No. See picture of a demounted NC500 core. Air gap visible in centre leg. The external legs enclose the coil no more than 50%.
I can imagine high frequency strayfield issues, but pure and correctly oriented toroids should not be worse.
Attachments
Sorry no. You seem to say that flux is "happening" in the coil and the core "encloses" it, but flux "happens" in the core.
The fact that toroidal transformers have low strayfields is because the core is "screened" by the coil (copper windings). Strayfield leakage in toroids is where terminal wires come off the core where it is not so well "screened".
Again you are saying what I said before. You may have misunderstood my suggestion that the coil screens the gap but once again you are stating the same in similar terms. We do actually agree with each other.
In terms of the coil screening the gap it is important to note that this is not an electrostatic effect. It is a magnetic effect. Stray H-Field concentration at the gap, the gap is treated as being distributed but it is not, is compensated by a reverse H-field in the winding. This is one of the reasons why losses in components carrying high AC currents experience higher copper losses specifically around the gap.
Maybe we are talking at odds with each other in respect of B and H.
The core acts as a short circuit to the B field so the windows carry little to no stray field. In magnetic, reluctance, terms the windows form a parallel circuit with the externally closed, short circuited, core.
The effective field is nominally, reduced in ratio to the effective permeability, Ue, of the core material which is something like 2,000 or more for this sort of ferrite. 300mT -> 150uT
In terms of the coil screening the gap it is important to note that this is not an electrostatic effect. It is a magnetic effect. Stray H-Field concentration at the gap, the gap is treated as being distributed but it is not, is compensated by a reverse H-field in the winding. This is one of the reasons why losses in components carrying high AC currents experience higher copper losses specifically around the gap.
Maybe we are talking at odds with each other in respect of B and H.
The core acts as a short circuit to the B field so the windows carry little to no stray field. In magnetic, reluctance, terms the windows form a parallel circuit with the externally closed, short circuited, core.
The effective field is nominally, reduced in ratio to the effective permeability, Ue, of the core material which is something like 2,000 or more for this sort of ferrite. 300mT -> 150uT
So how about using an air coil instead a pot cored coil wrt the "hysteresis distortion"?
Hard to get an air-cored coil of sufficient inductance without making it really big and susceptible to induced noise.
Please read former posts....no problem to make an air-cored coil of sufficient inductance; when necessary mumetal screened.
The important part of Bruno's solution appears to be post-inductor feedback. Getting that control system stable is where the real IP lies. By including the inductor in the feedback loop, all the nonlinearities of the inductor are attenuated by the loop gain.
I seem to recall Bruno/Philips having a patent on that, though, that could be close to expiry by now.
Tom
Yes. Including the filter inductor in the loop acts on distortion components that may result from hysteresis effects but I don't think the UCD patent was as much to do with post filter feedback as being that post filter feedback drops out of the equations.
The hard part is deriving those equations and then adapting them to work well. UCD appears to be very forgiving in that respect because the modulator gain is influenced by the load compensating for variations in that load.
In the case of Purifi the claim appears to be that the loop includes higher orders of integration and as a result loop gain which tames the issue{s} further.
The hard part is deriving those equations and then adapting them to work well. UCD appears to be very forgiving in that respect because the modulator gain is influenced by the load compensating for variations in that load.
In the case of Purifi the claim appears to be that the loop includes higher orders of integration and as a result loop gain which tames the issue{s} further.
The UCD innovation was to tame/control the feedback loop to allow the inductor to be part of the feedback loop (thus allowing the feedback to compensate away the effect of the inductor on output impedance. The purifi improvements are incremental, allowing for, as you write, higher orders of compensation.
Today I got the Neurochrome Universal Buffer.
The first impression is that it is smoother and the sound more "plastic" than with the EVAL1 buffer, or bypassed.
The noise floor is also lower (my test was not too scientific but it is still a measurement: I bypassed the autoformer on the tweeter of the B&C coaxial so that it has a 110Db sensitivity, then measured the noise with my umik microphone and, with gain matched, I get a measurement of the hiss which is almost 10Db lower).
With the gain at unity it sounds relaxed and with slightly understated dynamics – with the gain at 7.3 or 13 it is much more dynamic. I chose to keep 7.3. I matched the level with the DAC's control, so this may also be due to the DAC behaving in a different way at various attenuation levels, but I would consider this case unlikely.
The board is quite small - yes, we read the specifications on Tom's site, but in the flesh (well, in the silicon and copper and fiberglass) it really looks diminutive. But it punches strong.
Adding it to my amp was really easy, the only problem is that I burned one of my DIYinHK superregulators (do not ask) so I had to shuffle a few things in order to get power supplies for everything. Everything is still on a wooden board because the case has not arrived yet. But luckily with my speakers I barely use the very first watts.
Maybe (maybe) a buffer designed around the Weiss OP-2 discrete opamps can do better (on paper it has a -150Db SINAD or something like that), but ONE opamps costs almost like two Neurochrome Universal Buffers. For the quality, it is almost a bargain!
And now also two pics.
The first impression is that it is smoother and the sound more "plastic" than with the EVAL1 buffer, or bypassed.
The noise floor is also lower (my test was not too scientific but it is still a measurement: I bypassed the autoformer on the tweeter of the B&C coaxial so that it has a 110Db sensitivity, then measured the noise with my umik microphone and, with gain matched, I get a measurement of the hiss which is almost 10Db lower).
With the gain at unity it sounds relaxed and with slightly understated dynamics – with the gain at 7.3 or 13 it is much more dynamic. I chose to keep 7.3. I matched the level with the DAC's control, so this may also be due to the DAC behaving in a different way at various attenuation levels, but I would consider this case unlikely.
The board is quite small - yes, we read the specifications on Tom's site, but in the flesh (well, in the silicon and copper and fiberglass) it really looks diminutive. But it punches strong.
Adding it to my amp was really easy, the only problem is that I burned one of my DIYinHK superregulators (do not ask) so I had to shuffle a few things in order to get power supplies for everything. Everything is still on a wooden board because the case has not arrived yet. But luckily with my speakers I barely use the very first watts.
Maybe (maybe) a buffer designed around the Weiss OP-2 discrete opamps can do better (on paper it has a -150Db SINAD or something like that), but ONE opamps costs almost like two Neurochrome Universal Buffers. For the quality, it is almost a bargain!
And now also two pics.
Plastic as in plastic arts, in particular sculpture, not as in the synthetic material. I meant that the soundstage is very precise and natural, wide and deep, with very precise layers and everything has the right size and placement. As if sculpted.
Roberto,
Thanks for the review and nicely written. I'll be picking up a UB along with a few other niceties from Neurochrome fairly soon. Like you said, at the price being offered, the UB's performance is mind boggling. The Weiss OP-2 opamps are indeed quite nice, but I will say that a dac that I had listened to with the Weiss OP-2 in the output stage (Mivera Audio Purestream) was beaten without prejudice by the Matrix Audio X Sabre (MQA) and Matrix Audio Element X (the output stages are opamp based for these units). And a colleague of mine who had an Auralic Vega found the Purestream to be lacking in ultimate detail and PRAT.
At the end of the day, it’s about design execution and it appears to be well done here.
Best,
Anand.
Thanks for the review and nicely written. I'll be picking up a UB along with a few other niceties from Neurochrome fairly soon. Like you said, at the price being offered, the UB's performance is mind boggling. The Weiss OP-2 opamps are indeed quite nice, but I will say that a dac that I had listened to with the Weiss OP-2 in the output stage (Mivera Audio Purestream) was beaten without prejudice by the Matrix Audio X Sabre (MQA) and Matrix Audio Element X (the output stages are opamp based for these units). And a colleague of mine who had an Auralic Vega found the Purestream to be lacking in ultimate detail and PRAT.
At the end of the day, it’s about design execution and it appears to be well done here.
Best,
Anand.
Last edited: