´t clear to me that you meant "dynamics" in a way of sound description.
I also think that it´s not the best idea to drive op amps. too close to their limits, you always have to find a way between noise, THD and some more parameters, but in the end what counts will be of course the sound performance.
I also think that it´s not the best idea to drive op amps. too close to their limits, you always have to find a way between noise, THD and some more parameters, but in the end what counts will be of course the sound performance.
No worries, its an easy mistake to make - I am looking forward to a day when I can present a measurement which correlates well with subjective dynamics. But not much progress to report towards that goal so far. I did have an idea to use a subtraction circuit to find the difference in output between a very lightly loaded and more heavily loaded opamp, perhaps using AD8129 which has a very high CMRR (100dB or more at LF).
The use of a subtraction circuit is an interesting approach, I also had many thoughts about correlation between the measurements and the subjective sound perception. Surprisingly, you will not find a lot of research studies concerning this topic. There is still a lot of work to do, and maybe some exceptional results are waiting for us. Remember, it was a long time "natural law" that distortions below 1% are inaudible, meanwhile it is stated 0,1% or lower.
So, keep it rolling and please present your results on diyaudio 🙂.
So, keep it rolling and please present your results on diyaudio 🙂.
Hi Cidious,
I had a look over your posts now, and some things made me wonder. Of course, changing parts (and burn in time) will influence the sound but not in that dramatic way you occasionally described. Especially it should not affect channel balance our sound level at all. Hot swollen caps?- it should not even happen to the cheapest standard caps. Audible distortions after slight changes?
This all tells me that something is not working in the way as it should do.
My first conclusion was, that there something is oscillating in the circuit. It is possible that the oscillation is not stable itself and the amplitude changes for example with temperature or it stops sometimes completely. This could explain why the sound performance is that "unstable".
On the other hand, I had a look to the datasheets of your preferred OPAs and it seems they are not problematic regarding stability. So I don´t know.......
Another possibility will be, that the DIP-switches for gain are not working well, but it´s unlikely that both will fail in the same way.
I had a look over your posts now, and some things made me wonder. Of course, changing parts (and burn in time) will influence the sound but not in that dramatic way you occasionally described. Especially it should not affect channel balance our sound level at all. Hot swollen caps?- it should not even happen to the cheapest standard caps. Audible distortions after slight changes?
This all tells me that something is not working in the way as it should do.
My first conclusion was, that there something is oscillating in the circuit. It is possible that the oscillation is not stable itself and the amplitude changes for example with temperature or it stops sometimes completely. This could explain why the sound performance is that "unstable".
On the other hand, I had a look to the datasheets of your preferred OPAs and it seems they are not problematic regarding stability. So I don´t know.......
Another possibility will be, that the DIP-switches for gain are not working well, but it´s unlikely that both will fail in the same way.
I got swollen/bulging caps(two of them) in my first BCL clone.
It was an already assembled board.
This happened after about 4-5years of use.
Opamps, from what I have read in other Posts(IIRC), there are not that many that are stable in this circuit.
It was an already assembled board.
This happened after about 4-5years of use.
Opamps, from what I have read in other Posts(IIRC), there are not that many that are stable in this circuit.
Were it the caps beside the opamp?
By the way, I had a closer look to the pics of your pcb; you did a cut through the groundplane. Is there still a connection between the planes? I can´t see the bottom side, so it might be there. If not - you got a problem ;-).
By the way, I had a closer look to the pics of your pcb; you did a cut through the groundplane. Is there still a connection between the planes? I can´t see the bottom side, so it might be there. If not - you got a problem ;-).
No, it was the 2nd and 3d from the left seen from the front of the PCB.
I've made a star-gnd under the PCB using a ferrite bead parallelled with a 100nf cap between the GND points and the star-gnd.
Writing from my phone, don't have pics of that on it.
Edit:
Found a bad pic, but you get the point.
Attachments
Last edited:
You are right, caps can fail and they do. What made me wonder was that the caps in Cidious´ amp. failed after a short time. I have seen a couple of swollen/bursted caps on computer mainboards or fast switching psu. In such "high speed" applications caps get highly stressed, but in the circuit we are talking about, the stress to the caps is limited (imo).
I am not sure whether this is remarkable or just normal fatigue.
This is a simple circuit but it might have some pitfalls.
To push on my own project with this circuit I have still to do some tests and experiments also regarding stability. Maybe I can lift some secrets of it 😕
Wierd thing is that those two caps got much warmer than the others after swapping to Pana FC 35V ones, as I tried fixing the first BCL before giving up on it.
Now I am in no hurry to get the mkII done as I have a XP7 clone that sounds very nice.
I'll compare the two when the mkII is done.
I also have a mod to try on the XP7, putting 22uH very low impedance inductors in series with the supplies to the AD797's.
In ltspice it is an improvement, remains to be seen if it is in real world application as well.
Indeed, it is remarkable. At the moment I´ve no idea what could be the reason for it.
I plan to use AD797 in my hp amp. project, so please can you tell me what will be the improvement using the series inductors in the supply rails.
I got "nicer" FFT graphs when using them, less distortion.
It was a while ago I did the simulations though.
I'll run new simulations before modding the XP7.
Thanks. This makes me curious, I would be very pleased if you could post the graphs when you have completed the simulations.🙂
It's was a weird thing. But the channel imbalance could have come from anything. After I've rechecked the board for bad solder points and only replaced the resistors in between the mundorf caps for metal films, the imbalance was gone.
What wasn't solved was the low volume and the under powered sound. But while I was looking on the internet I had my headphones on and the device on full gain and full volume which was still nothing near very loud sound levels. After a few minutes the sound became louder and eventually I had to turn down the volume to not let it hurt my hears. After some more minutes I could turn down the gain switches one step and had my old volume levels back..
Like I said.. A weird problem.. but not one I haven't experienced before. I had this with my REL STORM 10" (30kg) sub woofer. I bought it from a electrostatic freak that wanted to go back to 2.0. He hadn't used it for 4 years. All this time it stood idle on the attic. When I powered it on, it was not producing any sound. I had to keep it on the power first and put it on some test tones first to get everything beefed up again. After one night on the power everything came to life and was like it was supposed to be 🙂 This puppy had some really big passive components.. But like I said. A sort like problem happened to me before..
What wasn't solved was the low volume and the under powered sound. But while I was looking on the internet I had my headphones on and the device on full gain and full volume which was still nothing near very loud sound levels. After a few minutes the sound became louder and eventually I had to turn down the volume to not let it hurt my hears. After some more minutes I could turn down the gain switches one step and had my old volume levels back..
Like I said.. A weird problem.. but not one I haven't experienced before. I had this with my REL STORM 10" (30kg) sub woofer. I bought it from a electrostatic freak that wanted to go back to 2.0. He hadn't used it for 4 years. All this time it stood idle on the attic. When I powered it on, it was not producing any sound. I had to keep it on the power first and put it on some test tones first to get everything beefed up again. After one night on the power everything came to life and was like it was supposed to be 🙂 This puppy had some really big passive components.. But like I said. A sort like problem happened to me before..
Last edited:
I ran a quick simulation now.
I'm the first to admit that I'm less than an expert on designing audio or using LTSpice.
For example I used two resistors as a "simulated" pot...
