3D "AIR" EQ - "Night EQ" PCB's Complete!

[fucanay]

[quote author="jdbakker"][quote author="khstudio"]Good idea but your NEVER going to fit 12 Lorlins acrossed the front... what else is there that's cost effective?[/quote]
Has anyone ever used the Grayhill 56-series (or other half-inch rotaries) ? It would appear that Purple Audio use 'em, so they might not be total crap. Digi-Key stock 'em (GH5601-ND). At US$10.61@1 / $8.48@25 they're not as cheap as the Lorlins, but not quites as expensive as the Elmas.

JDB.
[the ITT/Cannon half-inch switch (Digi CKN6049-ND) is somewhat cheaper][/quote]

I just upgraded my green pres to use the grayhills, but I'm having a hard time finding decent knobs for the 1/8 inch D shafts. They are really small and kind of a pain to solder the resistors to, but not impossible. I wouldn't want to have to pay $120 for switches on this though.

Matt

 

[clintrubber]

A quite poor solution to the rotary-size problem would be a 2RU case... Just mentioning it....

 

[Purusha]

I have good news for ya. You can squeeze all 12 lorlins on my front panel design :thumb:

I measured it so it's possible... if 12 pos is enough.

 

[Slenderchap]

Checked a few 470k rev log P16's yesterday.....

Track value 440k - 490k
Centre value - 45-53k

(centre is "by eye" as these pots didn't have a centre ident)

Colin
www.audiomaintenance.com

 

[khstudio]

[quote author="Slenderchap"]Checked a few 470k rev log P16's yesterday.....

Track value 440k - 490k
Centre value - 45-53k

(centre is "by eye" as these pots didn't have a centre ident)

Colin
www.audiomaintenance.com[/quote]

That center value looks VERY good... are you sure?
If so, they seem like excellent candidates.

 

[clintrubber]

[quote author="Purusha"]I have good news for ya. You can squeeze all 12 lorlins on my front panel design :thumb:

I measured it so it's possible... if 12 pos is enough.[/quote]
Hey that's good news.

Now for the next hurdles, that 12-pos is indeed a bit limited. Let's see if these could be taken as well:

Range is -19 .. +20 dB (based on the 5k62, 56k2 & 500k-pot). We want to have a neat 0dB mid-position so a 12-position rotary could do for instance (first simple approach replicating the range the pot gives) boost in (6) 3.33dB steps and cut in (5) 3.8 dB steps.
I think that s*cks, not ?

Alternative is to say that we never need the full range and go for instance for -10 .. + 12 dB in 2 dB steps.

Or some trickery with an added 'range'-switch (say a SPDT on the 56k2 or ...) which can give all kind of possibilities but loses the elegance of the design as it is now.

Hmmm.....


Regards,

Peter

 

[mick]

Am I missing something, didn't Kevin just use switches on his 'air band' only unit, that way you know only the bands your using are in, when you've got 6 bands if there not on the spot for '0' it will be a pain to set up from scratch... wouldn't it ? I don't know that's why I'm asking.

Mick

 

[clintrubber]

[quote author="mick"]Am I missing something, didn't Kevin just use switches on his 'air band' only unit, that way you know only the bands your using are in, when you've got 6 bands if there not on the spot for '0' it will be a pain to set up from scratch... wouldn't it ? I don't know that's why I'm asking.

Mick[/quote]
To prevent any eventual confusion: the rotary switches I've been mumbling about concern replacements for the pots, not any enabling or disabling of EQ-bands.

Reasons for wanting to use rotaries:
* 'proper' stereo-use of this EQ
* rev-logs would require 'tuning' to have a proper flat-band mid-setting

Problems with rotaries:
* economical models give (too) large stepsizes
* cost

Discussion:
* can we live with 2 or 3 dB stepsizes or is this an EQ that you rarely want to use for more than some +/- 10dB ?

Bye,

Peter

 

[Slenderchap]

F.Y.I.

Sample batch of 20 x 470k rev log P16's

Average value 457.7k
Centre point average value 52.7k (centre "by eye")

Colin
www.audiomaintenance.com

 

[clintrubber]

[quote author="Slenderchap"]F.Y.I.

Sample batch of 20 x 470k rev log P16's

Average value 457.7k
Centre point average value 52.7k (centre "by eye")

Colin
www.audiomaintenance.com[/quote]
Thanks for the stats. What's the spread ?

Regards,

Peter

 

[mick]

Yes sorry Peter I was talking more about pots and centre indents, but I believe we seem to be addressing the same point.

Mick

 

[fucanay]

Well, since I already have the small bear pots, I'm just going to use those and twist the knobs until it sounds good. It' won't be surgical EQ, but it'll be fine for my uses. If I get real ambitious, I may look for the 47K centers and make my knobs align it to center and live with the results. Personally, I was looking for a cheap EQ and if I get caught up in expensive pots it won't be that anymore.

Still interesting to hear how this all pans out though. There is always that one elusive part and it's cool to see how people get around it.

Matt

 

[khstudio]

[quote author="fucanay"]Well, since I already have the small bear pots, I'm just going to use those and twist the knobs until it sounds good. It' won't be surgical EQ, but it'll be fine for my uses. If I get real ambitious, I may look for the 47K centers and make my knobs align it to center and live with the results. Personally, I was looking for a cheap EQ and if I get caught up in expensive pots it won't be that anymore.

Still interesting to hear how this all pans out though. There is always that one elusive part and it's cool to see how people get around it.

Matt[/quote]

This is where I stand also... I've already bought 50+ Alphas from Small Bear.

look for the 47K centers and make my knobs align it to center

My thoughts too + keeps the cost of this Q VERY low.

 

[peterc]

Looking at the cct, there is a way of getting the centre detented pots to be flat, no matter what the tolerance.

My maths might not be correct, so if you know better, please chime in.

The gain of each stage is set by the combo of the 5k6 + the centre value of the 500k pot divided by 56k2 (R30), if this is an adder, right? So (5k6 + 47k)/56k2 = 0.94. Nearly a gain of 1. (this measurement actually gives a 1/2dB error)

That why the 47k centre position that Kevin measured was important.

Well, from the above sentence, we could vary the 5k6 resistor to compensate for the pot tolerance. So if one stuck a 10k trimmer in instead of the 5k6, one'd be able to trim each stage's gain to as close to 1 as dammit. Pull out the trimmer & put in the closest resistor value.

Ideally the centre position should measure 50k6 for a gain of 1.

So we could use Omegs' fine 470k Rlog pots & still be sure that the centre position would be flat. The gain & cut at the extreme ends of the pot would not be exactly the same but if you are boosting 15db with this wide Q cct, you'd better have some real headroom in your system!

Peter

 

[clintrubber]

[quote author="peterc"]Looking at the cct, there is a way of getting the centre detented pots to be flat, no matter what the tolerance.

Well, from the above sentence, we could vary the 5k6 resistor to compensate for the pot tolerance. [/quote]
Sure, as implied more or less clearly in the various previous discussions, this will work.
It's a 'calibration' in one spot, obviously no guarantees about the 'law' to the left & the right (the 'law-signatures' might be shifted w.r.t. each other; note that 'loggish laws are usually 'piece-wise lineair').
But it's a calibration in the most important spot (0 dB), so that's definitely nice.
The max boosts & cuts will also vary between bands though, but that won't be too big a problem.

So this might all be fine enough and/or hardly noticable. It's also about the best that could be done given the situation, so nice that that 'best reasonable effort' might most likely be enough :wink: :thumb:


BTW, I feel this whole circuit could have better used some other boost/cut scheme than the present method. For instance summing the bands 'volumepot-wise' could make the summing circuit less sensitive to the pot-value of each band. Most likely even totally insensitive to it. And if the circuit as is sounds nice I guess that 'converted' one will as well.

But that's all for another day; I'll be happily building this EQ as it is (thanks all you-know-who for making it happen).

Regards,

Peter

 

[PimD]

[quote author="peterc"]I used Kevins last schematic....... It may sound a bit general, but thats all I have!

I added
4 x22R resistors,
4 x 22uF/25v caps,
12 x .1uF caps (2 per IC) &
2 x 22uF/63v caps for the input.

Peter[/quote]

I had a look at the schematic and pcb overlay, but i couldn't find the 4 22uF/25V caps. where do they go, and wich type is it?
And also on the schematic it says 9x .1uF caps, adding 12 would make 21, i can only find 19.
Will be ordering parts soon, at musikding.de
can i use a Silver Mica for the 22pF (it's the only one they have)

Pim

 

[clintrubber]

[quote author="PimD"]can i use a Silver Mica for the 22pF (it's the only one they have)

Pim[/quote]
I didn't check the schematic, but how many are needed ? IIRIC these silver mica's are not cheap. The tiny small ceramic ones will be fine (black stripe: zero temp-co), PM me if you can't find these.

Bye,

Peter

 

[PimD]

I need only 2, so the costs won't be too big, (about €1) so not a real problem then, just wanted too know if could use them.

Pim

 

[dissonantstring]

hi peter c.
thanks for the explanation which seems very reasonable as a "calibration" procedure for the 0dB center point of each cut/boost pot. would it be reasonable to assume that it would be the same to measure the center detent of the pot and just calculate the difference in resistance needed to obtain the "ideal" 50K6 ohm for a gain of 1 - then replace the 5K6 with whatever your calculation result is?
for example:
i have a 470K rev.log pot with a center detent at 43.2K.
50K6 - 43.2K = 7K4 ohms resistance to obtain a gain of 1.

or would a trim pot be necessary in this case for reasons i'm just too "green" to see? :green:
and what happens if the center detent of the pot is more than 50K6?
i know there is probably too much discussion about this and we should really build and test before going off on a tangent, but i'm curious.
thanks.
-grant

 

[peterc]

i have a 470K rev.log pot with a center detent at 43.2K.

The total of this value + the series resistor (currently 5k6) should be 56k2. So in this case the 5k6 should be changed to 56k2 - 43k2 = 13k.

and what happens if the center detent of the pot is more than 50K6?

As long as the centre value is not more than 56k2, this method will work.


Peter