anatech said:
Chris,
I mention this solely in the spirit of clearing up any possible misunderstandings, and not with any intent to provoke or extend any arguments.
No R-core transformer I have seen appears to have any core "gap" as you suggest here, and I wonder if you are thinking instead of C-cores. All the R-cores I am familiar with are constructed in the same manner as toroids being 'wound' from a long thin strip of steel, but in this case the strip is tapered towards each end so that the final result is virtually circular in *cross-section* through any part of the core.
However, unlike toroids, instead of also being circular in 'plan' view like a donut, R-cores look more like a Race-course (hence the term 'R', I was led to believe, but this may not be so!) with two longer parallel straight sides and curved ends. I haven't destructively tested their constructions to be 100% certain, but from the dual-part plastic bobbins used, I assume that these cores are not split and then later joined (like C-cores are), but as always, I could be wrong.
Secondly, as I mentioned before, after discovering the sonic differences with alternative mains orientation, in order to decide which way to connect my transformers up *before installing them*, I do the check on the bench to avoid later re-wiring, and at that stage they are merely loaded with a resistance. Accordingly, no diodes or anything else for that matter can affect this massive change in radiated fields which I can measure, albeit not quantitively. As I also said, the fields are actually worse when the transformers are loaded, in my experience, when compared with unloaded or 'freewheeling'.
Thirdly, I have never come across any readily-available 'standard' (generic) type of toroid which has an electrostatic shield, and most (unwitting) users will doubtless purchase such items because they mention this (and the manufacturer's/supplier's names) frequently on this Forum. Additionally, I look at numerous catalogues/websites (from all over the world) supplying these items, also some specialist audio components suppliers, and it is clear from the illustrations they show that these also do not include an electrostatic interwinding shield which is vital to reduce (regrettably, not eliminate entirely) capacitive coupling of mains hash etc.
For any such interwinding shield to be effective it must be grounded, and accordingly a suitable grounding wire is required, and one never sees these additional wires except for rarely when the transformers are deliberately made in this manner. I need to have my toroidal transformers bespoke made in the UK to enjoy these kinds of benefits, including a GOSS band to supposedly reduce overal emissions.
More recently, using R-cores has proved to be sonically preferable when directly compared with either "E-I" or toroidal transformers, and I tend to favour these now, although they are not so freely available, and certainly there appears to be none fabricated in the UK unless my extensive enquiries have failed me. This is very likely to be a licensing/patents issue as far as my enquiries show.
Regards,
Bobken said:
Chris,
I mention this solely in the spirit of clearing up any possible misunderstandings, and not with any intent to provoke or extend any arguments.
No R-core transformer I have seen appears to have any core "gap" as you suggest here, and I wonder if you are thinking instead of C-cores. All the R-cores I am familiar with are constructed in the same manner as toroids being 'wound' from a long thin strip of steel, but in this case the strip is tapered towards each end so that the final result is virtually circular in *cross-section* through any part of the core.
However, unlike toroids, instead of also being circular in 'plan' view like a donut, R-cores look more like a Race-course (hence the term 'R', I was led to believe, but this may not be so!) with two longer parallel straight sides and curved ends. I haven't destructively tested their constructions to be 100% certain, but from the dual-part plastic bobbins used, I assume that these cores are not split and then later joined (like C-cores are), but as always, I could be wrong.
Secondly, as I mentioned before, after discovering the sonic differences with alternative mains orientation, in order to decide which way to connect my transformers up *before installing them*, I do the check on the bench to avoid later re-wiring, and at that stage they are merely loaded with a resistance. Accordingly, no diodes or anything else for that matter can affect this massive change in radiated fields which I can measure, albeit not quantitively. As I also said, the fields are actually worse when the transformers are loaded, in my experience, when compared with unloaded or 'freewheeling'.
Thirdly, I have never come across any readily-available 'standard' (generic) type of toroid which has an electrostatic shield, and most (unwitting) users will doubtless purchase such items because they mention this (and the manufacturer's/supplier's names) frequently on this Forum. Additionally, I look at numerous catalogues/websites (from all over the world) supplying these items, also some specialist audio components suppliers, and it is clear from the illustrations they show that these also do not include an electrostatic interwinding shield which is vital to reduce (regrettably, not eliminate entirely) capacitive coupling of mains hash etc.
For any such interwinding shield to be effective it must be grounded, and accordingly a suitable grounding wire is required, and one never sees these additional wires except for rarely when the transformers are deliberately made in this manner. I need to have my toroidal transformers bespoke made in the UK to enjoy these kinds of benefits, including a GOSS band to supposedly reduce overal emissions.
More recently, using R-cores has proved to be sonically preferable when directly compared with either "E-I" or toroidal transformers, and I tend to favour these now, although they are not so freely available, and certainly there appears to be none fabricated in the UK unless my extensive enquiries have failed me. This is very likely to be a licensing/patents issue as far as my enquiries show.
Regards,
Have you ever compared an R-core to a toroid with electrostatic shield? It is easy enough to have a toroid custom made with a shield.
Regarding R-core, I always assumed that "R" stood for round cross section, as that apppears to be the only fundamental difference I see over other types. This allows the bobbin to be spun around the core while winding, and also provides a more even flux density through the core.
Also, I was under the impression that the primary was wound on one bobbin, while the secondary was wound on the other side. I guess it doesn't have to be done this way, but it would seem that it would have to be done this way to reduce capacitive coupling between the primary and secondary, which is said to be an advantage of R-cores over toroids. Otherwise, a shield would be the alternative, and that could be done with a toroid.
It is well known in diy circles to orient the polarity of a transformer. The leakage capacitance of the windings to ground is different for each polarity. This can be easily measured by disconnecting the center tap from ground, connecting an ammeter between the center tap and the grounded chassis, and orienting the primary polarity for the lowest ground current.
Bonsai said:
Hi Bonsai
Having experimented briefly with this i found the effect audible but not earth-shattering. Of course, as it costs practically nothing, there is no excuse for a high quality system not to take advantage of it.
Possible "explanations"
1. As the effect of the dielectric does not stop at the cable boundaries it makes sense that surrounding materials will also have an effect. Apart from dielectric effects placing a cable on the floor will probably deform the electric field as well.
2. Cable acts as a mechanical antenna and feeds air borne vibrations to sensitive boxes, like preamps and sources. I can imagine this being a real issue with heavy duty interconnects feeding well isolated components.
3. Something to do with the electrostatic potential of the carpet. I honestly don't understand this one.
The speaker cable i mostly experimented with was an ultralight diy copper ribbon with very thin isulation suspended on air-filled bladders. I seriously doubt vibration feed was of any importance but having the carpet as part of the dielectric did not sound so good.
Sorry Scott, that incident happened over 20 years ago, and under Dave Wilson's direction. It was a surprise at the time, and I can't even remember what amps were used (probably Krell) and it just might have been a fluke. I am talking about TODAY. But as IC designers tend to starve their output stages, more and more, it looks like, and that is probably the problem, and it can't be easily fixed. I am spoiled, somewhat, in that I can NOW measure the harmonic series, easily with an FFT based analyzer. I couldn't back then.
However, the discrete design that I made for Dave Wilson, 20 years ago, works just fine, and doesn't have higher order harmonics to anywhere near the degree that the IC's have, and I have just compared the two within the last few days with the same test equipment, probes and power supplies included. I guess that is what is disappointing.
What is important here, is that I am comparing THD measurements between Analog Device IC's and a good discrete design, (like Dave Wilson's mike to line amp, or the Parasound JC-2 line amp) and getting poorer MEASURED results.
However, the discrete design that I made for Dave Wilson, 20 years ago, works just fine, and doesn't have higher order harmonics to anywhere near the degree that the IC's have, and I have just compared the two within the last few days with the same test equipment, probes and power supplies included. I guess that is what is disappointing.
What is important here, is that I am comparing THD measurements between Analog Device IC's and a good discrete design, (like Dave Wilson's mike to line amp, or the Parasound JC-2 line amp) and getting poorer MEASURED results.
Andrew,
While my morning caffeine has not quite kicked in, yet, the two may not be 100% linked. A stray field in itself means nothing if consideration as to how it is oriented is not taken. In other words, a stray field may couple to the circuitry directly, or inject current into the ground via coupling to the chassis. How the transformer is oriented may change the relative effect.
The measurement I spoke of is used to measure the ground current in the final position of the transformer in the chassis, and how it capacitively couples to the chassis. In that scenario, you are pretty much limited to turning the transformer over on the same bolt position, and reversing polarity. This would presumably be done after the location and orientation of the transformer in the chassis is determined with respect to the PCB based on minimizing the coupling of the stray field to the PCB.
While my morning caffeine has not quite kicked in, yet, the two may not be 100% linked. A stray field in itself means nothing if consideration as to how it is oriented is not taken. In other words, a stray field may couple to the circuitry directly, or inject current into the ground via coupling to the chassis. How the transformer is oriented may change the relative effect.
The measurement I spoke of is used to measure the ground current in the final position of the transformer in the chassis, and how it capacitively couples to the chassis. In that scenario, you are pretty much limited to turning the transformer over on the same bolt position, and reversing polarity. This would presumably be done after the location and orientation of the transformer in the chassis is determined with respect to the PCB based on minimizing the coupling of the stray field to the PCB.
Well Scott, in attempting to reduce the residual harmonic series in my Sound Tech this morning, I added an output cap across the gen out, and reduced the harmonic series considerably. I still have some nagging 7th harmonic that seems to come from the input stage of the analyzer. I might, in future, force the IC's on the front end to be class A. What do you think? PS I might do that to the oscillator too!
maxpou said:
Hi,
Regrettably I cannot speak from experience of this particular range of transformers, but I would be extremely surprised to learn that these are any different from a large range I have experienced from suppliers from all over the world.
I am not quite so familair with E-I types, if that is what this series consists of, as I haven't used E-Is for many years, but those I have tested were quite similar in exhibiting different fields (and sonic results) with alternate connections from the mains to their primaries.
Regards,
pooge said:
Hi,
I haven't ever had the chance to stand these two types of same VA transformers side by side at the same time, unfortunately, to make any direct comparisons. I only went over to using R-cores more recently, and the opportunity will arise at some time when I have one of each of same ratings together, so I will be able to compare at that time.
Also, as admitted before, my tests are not definitive nor calibrated (to satisfy the meter readers) sufficiently to accurately relate from one test to another which was carried-out some time later.
However, my overal impression is that R-cores are generally quieter than toroids, when compared like for like, if that helps.
You may be right about the origin of the 'R' in R-core, and I suggested that my impression might be wrong here.
I also believed that primaries were wound on one side and secondaries on the other for precisely the same reasons which you mention. However, as mentioned, I haven't (yet!) tried to take one apart or modify one in any way, and the origin of the wiring in those I am aware of was hidden completely under some insulating shielding, so I cannot be sure.
Regards,
AndrewT said:
Hi,
No, regrettably I cannot as I have not seen any reason to make this check before, assuming I fully-understand what is suggested here. I'm not sure right now if I have got my head around this, though, and I don't usually use centre-tapped transformers.
Given time, I will see what I can come up with and let you know.
Regards,
Thanks Chris, for the 'tip'. I contacted the 'right guy' and got the info I needed (mostly).
Bobken, I'm with you, all the way. I don't want to go away, or anything, but it sure gets old, sometimes.
SY, you shouldn't take things so personally, IF I did, I would feel REALLY insulted on this forum, on a continuous basis.
Bobken, I'm with you, all the way. I don't want to go away, or anything, but it sure gets old, sometimes.
SY, you shouldn't take things so personally, IF I did, I would feel REALLY insulted on this forum, on a continuous basis.
Bobken said:
I used the measurement at center tap because that is where most are grounded. However, you can measure from whatever point your transformer or power supply is grounded.
The effect I speak of is not necessarily related to stray external fields. The transformer core is usually grounded to the chassis. The windings of the transformer are capacitively coupled to the core. Any such coupling from the windings to the core finds a loop through the chassis and the ground back to the transformer, which noise current can be conductively coupled to signal ground.
If the windings are wound in, for example, series connected layers with the hot connected to the end of one layer and the neutral connected to the end of another layer, the coupling to the core would be different than if the hot and neutral were reversed. One polarity will usually have less current in this loop. This is why a good transformer will be bifilar wound. This helps to ensure that the capacitance of each winding is the same, and help to cancel these effects in the loop.
- Status
- Not open for further replies.