Hi Jorgen.
You did not miss anything. I have not posted OLG, gain/phase margin and also small/large signal squarewave with(out) capacitive load.
The instability problem that arose in January 2019 and later got solved by installing mosfet drain-gate capacitors was due to the unusual layout (indeed this is not the ideal layout for fast amps), parallel placement of mosfets, and long copper traces at VAS output (which has been loaded to ground via two 47pF since revision 1). Is this the solution you point to? Because there was an earlier solution posted by myself, akin to P101 gate-source capacitor solution, but didn't work in v4h.
Hi again
I have read your thread from end to end of interest, not because I was in need for a solution to a problem. I am looking for an amplifier to build, and I like what I see. No unobtaniums
I have 10 sets of 2SK176/2SJ56, and I think that a couple of them would be a good match for your design. I like the sound the old mosfets have. They sounded 'warm' to me.
After reading about the stability issue and the solution, I simply got curious about the numbers and square waves, as they can hint of potential problems. I was hoping for some before/after images/numbers. Nothing more
Keep up the good work, looking forward to read about your preamp
Jørgen
I have read your thread from end to end of interest, not because I was in need for a solution to a problem. I am looking for an amplifier to build, and I like what I see. No unobtaniums
I have 10 sets of 2SK176/2SJ56, and I think that a couple of them would be a good match for your design. I like the sound the old mosfets have. They sounded 'warm' to me.
After reading about the stability issue and the solution, I simply got curious about the numbers and square waves, as they can hint of potential problems. I was hoping for some before/after images/numbers. Nothing more
Keep up the good work, looking forward to read about your preamp
Jørgen
Hi again
I have read your thread from end to end of interest, not because I was in need for a solution to a problem. I am looking for an amplifier to build, and I like what I see. No unobtaniums
I have 10 sets of 2SK176/2SJ56, and I think that a couple of them would be a good match for your design. I like the sound the old mosfets have. They sounded 'warm' to me.
After reading about the stability issue and the solution, I simply got curious about the numbers and square waves, as they can hint of potential problems. I was hoping for some before/after images/numbers. Nothing more
Keep up the good work, looking forward to read about your preamp
Jørgen
Hi.
I appreciate your interest and patience. Happy to know you like my no-unobtainium approach, although I am a little worried about the fate of lateral devices. In an online search one can find that there are a lot of research that have been done on new lateral designs based on silicon carbide. If one pair hits the market as power devices with similar characteristics as the present ones, will be superb. The metal mosfet pair you mention looks like genuine unobtainium though, you're probably lucky to have them. I agree they have a warm note left on the music. Don't know how you will fit them on the board though.
I know OLG, phase/gain margin can indicate potential problems. As I mentioned in a recent post a few weeks ago, I have driven fast squares into the amp after the stability upgrade and it remained stable regardless of output level. I hope you are not worried about potential problems anymore. But if you still are, there is always an inductor that you can use to run the amp into capacitive loads safely.
Thanks for the kind words.
Hi.
Check all the resistor values next. Also the LEDs for normal voltage drop near 1.8V.
Yet another small problem
Hi Shaan !
I received my speaker protections, made my 2 GLB (ground loop breakers) circuits, and finally installing my 2 V4H boards in their enclosure, dual mono setup.
I previously did all the setup procedure for both amplifier boards without the GLB installed, and everything was fine tuned, offset <2mV.
But now, with its GLB installed, the offset of one of the boards goes up to -165mV !!
GLB removed, back to 2mV. The GLB itself is not faulty, as I swapped it with my other GLB and the offset also going up to 165mV.
The other amplifier board is normal, 2mV offset with or without GLB.
Of course I could leave the Ground (OV) floating on this side, for this channel, the offset would stay 2mV ............... but ..............
What's wrong ?
Thanks a lot for your help.
JM
Hi Shaan !
I received my speaker protections, made my 2 GLB (ground loop breakers) circuits, and finally installing my 2 V4H boards in their enclosure, dual mono setup.
I previously did all the setup procedure for both amplifier boards without the GLB installed, and everything was fine tuned, offset <2mV.
But now, with its GLB installed, the offset of one of the boards goes up to -165mV !!
GLB removed, back to 2mV. The GLB itself is not faulty, as I swapped it with my other GLB and the offset also going up to 165mV.
The other amplifier board is normal, 2mV offset with or without GLB.
Of course I could leave the Ground (OV) floating on this side, for this channel, the offset would stay 2mV ............... but ..............
What's wrong ?
Thanks a lot for your help.
JM
Interesting.
With the GLB installed what voltage do you measure across the GLB for each channel?
OK, with GLB installed:
0 mV on the "good" side (left channel), black probe of the DMM on the Earth, red probe on the Ground 0V
-35mV on the "bad side" : black probe of the DMM on the Earth, red probe on the Ground 0V
Could the smps of the bad side be responsible ?
The GLB are connected Ground directly to Ground Earth of AC connector and Ground Earth of smps, 0V directly to 0V of the smps.
JM
Do the SMPSs have their own mains earth connection along with mains live and neutral? If yes, are they connected to mains earth?
If not, one more test. Power each channel from the SMPS of the other channel and check if the same problem occurs with the previously 2mV reading channel, or the same channel reads high offset again.
Each smps has its own connection (earth live & neutral) to a common plug.
Mains earth is OK in this house.
Maybe I should investigate deeper this Earth question in connectors, plugs etc
JM
Is the mains earth terminal of the smps connected to its 0V output on-board? You can find it via probing with continuity tester. If yes, then remove the earth connections from both smps. Then connect the 0V outputs to chassis earth only via GLBs. Then check again for offset.
If not, then there may be a problem with the 0V output powering the channel that's showing high offset.
I've already checked this. 0V output terminal of smps is not connected to earth terminal of smps. It's connected only through Glb.
Ok, I'm going to swap both smps and see.
Thank you
JM
More details
The other 'good side' is perfectly ok with the suspected smps and glb and cords, offset 3mV.
Then I guess there is no offset problem with this smps.
One more thing I should do is switching the glbs I guess ....
JM
Ok, I just switched the glbs, they are innocent .....
Any idea?
Tuning this board without glb and according to the procedure is not a problem.
As soon as the glb is connected, offset jumps ....
The other 'good side' is perfectly ok with the suspected smps and glb and cords, offset 3mV.
Then I guess there is no offset problem with this smps.
One more thing I should do is switching the glbs I guess ....
JM
Ok, I just switched the glbs, they are innocent .....
Any idea?
Tuning this board without glb and according to the procedure is not a problem.
As soon as the glb is connected, offset jumps ....
Last edited:
Ok. Last check and we will move to the amp modules. Previously you said there is about 35mV across the GLB. Is this voltage still there when 0V is connected to GLB but amp board is not connected to smps? Or does this occur only when the amp board is connected?
The thing is that the offset that is happening at the amp output is a result of something happening elsewhere either on the amp board or the 0V on the SMPS, the offset cannot physically cause the 35mV drop you previously measured across the GLB if all connections are done correctly, but the 35mV drop can surely cause the offset. I guess the heatsinks are connected to earth directly via chassis metal and not connected to the 0V on the amplifier boards?
So yes, please check for shorts to heatsink on the amp board if the 35mV drop is not present when amp board is not connected to smps.
The thing is that the offset that is happening at the amp output is a result of something happening elsewhere either on the amp board or the 0V on the SMPS, the offset cannot physically cause the 35mV drop you previously measured across the GLB if all connections are done correctly, but the 35mV drop can surely cause the offset. I guess the heatsinks are connected to earth directly via chassis metal and not connected to the 0V on the amplifier boards?
So yes, please check for shorts to heatsink on the amp board if the 35mV drop is not present when amp board is not connected to smps.
Ok. Last check and we will move to the amp modules. Previously you said there is about 35mV across the GLB. Is this voltage still there when 0V is connected to GLB but amp board is not connected to smps? Or does this occur only when the amp board is connected?
I've checked this: there is OmV across the GLB when the amp board is not connected.
It only occurs when amp board connected.
I guess the heatsinks are connected to earth directly via chassis metal and not connected to the 0V on the amplifier boards?
Yes, it is, there is a very good Earth cnx on the heatsink.
So yes, please check for shorts to heatsink on the amp board if the 35mV drop is not present when amp board is not connected to smps.
I don't know how to perform this with the amp board fasten on its heatsink.
I've had this amp board working and singing very well for an hour (without GLB)
and everything seemed normal ...........
Thanks for the help !!
JM
I've checked this: there is OmV across the GLB when the amp board is not connected.
It only occurs when amp board connected.
I guess the heatsinks are connected to earth directly via chassis metal and not connected to the 0V on the amplifier boards?
Yes, it is, there is a very good Earth cnx on the heatsink.
So yes, please check for shorts to heatsink on the amp board if the 35mV drop is not present when amp board is not connected to smps.
I don't know how to perform this with the amp board fasten on its heatsink.
I've had this amp board working and singing very well for an hour (without GLB)
and everything seemed normal ...........
Thanks for the help !!
JM
You need to test for short with the amp mounted on the heatsink. Just remove all connections to SMPS, then place one probe of the continuity tester to chassis earth and the other probe to the five terminals one by one. There should be infinite resistance shown, neither short or any particular resistance value.
If the test show infinite resistance i.e. no short or finite resistance, then possubly the channel is oscillating. If you have an oscilloscope then check for this with the amp powered up. If you don't have a scope then take a 1N4148, a 10ohm 0.25watt resistor, and a 1uF film capacitor. Place all three in series respectively, if the high offset was positive in polarity then keep the anode pin of the diode facing the amplifier output terminal. Connect one of the open end of this network to amplifier output and the other end to 0V. Turn on amplifiee and measure the voltage across capacitor. If there is oscillation it should read 700mV or higher.
If no oscillation is present, then it gets a little harder to help you debug for me remotely. But we will see about that.
Happy to help.
If the test show infinite resistance i.e. no short or finite resistance, then possubly the channel is oscillating. If you have an oscilloscope then check for this with the amp powered up. If you don't have a scope then take a 1N4148, a 10ohm 0.25watt resistor, and a 1uF film capacitor. Place all three in series respectively, if the high offset was positive in polarity then keep the anode pin of the diode facing the amplifier output terminal. Connect one of the open end of this network to amplifier output and the other end to 0V. Turn on amplifiee and measure the voltage across capacitor. If there is oscillation it should read 700mV or higher.
If no oscillation is present, then it gets a little harder to help you debug for me remotely. But we will see about that.
Happy to help.
You need to test for short with the amp mounted on the heatsink. Just remove all connections to SMPS, then place one probe of the continuity tester to chassis earth and the other probe to the five terminals one by one. There should be infinite resistance shown, neither short or any particular resistance value.
I get infinite R on all terminals, except for SGND where I get 41MOhm (next to infinite on my DMM)
If the test show infinite resistance i.e. no short or finite resistance, then possubly the channel is oscillating. If you have an oscilloscope then check for this with the amp powered up. If you don't have a scope then take a 1N4148, a 10ohm 0.25watt resistor, and a 1uF film capacitor. Place all three in series respectively, if the high offset was positive in polarity then keep the anode pin of the diode facing the amplifier output terminal. Connect one of the open end of this network to amplifier output and the other end to 0V. Turn on amplifiee and measure the voltage across capacitor. If there is oscillation it should read 700mV or higher.
I have an oscilloscope. How should I proceed ?
Amp powered up, and checking on OUT with a probe ?
JM
I get infinite R on all terminals, except for SGND where I get 41MOhm (next to infinite on my DMM)
If the test show infinite resistance i.e. no short or finite resistance, then possubly the channel is oscillating. If you have an oscilloscope then check for this with the amp powered up. If you don't have a scope then take a 1N4148, a 10ohm 0.25watt resistor, and a 1uF film capacitor. Place all three in series respectively, if the high offset was positive in polarity then keep the anode pin of the diode facing the amplifier output terminal. Connect one of the open end of this network to amplifier output and the other end to 0V. Turn on amplifiee and measure the voltage across capacitor. If there is oscillation it should read 700mV or higher.
I have an oscilloscope. How should I proceed ?
Amp powered up, and checking on OUT with a probe ?
JM
I get infinite R on all terminals, except for SGND where I get 41MOhm (next to infinite on my DMM)
This is a potential source of the problem. It should read infinite as all the other terminals, because compared to the other terminals this one has the lowest area of copper trace, so even if the other terminals gave 41Mohm this one should give infinite. Check the insulation, also see if there is any component leg under the board that is very close to the heatsink. Cut it short.
I have an oscilloscope. How should I proceed ?
Amp powered up, and checking on OUT with a probe ?
Yes.
This is a potential source of the problem. It should read infinite as all the other terminals, because compared to the other terminals this one has the lowest area of copper trace, so even if the other terminals gave 41Mohm this one should give infinite. Check the insulation, also see if there is any component leg under the board that is very close to the heatsink. Cut it short.
I have an oscilloscope. How should I proceed ?
Amp powered up, and checking on OUT with a probe ?
Yes.