A question just popped into my noggin'. Should an audio transformer can/clamp/outer enclosure be grounded? I don't mean just the screen, but the Clamp around the lams and/or the "can" a lot of input transformers are potted into.
Should the audio transformer can be grounded?
- Thread starter Ethan
- Start date
Help Support GroupDIY Audio Forum:
Yes, the cans of the input transformer should be grounded. Input transformers are very sensitive to noise from all that magnetic, high frequency, 50/60 Hz crud floating around. Some input transformers have three cans..... one inside another...... to shield the transformer and its low level signals. Attaching the outer shield to earth allows the blocked crud to trickle away peacefully.
SSLtech
Well-known member
Difference between electromagnetic shielding and electrostatic sheilding...
Electrostatic interference seeks a path to ground. If you have a strong electromagnetic source nearby that you need to shield something from, you don't actually need to ground the shield. If you are shielding from an unknown, potentially RF. electrosatic type pickup, then no amount of shielding will be very effective without a path to ground.
Keith
Electrostatic interference seeks a path to ground. If you have a strong electromagnetic source nearby that you need to shield something from, you don't actually need to ground the shield. If you are shielding from an unknown, potentially RF. electrosatic type pickup, then no amount of shielding will be very effective without a path to ground.
Keith
G
Guest
Guest
[quote author="Ethan"]A. Should an audio transformer can/clamp/outer enclosure be grounded? [/quote]
Yes, you got electrostatic (Faraday) shield for a free.
Why not to do ?
xvlk
Yes, you got electrostatic (Faraday) shield for a free.
Why not to do ?
xvlk
The ones I've seen are usually grounded through the mounting........... When mounted on a board they seem to run a track to earth. I guess it depends on the designer's intention and what problems s/he encountered. Thanks for clarification re: magnetic fields.
SSLtech
Well-known member
xvlk is right. Why not do it? -It's a free electrostatic i'nsurance'. Like I said, it's not essential, if all you're seeking is electromagnetic 'insurance' (and Mu-Metal is knowm for its electromagnetic properties more than anything) but anything conductive which evelopes a conductor can be an electostatic shield, so why not use the opportunity?
Well, one possible thought... excess capacitance under exceptional circumstances, but that's primarily a theoretical objection as I see it. Most of the time it's a free insurance without penalty. Not always essential, but usually very worthwhile.
Keith
Well, one possible thought... excess capacitance under exceptional circumstances, but that's primarily a theoretical objection as I see it. Most of the time it's a free insurance without penalty. Not always essential, but usually very worthwhile.
Keith
Larrchild
Well-known member
I use sowters in my fairchild clones and they are mumetal with a brass mounting collar. That is to decouple them magnetically from a ferrous chassis but maintain a ground. So clamps to mumetal cans should also be nonferrous.
Lar
Lar
Cool thread.
Transformers have a lot of "things" inside that may or may not reduce hum and noise. Try some experiments with the leads and can grounds. never know which will be best.
"things" from the center and moving outwards:
1) core ground-usually a wire solderd directly to a lam, which is then brought out to a terminal connection that may have other grounded "things" attached to it. Or it may be soldered internally to the can.
2) shield between pri and sec- can either be foil that either has a lead soldered to it or just a wire pressing against it held by tape, or the shield can be a bunch of turns of copper wire that are un-terminated on one end, this is your typical UTC shield. The wire is brought out and tied to the core lam ground in most cases, so there is really no way to use the core ground and e-shield ground seperately in this case
Note that foil shields must not have the ends touching, as this could turn into a big shorted turn that the primary would see, which might make the transfomer useless.
3) outer foil shield like API-looks cool but virtually useless.
4) internal shielding cans- these can be real mu metal, or imitation stuff, like they use in UTC.
Triad.... UTC....Telefunken etc sometimes use copper and/or brass foil or thicker copper sheet metal inbetween mu cans for extra boogie woogie.
5) core brackets-not used much in line level stuff but sometimes seen on after market outputs, like Sowter. Used a lot on power outputs. Not the most reliable ground because of paint, metal oxidation, lam coatings, loose bolts, etc.
6) external case- nice steel cans like UTC LS series, round mu cans like most stuff being made today, extruded rounded rectangular steel like old Triad and Peerless...usually a wire is soldered to the can directly or soldered to a lug that had been spot welded to the can. So you have two possible paths to ground here, thru the case mounting hardware, or "terminal 11"...
the "earthy" stuff is most important at mic input levels, important at line in and innerstage levels, and not as important in output levels.
Sometimes the only way to have access to all of these earthing points is to wind your own transformer. Some x-former makers have seperate grounds for e-shields amd cans.
Transformers have a lot of "things" inside that may or may not reduce hum and noise. Try some experiments with the leads and can grounds. never know which will be best.
"things" from the center and moving outwards:
1) core ground-usually a wire solderd directly to a lam, which is then brought out to a terminal connection that may have other grounded "things" attached to it. Or it may be soldered internally to the can.
2) shield between pri and sec- can either be foil that either has a lead soldered to it or just a wire pressing against it held by tape, or the shield can be a bunch of turns of copper wire that are un-terminated on one end, this is your typical UTC shield. The wire is brought out and tied to the core lam ground in most cases, so there is really no way to use the core ground and e-shield ground seperately in this case
Note that foil shields must not have the ends touching, as this could turn into a big shorted turn that the primary would see, which might make the transfomer useless.
3) outer foil shield like API-looks cool but virtually useless.
4) internal shielding cans- these can be real mu metal, or imitation stuff, like they use in UTC.
Triad.... UTC....Telefunken etc sometimes use copper and/or brass foil or thicker copper sheet metal inbetween mu cans for extra boogie woogie.
5) core brackets-not used much in line level stuff but sometimes seen on after market outputs, like Sowter. Used a lot on power outputs. Not the most reliable ground because of paint, metal oxidation, lam coatings, loose bolts, etc.
6) external case- nice steel cans like UTC LS series, round mu cans like most stuff being made today, extruded rounded rectangular steel like old Triad and Peerless...usually a wire is soldered to the can directly or soldered to a lug that had been spot welded to the can. So you have two possible paths to ground here, thru the case mounting hardware, or "terminal 11"...
the "earthy" stuff is most important at mic input levels, important at line in and innerstage levels, and not as important in output levels.
Sometimes the only way to have access to all of these earthing points is to wind your own transformer. Some x-former makers have seperate grounds for e-shields amd cans.
Brian Roth
Well-known member
Just my 2 cents....
If you elect to tie a can/shell/electrostatic shield to ground, do NOT use the audio "0V" line. Tie that can/shield to chassis ground via the shortest possible path.
Reason: the "0V" rail is often at a relatively high impedance at RF frequencies. For the same reason, XLR pin 1 should hit chassis as soon as possible with a very short lead and NEVER directly mingle with audio "0V".
This also can minimize "pin 1 buzz" problems.
Bri
If you elect to tie a can/shell/electrostatic shield to ground, do NOT use the audio "0V" line. Tie that can/shield to chassis ground via the shortest possible path.
Reason: the "0V" rail is often at a relatively high impedance at RF frequencies. For the same reason, XLR pin 1 should hit chassis as soon as possible with a very short lead and NEVER directly mingle with audio "0V".
This also can minimize "pin 1 buzz" problems.
Bri
Re: soldering to lamination. Somebody mentioned that one to me once. I don't get it. My understanding is that the laminations have a layer of oxide to insulate each lam from the other. Otherwise the lam would appear.... electrically..... as a single block....... which would screw up high freq performance.
mikka, your right. it is confusing to me also.
Your typical M6 silicon steel that is used in power applications and some audio outputs is insulated with a coating.
On the Ni lams, I just checked some from Mag Met and they read zero ohms on the surface.
.
.
.
Usually your silicon steel lams get shorted at the ends due to either the burrs from stamping, or some jackass pounding on them too hard with a hammer and core tool. (that would be me about twenty years ago).
But as long as they are insulated from each other on the main plane, it does not seem to matter if the ends are shorted. You can check your typical power transformer by scraping through the varnish with your ohm meter probes. Some will show a short, some will not.
The explanation given to me by a transformer engineer at the old place was that the current does not have time to travel all the way to the ends of the laminations to connect with each other. So that even if the ends are shorted, the core still resembles that of insulated laminations.
On the nickel thing, all I can think of is that when they are wax impregnated, a thin layer of wax gets in between each lam, and thus insulates each lam from the other. But some transformers are not impregnated, like the Tab V76 input iron that I bought for a DIY mic pre.
I will see if I can find out more on this.
cj
Your typical M6 silicon steel that is used in power applications and some audio outputs is insulated with a coating.
On the Ni lams, I just checked some from Mag Met and they read zero ohms on the surface.
.
.
.
Usually your silicon steel lams get shorted at the ends due to either the burrs from stamping, or some jackass pounding on them too hard with a hammer and core tool. (that would be me about twenty years ago).
But as long as they are insulated from each other on the main plane, it does not seem to matter if the ends are shorted. You can check your typical power transformer by scraping through the varnish with your ohm meter probes. Some will show a short, some will not.
The explanation given to me by a transformer engineer at the old place was that the current does not have time to travel all the way to the ends of the laminations to connect with each other. So that even if the ends are shorted, the core still resembles that of insulated laminations.
On the nickel thing, all I can think of is that when they are wax impregnated, a thin layer of wax gets in between each lam, and thus insulates each lam from the other. But some transformers are not impregnated, like the Tab V76 input iron that I bought for a DIY mic pre.
I will see if I can find out more on this.
cj
Hey thanks CJ. I seem to remember checking Grain Oriented Lams for resistance at the surface. Some showed continuity, some didn't. I thought wtf.... and put it down to manufacturing inconsistency. It all hurt my head too much so I just put it all together and it worked.
Re: soldering an earth lead to the lams....perhaps if a small electrical current was flowing in the lams....... causing noise....... but then there is the insulation issue. Dunno. ..... Whatever you can find out would be good to throw in the pot.
Thanks to Brian Roth also for those comments.
Re: soldering an earth lead to the lams....perhaps if a small electrical current was flowing in the lams....... causing noise....... but then there is the insulation issue. Dunno. ..... Whatever you can find out would be good to throw in the pot.
Thanks to Brian Roth also for those comments.
