If you had to put an effect in with reverb and delay, it would be...

[Ben N]

I guess another question I have would be should I put the orange squeezer and the highpass filter before or after the delay section?

Assuming these are all in the same box, I would think the compressor would go before the delay and the highpass after. But that may depend on how you use them, i.e. if you intend on using the delay separately from the reverb, you may want to be able to have the compressor off when the reverb is off.

Actually, upon reflection, the compressor and highpass should be spliced into the reverb circuit between the input buffer and the driver--Otherwise you are cutting dynamics and bass out of the dry signal, which you do not want to do (see Mark's posts above). That would be around R4 and C4, looking at the GGG schematic of the Anderton circuit. Somebody smarter than me will have to tell you exactly how to achieve this. Also, a better place for a tone control may be after the dwell control, where it only affects the reverb and not the dry--especially if the whole shebang is going to get fed into an amp input, where you presumably have perfectly adequate "master" tone controls. A passive single control would be easy to splice in here, since you are already running wires from the board to the dwell pot, but I don't know how bad the signal losses would be--you may need to increase the value of R9 or add a recovery stage after the tone stack.

One more thought, now that I am looking at the schematic: If you run the delay in series with the reverb then both the dry and delayed signals get reverbed. An alternative that might be sonically useful is essentially to run the delay in parallel with the reverb: Feed the delayed signal (delay only) into the reverb circuit at the mixing node (the junction of R1, R3 and R5). If you do that you can leave out the delay unit's own input buffer and mixer altogether (although you lose the ability to put reverb on the echoes). In this scenario, the order is:

Stage Center input buffer (IC1A), which feeds, in parallel, the following three paths:
     > Orange Squeezer > hi-pass filter > reverb driver stage (IC1B) > reverb pan > recovery stage (IC1D) > dwell & tone controls
     > delay (without internal dry signal path) > 50k delay mix pot
     > dry signal > 50k dry mix pot (R10)
all of which converge at the inverting input of IC1C.

There are a lot of possibilities--including the "normal" of just running everything in series--this is just an alternative suggestion.

Also, can I just make the highpass filter a switch, basically more or less bass, since I'll be having at least a treble and bass control in the tone stack anyway? Or is it more important to have exact control over the bass going into the reverb?

Yes and yes. However, you should probably first play with it using either a pot or various fixed resistor values so that the amount of cut you get with the switch is optimized. Maybe a good compromise would be an internal trimpot that you could set in advance, and a switch on the panel to switch the pot in or out of the circuit.

Edit: I just read Mark's response, which was posted while I was writing the above. I don't see any contradictions. I hope it isn't too confusing... 

[mth5044]

Haha, you can hope all you want, but it sure is confusing. For me atleast. I'm still looking over the reverb schematic to locate where all this stuff is going.

I guess I should make some clarifications. I want the delay to be totally independant of the reverb. So when considering this stuff, it should be only the compression, eq and the reverb. I guess it would also be cool to have more of an eq 'section' (which would include the compressor, the tone stack and the highpass filter) where there would be a switch for the reverb, then a switch for the eq section of the reverb. It would also be cool if I could just have the eq section be on without the reverb, but that is not necessary and will probably make it more complicated because the 'section' will be inside the buffers of the reverb circuit (if I understood what you guys are saying). So I guess it would have the options of; Reverb off Eq off, Reverb on Eq off, and Reverb on Eq on.

That would be as simple as having a dpdt switch, where the output of the input buffer on the reverb goes to the a middle lug, which switches to either the compression/highpass (which then goes to the reverb driver) or just to the reverb driver, giving the option of w/wo compression/highpass. The other side of the dpdt would go right after the dwell pot, then split to either the tone stack (which then goes to the reverb recovery stage) or right to the recovery stage.

This way, one switch will turn on/off the entire eq section. Also, I'm planning on using two trimmers from the compressor, so it will just be an on/off switch like they used to be, once the volume and compression is set on the inside. They will also be easily adjusted because the top of the box can be opened with two latches. So, with the 4 knob spaces, 1 switch, and 1 on/off switch areas avaliable, I will fill them with-

on/off for eq section
on/off for orange squeezer
highpass filter knob
treble
mid
bass

Oh God I hope I understood what you guys are saying. If so, I'm going to have to study these schematics some more. I can get the orange squeezer pretty easily, I have no idea about where the buffers, recovery and stuff like that is on the reverb (I know you said what they were, but I'm still trying to get it into my head), and I'm still a bit confused about the highpass filter. I know a basic tone control is just an input into a pot, 2nd lug is out and third has a cap to ground. You said something about a dualpot and stuff, so I'm going to have to re read and look around.

Thanks again for all your help. I have some thinking to do 

[mth5044]

hopefully this isnt too big. This is my understanding of what the entire reverb section should look like. I like pictures, so this will definitly help 

[Ben N]

I'll try to be a little clearer by explaining parts of the circuit.

By way of background, you should read up on opamps--there is a good article at Geofex.

Look at the schematic (http://www.generalguitargadgets.com/diagrams/stage_center_reverb_sc.gif).
From the input jack, the signal passes the usual input cap and resistor on its way to the first opamp, labeled IC1A. The signal enters through the input labeled (+), so we know that it is a non-inverting opamp configuration. Now note the line between the output terminal of the opamp and the inverting input (-)--there is no resistor there. That line is called the feedback loop, and where there is no resistance in the feedback loop of a non-inverting opamp, the gain is 0. The main purpose of that stage is to serve as a buffer, or current source--to provide a low enough output impedance to be able to drive all the stuff that follows--and to isolate the circuit from what comes before, like a pickup. (There are a couple of great articles on buffers at AMZ.) Note that the signal splits after the buffer, and that is where you'd have havoc if there was no buffer to guarantee that each signal path sees a consistent, low impedance feeding it.

Then part of the signal goes "up" to the Mix pot--that is the dry signal path. The other part goes "down", through a coupling cap and resistor to the next opamp, the one labeled IC1B--that is the reverb signal path. That cap is already restricting the low-end content that gets through, and you may be able to get away with just tweaking that cap by using a higher value to remove more bass, although a true filter, consisting at a minimum of a cap + a resistor to ground, will give better performance. This second opamp is inverting, but it has a resistor in the feedback loop, which tells you that the gain is greater than 0. The gain (not taking account of the caps) is calculated by dividing the feedback resistor value by the input resistor. 470k/22k=21.36--not a lot, but the main issue here is provide current drive. The cap in the feedback loop has the effect of filtering out very high frequencies (because the frequencies that pass through the cap to the input without being blocked by a resistor are essentially canceled out), which cuts noise and improves efficiency. That opamp actually provides the push to the transducer at the input of the pan that is strong enough to send a signal through the spring.

At the other end of the spring there is another electromechanical transducer (or pickup) that turns the movement of the spring into a little electrical signal. That needs to be amplified to be useful, so it goes to IC1D, another inverting amp with a gain of about 47. The "Dwell" control is for how much of the reverb signal is going to be used--really, a mixing fader. Note that the reverb signal is now in phase with the dry signal since it has been inverted twice.

Both the dry and wet signal paths meet at the inverting input of IC1C, which is configured as a mixer.

Don't worry about what I said about a state variable filter--you don't need that. A simple hi-pass filter (series cap, resistor to ground) will do just fine, or maybe a 2d-order (two of those in a row) for sharper cutoff. You'll have to do some math and/or experiment to get the values right, and you may have to adjust R7 for more gain.

As for your controls, I don't understand why you need both an eq switch and a highpass knob--seems to me you'd want the knob for tweakability, or the switch for simplicity, but not both. As Mark said, a low-pass filter is really all the eq you need on the way in. The idea here is not to have an EQ (a seprate effect is better for that) but to have an eq that optimizes your reverb, and for that the eq should be in the reverb signal path only.  

Anyway, it looks like you are on the right track with the switching. The diagram looks good, except that I would put the 1/2-DPDT after the OS/HP, on the way into the reverb driver, rather than on the output of the buffer. That way, when the OS/HP are not in use, they are not adding noise.

Cool project. Good luck.

[Mark Hammer]

Same schematic.  R4/C4 yield a 1-pole highpass function (decline of 6db per octave) with rolloff starting below around 330hz.  That already takes out some of the bass, but a 1-pole highpass still leaves much of the bass in there.  Remember, the content at 165hz is only 6db lower with that filter.

Flip the order of R4 and C4.  Now make C4 two .022uf caps in series.  If you left it like that, those two caps in series would be equal to .011uf, giving you a rolloff starting around 660hz; a tad on the high side.  What we'll do is tack on a 100k resistor to ground after each one of those caps.  That gives a 2-pole rolloff (12db/octave) below 72hz.  Low, but that's your full bandwidth.  Now stick a second 100k resistor in parallel with each of those 100k resistors.  That gives a combined parallel resistance of 50k for each pair.  Your bass now rolls off at 145hz.  Okay, now replace the paralleled 100k with a paralleled 47k.  That makes your effective parallel resistance 32k in stead of 50k.  That now gives a rolloff starting around 226hz.  That may seem lower than the original rolloff provided by the schematic, but actually because of the sharper slope it reduces bass even more.

If you wire up a 3-position DPDT toggle to either switch in the pair of extra 100k resistors or extra 47k resistors, or nothing (centre position), you'll have 3 different bass rolloffs at 72hz, 145hz and 224hz.  If those .o22uf caps were .015, you'd have rolloffs starting at 106hz, 212hz and 331hz.  Again, note that the rolloff of a 2-pole filter at 331hz will be steeper/sharper than that provided by the 1-pole filter shown in the schematic.

Since a lot of the volume lives in the bass region, trimming more bass, or letting more bass back in, means you have to adjust the gain of the stage that drives the springs appropriately, so that it is, in Goldilocks' terms, not too hot and not too cold - juuuuust right.  If you replace the 470k resistor shown with a 150k fixed resistor in series with a 500k pot (wired up as a variable resistor), you'll be able to adjust the gain from 6.8x to 29.5x.  Currently, it is set at 21.4x.

Note that in the original Fender tube reverb design, what got called "Dwell" was actually a control similar to the one I just described.  That is, it adjusted how hard the springs were pushed by adjusting the drive gain.  I gather they called it "Dwell" because if you smack the springs hard, they wobble for a longer period of time before they die out.  Not quite a "Decay time" control, but somewhere in the neighbourhood.

Because the treble rolloff of the signal driving the springs depends on both C1 and R7, you'll need to reduce the value of C1 to compensate for the manner in which we've changed R7.  You probably want to go with 150pf.

IC1d and IC1c have no capacitors in the feedback loop (in parallel with R9 or R5).  Since the spring mechanism is not a particularly efficient means of transmitting signal, you need a lot of gain to recover it, and some of the "signal" you get from that stage will be hiss.  So, it is a good idea to stick a 33pf cap in parallel with R9 and a 390-470pf cap in parallel with R5.  That will still give you plenty of breathing space for whatever tonestack you have planned to do some good.

[mth5044]

So.. another question.

Tone stack.

ROG Tonemender?

http://runoffgroove.com/tonemender.html

I've been reading around about the tonestacks and how they all have volume drop problems. Is putting this buffer into the signal before the output buffer of the reverb circuit going to cause difficulties, or am I going to be able to just drop this in there after the dwell pot?

Finally, I appreciate the incredable write up on the highpass switch, but since I'm just making the compresser a switch and putting the trimpots on the pcb, I will have another pot hole opened for a knob for the highpass thing. So now I have to figure out exactly what that looks like, and if that Tonemender works, I'll have all the parts I need 

Then comes the task of figuring out where exactly everything goes from your guys' excellent descriptions
Thanks again.

[mth5044]

Well, after a bit of mulling around, I desided to try and get this project over with. So I'm making more pictures to try and understand.



So far, I have the tonemender right after the dwell, as was established a while back. I'm not totally sure where the orange squeezer and the highpass filter will go. Like you said, somewhere around C4 and R4, but where exactly? Also, I'm not 100% (still) how the highpass filter is set up, so a simple google came up with that. While I'm not completely familiar with schemaitcs, I've only seen those triangle things in IC chips and whatnot, so that may be wrong. I know you have already given me how it looks in words and with examples, but I'm really dense atm 

So, hopefully someone can come along and say exactly where the orange squeezer and hp filter would go. I could experiement, but with the PCB all populated and whatnot, soldering and desoldering components isn't up my alley.

[hoyager]

Sorry to resurrect an old thread

Mark if you're out there, or any else who would know, is this the way to wire up the 2 pole high pass filter after the input buffer, and before the tank driver?



the TL072 is the input/output, lefthand 5532 tank driver and follower, RH 5532 1 band of the PAIA 4 band eq

http://www.harpamps.com/schematics/4bandeq.pdf

And a low pass filter in form of the SWTC.

Top right is the bipolar 9v to +/-15v charge pump / voltage doubler.

Andy

[oldschoolanalog]

oops!