Inductors? What I added was what I though correctly represented the self inductance of the leads/traces, 10nH per cm.
Many amp designs control the ground loop voltages/currents by strategically placed resistors in the ground loops. Obvious places are between AGND and PGND and/or between PGND and PE. I wonder if it is high loop currents that are being transferred to the audio signal of an interconnect though inductive means?
If you measure the capacitor currents of your PSU, you can see that most of the charging current flows from the positive rail to the negative rail and not through the star . What would happen if the current demand was less symmetrical?
Many amp designs control the ground loop voltages/currents by strategically placed resistors in the ground loops. Obvious places are between AGND and PGND and/or between PGND and PE. I wonder if it is high loop currents that are being transferred to the audio signal of an interconnect though inductive means?
If you measure the capacitor currents of your PSU, you can see that most of the charging current flows from the positive rail to the negative rail and not through the star . What would happen if the current demand was less symmetrical?
In Grounding test2, I increase the wire resistance for the return and the two rails from 0.01 to 0.1ohm to reduce the charging current to 0.1 of the L side. You can see the simulation result is the same, 17mV
It is a little troubling to me when I remove the 1.5V generator and set the 3KHz to 0V and look at the just the 0.0001V generator alone. It couple over more than 3KHz. I don't know why. Look at Grounding test3.
All you have to do is run and read across the dummy load resistor of the R side exactly the same as before.
I know those inductance are wire inductance.
When you connect AGND and other grounds even with a resistor. Once you connect them, there will be current flow between them. There is no way out. It is a mistake to play the grounding ONLY to lower the noise. You are not going to get rid of it. The KEY is to make it common mode and use the common mode rejection of the power "opamp" amp to reject the noise. Read Cordell p380 section 18.4 Input Ground Loops. This is how you make the noise pseudo common mode. That's why I said from the beginning that this is a system design, not just grounding. You cannot ground the noise out.
I drew the way I would do it post #87 if I were to design for OP. I would use FULL DIFFERENTIAL and break the ground right at the input of the preamp.
This is pretty standard way in the industry. That's the reason you'll find a lot of real differential amplifier IC on the market. They put all the resistor inside with the opamp for better matching EXACTLY like what I drew in post #87. This is even better than the pseudo differential design.
It would be interesting if you use my .asc file to start, then change to all the other grounding schemes discussed here and see how they fare. I did the hard work of putting a real power amp into the file already, all you have to do is add the preamp and change the grounding scheme to the other configuration and inject the 1.5V signal to see how they work.
It is a little troubling to me when I remove the 1.5V generator and set the 3KHz to 0V and look at the just the 0.0001V generator alone. It couple over more than 3KHz. I don't know why. Look at Grounding test3.
All you have to do is run and read across the dummy load resistor of the R side exactly the same as before.
I know those inductance are wire inductance.
When you connect AGND and other grounds even with a resistor. Once you connect them, there will be current flow between them. There is no way out. It is a mistake to play the grounding ONLY to lower the noise. You are not going to get rid of it. The KEY is to make it common mode and use the common mode rejection of the power "opamp" amp to reject the noise. Read Cordell p380 section 18.4 Input Ground Loops. This is how you make the noise pseudo common mode. That's why I said from the beginning that this is a system design, not just grounding. You cannot ground the noise out.
I drew the way I would do it post #87 if I were to design for OP. I would use FULL DIFFERENTIAL and break the ground right at the input of the preamp.
This is pretty standard way in the industry. That's the reason you'll find a lot of real differential amplifier IC on the market. They put all the resistor inside with the opamp for better matching EXACTLY like what I drew in post #87. This is even better than the pseudo differential design.
It would be interesting if you use my .asc file to start, then change to all the other grounding schemes discussed here and see how they fare. I did the hard work of putting a real power amp into the file already, all you have to do is add the preamp and change the grounding scheme to the other configuration and inject the 1.5V signal to see how they work.
Last edited:
The result is only about 1mV of noise which is not bad.
But the way you drive it, you defeated the pseudo differential input of the IPS that I put in. What I am trying to do is to make the ground noise from the earth ground common mode.
I am very busy, or else I'll incorporate the preamp circuit into the simulation. I am powering up my 3EF diamond OPS and I have oscillation in the pre-driver and driver stage and I am busy fixing that. My contract work is going to start again next week
. I'll see whether I can find some time modify the .asc file with the preamp.
I don't change the source, those are store bought. You can do the same grounding using a 10ohm resistor to separate the source ground to the preamp ground to limit the current. BUT after that, you use the source ground as the -ve side of the differential signal. Look at post #87, the source ground connects to X1-2 the source signal connects to X1-1. Now the differential amp will cancel out the ground noise. I don't do a thing to the source.
Remember the ground noise from the source also is in the signal, the ground noise will be common mode noise.
This is how the system works, the signal at the output of the source is referenced to the source ground. It is the noise on the source ground WITH RESPECT to the preamp ground that is giving problem, this is how your original .asc file showed, that you drive the preamp ground and it is common mode signal. The signal of the source got to have the ground noise of the source ground.
With the true differential input, you kill the ground noise right at the input. I really have to find some time to put in the preamp and source into the .asc file.
Edit
I drew the system design. The CD player has signal and source ground. I put the noise between the source and preamp ground like you did. I tied the source and preamp ground together by a 10 ohm. Then I use the source ground as the -ve signal of the differential signal.
I labeled the full differential for the preamp and pseudo differential between the preamp and the power amp.
Attachments
Last edited:
OK, I added the full differential preamp and the source for both sides. I labeled L and R preamp, L and R Source. I also created the "source GND" and I put the 1.5V 60Hz generator between the source GND and Preamp GND. I use only a single opamp as preamp, it will be the same if you add the tone control and the second opamp exactly like what the OP wanted. The key is to stop the noise right at the input and the noise doesn't go any farther than the first differential amp. I would use a ground plane for the preamp GND and forget all the star and all on the pcb. It is important to have 22ohm in series with the power rail to the preamp pcb and the IPS. Then follow up with 1000uF cap for filtering out the ripple. The 22ohm current limit the ripple so you would not contaminate the quiet preamp ground and the IPS/VAS ground.
This is what I tested so far. I am sure you guys have more creative ways to test it.
1) 2V 2KHz L-generator on, 1.5V 60Hz noise on, I got about 70uV of feed through of 2KHz from the L side.
2) I turned off the 2KHz L-generator, I got 80uV of 60Hz ripple. I turned off the noise generator, I pretty much get the same amount. I believe that is from the main rectifiers, not from the noise generator. I think the full differential preamp is doing a very good job in eliminating the noise.
Now you can see exactly what I meant about full differential preamp input that has NOTHING to do with the source. I use pseudo differential as per Mr. Cordell at the input of the IPS.
Anyone interested is encouraged to play with it. I attached the model file which I posted before. You have everything needed to play with this. Remember, it is very important you don't add the noise generator between the low side of the source and the source GND. It does not happen like that in real world ever. The noise source is to simulate the ground noise between the source and the preamp, so the noise have to be in the signal of the source also, not exception on this one.
Also, it is very easy to modify this file to test out all the grounding designed posted in this thread since I input the whole typical power amp for you already. Then you can really compare all the ideas.
This is what I tested so far. I am sure you guys have more creative ways to test it.
1) 2V 2KHz L-generator on, 1.5V 60Hz noise on, I got about 70uV of feed through of 2KHz from the L side.
2) I turned off the 2KHz L-generator, I got 80uV of 60Hz ripple. I turned off the noise generator, I pretty much get the same amount. I believe that is from the main rectifiers, not from the noise generator. I think the full differential preamp is doing a very good job in eliminating the noise.
Now you can see exactly what I meant about full differential preamp input that has NOTHING to do with the source. I use pseudo differential as per Mr. Cordell at the input of the IPS.
Anyone interested is encouraged to play with it. I attached the model file which I posted before. You have everything needed to play with this. Remember, it is very important you don't add the noise generator between the low side of the source and the source GND. It does not happen like that in real world ever. The noise source is to simulate the ground noise between the source and the preamp, so the noise have to be in the signal of the source also, not exception on this one.
Also, it is very easy to modify this file to test out all the grounding designed posted in this thread since I input the whole typical power amp for you already. Then you can really compare all the ideas.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.