when you disconnect one channel you MUST also disconnect all the ground links of that channel. Otherwise the disabled channel could still create a ground loop.
A ground loop through the disabled channel will affect the readings that you get from the enabled channel.
I realise your readings may be affected by low battery.
Look at your mVdc offset values for one channel.
They seem to be varying depending on the input conditions. That should not happen.
With an open interconnect attached, your readings for mVac are all very high. This proves that your open cable is acting as an aerial trying to pick up everything passing through your house. Just use an open RCA socket, do not attach a cable.
A ground loop through the disabled channel will affect the readings that you get from the enabled channel.
I realise your readings may be affected by low battery.
Look at your mVdc offset values for one channel.
They seem to be varying depending on the input conditions. That should not happen.
With an open interconnect attached, your readings for mVac are all very high. This proves that your open cable is acting as an aerial trying to pick up everything passing through your house. Just use an open RCA socket, do not attach a cable.
Last edited:
the Chassis Ground note (at bottom of page) is unnecessary.
The IEC PE wire should be connected directly to chassis close and permanently.
No other wire must be allowed to disturb this PERMANENT Safety connection.
The Main Audio Ground (MAG) can be connected directly to Chassis next to the MAG, or can be connected to Chassis via the Disconnecting Network to a chassis point next to the MAG .
The MAG to Chassis connection does NOT need to be near the Safety PE to Chassis connection.
I prefer to see the dual rectifier Zero Volt wires connected to each other. Then a single wire to the MAG. As shown you have two Zero Volt wires going to MAG. This can create a small voltage differential across the joint and could make for pulse interference in the other MAG connections.
Tauro this looks like a scrubber circuit. I made and talk about one in this thread.
http://www.diyaudio.com/forums/digital-source/234612-sony-dvp-s9000s-loud-humming.html
If all else fails I might give it a go. Your schematic looks great btw, it is exactly how it is in my amp!
Andrew
I did not disconnect all the wires, and it might still have a ground loop. I will do that and see what happens.
I might be able to hook up an external PSU today. Lets see if I am able to.
http://www.diyaudio.com/forums/digital-source/234612-sony-dvp-s9000s-loud-humming.html
If all else fails I might give it a go. Your schematic looks great btw, it is exactly how it is in my amp!
Andrew
I did not disconnect all the wires, and it might still have a ground loop. I will do that and see what happens.
I might be able to hook up an external PSU today. Lets see if I am able to.
Looking at the diagram above and then comparing to my amp it would seem that my psu is floating with a nominal star earth point existing on the mounting block above the transformer where all the 0v from the caps and the speaker returns effectively end up. This is the 0v point of the transformer, but it's not connected back to a DDRC and then back to the chassis earth at any point.
Am I ok running a wire back from the 0V point on the mounting block to the chassis earth point via a single wire and a DDRC?
An externally hosted image should be here but it was not working when we last tested it.
Am I ok running a wire back from the 0V point on the mounting block to the chassis earth point via a single wire and a DDRC?
Tauro/sq,
The snubber (DDRC) circuit can mask problems with the system. When I had to use a snubber circuit, my problem was actually caused by a badly wired power cable going into a power strip. The power cable wasn't even connected to the CD player that I thought was giving me the problem!!! turned out it wa a power cable and the CD player just "helped" the issue get even worse by making another ground loop. So yeah use those with a grain of salt :/
Anyhow,
I took your advice and disconnected the board completely from the RCA Plug. I unsoldered the ground and signal connections. I also disconnected the Speaker Ground connection so the board should have been completely disconnected from the circuit. The board was however screwed into the chassis so that's that.
The testing did not make any difference. I also took the advice and plugged in an open RCA plug instead of an open RCA Cable and you were right, the noise decreased to the levels as if I didn't have the plug and the RCA input was open. So the cable is picking up ALOT of noise!!!!
Unfortunately I couldn't find a suitable +-60 VDC power suply :/ or a dual AC Power supply to isolate the transformer.
I really think this is a component issue now.
The snubber (DDRC) circuit can mask problems with the system. When I had to use a snubber circuit, my problem was actually caused by a badly wired power cable going into a power strip. The power cable wasn't even connected to the CD player that I thought was giving me the problem!!! turned out it wa a power cable and the CD player just "helped" the issue get even worse by making another ground loop. So yeah use those with a grain of salt :/
Anyhow,
I took your advice and disconnected the board completely from the RCA Plug. I unsoldered the ground and signal connections. I also disconnected the Speaker Ground connection so the board should have been completely disconnected from the circuit. The board was however screwed into the chassis so that's that.
The testing did not make any difference. I also took the advice and plugged in an open RCA plug instead of an open RCA Cable and you were right, the noise decreased to the levels as if I didn't have the plug and the RCA input was open. So the cable is picking up ALOT of noise!!!!
Unfortunately I couldn't find a suitable +-60 VDC power suply :/ or a dual AC Power supply to isolate the transformer.
I really think this is a component issue now.
Hi,
My advice it is to scope all the transistors base one by one and see if you can determine if one have any issues with noise. I think it would let know if you have a component problem. The base should be at least clean. One thing make sure the input it is shorted. It is very hard to troubleshoot this problem without schematic. Give it a try and see what happen.
My advice it is to scope all the transistors base one by one and see if you can determine if one have any issues with noise. I think it would let know if you have a component problem. The base should be at least clean. One thing make sure the input it is shorted. It is very hard to troubleshoot this problem without schematic. Give it a try and see what happen.
Did you scope across each transformer secondary? As gmphadte said, it looks like you should install a snubber on each of those. A snubber is a resistor, usually with a series capacitor to limit the frequencies that the resistor snubs, to only the unwanted frequencies.
If you see any ringing or overshoot, there, you can try the following procedure, to find the optimal R value, and the series C value to go with it:
http://www.diyaudio.com/forums/powe...lm-caps-electrolytic-caps-30.html#post2828689
This will not solve the entire problem. But it might take away the buzzing, leaving only the hum.
By the way, it is difficult to imagine your that hum problem could be caused by components, unless perhaps it is failed electrolytic decoupling capacitors, in both channels. Otherwise, how likely is it that components would be causing hum in BOTH channels?
If you see any ringing or overshoot, there, you can try the following procedure, to find the optimal R value, and the series C value to go with it:
http://www.diyaudio.com/forums/powe...lm-caps-electrolytic-caps-30.html#post2828689
This will not solve the entire problem. But it might take away the buzzing, leaving only the hum.
By the way, it is difficult to imagine your that hum problem could be caused by components, unless perhaps it is failed electrolytic decoupling capacitors, in both channels. Otherwise, how likely is it that components would be causing hum in BOTH channels?
Last edited:
Are you familiar with Faraday's Law?
It relates time-varying magnetic fields and currents in conductive loops.
Any time-varying field will induce a corresponding time-varying current in every conductive loop, with a magnitude (if all else is equal) that is proportional to the geometric area enclosed by the loop. Also, conversely, any time-varying current in a loop will create a time-varying magnetic field.
Think of every pair of transformer wires as a loop, both primaries and secondaries. To make their loop area as small as possible, the two wires must stay very close to each other, everywhere. The usual way is to twist them tightly together, all the way to each end.
The rectifier output pairs also have such time-varying currents. So the wires in those pairs should also be kept as close together as possible, everywhere.
If one of your AC mains input wires goes to a switch and then to one of the transformer primaries, then the wire that goes from the input to the switch must follow the other switch wire (the other transformer primary) all the way to the transformer (twisted with it, ideally), and then follow the wire from the transformer to the switch (twisted with it, ideally). If this is not done, a large "enclosed loop area" is created and it will induce hum in every other loop that has enclosed loop area.
If your RCA input wire pairs get separated from each other, that would form some enclosed loop area, which would be an antenna for the fields that the AC and rectified AC have radiated, putting hum into the inputs of a high-gain power amplifier. As gmphadte has said, the wiring from the RCA inputs to the PCBs should be shielded twisted pair. The shield should NOT carry the audio ground. The shield should be connected at one end only, to the chassis. Typically, the input signal ground should go to the ground end of an input resistor, on the PCB, and THEN to the star ground, with its own separate star ground connection. Yours seems to be done differently. It might be creating a large "enclosed loop area", making an antenna for hum, if the input resistor's ground is ALSO connected to the star ground, from the PCB.
It relates time-varying magnetic fields and currents in conductive loops.
Any time-varying field will induce a corresponding time-varying current in every conductive loop, with a magnitude (if all else is equal) that is proportional to the geometric area enclosed by the loop. Also, conversely, any time-varying current in a loop will create a time-varying magnetic field.
Think of every pair of transformer wires as a loop, both primaries and secondaries. To make their loop area as small as possible, the two wires must stay very close to each other, everywhere. The usual way is to twist them tightly together, all the way to each end.
The rectifier output pairs also have such time-varying currents. So the wires in those pairs should also be kept as close together as possible, everywhere.
If one of your AC mains input wires goes to a switch and then to one of the transformer primaries, then the wire that goes from the input to the switch must follow the other switch wire (the other transformer primary) all the way to the transformer (twisted with it, ideally), and then follow the wire from the transformer to the switch (twisted with it, ideally). If this is not done, a large "enclosed loop area" is created and it will induce hum in every other loop that has enclosed loop area.
If your RCA input wire pairs get separated from each other, that would form some enclosed loop area, which would be an antenna for the fields that the AC and rectified AC have radiated, putting hum into the inputs of a high-gain power amplifier. As gmphadte has said, the wiring from the RCA inputs to the PCBs should be shielded twisted pair. The shield should NOT carry the audio ground. The shield should be connected at one end only, to the chassis. Typically, the input signal ground should go to the ground end of an input resistor, on the PCB, and THEN to the star ground, with its own separate star ground connection. Yours seems to be done differently. It might be creating a large "enclosed loop area", making an antenna for hum, if the input resistor's ground is ALSO connected to the star ground, from the PCB.
Star grounding:
Signal input grounds should have their own separate connection to the single star ground point, and should not share any length of conductor with any other ground-return current, from any other part of the circuit.
All time-varying ground-return currents WILL induce time-varying voltages that are distributed across the ground-return conductors, themselves.
Any "ground" point that is connected to any ground-return conductor that SHARES any length of conductor with any other ground's ground-return current, will have some non-zero time-varying voltage component from the other ground's return currents. i.e. the "ground" voltage, back at the PCB end of the ground return conductor, will be a non-zero time-varying voltage; a "bouncing ground".
You could use a separate conductor for every ground. But some are much less sensitive than others. The main ones to worry about are the signal input grounds, right at the amplifier inputs.
If the "ground" voltage at the ground end of the input-to-ground resistor (at the amplifier's input) is a time-varying voltage, that time-varying "ground" voltage will be effectively arithmetically summed with the input signal voltage!
The star ground point: You would not want to have any PCB or audio ground points connected anywhere BEFORE any of the reservoir/smoothing capacitors. Otherwise, the voltages induced across the inductance and resistance of the ground conductors, themselves, by the charging pulse currents, would get into the entire ground network. So your star ground should be at least a little bit downstream from the ground pin of the last reservoir cap (i.e. not before the output of the power supply).
Signal input grounds should have their own separate connection to the single star ground point, and should not share any length of conductor with any other ground-return current, from any other part of the circuit.
All time-varying ground-return currents WILL induce time-varying voltages that are distributed across the ground-return conductors, themselves.
Any "ground" point that is connected to any ground-return conductor that SHARES any length of conductor with any other ground's ground-return current, will have some non-zero time-varying voltage component from the other ground's return currents. i.e. the "ground" voltage, back at the PCB end of the ground return conductor, will be a non-zero time-varying voltage; a "bouncing ground".
You could use a separate conductor for every ground. But some are much less sensitive than others. The main ones to worry about are the signal input grounds, right at the amplifier inputs.
If the "ground" voltage at the ground end of the input-to-ground resistor (at the amplifier's input) is a time-varying voltage, that time-varying "ground" voltage will be effectively arithmetically summed with the input signal voltage!
The star ground point: You would not want to have any PCB or audio ground points connected anywhere BEFORE any of the reservoir/smoothing capacitors. Otherwise, the voltages induced across the inductance and resistance of the ground conductors, themselves, by the charging pulse currents, would get into the entire ground network. So your star ground should be at least a little bit downstream from the ground pin of the last reservoir cap (i.e. not before the output of the power supply).
Where are the testing measurements that led you to that conclusion?
You should have 8 DC voltage measurements and 8 AC voltage measurements.
No.
The mains wiring looks to have untwisted pairs.
That could be rewired to measure if that makes any difference.
The secondary wiring to the rectifiers looks to have untwisted triplets.
The PSU wiring from the rectifiers looks to have untwisted triplets.
Reducing the radiating areas of all these cables could make a very big difference to what the input wiring sees as interference.
The input wiring could be saved for a stage 17 modification.
Is there any RF attenuation at the input sockets, or between the input sockets and the input transistors?
The mains wiring looks to have untwisted pairs.
That could be rewired to measure if that makes any difference.
The secondary wiring to the rectifiers looks to have untwisted triplets.
The PSU wiring from the rectifiers looks to have untwisted triplets.
Reducing the radiating areas of all these cables could make a very big difference to what the input wiring sees as interference.
The input wiring could be saved for a stage 17 modification.
Is there any RF attenuation at the input sockets, or between the input sockets and the input transistors?
Last edited:
There s a common mistake that is often done and it seems
to be the case in this amplifier and it s likely to be systematicaly
the case with a wiring such as in post 42.
Connecting the RCA ground to the star ground is erroneous
as the amplifier board is also connected to this star ground,
hence the zobel network wich goes to ground in the PCB
will induce a variable AC current running from PCB to star
ground at high power , thus there s no more equipotential
grounding since the voltage induced in said wire will be seen
as being in serial with the input signal , increasing THD and noise
at high power.
Set apart this point the problems encountered by the OP
are typical of an erroneous grounding scheme.
to be the case in this amplifier and it s likely to be systematicaly
the case with a wiring such as in post 42.
Connecting the RCA ground to the star ground is erroneous
as the amplifier board is also connected to this star ground,
hence the zobel network wich goes to ground in the PCB
will induce a variable AC current running from PCB to star
ground at high power , thus there s no more equipotential
grounding since the voltage induced in said wire will be seen
as being in serial with the input signal , increasing THD and noise
at high power.
Set apart this point the problems encountered by the OP
are typical of an erroneous grounding scheme.
mmm, just measured across the transformer 0v to the earth on the IEC and I get 34 volts (psu is +/-55v)
I've tried the rectifier to cap wires twisted- makes no difference.
I've tried the mounting block to rectifier wires twisted, makes no difference.
I haven't tried twisting the wires from IEC to traffo yet.
I've tried wiring the 0v from the caps to the mounting block from the top cap, from the bottom, from both a with a link between caps and from both without the link.- no difference.
Waggling either the input 10core IDC wire, or the one that runs from pre to right channel, anywhere in the case- huge difference in noise.
If I disconnect the IDC that runs from the input board to the pre I get no noise at all- I guess the pre stage mutes on sensing no input
I've tried the rectifier to cap wires twisted- makes no difference.
I've tried the mounting block to rectifier wires twisted, makes no difference.
I haven't tried twisting the wires from IEC to traffo yet.
I've tried wiring the 0v from the caps to the mounting block from the top cap, from the bottom, from both a with a link between caps and from both without the link.- no difference.
Waggling either the input 10core IDC wire, or the one that runs from pre to right channel, anywhere in the case- huge difference in noise.
If I disconnect the IDC that runs from the input board to the pre I get no noise at all- I guess the pre stage mutes on sensing no input
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.