Some nice inductor shielding pics at: Selection Guide for Power Inductors in Consideration of Leakage Flux | Inductors (Coils) | TDK Product Center
I wouldn't use molded metalic types in audio filters at all. Also a resin shielded type brings the same reservations. A tiny air gap between drum core and a shield core helps with linearity, so I would rather chose CLF-NI type given in this example, if I had to...
It returns to my question: Do we need shielded inductors? There are good for power supplies. In our example, when using multiple coils, I would use linear ferrite cores coils and stack components in line with a leakage magnetic field. It would increase a summary inductance a bit which is a bonus. If external fields are concern, there is an old-style metal cage as a solution around inductors section.
It returns to my question: Do we need shielded inductors? There are good for power supplies. In our example, when using multiple coils, I would use linear ferrite cores coils and stack components in line with a leakage magnetic field. It would increase a summary inductance a bit which is a bonus. If external fields are concern, there is an old-style metal cage as a solution around inductors section.
Last edited:
Would you give an example - the particular parts you'd use and how you'd control and design for the mutual inductances? I'd be interested in seeing a practical solution as I discarded unshielded (cheaper) inductors years ago.
@pelopidas - things seem OK where I'm at but that's definitely not normal as there are official-looking people checking temperatures and in my case they told us we need to stay indoors for 7 days as my wife recently returned from another city which is now under lockdown. Sounds draconian but actually from my perspective reasonable that they're taking such precautions and me and Mrs were staying indoors anyway, just going out for a short while every couple of days to buy fruit & veg. Now we'll rely on delivery guys to bring those things I guess.
Business is definitely in hibernation though I wasn't seeing any major shortages at the local supermarket except bananas got sold out before we got there! The original return to work date of Feb 2nd has been put back a couple of times, think its Feb 10th now but I would guess that's also going to slip again. As regards international shipping the agents handling that aren't back at work yet which means while internal (domestic) shipping has restarted, internationally its going to take longer.
Business is definitely in hibernation though I wasn't seeing any major shortages at the local supermarket except bananas got sold out before we got there! The original return to work date of Feb 2nd has been put back a couple of times, think its Feb 10th now but I would guess that's also going to slip again. As regards international shipping the agents handling that aren't back at work yet which means while internal (domestic) shipping has restarted, internationally its going to take longer.
Not sophisticated at all if all inductors Lx are connected in series and probably deviation of total inductance is minimal. It can be simulated in software.
It is counterproductive to shield individual components when to the external world it is visible as a one inductor with a total inductance of sum of Lx's. Now a lesson form PhiDAC is learnt that series of shielded MLF components give not the best sound. When a chain of inductors is replaced with a single coil, a shielded version becomes practical, it is obvious. However as I said before, not every shielding type is audiophile friendly and all modern molded methods that attempt to minimise a leakage are not. A design where there is a tiny air gap between drum core and a shield core helps with linearity, this one I would be looking for. This is a main point I wanted to say.
Last edited:
You missed my invitation to show an example? I'm aware that coupled inductors can be simulated in software, that's something that I've been doing for a long time now. The tricky bit is working out what the coupling (K value in LTSpice for example) is, based on geometry. That's what I'd appreciate you showing me how to do.
Did you miss also my reason for using multiple inductors? Its because that lowers the effective tolerance meaning that DIYers who build don't need to buy lots and select the most suitable one with an inductance meter.
So you're saying the one I've selected as the best bet at present isn't suitable? Based only on the fact that its got a molded shield? I do plan to try it, you might be right but only experiment will tell for sure.
That's the ideal kind I would agree but so far I've not found a suitable one with a close enough tolerance and usable Q value. I have lots of P14 cores here which fit the bill (the same type I'm using on GrossDAC) but they're rather large - there might be enough space though without going beyond 10*10cm. The disadvantage for builders is they won't be off the shelf, people will either have to buy from me or wind them themselves, assuming the cores and bobbins are available from Mouser.
I understand it fully, but I asked a question "why we need shielded inductors" in this particular case of multiple inductors. From a DIY perspective it is easy to build a shield around all of them (if it is needed after all). Of course the ultimate test would be to hear it.
Yip, in general shielded inductors are counterproductive in this scenario, I didn't focus on any particular type of design. Maybe situation would change if we found a great audio-grade shielded type in a tiny size 0805 at a low price, but there is no such things, I am sure.
There are two options:
1. In the existing PhiDAC design switch to the unshieided inductors and build two DIY cages around them (L&R channel). Test whether it brings improvements. It would be the best option for those who source components themself from Mouser and do not have an access to the LCR meter.
2. For those who chose to order kits from @abraxalito, there will be an increased cost of a carefully selected (or self made) single inductor with a required tolerance. I am willing to pay extra, but I am partially disabled, so a poor DIY.
Looks like you ignored my request for now a second time. There must be a reason for that - which could well be that your claims were pure bluster.
And I mentioned the reason - that I'd done experiments with unshielded and found the interactions between them meant the behaviour wasn't predictable enough to design with them. They are cheaper which is what initially attracted me to try them out - if I'd come to the conclusion that they were workable in practice, I'd already be using them. But then you came along and claimed (but significantly, did not show) that it wasn't difficult as it could be simulated. So I asked you to demonstrate to me how their behaviour could be successfully modeled, but to no avail. That information from you is still MIA hence without it I'll just continue on my own merry way.
The precise position of the shield relative to the unshielded inductors is certain to affect their mutual inductance - for example if it bows a bit in the middle taking it closer to the open ends of the bobbin cores. A bit of a big wildcard for DIYers, one I'd personally prefer to avoid by using shielded types. You're free to fork off your own design, based on my published schematics, but with unshielded inductors though.
I take it then that this design won't appeal to you personally, hopefully there will be others not put off by my choice of shielded inductors.
Option 1 I've said already isn't a practical one in my view.
Option 2 might be a goer depending on my results from the 1210 inductors I hope to order.
So I take it you're not in favour of option 1 yourself as you say you're willing to pay more? Is there anyone else in favour of option 1 out there?
And I mentioned the reason - that I'd done experiments with unshielded and found the interactions between them meant the behaviour wasn't predictable enough to design with them. They are cheaper which is what initially attracted me to try them out - if I'd come to the conclusion that they were workable in practice, I'd already be using them. But then you came along and claimed (but significantly, did not show) that it wasn't difficult as it could be simulated. So I asked you to demonstrate to me how their behaviour could be successfully modeled, but to no avail. That information from you is still MIA hence without it I'll just continue on my own merry way.
The precise position of the shield relative to the unshielded inductors is certain to affect their mutual inductance - for example if it bows a bit in the middle taking it closer to the open ends of the bobbin cores. A bit of a big wildcard for DIYers, one I'd personally prefer to avoid by using shielded types. You're free to fork off your own design, based on my published schematics, but with unshielded inductors though.
I take it then that this design won't appeal to you personally, hopefully there will be others not put off by my choice of shielded inductors.
Option 1 I've said already isn't a practical one in my view.
Option 2 might be a goer depending on my results from the 1210 inductors I hope to order.
So I take it you're not in favour of option 1 yourself as you say you're willing to pay more? Is there anyone else in favour of option 1 out there?
Hi @abraxalito. You are not my supervisor, so it will be no penalty, right?
And the same regarding yourself. You do what is the best for you. Now while our basic relationship is sorted out, there is always a common sense. We know for sure that multiple MLF's are poisonus to the sound. It was your own discovery and now you are asking for a proof from me. The same asking for a scientific method of simulation so I could to teach you how to do it, instead of a simple check by putting 15 inductors in place and do an ultimate test. Reality check, please. Sorry, I have to say things straight instead of ignoring, but I had been requested multiple times.
Why is this assumption? Everybody can try option 1, it is why I speak about. If I wanted to convince you to do it yourself, I would bring it in private.
If you want to know my personal preference, it would be purchasing a proper LRC meter and going back with LingDAC instead which has a higher order filter. I neded a meter already many times and I consider it a negligence on my side that I didn't do it so far. I am afraid that multiple 1210 inductors will bring a little effect, I would rather go straight away with a single quality inductor.
Hi @abraxalito. You are not my supervisor, so it will be no penalty, right?
I'm sure there is a penalty for BS-ing, I might call it karma. Others may give it another name like 'what goes around, comes around'. Me not being your supervisor doesn't affect karma.
Can't follow here - are you saying BS-ing on a public forum is the best for you?
I can only assume you're using 'we' like the Queen does. I don't 'know' this, I merely noted that I can measure HF distortion on the PhiDAC Quad when MLFs are used which doesn't show up when inductors with higher current rating are used. Whether its audible hasn't been determined yet but I prefer not to compromise the measurements if there's a workable alternative.
And I mentioned that I did do such experiments years ago and rejected the approach. So who here needs the reality check? It seems that despite multiple explanations in multiple posts you still haven't gotten the message.
Why would anyone want to try it? When there's a solution where the experiment has already been done and results shown (as I hope in the future)? Since you've only claimed it works but not shown any result, why would anyone want to take the risk of time wasted on such an approach?
There are posts on this thread saying that LingDAC sounds better that PhiDAC. There are never ending debates about inductors for audio, as it always matters. You made your own contribution in one of these debates saying that ferrites sound good as long there is no DC component. Despite of your own assesment you made a project in which DC has a value exceeding a maximum amplitude of audio signal, that is a hudge violation of your own rules. In this situation I don't have to prove anything, you already did. This project requires a quality inductors, oversized due to a DC, it is a critical part of the project. It would be wise to abandon these tiny sound poisoning multiple shielded inductors, but you say that no-one is going to follow my advise. That makes it end of the story, I have nothing more to say in this matter.
Probably for the best.
Normally I would correct the errors in your perception but since you've so far proven completely impervious to correction, I'll leave them. If anyone else feels @sajunky has a valid point which I've not addressed, do please post.
Strange noise question
Right now I am feeding the Phidac from an AK4118 Spdif to I2S converter. Both have clean 12v filtered with a LT3045. I play flac files via Foobar2000.
So, when no song is playing, the background is silent, nothing at all. But when I start up a song, there is split second before the music starts where there is a hissing sound, like tape hiss.
My question is, is this because the music was recorded on tape first, or is this a problem in the AK4118 or the dac?
Right now I am feeding the Phidac from an AK4118 Spdif to I2S converter. Both have clean 12v filtered with a LT3045. I play flac files via Foobar2000.
So, when no song is playing, the background is silent, nothing at all. But when I start up a song, there is split second before the music starts where there is a hissing sound, like tape hiss.
My question is, is this because the music was recorded on tape first, or is this a problem in the AK4118 or the dac?