I am building a simple EL84 based SE amp in the course of which I have experienced magnetic coupling between the power transformer and the output transformers. The transformers I have are unshrouded. Shrouded transformers of course look better but I expect they also have some purpose in reducing extraneous magnetic fields. However, shrouding generally only covers the winding bobbin. The laminations are still exposed and I would expect a lot, of not the majority, of extraneous flux to exit from the laminations. So, my question is, wwhat is the purpose of shrouding and how effective is it?
Cheers
ian
Cheers
ian
Place them 90 degrees apart and as far away as possible from each other. Shrouding is only for looks unless completely covered in mu metal.
Thanks guys. That confirms my initial thoughts.
There is limited space on the chassis so moving them further apart is not really an option. They are already at 90 degrees to each other. I do know that soft steel is quite a good screen for magnetic fields. Not as good as mu metal of course but I have used it in the past to kill a small extraneous field from a toroidal mains transformer. The chassis I am using is steel so I am wondering if that is acting as a magnetic conductor. I toyed with the idea of placing the mains transformer on the opposite side of the chassis from the output transformers but there is not really enough space to do this. In the meastime I have ordered a toroid mains transformer which will fit underneath the chassis.
Cheers
ian
There is limited space on the chassis so moving them further apart is not really an option. They are already at 90 degrees to each other. I do know that soft steel is quite a good screen for magnetic fields. Not as good as mu metal of course but I have used it in the past to kill a small extraneous field from a toroidal mains transformer. The chassis I am using is steel so I am wondering if that is acting as a magnetic conductor. I toyed with the idea of placing the mains transformer on the opposite side of the chassis from the output transformers but there is not really enough space to do this. In the meastime I have ordered a toroid mains transformer which will fit underneath the chassis.
Cheers
ian
When you refer to "shrouding" are you referring to a metal box around the xfmr or the big copper strap that is sometimes rapped around the whole xfmr (parallel to the winding turns but outside the E-I core). The copper strap prevents magnetic flux (shorted turn effect) from leaving the core, which could couple to other xfmrs or circuitry. It works in reverse too. Usually one sees these on power xfmrs since they are the worst offenders, but it should work on an OT also. I have seen some xfmrs on "E-fleece" for real cheap that had the copper strap installed wrong. (90 degrees) Pretty good proof it works when done right.
Take a piece of sheet steel, like the cover of something or other item, and make it touch the chassis as you hold it between the two transformers. if that reduces your hum, then shielding might be sufficient. As an example, some recent era high gain Marshall guitar amps have the chassis formed to act as a barrier around the output transformer under the chassis.
Steel of any kind is virtually useless as a magnetic shield at AF. If you had any sucess at all with steel then it was because of the electric field shielding it provided, not magnetic (or very possibly a change in the eddy current paths, though this is fairly unlilkely).
1) What kind of chassis? Steel or aluminum?
2) What kind of PTX? "Stand up" or "drop in"?
3) How severe is the problem?
Did you o'scope the output to see what the problem actually is? I did a project with a drop-in type PTX on a steel chassis, and got some mag coupling. The output looked like a slightly distorted version of the mains waveform, and at the mains frequency. If your noise looks more like a sawtooth at twice the mains frequency, then it's a ground loop problem. For this particular project, the level of the noise leaking into the OPT wasn't enough to cause any problems worth worrying about.
If you have a drop-in PTX, you could try grommets to add some separation between the lamms and chassis. You could also experiment with OPT placement. Some orientations may provide better isolation.
Thus, you mean that when all the valves are removed whilst the unit is powered, you ear a 50Hz hum?
I am not 100% sure of the correct term. I think they are sometimes called bell ends - like the blue part of this transformer:
https://edcorusa.com/gxse10-4-5k
Cheers
Ian
I am puzzled by this. Grain oriented silicon steel (GOOS) bands are habitually used around toroid transformers to reduce stray flux.
Cheers
Ian
As I mentioned earlier, it is a steel chassis.
Both the power and OP transformers are 'stand up' types.
The hum is clearly audible immediately you switch on before the valves have had time to warm up. It is worse in the channel whose OP transformer is nearest to the mains transformer.
I will try separating the mains transformer from the chassis.
Cheers
Ian
Have not done that particular test. However, I raised the mains transformer on a2.5mm thick aluminium plate and the hum reduced significantly. I then removed the mains transformer and placed it beside the chassis as far away as possible from the the OP transformers and the hum has almost completely disappeared.
Cheers
Ian
Wrong.
*Any* magnetic metal surrounding a transformer will more or less magnetically shield (wouldn't use the word "kill" in any case), including mu-metal but also all others, including humble steel.
The difference between mu-metal and others is only of degree, since it has some 8-10X the permeability of plain low grade steel but the working principle is exactly the same.
Plain steel can and will reduce magnetic interference appreciably, you will just have to use more of it.
Of course, a thin, 18 ga or less shroud made out of sheet iron will have a smaller effect, but mu-metal isn't a "magic" material, it still follows Physics Laws ...... and a thin mu-metal sheet metal shroud won't do much either.
But apparently this is just about steel endbells, which won't do much to shield magnetic fields since they are largely in the wrong place for that.
End bells or transformer shrouds are both acceptable usage around here, they provide mechanical protection for the windings and terminations and isolation from E fields.
I have had some luck reducing hum from power transformers by putting them in steel cans, (as MerlinB pointed out this is of limited effectiveness, but might get you a few dB) adding a copper belly band and by raising the transformer roughly 0.5 - 1.0cm off of the chassis on rubber bushings, particularly if a steel chassis was used.. (This last also helps with mechanical buzz coupling into the chassis if it is a problem.)
I have had some luck reducing hum from power transformers by putting them in steel cans, (as MerlinB pointed out this is of limited effectiveness, but might get you a few dB) adding a copper belly band and by raising the transformer roughly 0.5 - 1.0cm off of the chassis on rubber bushings, particularly if a steel chassis was used.. (This last also helps with mechanical buzz coupling into the chassis if it is a problem.)
Yes, that works by encouraging eddy currents to flow in the belly band, which weakens the field outside the band (at 50Hz the effect is fairly unimpressive. It's better at higher 'hash' frequencies). You generally have quite a thick band, fully encircling the transformer core. It's difficult to acheive the same effect with an EI transformer owing to its awkward shape. It also needs to be a continuous band, so simple steel 'bulkheads' inside the chassis will be of little use for magentic shielding (OK, in really 'knife-edge' scenarios even a small change in the chassis eddy currents might cure the hum, but that's a special case).
*sigh* Yes, thank you, Dr Pedantic. The point I was making was that sort of steel that any rational person might use in a practical amplifier has such a pathetic impact on magnetic shielding that it is basically useless. GOSS is as good as steel gets, but even that has to be a seamless band all around the transformer to be useful, and if it's an EI transformer then you can be sure the leakage flux is actually spewing out of the corners that you can't easily encircle.
Not sure how thin you mean, but 0.15mm thick mu-metal has a powerful effect on magnetic fields. The difference is often an easy 6dB in terms of hum picked up in other circuitry. Mu-metal has 100x the permeability of steel, not just 10x as you suggest
Last edited:
What is the present separation between PT and closest OT? Is the separation closest core to core, or are the windings closest? Perhaps a photo would be great.
When you say 90 deg rotated, do you mean each OT is horizontal mounted (as per the Edcor link), and the PT is vertical mounted (like the CXSE Series from the Edcor link) ? Or that the PT and OT are both horizontal mounted but rotated 90 deg.
The strongest stray flux is from the exposed winding bobbin area - perhaps best envisaged by thinking of the outer winding layers as a solenoid and recalling the flux distribution around a solenoid winding. Better coupling between two adjacent solenoids would be when they have the same axial orientation, and are effectively side by side.
Without any more appreciation of the situation, I'd be thinking the bell ends may well divert enough of the stray flux egressing the PT, and flux ingressing the OT windings, to be of noticeable benefit.
The PT heater winding(s) is usually the outermost winding on the former in old transformers - is that the case for your PT ? The outer winding would be more likely to cause stray flux.
When you say 90 deg rotated, do you mean each OT is horizontal mounted (as per the Edcor link), and the PT is vertical mounted (like the CXSE Series from the Edcor link) ? Or that the PT and OT are both horizontal mounted but rotated 90 deg.
The strongest stray flux is from the exposed winding bobbin area - perhaps best envisaged by thinking of the outer winding layers as a solenoid and recalling the flux distribution around a solenoid winding. Better coupling between two adjacent solenoids would be when they have the same axial orientation, and are effectively side by side.
Without any more appreciation of the situation, I'd be thinking the bell ends may well divert enough of the stray flux egressing the PT, and flux ingressing the OT windings, to be of noticeable benefit.
The PT heater winding(s) is usually the outermost winding on the former in old transformers - is that the case for your PT ? The outer winding would be more likely to cause stray flux.
Last edited:
"Dr Pedantic" must be a honorific distinction, if it means correcting conceptual errors and wrong blanket statements 😉
From K&J Magnetics, "Sellers of the World's strongest magnets" https://www.kjmagnetics.com/blog.asp?p=shielding-materials who "should" know what they are talking about, a few selected cut and paste tidbits:
and the final conclusion is:
Going to the source, those who do know, is better than reading 1000 Forums and picking little bits here and there .
FWIW I have *actually* developed usable shielding for a friend who made custom channel strips for Recording (think clones of classic Rupert Neve designs, etc.) and in one experiment, just for the heck of it, cut slices of plain galvanized steel pipes, the kind used for gas and cold water distribution 😱 and placed them around unshielded input mic transformers (600r>10k) .
Shorted the input winding and put a magnetic interference generator close enough (a power transformer driving a resistive load so it radiated a nice sinewave magnetic field) so we got a strong, easy to read hum signal.
By putting the "donut" around it, we could *measure* a hum voltage reduction between 10-15dB 😱 ... better than the 6dB achieved with your mu-metal shield.
Yes, wall thickness was way higher , around 1.5mm ... so what?
Yes, I know that for many "more exotic and expensive is better" ....... but .... me? ...... I go by the actual bench tested experiment results 🙂
- Status
- Not open for further replies.