Comments on EQ schematic

[sahib]

To be perfectly honest with you I did not think of selling them so soon. So please bear with me for a day or two but I'll work out a mutually satisfactory figure.

Indeed you can compensate by changing the capacitor values. Jeane is going to simulate standard cap values to see the response.

 

[letterbeacon]

No problem on waiting at all -my to-do list is ridiculous already!

 

[Jean Clochet]

...taps for the Pultec MEQ -am I right in thinking you can compensate for that by changing capacitor values?

You can compensate yes. 
It's best that you don't go too wildly away from an ideal cap value though or the bandwidth will be different for that particular frequency.
You could compensate for BW with a variable resistance in series with the L's & C's to change the Q like the Pultec "Bandwidth" control on the EQP1.  But it's best if all freqs. are about the same BW relative to each other to start with. IMHO.
Of course, having 2 or maybe 3 freqs. per L value is fine and isn't going to change the slope to any extent that you'd notice.  As long as the frequencies aren't too far apart that is.

To start with though, I wasn't even going to look at the Pultec frequency markings,  just match standard cap values to the taps on Cemal's inductors to see what there is as a baseline.

 

[sahib]

Apologies Jean. I did not mean to turn this into a full time job for you. As you clearly said, just to match the available coils with the standard value caps to see what potential we can get.

 

[mulletchuck]

Could someone post a schematic (for simulation purposes) of what a 600Ω input looks like?

i'm assuming this is incorrect:

 

[Jean Clochet]

Apologies Jean. I did not mean to turn this into a full time job for you. As you clearly said, just to match the available coils with the standard value caps to see what potential we can get.

Oh, no apology necessary.
I just mean that I'd match the cap values with inductor values such that Bandwidth/Q is similar for all frequncies.  I only mention keeping it simple at first with standard cap values for the reason of using an off the shelf single cap part rather than having to buy 3 values just to make an odd value.  We can add the odd values later after we know what the basic inductor will do in a Pultec type circuit. 

I have one other bit of work to finish up first but I'll be able to get to this in a day or so and it's not a bother at all Cemal.  8)

 

[Jean Clochet]

i'm assuming this is incorrect:

You could use 2 X 300 ohm resistors in place of your 2 X 600 ohms or remove one of your 600 ohms.  Or make the internal resistance of your generator be 600 ohms.  Any scheme would simulate a 600 ohm source impedance/resistance. 
If you're simulating a single-ended filter network like the one posted originally in this thread,  then I would opt for a single 600 ohm resistor fed from your generator to the network.  Just make sure that your generator is set for zero internal resistance in that case.

 

[shabtek]

nothing to add to the eq discussion--am interested in resonant iron thingies--put me down for some chokes.

 

[shabtek]

I would have to bow out.

 

[sahib]

Last night I scribbled something as an idea based on MEQ5 configuration. For those members who may have difficulty in understanding the schematic I simplified the boost and cut sections of MEQ5 with a single inductor and capacitor network.

For the moment just ignore the inductor/capacitor networks. Also assume that the potentiometers are just single value resistors. So, what we have is a standard resistive voltage divider. The resistor which the signal is applied to is the high side, and the one connects to the ground is the low side.

On the boost section the high side resistor is greater than the low side. Which means the signal is subjected to a greater attenuation.

Now replace the high side resistor with a potentiometer and introduce the inductor/capacitor network. Bear in mind that at the frequency of interest the network impedance is lower than the potentiometer.

When VR1 is turned full to the right the network is shorted. So all frequencies are subjected to the same amount of attenuation set by the full resistive divider.

But as VR1 is turned to the left the lower inductor/capacitor network impedance  starts  to take effect  and the frequency of interest follows the shorter path through the network and reaches to the output with an attenuation rate set by VR1, while the others have to go through the full resistive part of the voltage divider and receive the full attenuation.

On the cut it is opposite. The high side resistor is lower in value and the low side is higher. So the signal is subjected to less attenuation. Again when VR1 is turned full to the bottom the network is shorted and all frequencies receive the same amount of attenuation. But as the VR1 is turned up the low network impedance starts to take effect and the frequency of interest follows the shorter path through the network and to the ground, receiving greater attenuation.

The VR2 is a series resistor which increases/decreases the DC resistance of the inductor hence affects the Q of the network (as Jean mentioned previously).

Below is my scribble. I have introduced a push button switch (can be rotary) to change the function of the circuit. In normal position the circuit is on Boost mode and when the button is pressed it switches to Cut mode.

On Boost mode VR1 and R1 forms the resistive voltage divider with R1 being on the low side. VR2 is shorted.

On Cut mode VR1 and R1 swaps places. Now R1 is on the high side. R2 also enters the game and shunts R1, lowering its value. The impedance network now also wraps around VR1 on the low side. VR2 also joins in series and controls the Q.

I know I have two interesting valve based bandpass eq schematics. One is I think in Wireless World and unusually had the high cut (low pass) in the front but with a sweepable frequency using a 3 gang variable capacitor (for 3rd order).  The other one is the service manual of a Krohnhite bandpass eq. Unfortunately all my archive is in boxes and trying to find them is going to take some time. But we can borrow some ideas and patch up some Frankenstein eq. I'll let Jeane to carry the brain though as he probably has steadier hands.

 

[mulletchuck]

thanks for this post and the descriptive explanation.  I need to read it a bunch of times before it makes sense. 

How does the boost circuit actually "boost" anything?  You described it using the word "attenuate" which means to lessen.  So, is the boosting really just a fake effect of attenuating all of the frequencies that you DON'T want cut, and using a gain stage afterwards to increase the signal level by an arbitrary amount?  Like how the two examples in the attached picture result in the same curve, even though they're achieved via different means.  the one on the left is a Hi-Shelf 15dB boost at 3Khz, and the other is a lowShelf 15db CUT at 3Khz, followed by a 15db overall boost.

 

[sahib]

Chuck,

What you say is correct.

The confusion comes from the fact that the term "boost" is also used for amplification.

But the reason that it is used here is because we want to emphasise the frequency of interest. Otherwise there is no amplification.

 

[leadbreath]

hey sahib

£95.00 is way too steep for my poor arse, i would be willing to pay up to £50.00 max, £100 for 2, but that comes down to you obviously.

mick

 

[mulletchuck]

So, no matter what, this circuit will actually decrease the voltage of the AC signal, and will need to be followed with a make-up gain stage to raise the voltage back up?  Is this how all EQs actually work?

Chuck,

What you say is correct.

The confusion comes from the fact that the term "boost" is also used for amplification.

But the reason that it is used here is because we want to emphasise the frequency of interest. Otherwise there is no amplification.

 

[sahib]

Naturally, as it is a voltage divider.  So, as you rightly said you have to make up for the loss afterwards by amplification.

Not all EQs are like this one. This is a passive network and it does not have a mechanism to make up for the loss it presents to the signal. 

On the other hand you can have active networks in which you can maintain unity gain or even increase gain. However, there is a limit in terms of the amount of gain applied in active networks. If greater than what is permitted required, again a seperate gain stage is used at the end of the chain.

 

[opacheco]

Last night I scribbled something as an idea based on MEQ5 configuration. For those members who may have difficulty in understanding the schematic I simplified the boost and cut sections of MEQ5 with a single inductor and capacitor network.

For the moment just ignore the inductor/capacitor networks. Also assume that the potentiometers are just single value resistors. So, what we have is a standard resistive voltage divider. The resistor which the signal is applied to is the high side, and the one connects to the ground is the low side.

On the boost section the high side resistor is greater than the low side. Which means the signal is subjected to a greater attenuation.

Now replace the high side resistor with a potentiometer and introduce the inductor/capacitor network. Bear in mind that at the frequency of interest the network impedance is lower than the potentiometer.

When VR1 is turned full to the right the network is shorted. So all frequencies are subjected to the same amount of attenuation set by the full resistive divider.

But as VR1 is turned to the left the lower inductor/capacitor network impedance starts to take effect and the frequency of interest follows the shorter path through the network and reaches to the output with an attenuation rate set by VR1, while the others have to go through the full resistive part of the voltage divider and receive the full attenuation.

On the cut it is opposite. The high side resistor is lower in value and the low side is higher. So the signal is subjected to less attenuation. Again when VR1 is turned full to the bottom the network is shorted and all frequencies receive the same amount of attenuation. But as the VR1 is turned up the low network impedance starts to take effect and the frequency of interest follows the shorter path through the network and to the ground, receiving greater attenuation.

The VR2 is a series resistor which increases/decreases the DC resistance of the inductor hence affects the Q of the network (as Jean mentioned previously).

Below is my scribble. I have introduced a push button switch (can be rotary) to change the function of the circuit. In normal position the circuit is on Boost mode and when the button is pressed it switches to Cut mode.

On Boost mode VR1 and R1 forms the resistive voltage divider with R1 being on the low side. VR2 is shorted.

On Cut mode VR1 and R1 swaps places. Now R1 is on the high side. R2 also enters the game and shunts R1, lowering its value. The impedance network now also wraps around VR1 on the low side. VR2 also joins in series and controls the Q.

I know I have two interesting valve based bandpass eq schematics. One is I think in Wireless World and unusually had the high cut (low pass) in the front but with a sweepable frequency using a 3 gang variable capacitor (for 3rd order). The other one is the service manual of a Krohnhite bandpass eq. Unfortunately all my archive is in boxes and trying to find them is going to take some time. But we can borrow some ideas and patch up some Frankenstein eq. I'll let Jeane to carry the brain though as he probably has steadier hands.

Dear sahib,
Could you upload your schematics images again please?

Thanks in advance
opacheco

 

[sahib]

Hi Opacheco,

Thank you too.

That was nearly 13 years ago. I replaced my PC twice since then. So, I really have to dig into the archives. However, the magazine article with schematic for the tube bandpass filter I mentioned is attached.