Filters

All general audio posts go here.
User avatar
karatestu
Posts: 5945
Joined: Sun Jan 08, 2017 4:40 pm
Location: North Yorkshire
Has thanked: 1865 times
Been thanked: 1378 times
Great Britain

Filters

Unread post by karatestu »

The Doc was very certain in his view that all filters are bad for the music. Although we do not have the vast knowledge and experience (over many decades) that he had to come to this conclusion many of us here have witnessed the effects of removing them from our equipment and reported our findings. Just because we are not EE's , have only used or ears and not connected up an oscilloscope to produce some lovely graphs does not make our findings any less relevant to us or like minded individuals.

Mostly our exposure to this in a diy sense has been removing passive xovers (electrical filter) from speakers, doping them (adding mechanical filter) and removing wadding (removing mechanical filter). In a non diy setting people have preferred the NVA products over other brands which most likely had all the filters inside.

I have also witnessed the improvements that my ear has detected from removing amplifier based filters. I removed first an input stage high pass filter that is there to mostly block DC from proceeding equipment and then listened. Next i removed the zobel network on the output and listened. Lastly i removed the output inductor along with parallel damping resistor and listened. Yes, the Doc was right about filters and with super low capacitance nva speaker cable my amps did not audibly oscillate and self destruct.

It is easy to become engrossed and addicted to the removal of filters (yes me) and thoughts wander off to where else they can be removed or made less intrusive. NVA amps still have a low pass filter in the input stage with a corner frequency higher than most others i have seen but it really is a necessary evil to stop radio frequencies being amplified. It has been said to me that the lack of an input capacitor and use of passive volume stage means that this corner frequency changes as the volume is changed. I don't know about that as i am not well versed in electrical theory.

But The Doc extended his filter theory to the power supply also. The only place you will find a regulator in a piece of NVA equipment is in the phono stage. Even then they are as simple as you could find. LM3XX 3 pin regulators with voltage setting resistors and a 0.1uf cap on the input and 0.1uf on the output. Everything i have seen over the years told me no problem with the 0.1uf caps on the input. But 0.1uf on the output is a different matter. Usually you see much higher values. Doc must have chosen the value to be less intrusive as regards phase and keep it simple stupid is best.

Everything i had learned up to joining HFS (and Doc told me through PM's to forget everything i had learned so far) had involved (amongst other things) swapping out these "agricultural" regulators for more advanced, better measuring regulators. Indeed at the time the results were mind blowing to me but the more i think about it the more i realise at the time i was not listening for what i do know. It was all Naim modding and i was wanting more hifi i think - you know, detail at all costs, sweeter high's, tuneful bass and all that bollocks. I got lost very far up my own arse along the way somewhere.

Things have certainly changed for the better. I am using bits of NVA (amp boards, speaker drivers, speaker cables) with Richards design philosophy used as much as possible to make the rest. Richard is not here now to keep me on the straight and narrow (still miss the old grump) so it has become a rather lonely place in karatestu diy land. Instead of having someone who has done it before and can tell you not to bother doing a certain thing i now have to try all the things that come in to my head to find out for myself. Not a bad thing you might say but it takes a lot of time and chin scratching for someone with patchy electrical knowledge. I miss the fun i had with him and the comments like "i am surprised you haven't blown them up yet" or "don't be daft do it properly or not at all". I must admit i did try some things and not mentioned it on forum for fear of the reaction. But at least i tried it and came to my own conclusions.

So winter is coming and a quieter time at work in katartestu land. Thoughts turn to what diy i am going to do and not get Richard spinning in his grave. I must admit i do have regulators on the front end's of my power amps at the moment. Not sure they will stay or if they do what form they will stay in. Its not a quick easy job to take them out to compare to not having them. It will be done at some point but i will need new transformers as the voltage on the front ends are too high.

I think about my Naim CD3.5 and all the things i have done to it over the years. Yes, much of what i did was in NVA style before i even found HFS. That was lucky. Adding independent psu's for all three of the dac supply pins was a major mod which i loved. Then there was moving transformers away from the sensitive circuitry - another big musical win although i didn't think in "musical" terms at the time. Bypassing the players anologue stage was one of the things i did whilst a member of hfs. Nobody told me to do it but it was a wacking great 7 pole bessel filter comprising of hundreds of components with output buffer and other bells and whistles. Best thing i ever did was get rid of that and the associated psu's which had a tendency to breed like rabbits after nothing more than a look in the pfm diy room.

But some of the regulators i put in the CDP definitely not NVA style. Ok the ALWSR super regs are gone because they were powering the output stage but there are lots of other things i put in. My only defence against ripping them all out is the fact that this is digital and the Doc did say he knew little about digital electronics. I would be loathed to take it all out and start again but the thought continues to nag in my head.

What else to do ? The tweeters have a series capacitor as there is no getting away from it really. Feed a tweeter a full range signal and it wont last long. It also protects against any DC from the amps output. We all remember the Doc's thoughts on active xovers or you ought to do he said it often enough. So i will not go there apart from having a play with them in my PA system. How else can this high pass filter be done passively ? Well of course - the input capacitor and associated resistor to ground which is not in a nva amp by design could be inserted and so replace what the cap does before the tweeter. All sounds good in theory but you need the tweeters to have their own amplifier otherwise you will have no bass. Who knows if there is actually any improvement to be had and it may sound even worse - is there any point ? At least it would make the amp's low pass filter's corner frequency rock solid and not moving with the volume control.

Sorry it turned in to a bit of an essay which may be a load of :Bllocks:
DIY FREE ZONE

Ordinaryman
Posts: 475
Joined: Tue Apr 16, 2019 4:50 pm
Has thanked: 100 times
Been thanked: 77 times
Great Britain

Re: Filters

Unread post by Ordinaryman »

WOW!!..great post 8-) more?

User avatar
karatestu
Posts: 5945
Joined: Sun Jan 08, 2017 4:40 pm
Location: North Yorkshire
Has thanked: 1865 times
Been thanked: 1378 times
Great Britain

Re: Filters

Unread post by karatestu »

More you say ? Lets see what i forgot to put in that essay.

If you remove the series cap from before the tweeter and do the filtering before the amp then you have nothing between the output of the amp and the fragile tweeter. If the amp goes DC for any reason then there is nothing to protect it and you will be replacing tweeters as well as mid bass drivers. But at least you are still doing the filtering passively rather than actively with opamp's and psu's. Would the Doc agree ?

There might be insertion losses but then the cap before the tweeter may have that too. On the note of insertion losses (decrease in signal voltage due to adding a component) there is a padding resistor before the tweeter to equalise the levels between it and the mid bass. Say you wanted to get rid of that as well. To my tiny mind there would be only three ways to do it and it would require a dedicated tweeter amplifier.

Could do the attenuation by a potential divider before the power amp but that is just moving the resistor (and adding another) to a different location. It does remove everything between amp and loudspeaker though. Or you would need a dedicated volume pot / stepped attenuator in the passive pre for both the mid bass and tweeter - biamping still required but it would be nice to have independent control. Third option is to lower the gain of the power amplifier on the tweeter. This is a world of pain though and although it could be done with the swapping of one resistor in the feedback network, the amp could go unstable and turn into an oscillator very easily. Further work would have to be done which is beyond my pay grade. :grin:

The Doc may have relied on the high PSRR (power supply rejection ratio) of his circuits to enable him to get away with not having regulators or very simple ones as in the phono amps. However the circuits do respond very well to increasing the VA of the transformer which makes the power supplies regulation better and lower impedance. A higher impedance psu reacts more with the load therefore producing more ripple which we don't want.

There are ways to objectively improve even the basic regs as in the phono amps just by the substitution of one component. However i have only tried it with 5V supplies and it did work well there. With 20V we need to use a different component which i haven't tried. It lowers noise and output impedance.

Just need Lurcher to correct my mistakes but that is how i see things. A lot of this could be a solution looking for a problem :lol:
DIY FREE ZONE

User avatar
karatestu
Posts: 5945
Joined: Sun Jan 08, 2017 4:40 pm
Location: North Yorkshire
Has thanked: 1865 times
Been thanked: 1378 times
Great Britain

Re: Filters

Unread post by karatestu »

Found a quote from Doc.

"Motives for diy 1) get it on the cheap 2) overkill.

How true but he forgot one. 3) bespoke. If you buy ready made product you are stuck with a standard product (unless you modify it and destroy resale value) that is designed to work with a lot of different equipment. Take the example earlier about moving the tweeter filter capacitor to line level . That has to be done only for the tweeter amp. No manufacturer i know would make amps just for tweeters or just for mid bass use.

I intend to make up my own stepped attenuators using resistor values in a lower range and with less volume difference between steps. Only with diy could you get this. Volume pot for each driver -sure thing if you are doing it yourself. So you can make things work better with what you have got, rather than cover all bases.
DIY FREE ZONE

montechristo358
Posts: 278
Joined: Sat Mar 03, 2018 2:35 pm
Has thanked: 0
Been thanked: 19 times
Great Britain

Re: Filters

Unread post by montechristo358 »

Stu, assuming a stepped attenuator and amp per driver

Would you put the sa liine side or speaker side of the amp?
Bedroom Rig
PiCorePlayer Squeezebox clone, Allo Boss Dac, NVA AP10H with Stepped Attenuator upgrade and external PSU, Audioquest Nightowl Cans, NVA mini BMU

Office Rig
PicorepLayer Squezeebox clone with HifiBerry DIGI+, NVA SSP Digital, Arcam AV9. NVA SSC, NVA AP10P, NVA LS5, NVA Cubettes, NVA Mini BMU

Kitchen Rig
PicorepLayer Squezeebox clone with HifiBerry DAC+, NVA SSC, NVA P50, Project AMPBox, Sonance Ceiling Speakers

User avatar
karatestu
Posts: 5945
Joined: Sun Jan 08, 2017 4:40 pm
Location: North Yorkshire
Has thanked: 1865 times
Been thanked: 1378 times
Great Britain

Re: Filters

Unread post by karatestu »

montechristo358 wrote: Thu Sep 26, 2019 9:02 pm Stu, assuming a stepped attenuator and amp per driver

Would you put the sa liine side or speaker side of the amp?
Line side if it was me. I have thought about it a lot and it seems the ultimate way to tweak the sound in your room to taste. I guess it would work best with speakers like cubes and mine where the drivers are not matched for efficiency and there are no low pass filters. It just becomes a pain in the arse having to alter several volume controls . Six for the ultimate (3 way speakers) with balance control too or three if stereo controls were used.

You could argue that flat frequency response will not be achieved but then since when have most of us on here been too bothered about that.
DIY FREE ZONE

User avatar
karatestu
Posts: 5945
Joined: Sun Jan 08, 2017 4:40 pm
Location: North Yorkshire
Has thanked: 1865 times
Been thanked: 1378 times
Great Britain

Re: Filters

Unread post by karatestu »

I guess we are lucky that we are dealing with a simple 1st order high pass filter comprising just one capacitor. That will make it easier to recreate the filter in the amplifier at line level. Another plus of putting the high pass filter in the amp's input stage rather than just before the tweeter is that the tweeter amp will have an easier time if it does not have to amplify all that low frequency content.

All IMO of course.
DIY FREE ZONE

Lurcher300b
Posts: 933
Joined: Tue Jun 14, 2016 1:58 pm
Has thanked: 0
Been thanked: 16 times

Re: Filters

Unread post by Lurcher300b »

karatestu wrote: Fri Sep 27, 2019 8:15 am I guess we are lucky that we are dealing with a simple 1st order high pass filter comprising just one capacitor. That will make it easier to recreate the filter in the amplifier at line level. Another plus of putting the high pass filter in the amp's input stage rather than just before the tweeter is that the tweeter amp will have an easier time if it does not have to amplify all that low frequency content.

All IMO of course.

I would be careful, you have basically just created an active filter system. Much of what you write about "filters" is sort of correct, but the devil is always in the detail. The logical place to go with that approach is where a lot of us did on Audio Talk about 10 or so years ago, you end up with single ended amplifiers driving single driver full range speakers. Much as I liked and miss Richard, he always did have a habit of blurring opinions and facts, and his love for splitting the world into polar opposites often lead to him ignoring the complexity of the real that allowed him to view as dualities, systems that in reality were far more complex and nuanced.

However its a good path you are traveling, one that many of us have gone down, it will be interesting to see where it takes you.

Also remember, its not just DC that will kill a tweeter, any audio below its range will kill them or at the least massively add to the level of distortion they produce.

User avatar
karatestu
Posts: 5945
Joined: Sun Jan 08, 2017 4:40 pm
Location: North Yorkshire
Has thanked: 1865 times
Been thanked: 1378 times
Great Britain

Re: Filters

Unread post by karatestu »

Thanks Nick for your input much appreciated. I have no idea where it will take me.

I thought active filters were ones that required a power supply ?
DIY FREE ZONE

Lurcher300b
Posts: 933
Joined: Tue Jun 14, 2016 1:58 pm
Has thanked: 0
Been thanked: 16 times

Re: Filters

Unread post by Lurcher300b »

karatestu wrote: Fri Sep 27, 2019 12:16 pm Thanks Nick for your input much appreciated. I have no idea where it will take me.

I thought active filters were ones that required a power supply ?
Its all semantics again. If they are in their own own box with some buffers, or make them part of a feedback loop then its easy to call them an active filter. But put them in front of a power amp and claim they are not part of the power amp (which has a power supply and gain stages) then they magically are passive filters. In reality they are both the same. Just depends where you draw the lines. But strangely one is a good thing and the other evil.

Post Reply