Universal Passive EQ

Ahh I see now. So the pads marked with an "X" would need to be attached to the ground plane?
If this is the case then I think doing this board double sided would be fairly easy.

Dave

wave said:

Ahh I see now. So the pads marked with an "X" would need to be attached to the ground plane?
If this is the case then I think doing this board double sided would be fairly easy.

Dave

Click to expand...

You got it!. The PCB software I use shows the pads connected to the plane as an X.

Cheers

Ian

shabtek said:

thats close--can that second pole be made available,
and an extra pad per throw--maybe a 3rd row of 0.1" to accommodate daisychaining resistors.

i like where this is going

Click to expand...

How about this?

Cheers

Ian

wave said:

Ahh I see now. So the pads marked with an "X" would need to be attached to the ground plane?
If this is the case then I think doing this board double sided would be fairly easy.

Dave

Click to expand...

Turns out removing the ground plane and tracking the ground normally was quite easy. I even managed to get it on the same side of the PCB as all the other tracks so for home etchers it is a single sided PCB. Attached is a pic and here is a link to the zipped up Gerbers:

http://www.ianbell.ukfsn.org/EzTubeMixer/docs/EzTubeMixer/UniversalEQ/PCB/OSuniEQ.zip 

Cheers

Ian

ruffrecords said:

Turns out removing the ground plane and tracking the ground normally was quite easy. I even managed to get it on the same side of the PCB as all the other tracks so for home etchers it is a single sided PCB. Attached is a pic and here is a link to the zipped up Gerbers:

Cheers

Ian

Click to expand...

Awesome! Thanks Ian!

Dave

can you add the grayhill boards to the group buy; I am in for several (grayhill boards)

shabtek said:

can you add the grayhill boards to the group buy; I am in for several (grayhill boards)

Click to expand...

Are you happy with the new layout??

I have also had a request to do a Lorlin version. Is that of interest to anyone?

Cheers

Ian

shabtek said:

can you add the grayhill boards to the group buy; I am in for several (grayhill boards)

Click to expand...

Done. Let me know what you need and I'll update the group buy.

Cheers

Ian

ruffrecords said:

Are you happy with the new layout??

I have also had a request to do a Lorlin version. Is that of interest to anyone?

Cheers

Ian

Click to expand...

Hi Ian,
I like the new layout. I am def interested in the Lorlin version. If I have to etch them myself I will but I'm not very good at using Eagle.
If you wanted to knock a quick Lorlin layout together, I would be super appreciative!

Thanks,
Dave

Also,
What inductors would I need to buy? Is it all dependent on how I would build the EQ? Are there any examples of what the different inductors would give me in each position?

DAve

wave said:

Also,
What inductors would I need to buy? Is it all dependent on how I would build the EQ? Are there any examples of what the different inductors would give me in each position?

DAve

Click to expand...

It is very dependent on how you would build the EQ. The whole point of the universal EQ is that you can use a single PCB to make almost any EQ from the simplest to the most complex. On the one hand you can use it to make the 3 band Pultec or Helios EQs or the 2, 3 or 4 band versions of the poor/posh man's EQ. And then you can use it to build a unique EQ of your own.

As for inductors you can use most of the range from Carnhill or get some made by Chrion on this forum - the PCB is laid out to take either - or you could even wind your own.

If you let me know s bit more what you want from the EQ I could give you some more specific advice.

Cheers

Ian.

Here's a first draft of the Lorlin header PCB. It is 1.2 inches wide by 1.55 inches high. I have not checked whether the chosen orientation of the switch is the most convenient for an EQ switch. I simply chose the orientation that made the PCB layout simplest.

The Grayhill header PCB is 1.0 inches wide by 1.35 inches high.

Cheers

Ian

ruffrecords said:

...you wrote: ....

As you have discovered, the overall circuit impedance is the key to determining the size of inductors you need to obtain a given Q or bandwidth and the lower the circuit impedance the easier this gets which is one reason I suspect the original EQP1A works at 600 ohms. If people are going to drive this EQ from an op amp then that in itself should not be a problem but for a low cost tube mixer where you would want to perhaps drive it directly from a 10K fader your worst case source impedance is going to be 2.5K or more so you really don't want to load this with more than 10K which is exactly what the poor man's EQP1A tried to do. So if this EQ is going to be fairly universally applicable then that's probably where we need to stay. So yes, I agree with you we should aim fro a 10K bridging input impedance.

Click to expand...

Ian:

Sorry I am confused by this, would you mind explaining how I think about overall circuit impedance?  Is this the input impedance of the circuit (the input trafo or the input load resistor in parallel with the output impedance of the previous stage)?  Or is this the load on the output of the passive circuit.

I am trying to get my head around these passive EQ circuits and would like to understand how these effect the EQ.  I understand that the Q of the LC boost circuits is reduced as the impedance there is increased, does this also happen as the input impedance increases?

I have knocked together a point to point Pultec passive section to play around with and for the moment have driving it a really low impedance discrete op-amp based circuit that can drive pretty much any load.  I put a load resistor across the input but am trying to understand the effect of that load and what load I should be putting there.  On the output makeup gain stage I currently have a 10k load and a very high impedance op-amp input, but again I don't understand the implications of that, and what I should do there.

Any help on this would be appreciated, (I usually can pick things up but I am finding the passive circuits really befuddling)

bruce0 said:

ruffrecords said:

...you wrote: ....

As you have discovered, the overall circuit impedance is the key to determining the size of inductors you need to obtain a given Q or bandwidth and the lower the circuit impedance the easier this gets which is one reason I suspect the original EQP1A works at 600 ohms. If people are going to drive this EQ from an op amp then that in itself should not be a problem but for a low cost tube mixer where you would want to perhaps drive it directly from a 10K fader your worst case source impedance is going to be 2.5K or more so you really don't want to load this with more than 10K which is exactly what the poor man's EQP1A tried to do. So if this EQ is going to be fairly universally applicable then that's probably where we need to stay. So yes, I agree with you we should aim fro a 10K bridging input impedance.

Click to expand...

Ian:

Sorry I am confused by this, would you mind explaining how I think about overall circuit impedance?  Is this the input impedance of the circuit (the input trafo or the input load resistor in parallel with the output impedance of the previous stage)?  Or is this the load on the output of the passive circuit.

Click to expand...

There are several impedance involved.

I that case was referring to the impedance the EQ presents to whatever is driving it. This impedance varies with frequency and with the EQ settings. My aim was to make it generally no less than 10K ohms but it can get as low as 5K or be as high as around 27K.

The other important impedance is that of the load hung on the output of the EQ. This needs to be high enough not to load the EQ and not alter the curve shapes. Hoe high this is depends on the EQ topology.

The final impedance is the source impedance o feeding the EQ input. Lowere is better in this case. Op amds and cathode followers will be fine, but it is designed to be fed straght from a 10K fader which will have a worst case source impedance of about 2.5K ohms.

I am trying to get my head around these passive EQ circuits and would like to understand how these effect the EQ.  I understand that the Q of the LC boost circuits is reduced as the impedance there is increased, does this also happen as the input impedance increases?

Click to expand...

The Q in this type of LCR circuit is just 2*pi*f*L/R

Where f is the resonant frequency, L is the inductance and R the local circuit resistance. Clearly, if R is increased, L needs to increase by the same amount to maintain the Q. Conversely, if R is reduced then, if L stays the same, the Q gets bigger.

In the Universal EQ there is basically a pot divider made from a 22K series resistor and a 5K resistor to ground. This has a base level attenuation of about 15dB. If you shunt the 22K  with an LC circuit you reduce the attenuation at resonance so you get boost. The Q is determined by the local circuit resistance which is primarily the 22K. If you shunt the 5K with an LC circuit you increase the attenuation at resonance and you get a dip response. Since the local circuit resistance is 5K you need a smaller inductor to get the same Q. If you use the same LC circuit but just switch it from across the 22K to across the 5K, your dip Q is much higher than your boost Q. This is exactly what happens in the Helios 69 EQ.

I have knocked together a point to point Pultec passive section to play around with and for the moment have driving it a really low impedance discrete op-amp based circuit that can drive pretty much any load.  I put a load resistor across the input but am trying to understand the effect of that load and what load I should be putting there.  On the output makeup gain stage I currently have a 10k load and a very high impedance op-amp input, but again I don't understand the implications of that, and what I should do there.

Any help on this would be appreciated, (I usually can pick things up but I am finding the passive circuits really befuddling)

Click to expand...

There is really no need to put a load across the Pultec input - it provides the load to the op amp. If instead you used a 10K:10K transformer then the transformer prefers to work into a 10K load. The original Pultec EQP1A was designed to look like a load of 600 ohms, which is why the poor man's Pultec changed the circuit so it looked more like 10K so it could be used with modern 10K:10K transformers.

The Pultec output is very load sensitive and reall needs to be very lightly loaded - the original EQP1A should not really be loaded with less than 50K and the poor man's should not be loaded with less than 200K. In both cases you 10K is a little low. This will mostly cause the low cut to cut more than expected.

I am assuming your 10K load is on the output of the EQ and not the output of the op amp.

Cheers

Ian

Ian...

I can't thank you enough.  This is very clear, and gives me orientation for testing and a starting point to do some studying.  This helps a lot.

And it makes sense with respect to what I am seeing in my test.

Right now I have Op Amp - Passive circuit - load resistor - op amp... And I will remove the load resistor, but I am wondering about loading it with a transformer.

Does the inductance of an interstage transformer (like the one that loaded the original Pultec circuit) change the curves? (The tubetec PE-1s schematic I have labels it 10K:10K+10K, with DCR of 637 on the primary and 1026  on the secondary). It seems like it is in series with the other inductor (separated by a 2.2k resistor) and so with a transformer and its relatively large inductance the effective inductance would be higher, and the result of the Q equation would increase?

The g-pultec design increases the inductance somewhat, but only a little (175mH to 269mH). 

I know this is off topic, and please feel free to ignore the question I can probably figure it out. 

Thanks again for a really great and well thought out answer.

Bruce

Greetings Ian,
Do you have the Universal Passive EQ boards for sale?
Thanks!
Best,
Bruno2000

bruno2000 said:

Greetings Ian,
Do you have the Universal Passive EQ boards for sale?
Thanks!
Best,
Bruno2000

Click to expand...

Yes, I do. There is a group buy going on right now here:

http://www.groupdiy.com/index.php?topic=51131.msg649433#msg649433

I also had made interface PCBs for Grayhill and Lorlin switches so you can connect them to the Universal EQ PCB using ribbon cable.

Cheers

Ian

Greetings Ian,
What are the Carnhill part number(s) for the inductors?
Does ChrioN have these available as well?
Thanks!
Best,
Bruno2000

bruno2000 said:

Greetings Ian,
What are the Carnhill part number(s) for the inductors?
Does ChrioN have these available as well?
Thanks!
Best,
Bruno2000

Click to expand...

The Universal EQ is meant to allow you to design almost any EQ you fancy. As such, there is therefore no fixed set of inductors to use.. That said, the three inductor positions on the PCB are laid out to take most of the range available from Carnhill and they also take Chrion's inductors and he will make anything you want.

If you could be more specific about the EQ you want to build I could give you some more concrete advice.

Cheers

Ian

I was looking into Chrion's inductors as well for my Helios EQ boards I got from Ian and the last round of emails I had with Nicolas I believe he told me he wasn't doing the 9043 type inductors anymore.

Dave