solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Apr 19, 2017 12:15:40 GMT
The new pdf is up and the HD600 module is corrected.
The first version of the filter did not have the inductor and associated parts. I guess I forgot to add them in the BOM list.
With the parts in this new BOM list the filter should work as shown in the plot.
|
|
|
Post by tupisac on Apr 19, 2017 18:35:36 GMT
What about C4? Should I change them anyway? What was the correction curve without missing elements? Does it explain 1,8 gain?
I didn't really hear decrease in quality. Slightly reduced mids and highs, but I was explaining it with bigger bass covering them up.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Apr 19, 2017 21:32:08 GMT
I would change C4 + add the missing parts and change R1 to 560 Ohm as per the BOM list.
C4 may be a 10nF but as it is the wrong part anyway it could well be any value between 1pF and a few uF. Would not take chances and build the filter as it is supposed to be.
Without the missing parts the upper mids would be louder, the 1kHz would not be +5dB. More likely that is caused by C4 having a value >> 10nF
|
|
|
Post by tupisac on Apr 20, 2017 12:42:48 GMT
Managed to find such inductors: Only one shop in whole city had 100 mH - such value is considered exotic, like HV variant of LM317. 10% tolerance, 9x12 mm. I'll make them fit, but hope these are okay, because I don't want to pay for yet another shipment from Conrad.
|
|
|
Post by tupisac on Apr 20, 2017 14:59:00 GMT
Actually it looks pretty cool
|
|
|
Post by tupisac on Apr 20, 2017 15:02:42 GMT
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Apr 20, 2017 15:55:26 GMT
Yes, 100mH is about the largest value you can get. Id these are really 100mH and not 100uH then they are O.K. There seems to be no value written on them.
Conrad has the axial type inductors in case the 3kHz filter (the inductor) does not work. Check the FR with your meter, should drop slightly around 3-4 kHz opposite 1kHz.
How does it sound and is there still a level difference ?
|
|
|
Post by tupisac on Apr 20, 2017 21:13:15 GMT
Sorry, long day. Will measure things tomorrow.
Level seems all right. There is a slight difference to the other side - unfiltered seems like slightly louder. Might be highs, the new filter cuts them a bit. Strange, but I started to like the fudged filter - seems like it had mightier bass and less treble cut.
Is it possible to assemble the filter with just the sub-bass alteration? Like reaching to 100 Hz, or 1 kHz tops?
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Apr 21, 2017 5:07:03 GMT
Perhaps switch back the 470 Ohm for R1. Then see how you feel when switching. In that case it should be a tiny bit louder when listening to filtered.
The filter without the 3kHz filter section boosted the treble quite a lot. There should not be any treble roll-off. The upper treble is actually boosted by 4dB ! Only the 3kHz part (which some find too forward sounding) is cut by only 1.5dB. Takes away some of the 'shoutiness' some people claim it has. Also midbass is lowered just a tiny bit. You could build the first version I made for the HD600. That one only added subbass and a tiny amount of upper treble (air)
|
|
|
Post by tupisac on Apr 21, 2017 12:14:41 GMT
My definition of highs might be skewed - I have build-in roll-off and can barely hear above 14 kHz.
What I can hear with my wrecked hearing is that new filter makes the vocals less forward and removes some of the echoes and hisses. It makes the headphones slightly narrower, or less 'surroundy'. Some might describe it as warmer and smoother.
Now after hearing what fudged filter did I'm very curious about the inner workings and eager for experimentation. I'm very interested in your early filter and maybe some recipes for further personalization. I even started thinking that it would be fun to put something on a breadboard with variable resistors...
I've seen the "band", "high shelf" and "low shelf" descriptions in the schematics. Could you elaborate on that?
|
|
|
Post by tupisac on Apr 21, 2017 13:39:20 GMT
Is there an (moderately) easy way to generate FR plots for filtered vs. unfiltered (skipping headphones, just amp's output to input)? Which software should I look for?
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Apr 21, 2017 16:00:51 GMT
My definition of highs might be skewed - I have build-in roll-off and can barely hear above 14 kHz. What I can hear with my wrecked hearing is that new filter makes the vocals less forward and removes some of the echoes and hisses. It makes the headphones slightly narrower, or less 'surroundy'. Some might describe it as warmer and smoother. Now after hearing what fudged filter did I'm very curious about the inner workings and eager for experimentation. I'm very interested in your early filter and maybe some recipes for further personalization. I even started thinking that it would be fun to put something on a breadboard with variable resistors... I've seen the "band", "high shelf" and "low shelf" descriptions in the schematics. Could you elaborate on that? The early filter only had bass boost really and no treble boost so the one you have built now has more treble but a tiny bit less 'clarity/presence'. The filterboard has one 'notch' filter (a bit like an equalizer) which can either boost OR filter. The band is kind of limited and can roughly be set between 3kHz and 30kHz. The 'problem' is the amount of attenuation/boost and above all the bandwidth are kind of difficult to 'set' correctly to the headphones needs. Takes simulation software and measurement equipment to do this properly. There are also 2 high shelve and 2 low shelve filters which can be configured at almost any frequency. The filters are ONLY 6dB/octave and influence each other. The values need to be calculated depending on the notch filter used as the bandwidth of that notch determines the resistance values. For that reason making one with a few potmeters isn't really feasable as the gain would be all over the place. In other words... there is no 'generic' filter. Each filter is purposely designed to do a specific task and the result of measurements, listening tests, and simulation software (quite expensive software I use at work)
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Apr 21, 2017 16:04:11 GMT
Is there an (moderately) easy way to generate FR plots for filtered vs. unfiltered (skipping headphones, just amp's output to input)? Which software should I look for? Yes there is... and it is free. It is called REW. You need to register (free) at their website and then you can download it. www.roomeqwizard.com/You do need to 'set it up' correctly though as it is primarily intended for measuring/correcting speakers.
|
|
|
Post by tupisac on Apr 22, 2017 8:48:22 GMT
[...] In other words... there is no 'generic' filter. Each filter is purposely designed to do a specific task and the result of measurements, listening tests, and simulation software (quite expensive software I use at work) Well, you can judge by the FR plot that there is some advanced voodoo going on - especially when hump becomes dip in bass area of hd 600. I'm really glad I've kept fudged filter intact. When I get through REW manuals I'll share the results. It is a lot of reading tho
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Apr 22, 2017 9:07:24 GMT
I can easily simulate the filters and show what it does to the FR. Just not in the coming week.
You only need to do an FR sweep so really do not have to set a lot of parameters for that.
|
|