Some Discussion on Amp EQs
I get a lot of questions every now and then about different amp EQs so I figured I could do a big overview of a bunch of popular ones. I wrote most of this from my memory and also using simulations in LTSpice. Please check it out and let me know if I have made an error or something you think should be added to this post.
The following EQs I will discuss.
1. A James/Baxandall EQ with the example being the one from an Ampeg SVT.
(Note: Technically speaking I am going to discuss a James EQ which is not the same as a Baxandall but these 2 have been so mixed together when talking about amps that I am calling them the same thing. This is just a semantic point IMO)
2. The Active Mid Range control in the Ampeg SVT
3. A FMV tone stack from a 60s Blackface Fender
4. A FMV tone stack from a Marshall Superbass.
5. Hiwatt DR103 EQ (Plate Driven)
6. Hiwatt DR103 EQ (Cathode Driven)
Note I have moved the sections discussing Hiwatt's to the 4th page because I could not fit it into the character limit for this post.
Click this link to go directly to it
These 6 are not all of the possible EQs one can have in an amp, just the most commonly used ones in vintage amps such as Fender, Sunn, Marshall, Vox, Ampeg, Orange, Matamp, Mesa... ect ect.
I will be showing a # of graphics which show the input and output of these EQs in decibels as a function of frequency. This will show the frequency response of the circuit. For some added realness I will be including the actual tube driver stages so that the freq response plots reflect the loading of the tube stages for a more accurate plot. I will not include all of the gain stages in the amp tho, just the preceding tube driver stage and in the case of the active mid range control on the Ampeg, the 3 triode sections which comprise it.
1. The James/Baxandall Tone Stack.
Below is the schematic for the baxandall EQ as taken from an Ampeg SVT.
In the case of a Ampeg SVT, the EQ is driven by a 12au7 triode gain stage. The EQ has 2 controls Bass and Treble (hi and low). To summarize it is a passive EQ, meaning that the output signal level is never greater than the input signal level and in real life it is always attenuated to some degree due to the nature of filtering the signal. Thus there is only how much signal are you cutting. Baxandall EQs do have the illusion of being active, because unlike the FMV, when the controls are at noon, the EQ is ~flat, and the bass and treble operate like high and low shelving filters. Of course the EQ is not perfectly flat and can often be manipulated to be less than flat by changing the components. The Ampeg baxandall is exceptionally flat, possibly this is how the design was conceived to be as neutral as possible when the EQ is at noon. Ampeg did pull the wool over your eyes with the words Boost and Cut on the knob but this is purely their own semantic thing as the EQ is passive so the word Boost does not have an electrical meaning here.
The baxandall is probably my favorite EQ I should mention. It is primarily featured in only a few amps in the grand scheme of stuff. Ampeg SVTs, V4s, and B12s, Matamp and Oranges from back in the day, a few rare rare JCM 800 Bass units from Marshall and some others which escape me.
The Baxandall is actually 2 other circuits which are put together, one a high shelf and one a low shelf. If you take a look at the treble control and sort of blank out the bass you can probably get an idea of how it operates. if you turn the treble pot all the way, the circuit can be simplified to a high pass filter formed by the 1meg pot and the 47pF cap. This would result in a lot of high end at the output. If the treble was turned all the way down, then the circuit can be simplified to a LPF which is formed by the 1 meg pot and the 470pF cap to ground. This would have a lot off high end attenuated. However, the 47pF cap still filters which signal is going into the filter, so generally high end is only going into the control and only high end can be attenuated as little low end enters into this control.
On the bass control if the control is turned all the ways up, the circuit simplifies to a LPF formed by the 220k resistor, and the 10nF cap. The 1meg pot and 22k resistor also shape the filter but for now focus on those parts. The output would have a lot of bass as high end is attenuated while low end can go throw unattenuated. If the control goes all the to 0, the 1nF cap now forms a high end bypass where low end sees the 1meg resistance of the pot and thus, there is a HPF formed, resulting in heavy low end attenuation, and thus little low end in the output.
Now these two circuits are joined together by the 120k resistor that connects the wiper of the bass and treble controls. Think of this as a way to sum the circuit together. The resulting output which is taken from the wiper of the treble pot is the sum of these 2 filters. This circuit has minimal interaction between controls but there is some interaction.
Now the first graph of the day is the input and output of the EQ if the controls are at noon.
The blue signal is the input and green signal is the output.
So the input signal is ~0dB for all frequencies but there is some variance as the EQ loads down the tube gain stage at certain frequencies resulting in a little bit of loading. Now the green signal at the output (signal at the 10 meg resistor) is ~20dB less than the input signal. That is a lot of attenuation. Now the output also has some variance even at noon but its generally <3dB across the frequency band which is hard for one to detect if that is the case.
Now Im going to sweep the Bass or Treble with the other control being held at noon, then I will sweep the Bass and Treble together to show how the EQ works.
Here is the Bass sweep.
Blue line is again the input, while each green line represents a different turn of the pot and its resulting output while the Treble is at noon.
As you can see the sweep is quite large, going from ~-2dB to -35dB at the top of the shelving filter. Even at max bass, the output does not achieve unity with the input. There is always -2 to -3dB attenuation but that is not a ton of attenuation so its very much the best practical passive shelving filter.
Now for the treble sweep.
As with the bass there is a good range of sweep here. However the peak of the filter extends beyond 20 khz and I am not showing it on this sweep as that is beyond human hearing. The actual cutoff of the both the treble and bass controls is determined by the caps, resistors, and pot resistance values which comprise the circuit. If for example you wanted to increase the amount of high mids in the Treble control you would increase the 2 cap values on the Treble pot proportionally. So if you double or triple the values, more of the frequency band will be controlled by the Treble. The same is true of the Bass. However, too much increase in the cap values will result in more interaction between the Bass and Treble. So there is a trade off between amount of control and interaction.
The most likely reason Ampeg limited the treble frequencies is due to the fact that the active mid range extended to 3 khz in the 3 position of the mid select switch. That is right where the treble control no longer has any more control. So they were attempting to keep the controls as isolated as possible.
Finally, here are the treble and bass control swept together.
Some observations. As the controls are both increased towards max or decreased to 0, the mids are pretty much left alone. At max the EQ has a mid cut then and at minimum it has what appears to be a mid boost. This is one of the interesting things about this EQ. You can get psuedo mid boosting when the Bass and Treble backed off. However, this is as the cost of a ton of attenuation of highs and lows.
The reason the EQ has very good range is because the EQ is quite lossy. What I mean by lossy is that the EQ has a lot of insertion loss, ie, the output signal is always going to be heavily attenuated across all bands. The tradeoff with insertion loss is that you get more control if you increase it but more loss on the output. In a bass amp that is probably not a real big deal but in a guitar amp where you wanna get more clipping and distortion the result is that you need more recovery gain after the EQ to get the same level if the controls are ~ at noon.
One more important note, there is no mid control in this EQ and no easy way to put in some kind of true mid range control. That is also one of the trade offs. Of course, since the EQ has no mid cut at noon, its often interpreted as being more mid rangy then a FMV but if you do max out the controls, then a mid cut can be heard. However, I never run the EQ like this so its never a problem.
2. The Ampeg SVT Mid Range EQ
The Ampeg SVT Mid control (and Mid Selector) is one of the few true active mid range EQs and is the ideal (IMO) mid range control for bass and guitar. It offers approximately plus and minus 20dB of control on a specific mid range frequency. It is an active bandpass filter which involves 3 triode gain stages, and a tapped toroid transformer. The tapped toroid provides different inductance values for the filter which allows one to change the center frequency of the bandpass from 220 hz to 800 hz to 3 khz. The only say disadvantage of this EQ is that is requires a lot of parts, a lot of space, and design work to get it to work. I do not think its often worth the effort given how much real estate you need to take up. It has dozen plus parts in it, would require a weird 3 triode NOS tube and or combinations of 12ax7s, 12au7s, or 12dw7s.
Here is the schematic:
Suffice to say its a bit 2 much to explain here. Just understand it requires a lot of work to achieve.
Now the 3 way switch works to ground different parts of the tapped toroid which I am showing at the bottom of the schematic, represent as different inductors with different inductance values. I was able to reverse engineer their approx values since the schematics on the internet do not show their values.
Here are the 3 positions and their respective frequency responses.
The input of the EQ is flat at 0 dB for all frequencies. If the mid control is at noon, the output is essentially the same as the input. So the insertion loss is 0, there is none.
As the mid pot is turned the Q of the filter sharpens the response making the filter tighter around the bandpass center frequency.
This EQ is very impressive, and would rival any SS bandpass you could get. The fact that it is done with tubes is even more mojo.
3. Fender Style FMV
The FMV tone stack (FMV = Fender Marshall Vox) is a 3 control EQ which is the most popular EQ in Ampdom. I believe this is the case because most amp companies liked to copy popular designs and that this EQ is very classic to many people ears despite it being the least ideal from a outside viewpoint.
Here is the schematic for the Fender style FMV stack with component values close to a 60s Blackface Fender amp.
This type of EQ has the smallest part count, smallest footprint, and usually easiest wire up. The controls are from top to bottom, Treble, Bass, and Mids. Sometimes this EQ is called a TMB (short for treble mid bass) in some amp circles.
This is the input and output response of the EQ with the controls all at noon.
The most striking thing about this EQ is that it is hardly flat when the controls are at noon, and also a flat EQ cannot really be achieved in any sense. Thus this is why this EQ is far from the ideal EQ for an amp but is not a declaration that it is bad or good for an amp.
There is a fairly large insertion loss for Fender style FMVs but not as great as the baxandall. however, there is a lot of loss in the mid range as noted above.
Here is the sweep of the Bass pot
The sweep of the treble pot
The sweep of the mid pot
Some observations on this EQ. Well first off the mids can never be boosted or even brought back to flat. The amount of cut can be decreased but never fully removed. The Bass and Treble pot have a pretty good amount of swing but cant easily be flat. They are also more interactive as compared to the baxandall but its not extremely interactive. One particular thing to note is that a Fender FMV is driven from the plate of the triode driver stage and loads down the signal at the plate. This is one reason the insertion loss is nearly as high as the baxandall. The other reason would be the 10k mid pot. If the value of the mid pot was increased then the amount of insertion loss would decrease.
For reference here is the response with every control dimed
So the response is also quite far from what one would imagine it to be. Not to say that someone might like the response, that is for each to judge.
The response of the bass, treble, and mid controls can be adjusted by changing the values of the caps, resistors, and pot values in the circuit. This a wide range of adjustment can be made but the overall pros and cons will remain regardless.
4. The Marshall FMV
While I am calling this the Marshall FMV, the actual first incarnation of this circuit was done by Fender on the 50s Bassman amps which were copied by Marshall with the JTM45. The reason I am calling it the Marshall FMV is that Fender decided to change from this design to the 60s design while Marshall kept this design and became famous for the sound it created. However, one should give props to where it first appeared in amp history.
Here is the schematic for the Marshall FMV. this is from a Superbass. The Superleads have slightly different component values and different response.
Let me discuss some differences between this EQ and the Fender one. The first main difference is the added triode which is being used as a cathode follower. A cathode follower is a circuit which is used as a buffer, where the input impedance of the circuit is very high and the output impedance is fairly low. The original purpose of this stage is to decrease the insertion loss associated with the FMV. Since the cathode follower has low output impedance it is not loaded down by the FMV tone stack and thus is less lossy. Another difference is that there are different values for every component in the EQ, and a slight difference in the mid pot wiriing, and the pots have different types of sweeps, (log or linear).
Lets see what the EQ looks like with the controls at noon.
So the biggest difference you can see is that the insertion loss is much much lower. Almost 10 dB more across the band compared to the Fender FMV. However there is still the mid cut at noon on the controls, however it is centered a bit differently hear with the mid cut happening closer to 1k hz where as in the Fender it is broader and cuts more mid frequencies. The result is that there is more apparent low mids in the Marshall EQ.
For some further reference here is the Superlead controls at noon.
Even less insertion loss, and more low mids. This change occurs from changing the 56k resistor to a 33k resistor. This resistor is know as the slope resistor and as you decrease it the mids are more and more apparent. However decrease it to much and the whole EQ will not longer have much control since it will be practically bypassing everything.
now back to the superbass spec.
Here is the bass pot sweep
treble pot sweep
mid pot sweep
So since the EQ has less insertion loss, the Bass and Treble can almost go through without any cut while again the mids are limited here. The bass and treble have some decent range but generally people interpret the Fender tone stack as having more control over bass and treble. The Marshall FMV has the advantage of having more output signal tho, and more mids which often results in more gain and clipping. Something that is often desirable for guitarists.
Finally here is the Marshall FMV with all controls maxed
Same kind of thing as the Fender, inescapable mid cut but more output.
There are still a lot of other differences between Fenders and Marshalls beyond the EQ but I will not get into that here. Not the right topic.
Now the last graph here is how Lemmy EQs his amps, Bass and Treble to 0, Mids to 10.
The EQ almost flat, but there is a low end roll off style with a 3db cutoff around 80hz. There is also a ton of loss now with the controls down. This is not a huge problem if you do not need a ton of clipping and for bass overdrive this is probably not a bad way to rock the amp.
One thing to remember is that what sounds good is not always what is ideal. So if you like the Fender tone stack thats fine. If it sound good, then it is good. However I hope that a little bit of techno-babble here will help people understand what is going on underneath the hood. Thanks
Some very fine analytical work here, thanks very much for showing me this as it contains impossible to readily find information like the inductor values that Ampeg have used.:bassist:
so which EQ is best with 15 inch speakers :hiding: ;)
Whichever EQ you like the best Sam, thats the answer. :D
Are you posting this because you know I am GASing for a build from you?
but in all seriousness I was just trying to demystify some of the inner workings in amps.
Its all a very interesting discussion, I will have to direct my guitard here as he uses a JCM800 or a HD130.
I have also been thinking if I could add the SVT mid control into the GT120 preamp to add a true mid control to that amp.
The 2203 (jcm800) used the same EQ as the superleads while i am not sure about those HD130s.
I'll take both of those if you build them.
You don't want repeat customers?
I just want a quadraphonic Dunwich setup.
|All times are GMT -6. The time now is 11:29 PM.|
Powered by vBulletin® Version 3.6.12
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.