Question; AC induction into DC, and Vice-versa

[laatsch55]

OK, standard wiring practice states you do not run AC and DC in parallel, you are supposed to cross at right angles and be shielded and make sure the DC is twisted.
When looking at the transistor wall of a 700, all of these tenets are violated. Agee?

The collector bus bars arecarrying 100vdc+,-, while the base bus bars are unshielded, less than 1/4 inch apart and as parallel as you can get. Not to mention the emitters being ganged less than an inch away??

If standard wiring practice does not apply here----WHY??

 

[Skratch]

The 400 is not as bad as the 700, but I agree that they are not wired correctly. COST driven

 

[laatsch55]

How much induction occurs do you think and does it go AC to DC or the other way or both ways, and can I check and measure that?? Because if there is some induction going on then I have some SERIOUSLY ANAL things I'm gonna do to my own 700.

 

[mlucitt]

I always thought AC wires are to be twisted and DC wires are just sitting at potential. I also thought that as long as the signal wires are shielded (inputs) then it doesn't matter where they go or what they lay next to (shorter is better for signal loss though). The DC has some Electromotive Force (EMF) surrounding each wire but how much depends on the current level. I think the current levels on the transistor wall are fairly low. There is some magnetism associated with DC induction in coils and loops but as you point out - these are straight runs in the PL amps. With the copper wires, the aluminum chassis, and the tin and lead in the connections, there is not much of a chance for the magnetic fields to form, except around the transformer (and half of that is outside the chassis). Ground wire routing doesn't seem to make much of a difference.

People have been staring at that transistor wall for 40 years and it has not changed much save some ferrite beads and some capacitors on some units. I would not go crazy or chase superstitions. If you want to run a test with your fancy THD and S/N machine after making one change at a time, that might be a useful exercise. But I bet you get bored after awhile and seeing that your changes do not make much of a difference. But hey, what do I know? Joe?

 

[laatsch55]

It is believed by Dean and Ed that the wiring is responsible for parasitic oscillations. The 4uf and .33 uf caps strip off RFI and EMI that accumulates due to lenghth (and possibly theWAY) the output wiring is. If there was no induction these components would not be needed.

On another note. In the early 80's there was an analyzer that could measure signal to noise ratio down to -150db. Ed's immediate question was-"What does it have for transistors", because to measure that wouldn't you need something just as quiet ??????????? Or is that figure arrived at by extrapolation, or some logarithmic code???

 

[mlucitt]

-150db is the sound level of a gnat landing on a flower in a soft breeze in California measured from Wyoming.

There should be no external Radio Frequency Interference (RFI) (radio frequencies from 500KHz to 6GHz) inside the chassis (aluminum enclosure) if the chassis is properly grounded; and no internal RFI unless the amp is oscillating (and overheating in the process).

There will always be some ElectroMagnetic Interference (EMI) - rapidly changing electrical currents, not DC (DC makes no sound). The EMI is either emitted radiation from an external source or electromagnetic induction into the circuit. The sources of narrow-band and broadband emitted radiation are devices such as bug zappers, motor starting circuits, and leakage from microwave ovens. Not much we can do about those transient effects except shielding the sensitive wires and grounding the chassis. A characteristic of this noise is that it is stronger at lower frequencies and weaker at higher frequencies, although it can be modulated as in a television or computer switching power supply.

EMI from induction is caused when mutual interference between two (2) conductors generates an electromagnetic field that results in EMI. There will be some 60Hz (and harmonics of 60 Hz) EMI from the AC in the transformer but the distance from the signal wires helps to reduce the effect. Switching loads above 2 Amps is a typical source of induced electromagnetic interference, but that is not common in the PL amps. Integrated circuits can cause EMI, but generally there is a bypass capacitor near the IC to act as a snubber, but the snubber is not very effective over 2 Amps when electrical contacts are used. Twisting the wires can help cancel the effect of induced EMI but an EMI filter in the frequency band of concern is the only real way to suppress the EMI.

I agree with Ed and Dean, the high currents present in the output wires can cause EMI in the form of distortion or crosstalk and separating the wires physically may help.

 

[kevin]

MMMMaaaann ! Thats some heavy STUFF YALL TALKING RIGHT THERE !!!!! WOW !!!

 

[laatsch55]

A radio man fromn the Navy could probably answer while sound asleep in response to the Master chief's questions of that nature Kevin. When them Navy boys learn something--THEY F&^%$#% LEARN IT!!!!

 

[Gepetto]

No doubt that the current backwall could benefit from some optimization. It is not all that bad being point to point and most connections are well separated spatially. I wish that the AC to the thermal disconnect was not routed anywhere in the backwall area but it was simple and cheap for them to do this. If you aren't pushing your amp to the limits, a simple recommendation is to get the sensor and the wiring out of the way (with AC power off of course). Other benefits can be achieved by looking at the wire harness that is usually neatly tywrapped together and separating problematic wiring. For example, I always separate the low voltage AC wiring from the rest of the bundle and fly those 2 wires as a tight pair on thier own from the source up to the light board. Whenever you can tightly pair a DC rail distribution wire with its ground return you will lower the inductance effects of these distribution wires. Keeping the wiring from the RCA jacks to the control board well separated from other signal and power wires and tightly coupled to the adjacent signal ground wire. Lots of little things basically.

 

[laatsch55]

Ok , you've given me the impetus to be totally anal about this 700 Series I. I've been thinking on this for quite awhle. Ran it by Ed yesterday, and after one of those pregnant silences, i really thought we had been cut off, he said,as calmly as coulds be, " Have you thought of building a custom set of COPPER heatsinks?" Ed can be a very funny guy at times, well most of the time, and he was serious.

 

[mlucitt]

Lee is looking for mad amounts of copper.

 

[Gepetto]

Remember that you will quite effectively cancel the inductance of a DC power run by tightly pairing it with its respective ground return. So for example the positive 100V feed that runs from the bulk caps on the left side of the amp across the top of the amp towards the right hand side of the amp to supply B+ to the transistor output stacks (collectors of Q11, 13, 13, 17, etc.) should be tightly paired with a ground wire of same gauge and the right hand end terminated with a low ESR film cap. This was the intent of the 0.33uF caps you see on the rightmost barrier strip. However Phase ran the ground for these bypass in a small gauge wire down the center of the amp rear wall. Do the same for the B- feed. The most effective ground to pair these with is the speaker negative terminals which are ground wires and are the primary wires responsible for returning the B+ current back to the single point ground. Don't twist them, just tightly pair them using tywraps, waxed string(for you military guys) or similar. And get rid of the AC running in the midst of the signal wiring.

 

[laatsch55]

There will be no AC outside the transformer well, no rail fuses, the transistor sockets encased in copper boxes, buss bars wired point to point with loops, the outs will leave the chassis at the point they attach to the rails, the hi volt DC will be run outside the chassis in a copper channel to where it's needed then enter the chassis, no Hi thermal shutdown at all-eliminated, the board will have pins and will plug into a hard mounted socket with all wiring as you state Mr. Joe. the boards will be cut down to include the right side ONLY and stood off the copper boxes with snap-plastic standoffs, the Signal in will be directly in from wherever the board sets, it will be in copper tube till it leaves the chassis, NO RCA plugs, the wire from the board will converted to a male RCA plug long enough to reach the pre-amp........and some other things.

 

[laatsch55]

Imagine, when you look inside and see matching mono boards stood off of copper boxes with no wires that you can see going in or out.

 

[laatsch55]

I SAID ANAL!!!!

 

[laatsch55]

The transformer well will be a capacitor well, with the transformer being separate and away from the chassis.

 

[speakerman1]

There will be no AC outside the transformer well, no rail fuses, the transistor sockets encased in copper boxes, buss bars wired point to point with loops, the outs will leave the chassis at the point they attach to the rails, the hi volt DC will be run outside the chassis in a copper channel to where it's needed then enter the chassis, no Hi thermal shutdown at all-eliminated, the board will have pins and will plug into a hard mounted socket with all wiring as you state Mr. Joe. the boards will be cut down to include the right side ONLY and stood off the copper boxes with snap-plastic standoffs, the Signal in will be directly in from wherever the board sets, it will be in copper tube till it leaves the chassis, NO RCA plugs, the wire from the board will converted to a male RCA plug long enough to reach the pre-amp........and some other things.

Just remember that the longer the RCA run the more chance of picking up RF.

Larry

 

[laatsch55]

Yep, absolutely, and I figure if you do away with what historically is a problematic connection , it's a good thing.

 

[laatsch55]

Lee is looking for mad amounts of copper.

That never ends, and being in BIG industrial country, some of the scrap is simply amazing.

 

[mlucitt]

Joe, you mentioned "low ESR .33uF film caps" on the terminations of the DC rails. Would that be two caps on each rail as they terminate on the transistor wall totem poles for a total of four caps or is one cap on each + and - rail at the furthest end sufficient?

What would be some examples of those type of caps that we could get from Mouser?

Mark