Passive 6dB EQ

[ruffrecords]

Thanks Abbey. Useful input. Just to be clear, you are proposing for shelving EQs that the  frequency should be defined as where the boost/cut reaches 1/4 of the maximum in  dBs. Don't you think that would give misleadingly low values for most shelving EQs?

Cheers

Ian

 

[abbey road d enfer]

Thanks Abbey. Useful input. Just to be clear, you are proposing for shelving EQs that the  frequency should be defined as where the boost/cut reaches 1/4 of the maximum in  dBs. Don't you think that would give misleadingly low values for most shelving EQs?

Cheers

Ian

Very good comment!
In fact musicians think that the actual definition is misleading, because they hear their treble & bass controls acting on midrange.
Indeed, technicians/designers know that a 10k shelving EQ starts at about 1k, but we have to serve our customers.
In order to maintain consistency with usage, we would have to base the definition on 3/4th from max. Although in terms of operation, musicians would agree more with a definition based on "that's where I start to hear things changing".

 

[kambo]

hey Ian, what exactly is the difference between ur design and NYD design ?
kind a puzzled...
the only thing i noticed is recalculated resistor ladder  :

If you look closely at the NYD schematic, you will see the top half of the ladder is not always connected to the bottom half.

Cheers

Ian

ohhhh finally i c it
thanks Ian

 

[ruffrecords]

hey Ian, what exactly is the difference between ur design and NYD design ?
kind a puzzled...
the only thing i noticed is recalculated resistor ladder  :

If you look closely at the NYD schematic, you will see the top half of the ladder is not always connected to the bottom half.

Cheers

Ian

ohhhh finally i c it
thanks Ian

It took me ages to work out what is going on. There is a half ladder at the top which is connected to 0V via the 500 ohm load. This half ladder is used for boost. There is a second half ladder which is not in circuit when boosting but only becomes connected when cutting and adds resistance in parallel with the 500 ohm load. I don't know why NYD did it this way but his designs were always well thought out. All I have done is take the several ladders I used to have and combine them into a single ladder.

You will see the bottom half of my ladder adds up to 654 ohms and this is the maximum possible output impedance of the EQ. In practice the 654 ohms will be in parallel with the series combination of the transformer secondary impedance and the top half of the ladder. If the transformer secondary looks like 600 ohms the this series combination is 2010 ohms. This is parallel with the 654 ohms is 493 ohms. So in practice the output impedance  will be 600 ohms or less. This means that if you connect the EQ output to a 10K bridging network, it will lower the impedance of the bottom half of the ladder by about 5% which represents about 0.5dB difference in signal level. Bottom line is you should be able to feed the EQ straight into a 10K bridging balanced line input if you want to make a 100% passive EQ. To make it balanced in just add a good quality 600:600 transformer at the input.

Cheers

Ian

 

[kambo]

if i am hitting a tube gain stage on the output,
should i be using 10K grid resistor ?



edit:  thanks for the explanation btw, makes perfect sense...

 

[ruffrecords]

if i am hitting a tube gain stage on the output,
should i be using 10K grid resistor ?



edit:  thanks for the explanation btw, makes perfect sense...

No need for it to be 10K. It can be as big as you like, 100K or 470K; the important thing is to minimise the load on the EQ output. I only mentioned 10K in case people wanted to use an external gain make up amp like a console line input for example.

Cheers

Ian

 

[kambo]

sounds good.

hopefully i am gonna breadboard this today... 

i was experimenting with one inductor per band EQ with 12 caps/frequencies and it seems to be working very well on breadboard...
i will adopt my values to this.... same 600r input!
see if one inductor/12 caps will do the trick on this too....

 

[ruffrecords]

sounds good.

hopefully i am gonna breadboard this today... 

i was experimenting with one inductor per band EQ with 12 caps/frequencies and it seems to be working very well on breadboard...
i will adopt my values to this.... same 600r input!
see if one inductor/12 caps will do the trick on this too....

If you use a single value inductor and only change the value of capacitor to change the frequency then the bell shape will become narrower as the frequency rises. This may or may not be what you want. To maintain broadly the same shape at all frequencies you need to vary the inductor as well as the capacitor. Using a single valaue of inductor for each band certainly simplifies things.

Cheers

Ian

 

[kambo]

as frequency rises
for cut, i prefer its getting narrower  ; for boost i like getting wider

 

[ruffrecords]

Thanks Abbey. Useful input. Just to be clear, you are proposing for shelving EQs that the  frequency should be defined as where the boost/cut reaches 1/4 of the maximum in  dBs. Don't you think that would give misleadingly low values for most shelving EQs?

Cheers

Ian

Very good comment!
In fact musicians think that the actual definition is misleading, because they hear their treble & bass controls acting on midrange.
Indeed, technicians/designers know that a 10k shelving EQ starts at about 1k, but we have to serve our customers.
In order to maintain consistency with usage, we would have to base the definition on 3/4th from max. Although in terms of operation, musicians would agree more with a definition based on "that's where I start to hear things changing".

So, to summarise,  for bell shaped EQ bands you recommend quoting the bandwidth at the point where the response drops to one quarter the peak response (in dBs) and for shelving EQs, the point where it drops by one quarter from its final level.

It is interesting to note that in the Helios 69 EQ, the treble control actually does produce the the indicated boost/cut at 10KHz. The mid band boost/cut has a fixed 3dB bandwidth at maximum boost of about half the boost frequency. The low frequency bell boost curves are similar. Maybe this has something to do with its reputation for musicality. The Pultec EQP1A high boost bell has a similar 3dB bandwidth (although it is variable from about twice this width to half this width). Many of the classic EQs I have studied have similar parameters. Modern solid state EQs tend to produce much sharper curves.

Perhaps the old Q control should be labelled bandwidth and  labelled in percent from 100% down to about 30%.

Cheers

Ian

 

[kambo]

hey Ian,
i breadboarded ur version for max boost and max cut values, yesterday.
i wasnt able to get boost and cut at the same Q
cut was way sharper then the boost... boost was very broad, almost shelving type.
i think there were ~50r difference in total resistance, that must be it ?

( i actually added subtracted different resistors, Q was not changing
i tried very low resistance 15mh +33mh + 56mh +100mh with combination of cap values)

 

[kambo]

any chance of publishing the spread sheet to calculate the L and C 

 

[abbey road d enfer]

So, to summarise,  for bell shaped EQ bands you recommend quoting the bandwidth at the point where the response drops to one quarter the peak response (in dBs)

Yes; I know there would be endless debates between those who want 1/5th and others that want 1/3rd, but I think, if we could get a consensus on the principle, that would be one step in the good direction.

and for shelving EQs, the point where it drops by one quarter from its final level.

That's only for compatibility with legacy; if I wanted to be rigourous, I would keep a single definition for both bell and shelf. That would require a symbol, just like F(-3dB) is a symbol; if "sharpness" was symbolized by S(1/4th), it could be used for both, and satisfy musicians.
But I'm not in the AES anymore, so I don't think it will happen...

Perhaps the old Q control should be labelled bandwidth

I think it would make a little more sense than Q, but still, there is an academic definition of BW, that would collide with the actual EQ sharpness notion.

 

[ruffrecords]

any chance of publishing the spread sheet to calculate the L and C 

No problem. It is very simple. You just plug in the desired frequency and sharpness (Q) and it works out the values for L and C.

The spreadsheet is just a list of these so you can fill in each line and get the set of LC vales required for your EQ.

It is a Microsoft XL spreadsheet but you cannot attach those to posts so I have changed its extension to .txt. You will need to change it pack to .xlsx. If you prefer it in another format let me know.

Cheers

Ian

 

[ruffrecords]

hey Ian,
i breadboarded ur version for max boost and max cut values, yesterday.
i wasnt able to get boost and cut at the same Q
cut was way sharper then the boost... boost was very broad, almost shelving type.

The circuit resistance for cut is lower than for boost because of the the 10dB pot divider. This means the cut will be a lot sharper than the boost.  Some people think this is an advantage, The famous Helios 69 EQ does this. If you want the sharpness to be the same for cut as for boost you will need to use different LC values for cut and boost. To achieve this you need a two pole switch es for your frequency selection and ladder selection. This is basically what I did in the REDD EQ.

Cheers

Ian

 

[kambo]

txt extension worked well, thanks for the spreadsheet
sharper cut works for me anyway, i thought i was doing something wrong!

 

[kambo]

hey Ian,
i am trying to get cheapo inductors from mouser to test with, but anything over 100mh,
resistance of inductor jumping to around 200r... which will be  effecting Q

when i recalculate inductor/cap values, shall i add my inductor resistance to 500r in your formula ?

 

[ruffrecords]

hey Ian,
i am trying to get cheapo inductors from mouser to test with, but anything over 100mh,
resistance of inductor jumping to around 200r... which will be  effecting Q

when i recalculate inductor/cap values, shall i add my inductor resistance to 500r in your formula ?

Yes, the 500R assumes perfect inductors or inductors with a resistance that is small compared to 500 ohms. If your inductors have a largish resistance you should add it to the 500 ohms.

Cheers

Ian

 

[DerEber]

Dear Ian,

I want to implement your 6dB Eq in the Classic solo.
I want to do only highs and lows. No mids.
Now I try to figure out how to calculate the Eq for myself.
I am also playing with the Idea to use an spare shielded transformer as an inductor. Using one of the windings and leave the rest unconnected.

For the bass:

I know the good old f=R/(2*pi*L).

So what is R in this case?  Total Resistance surrounding L?
In your Exelsheet it says "Pot" in kOhm. Does this mean the Pot in front of the inductance?
I'd assume it to be the output pot.
I will have 2 Pots around the Inductance. One for the basscontroll and one for my preamp volume. As in the simple circuit from the beginning.
If the basscontroll-pot is in the equation then doesn't the frequency shift while turning?

best greetings from Germany and happy Christmas,
Stephan

 

[ruffrecords]

The spreadsheet is designed for use with the three band EQ and will give incorrect results for a two band version. The 'pot' in the spreadsheet refers to the total value of the potentiometer used for each band and is assumed to be the same for all three.

With shelving EQs like the bass and treble controls of the simple 6dB EQ, there are two values of R. The first affects the frequency at which the boost/cut starts and the second affects the frequency at which the response begins to shelve. In the 6dB EQ, these two points are pretty much the same and are determined by the 10K resistor or the 10K level control (depending on whether you are boosting or cutting).

So the basic R for your equation is 10K but this will give you the 3dB point of boost/cut not the final value. So if you plug in 10K and 2H of the original design you get a 3dB point at  just below 800Hz. It will take a couple of octaves for the shelf to flatten which gives an ultimate boost frequency of around 200Hz.

I have not simulated the original circuit to verify this but I will do so because in its original form it could be quite handy for the Classic Solo.

Cheers

Ian