Quickie in a metal box / grounding

[mcandmar]

Probably a basic question but i cant decide on the best approach.  I am working on building an AC powered quickie in a metal enclosure and have bolted the ground line of the IEC socket directly to the chassis for safety.

My question is at what stage should the ground bus (i.e. HV power supply - rail, RCA input and output grounds) be tied to the chassis.   Is it best to do it at the RCA inputs, RCA outputs, or at the power supply stage?

I went back over the SEX wiring for reference and found its connected to the chassis in two places, at the very start of the HV supply at the - diode / first filter capacitor, and at the very end of the chain at the headphone ground.  I was under the impression multiple ground points could cause ground loops so should i follow that example, or what is considered common practice for grounding amplifiers?

Thanks,

Mark

[Paul Birkeland]

The ground at the power supply is a reference for the heater supply, which is out of the signal path.

For your project, ground everything in the audio/power supply to a terminal between the tube sockets.

What's your plan for filament supplies? I should post some pics of my line powered Quickie sometime.

[mcandmar]

In the SEX amp i was referring to the black wire running from terminal 2 to 13 grounding the HV supply, which is then linked over to the right tube. I presume the intention is to tie the supply to ground at the start of the DC chain.

So basically i should create a star ground and run the HV negative and RCA grounds all to one point connected to the chassis, that makes sense and should be fairly easy to achieve.

For the filament supply i built this board with two fully independent supplies using LM317s outputting 1.4v. Each section is fed from a small transformer with two independent 4.5v taps.  Its about as quite a supply as i could make, and seems to work fine.  In testing it looked like i could have got away with a single rectifier and branch off two supplies from there, but for the final design i went for two completely independent supplies from start to finish.

I would love to see the AC quickie you made, as well as a few other projects of yours.  I've found some pics on the internet of very impressive looking preamp you built, and the amp in your signature intrigues me too.  You really need to get a gallery going

[Paul Joppa]

Each channel has its own chassis ground point, which are 13 and 23 if I've read the manual correctly. They are connected together with a wire as well, which does indeed violate the rules. The usually-stated risks are eddy currents in the chassis plate causing a hum voltage difference between the two grounds, and a ground loop of the chassis and wire. The first is minimized by placing it far from the power transformer, and the second by placing the wire against the chassis to minimize the loop area.

A technically superior method would be a central single-point ground between 13 and 23, with wires to 13 and 23, but of which would then be isolated from the chassis. Nearly equivalent would be to simply isolate 13 or 23 from the chassis. For the relatively low gain of a power amp, we thought this was not necessary. If it were a phono preamp, I might have insisted on "doing it right".

[mcandmar]

Thanks guys, i ended up creating a ground bus from each tube to a central post bolted to the chassis and it all seems to work perfectly.

I do have an issue with the small transformer for the filament supplies radiating noise into the amp.  I tried mounting it in every orientation possible and moving it around the limited space inside the case and i cannot find a sweet spot to stop it radiating into the chokes or output transformers.  As soon as i hang it outside the case it all goes quite.  The unit i am using is the first model in this datasheet, the 6va 2x 4.5v http://www.spiratronics.com/data/4430.pdf, its just a generic chassis mount unit.

Question is, are certain types of transformers less prone to radiating noise than others?  For example if i were to try one of the small encapsulated PCB mount transformers would it be any better?   Or do you have any other tips or tricks to get around this issue as i cant see a way to make this work.

Thanks again,

Mark

[Paul Birkeland]

Question is, are certain types of transformers less prone to radiating noise than others?  For example if i were to try one of the small encapsulated PCB mount transformers would it be any better?   Or do you have any other tips or tricks to get around this issue as i cant see a way to make this work.

Hello Mark,

Yes, certain transformers will radiate less than others.  PJ can fill in the details more accurately than I can, but one of his recommendations is to look for a 240V transformer with twice the output voltage that you need, then run it on 120V.

When I built my line powered Quickie, I used two Seduction power transformers (PT-1) with a copper tape loop around the outside of the transformer.  You could try that on your transformer, just wrap around the outside so that the copper tape makes a loop that follows the windings.  This essentially creates a shorted turn for the magnetic field outside the core. 

If you have output transformers on your Quickie, you might have a look at the SEX amp and how the power transformer and signal transformers are laid out, then mimic that as best as possible.

-PB

[mcandmar]

Hello Mark,

Yes, certain transformers will radiate less than others.  PJ can fill in the details more accurately than I can, but one of his recommendations is to look for a 240V transformer with twice the output voltage that you need, then run it on 120V.

When I built my line powered Quickie, I used two Seduction power transformers (PT-1) with a copper tape loop around the outside of the transformer.  You could try that on your transformer, just wrap around the outside so that the copper tape makes a loop that follows the windings.  This essentially creates a shorted turn for the magnetic field outside the core. 

If you have output transformers on your Quickie, you might have a look at the SEX amp and how the power transformer and signal transformers are laid out, then mimic that as best as possible.

-PB

Thats an interesting trick running 240v transformers @ 120v but unfortunately i live in a 240v world so i dont see how i could do that.

I'll try wrapping the transformer as i have a roll of copper tape somewhere.  I had thought of boxing it off in its own compartment but placing sheets of alloy and steel around the transformer made little to no difference which i dont really understand. It seems abnormally powerful as within 3-4 inches of the chassis plate the choke on the other side of the plate picks it up to the point its audible on headphones.  Hence i am wondering if another transformer is what i really need.

I did copy the BH recipe fairly closely with the main transformer (toroid) on top with the plate chokes on either side like the sex amp, then directly under them i have a power supply choke on one side, and the filament transformer on the other.  The two output transformers are tucked away in the front of the chassis as they are tiny little things.  I also used the same fiber insulators with solder tabs and wire grounding the case of each one back to the chassis ground point.  My mistake was locating the filament transformer directly under a choke, i thought it would radiate outwards in a doughnut shape, not up and through the chassis plate.  I discovered if i lay the transformer on its side it greatly reduces the hum so it seems to work the opposite way than i thought.

I will experiment tomorrow with the copper tape and see if i get anywhere..

Cheers,

Mark

[Paul Joppa]

In the SEX amp, the plate choke and output transformer do interfere with each other to some extent. They are however carrying the same signal, so any mutual pickup is not easy to notice or measure. But a power transformer carries only hum, so it would naturally induce that into the signal-carrying plate choke.

Toroids have quite a different field from E/I laminations, so you can't count on substituting one for the other keeping the same magnetic orientation. For the same reason, you can't simply apply the flux ring (copper shield) concept without figuring out the magnetic fields involved - and making sure you don't create a shorted turn inside the loop!

Our own power transformers are designed with a low magnetic induction, whereas commercial units usually have the maximum magnetic induction possible. The reason is that this has a powerful effect on size and cost.

Magnetic shielding also requires some knowledge of magnetic properties. For example, plain steel has a permeability for small signals that is much much smaller than mu-metal. I can't quote a number but it's on the order of 100 times. So for the same shielding effectiveness the steel would have to be 100 times as thick. At high field strength though, mumetal will saturate and loose its shielding effectiveness long before this happens to steel.

You could for example get an external 240:120v transformer to power the system, using 240-v primary transformers running at 120v. Just a thought.

[Grainger49]

Mark,

I think the power supply being outboard, maybe several feet away would be a good idea.  I did that on my Seduction and FP 2 builds.

[mcandmar]

In the SEX amp, the plate choke and output transformer do interfere with each other to some extent. They are however carrying the same signal, so any mutual pickup is not easy to notice or measure. But a power transformer carries only hum, so it would naturally induce that into the signal-carrying plate choke.

Toroids have quite a different field from E/I laminations, so you can't count on substituting one for the other keeping the same magnetic orientation. For the same reason, you can't simply apply the flux ring (copper shield) concept without figuring out the magnetic fields involved - and making sure you don't create a shorted turn inside the loop!

Our own power transformers are designed with a low magnetic induction, whereas commercial units usually have the maximum magnetic induction possible. The reason is that this has a powerful effect on size and cost.

Magnetic shielding also requires some knowledge of magnetic properties. For example, plain steel has a permeability for small signals that is much much smaller than mu-metal. I can't quote a number but it's on the order of 100 times. So for the same shielding effectiveness the steel would have to be 100 times as thick. At high field strength though, mumetal will saturate and loose its shielding effectiveness long before this happens to steel.

You could for example get an external 240:120v transformer to power the system, using 240-v primary transformers running at 120v. Just a thought.

Hi Paul,

The toroid is the HV supply, its a dual 18v tap wired in series for 36vac.  That one is not the problem, its the small filament supply transformer (dual 4.5v) that is causing the issue.  I proved this by disconnecting the transformer and ruining the filaments from D cell batteries.

I had a thought last night of using the 36vac output from the toroid into the primary side of a transformer, something with roughly a 7:1 ratio to give me ~5vac out.  Am i right in thinking if i got a 230v to 32vac transformer that was roughly a 7:1 ratio, and then run it at 36v to 5v it would have less magnetic radiation?

[Paul Joppa]

mcandmar, you are exactly right.

[mcandmar]

Splendid, we have a plan...

Thanks again!

[mcandmar]

Made a bit more progress.  I tried a few wraps of copper tape which didn't really do all that much.  Did a bit of reading on Mu-Metal tape which is really interesting, but fantastically expensive.

I tried a single tap 24v transformer as a test powered from 40vac giving a little over 4v out which did indeed run quieter so buying a dual tap 24v transformer is still an option as i only need a tad over 3.5v DC to run the regulators.

I then tried moving the existing filament transformer to the front of the case where the output transformer was and that got it far enough away from the choke to prevent it being picked up.  Delighted with that result i rewired, re soldered, and drilled a few new holes to mount the transformer in its new home.

I then set the output transformer into the case below the choke and it started humming again so at that point i swore a lot and chucked it all back on the shelf and walked away in a huff.

With the output transformers orientated on their side the noise is very small so we are getting close to a workable solution. Having thought about it i'm not sure if its the chokes/output transformers reacting, or if the output transformer is picking up the filament transformer.  If its the filament transformer a 230/24v transformer at 40/4v might be enough to be a workable solution. More testing needed as its teasingly close.

[Grainger49]

Mark,

I think the power supply being outboard, maybe several feet away would be a good idea.  I did that on my Seduction and FP 2 builds.

[mcandmar]

I hear ya, but that woud be admitting defeat