Joe's Improved Grounding Scheme

laatsch55

Administrator,
Staff member
Joined
Jan 14, 2011
Messages
75,287
Location
Gillette, Wyo.
Tagline
Halfbiass...Electron Herder and Backass Woof
#1
The following is a copy of Joe's post on the subject.

Here is what I did to achieve this.
1. Removed the white 22AWG chassis ground wire that runs from the copper bus pad between the input RCA jacks to the White Oak PL14_20 PCB, pin 5R.
2. Removed the 16AWG bus wire jumper that Phase normally installs between this same copper bus pad and the right channel speaker ground binding post.
3. I left the 2 signal grounds in the signal twisted pairs that originate at this copper bus pad between the input RCA jacks that run to the White Oak PL14_20 pin 2R and 2L respectively.
4. I added an 18AWG tinned solid copper bus wire jumper between the right and left channel ground (black) speaker binding posts.
5. I added a #10 ring lug with 2 white 20AWG wired into the ring lug on one end and into the White Oak PL14_20 PCB pin 5R and 5L respectively. The #10 ring lug was attached under one of the large bulk capacitor ground bus bar screws that connect the 2 bulk caps together.
6. I changed R2L from 56 ohms to 2.7 ohms.

This makes the large ground bus bar between the 2 bulk caps into the Single Point Ground for the entire amplifier. This wiring scheme eliminates the common mode noise from the large speaker ground from entering and affecting the low level input and low level ground reference on the White Oak PL14_20 PCB (caused by the bus wire jumper that Phase Linear installed between the right speaker negative binding post and the copper pad at the input RCA jacks)

Hope this is clear what I did. I will try to document in a diagram later on.
 

laatsch55

Administrator,
Staff member
Joined
Jan 14, 2011
Messages
75,287
Location
Gillette, Wyo.
Tagline
Halfbiass...Electron Herder and Backass Woof
#2
More on the subject from mlucitt in response to a question by speakerman1;

It is hard to retrofit some of the early SS or tube gear because the engineering back then was all about safety, not so much about the sound quality and most homes had two-prong receptacles, not three. With a metal cabinet there was the possibility of a shock hazard when connecting multiple pieces of gear together.

An example, if the amplifier power wire is zip cord and the non polarized plug (like those found on Phase Linear gear) is connected to the wall, the normally hot wire is on the right of the receptacle, so the wire on the left is neutral and connected to common back at the electrical box and grounded to the amp chassis. If the preamp is wired the same way and you turn the power plug over, you have just connected the hot wire to the chassis of the amp through the grounded interconnects - boom. For this reason the chassis was not grounded with any type of two conductor power wire.

To use an IEC power cable or jack, you have to be careful grounding the chassis. The reason is that the amplifier DC common floats above the AC ground by several volts, you can measure it by using a three-prong extension cord on a two-prong power cord and check the AC voltage between the ground of the extension cord and the chassis. You have to use a systems approach to grounding the amp and everything it touches, it is not a simple exercise, that is why Lee and I left it alone. We both may go back and rewire the amps for proper three-wire grounding but that will take some serious thought to avoid inducing DC voltage back into the electrical panel, placing voltage on the chassis if the grounding lug is not at absolute ground, or creating a ground loop between the power wires and the interconnect shields. Not as simple as placing a ring connector from the green wire on the chassis, in some cases.

Mark

More after a question by Fairchild
Jim, That is our point exactly. With a two-prong, non-polarized cord, you could flip it either way. "Hot" could be on the left wire or the right wire as it goes into the amp. On the older gear those two wires typically go to a fuse on one wire and a switch on the other wire and then right into the transformer - no polarity.

With the IEC, you have to get it right. You want to switch the hot wire not the neutral. Because if you switch neutral the rest of the amp is still hot. If you open it up to make a measurement, you could get fried between the chassis (now grounded back to the electrical box) and the transformer, fuse, or any part of the amp wiring that connects back to the transformer primary where the 120 VAC is sitting. If fact, it is best to switch both legs of the input power coming in to the amp on a DPST switch rated above the mains voltage and 150% of the current. The other possibility is a short in the transformer, it will probably go to ground and make the chassis live, not good unless the chassis is properly grounded and fused.

So on my Corcom IEC jack the hot wire is brown, the neutral wire is blue, and the ground wire is green with a yellow stripe. I will install a DPST switch on the brown and the blue wires, put the 10 Amp line fuse on the brown wire, and then wire it into the transformer. The blue wire will come off the switch and right into the transformer. I will place .1 mF, 275 VAC, X-type capacitors across the poles of the switch to absorb the arcing in the switch. The green ground wire will be bolted to the chassis and wired to a parallel 35A bridge rectifier as a path for fault currents, a 10 Ohm, 5 Watt resistor to reduce loop currents, and a .1 mF 250 VDC capacitor to block RF interference; then on to the ground connection of the internal electronics.

Mark
 

laatsch55

Administrator,
Staff member
Joined
Jan 14, 2011
Messages
75,287
Location
Gillette, Wyo.
Tagline
Halfbiass...Electron Herder and Backass Woof
#3
And gepetto's wrap up,

Hi Lee and Larry
NEITHER side of the AC line should EVER be considered ground and should ALWAYS be isolated from chassis (it is a violation of just about every safety agency to do otherwise). The 3 prong IEC C13 type connector provides for the 2 isolated AC inputs plus a 3rd safety (non-isolated) ground. The safety ground purpose is to tie the chassis to the ground stake in your house which usually is driven into the earth near your circuit breaker box. In the event that the insulation of one of the isolated AC inputs breaks down and shorts to the chassis, it provides a path for the resulting current to this earth ground which in turn pops your circuit breaker and preserves human safety. Without the safety ground, this AC isolation breakdown would end up energizing the chassis and presenting a human shock hazard.

With that said, using the third wire safety ground to tie to the chassis in your Phase Linear is asking for ground loop problems which usually always shows up as hum. This is because the DC ground in the Phase Linear is also tied directly to the chassis. To be successful in adding the IEC safety ground attachment to your Phase Linear chassis, you will have to isolate the amp DC ground from the chassis at the multiple tie points that are in the original design (I think there are 2 of them, the middle leg of the two small 3 terminal barrier strips on the amp back wall). Since I have never tried this, I do not know what "other" noise problems will result from isolating the Phase Linear chassis from DC ground and I would have to resort to the "your mileage may vary" phrase.

I don't know if this is helpful at all or not.
 

laatsch55

Administrator,
Staff member
Joined
Jan 14, 2011
Messages
75,287
Location
Gillette, Wyo.
Tagline
Halfbiass...Electron Herder and Backass Woof
#4
More from Joe(Gepetto);

All Phase Linear boards that I have run across have 2 different grounds entering the board near the bottom. Phase did not use a different ground symbol on their schematics which makes it confusing. There is a signal ground and a chassis ground. The signal ground is GENERALLY a black wire that originates from the ground connecting the 2 RCA input jacks, this also connects to the chassis near the RCA jacks. In the PL400 this ground is GENERALLY delivered to the preamp board using twisted pair wiring. In the PL700B it is GENERALLY via shielded coax cable. All the sensitive low level analog signal grounds on the board are connected to this ground node. These ground nodes are kept separate on the board one for the right and one for the left channel. The signal ground comes in via pin 2 on the bottom edge of each channel at the bottom of the PCB. Connections to this ground are: R3, C6, Q3-C for each channel.
The other ground is chassis ground and GENERALLY is connected to the PCB via a white wire that originates at the same chassis connection near the RCA jack connection. On Phase Linear boards both channels are connected to this same ground wire, the traces are combined into one for this ground type. The power bypass caps C3 on each channel of the PCB are connected to this ground as well as C10 on each channel as well as the cathode of D11 and anode of D12 (these are the diode clamps used in the output current limit protection circuitry). There is a resistive connection between signal ground and chassis ground on the PCB. The left channel uses a 56 ohm resistor to connect these 2 grounds and the right channel uses a 2.7 ohm resistor. I don't know why they are of a different value but they consistently are these values. The speaker returns GENERALLY go directly from the binding post negative back to the POWER ground node that is the copper bus bar between the 2 large bulk caps mounted on the transformer. There is GENERALLY a connection from this copper bus bar to the chassis ground via a small 3 lug terminal strip mounted to the aluminum chassis.
Phase Linear really does not have a true single point ground in any configurations I have seen. The aluminum chassis serves as this single point ground.

Hope this helps out rather than confuse. Ask any questions that you have. Thx.

More;
You will note on the White Oak board that the signal ground (pin 2) is paired with the signal input from the RCA jack (pin 1) and physically away from the other signals and connections. This separation is on purpose to keep the lowest signal levels on the board isolated from the larger signal areas. You will also notice that the input signal from pin 1 is a clean, straight run up to the top of the board with an adjacent signal ground run. All to keep the signal to noise ratio as good as possible.
 

laatsch55

Administrator,
Staff member
Joined
Jan 14, 2011
Messages
75,287
Location
Gillette, Wyo.
Tagline
Halfbiass...Electron Herder and Backass Woof
#5
More on the board grounding from Gepetto;

Without confusing this any further Lee, I'll mention that the White Oak board has the option to totally separate both Chassis (CH_GND) and Signal (GND) grounds on the PCB. The signal grounds coming in on pin 2L and pin 2R are ALWAYS totally separate. The Chassis grounds are joined by the 24AWB bus wire jumper that you installed on the back of the board between pad 5L1 and 5R1.

This was to allow flexibility to change the ground scheme used in the Phase Linear if desired. This option does not exist on the Phase Linear boards. I suspect that there is a better ground and power distribution scheme that can be configured within the Phase chassis and one day when I have some spare time, I intend to prove that The +75/100V rails and -75/100V rails are also totally independent on the White Oak board for this same reason. These are also conjoined by the other 2 bus wire jumpers (11L1 and 11R1 and 12L1 and 12L2)
 
Top