OPT grounding question

[Nihilist]

Is it necessary to connect the "negative" terminal of the OPT secondary side to chassis ground ? Seems as though the isolating properties of the OPT (isolating the load from the driver circuit) would be reduced or even eliminated.

Would it be OK to run the OPT's secondary connected only to the driver/load ?




Thanks for answering what will probably be a stupid question.




...............................Blake

 

[SY]

It depends. If you take feedback from the secondary then, yes, you need to ground the other end. If you're running open loop, no particular need.

 

[Geek]

If you're running open loop though, you'll still want a 1M resistor or so from one secondary leg to chassis to bleed the charge that develops from the inter capacitance of the windings.

Cheers!

 

[Nihilist]

I'm running open loop, triode tied 6V6's. FYI.


Geek said "If you're running open loop though, you'll still want a 1M resistor or so from one secondary leg to chassis to bleed the charge that develops from the inter capacitance of the windings."

Is this for safety reasons, as when handling the amp for servicing , or for performance reasons ?



Thanks guys...........................Blake

 

[JandG]

I am fairly sure if you try UL you need to slam the secondary ground to something. I supose UL is feedback of some sorts anyway. I remember some sqeal fests when doing some UL experiments when ground was not attached.
Joe

 

[Geek]

Geek said "If you're running open loop though, you'll still want a 1M resistor or so from one secondary leg to chassis to bleed the charge that develops from the inter capacitance of the windings."

Is this for safety reasons, as when handling the amp for servicing , or for performance reasons ?

Safety.

I've received a nasty shock from equipment without it, even ones with top of the line trannies.

Not even servicing... just speaker wire in one hand, chassis in the other.

Cheers!

 

[zigzagflux]

When running open loop, I've seem some designs that wire equal resistors from each end to ground, kind of like a balanced system.

 

[hemgjord]

Safety!

If for some reason you have a insulation breakdown between pri and sec you will in most applications have B+ at your speaker terminals.
I strongly recommend grounding the secondary, that way you will blow a fuse or something else before you blow your self, your children, pets etc.

/Olof

 

[Nihilist]

Thanks for all the input .


Olof,

Do you recommend grounding the secondary directly to the chassis, or with a resistor ?





............................Blake

 

[hemgjord]

I normally ground the opt directly to the central star ground point, no resistors to ensure that you trip the fuse in case of failure. I generally don't think it matters much exactly where you ground it since it is not low level circuitry. Others may think differently.

/Olof

 

[ray_moth]

I reckon Olof is 100% correct.

To drive speakers from an ungrounded source in which there is any possibility at all of B+ or the mains coming into contact with the speaker leads would be foolhardy and potentially dangerous.

The fact that the seondary of the OPT is insulated under normal conditions means nothing if the amp should ever be subjected to severe accidental physical damage due to impact or fire.

 

[Johan Potgieter]

OK, the inportant stuff has been said.

Just a further small point: Grounding the secondary gives a certain reference to the interwinding primary-secondary capacitance. Ungrounded the secondary floats somewhere between B+ and earth; even that can 'bite' you when touching loudspeaker wires.

Bottom line: While there is no technical 'demand' to ground the secondary in your case, there is nothing against it and every reason from several points of view to do it. (Direct, no resistor.)

 

[Wavebourn]

When 50W was a huge power amps had floating 100V output and a separate feedback winding.
Why?
Because they used to drive symmetrical lines through the whole building so grounding could cause undesired feedback on input lines.

 

[Johan Potgieter]

OK Wavebourn, admittedly a good reason!

I was thinking more of the domestic case, or at most a stage situation with relatively short cables. Yes, don't remind me. Running sound through a building can be a real nightmare!

 

[Nihilist]

Thanks again guys. I can see the safety point of it clearly now.

I was concerned that the back EMF of the speaker might somehow influence the smaller driver tubes by modulating the ground, forming a sort of undesired feedback circuit.

My thinking was that if I had the OPT floating (with regard to ground) , the OPT itself would act as a buffer/isolator between the driver and the rest of the circuit.

IIRC, sometimes industrial/commercial AC power transformers are case grounded, but windings are not .




..................................Blake

 

[Johan Potgieter]

I was concerned that the back EMF of the speaker might somehow influence the smaller driver tubes by modulating the ground, forming a sort of undesired feedback circuit.

Do recognise that grounding one side of a winding will not cause signals to be necessarily set up in the 'ground'. That will only happen when part of that output circuit is common to the rest of an amplifier. Thus it is important to ground at one point only so that no part of the OPT-loudspeaker circuit (carrying relatively high signal current) is common to the rest of the amplifier to any appreciable degree.

 

[rman]

The first amp I built from a diagram I found on the net did not show the secondaries of the opt's grounded So that was how I built it.
Then I built a slightly more powerful amp. I wanted to use the smaller one to drive rear speakers only in a Hafler matrix. Of course it didn't work because there was no complete circuit. I was confused at first, but I remembered seeing diagrams with one side of the secondaries grounded, made the change, and it worked. I realize now I could just have connected the black terminals on my amp together with the same result. After reading this thread and the safety aspect. I sure am glad I grounded those suckers!

Rolf.