Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: Joel on December 07, 2017, 05:57:07 am
-
A friend asked me to look at building him a Princeton Reverb - but without the reverb or trem and having an FX loop instead. The idea being to have all the delay/reverb/etc off board using stompboxes.
I started by borrowing an FX loop from one of Tubenits ODS amps and grafted it into a AA1164 schematic where the reverb section would normally go - as shown in the first schematic. In the v2 schematic I simplified the loop a bit so it only has one pot.
Will the v2 circuit work well? I'd love some opinions from the experts. Got any suggestions? Opinions welcomed.
Cheers.
-
Doesn't look like the experts are going to render any opinions, so I'll respond.
I don't like parallel loops because a lot of FX devices already have mixing capabilities. But if you just gotta have one, here are some points:
The 12AU7 is going to consume 4.5 to 5mA, so the 18K dropping resistors need to be changed to 10K.
The signal coming off of V1b is going to be too large for the FX Send unless the Volume is turned way down, so I put a voltage divider in front of V3a. It's not much of a divider, but you might want to have the Volume very low and still have enough for the Send signal. Presumably your FX device would have an input level control. The voltage divider also prevents the V3a cathode follower from clipping if you really drive V1b.
The dry signal needs to be lower at the V2a grid in order to have more wet signal and still not slam V2a. If you want to turn the amp up, you will probably have to adjust just about everything.
You wouldn't have the loop being a second dry path with no FX device connected, so I rearranged that.
I think you would want the loop to have a flat frequency response, so I eliminated the treble boost.
-
I think you can have a fixed resistor for FX send instead of a pot. Same thing for FX return. I would want a pot to adjust FX level volume at the end of the FX signal chain. I'd probably look at a 250ka pot. On my amps with active FX that's just about all I adjust on the FX.
You could also try a 12AY7 for FX.
With respect, Tubenit
-
I've always used line level rack-mounted effects in my FX loops and never paid much attention to stompboxes. Now that I have looked at a bunch of stompboxes, input level control is a little bit rare. So I designed a stompbox friendly Princeton with no reverb or vibrato.
-
That could work.
But why do you loss-down, gain-up, loss-down? Image 1.
I approve of buffered output. But if it only drives guitar-cord stuff, a 100K pot is low-enough impedance.
This leads to a far simpler plan, image 2.
If you actually want the small distortion of the added stage, OK. But seems to me the effect loop is about the effect, and the main amp, and should not add any color of its own.
At the other end.... the power stage driver needs 0.3V-0.6V RMS at least to get the power bottles full-roar. This is at the upper end of what some guitar-cord boxes give. It is fairly common to have a 0.6V (peak) clipper and then some loss to get toward the more typical 0.2V on a guitar cord and main amp input. I'm wondering if you need a small gain after the loop.
-
That could work.
Really? Thanks.
But why do you loss-down, gain-up, loss-down? Image 1.
I loss-down and gain-up because that is what a Princeton does. I then loss-down for a stompbox level signal.
I approve of buffered output. But if it only drives guitar-cord stuff, a 100K pot is low-enough impedance.
10K is my general rule. A 30ft cable with 30pF/ft will noticeably shear off the highs around 2KHz with 100K. The buffered output will drive the whole spool.
If I change the load on V2A to a 100K resistor in series with a 10K pot, it will load V2A down too much as compared to what happens in a Princeton. I could change the Rp on V1B from 100K to 10K in series with 91K and take the 100K/10K divider from the 10K/91K junction for the FX send and take the stock bypass from the plate, I suppose.
This leads to a far simpler plan, image 2.
Which brings me to the second reason for my plan: It just LOOKS a lot more hi-tech.
If you actually want the small distortion of the added stage, OK
OK
But seems to me the effect loop is about the effect, and the main amp, and should not add any color of its own.
You must consider V2A to be part of the effect loop. I considered it to be one of the two ways to get the Princeton third stage so that you have the ability to overdrive the third stage with the bypass switch in either position. It's the third stage that gives the Princeton its tube color and that is why I went through so much trouble to get it pre-loop and still have the ability to revert to a stock Princeton.
At the other end.... the power stage driver needs 0.3V-0.6V RMS at least to get the power bottles full-roar. This is at the upper end of what some guitar-cord boxes give. It is fairly common to have a 0.6V (peak) clipper and then some loss to get toward the more typical 0.2V on a guitar cord and main amp input. I'm wondering if you need a small gain after the loop.
Yeah, that bothered me, also. But then again, 0.2V is going to make a loud noise in your bedroom. If you need more, you have the wrong amp and the wrong venue.
Other than clippers, a lot of those boxes will do considerably more. I didn't really want to add another gain stage, being too hi-tech already.
_ _ _ _
Now that I actually looked at a Princeton schematic, I noticed that the reverb setup is slightly different than the bulk of the Fender's. So I revised my schematic.
-
Some more FX loop ideas (from Merlin's 1st designing tube pre-amps book), using 1 x 12AX7
-
Some time ago I give help to a friend that build a modedd 5E3 with FXLoop (it is from Merlin's book with reference to the Engl Fireball)
and .... it works fine
(https://i.imgur.com/9XS2bAx.jpg)
Franco
-
Wow! Thanks guys! There is some really good info here.
I've also looked at a Dumbleator schematic and a LND150 loop that's supposedly the Mojotone FX loop.
I think I'll just need to choose something and build it and see how it goes. At least now I'm slightly more informed. I may need to muck about with some level pots and a scope to set the loop to unity gain.
Really appreciate the input guys. Thanks. :worthy1:
-
Mr Merlin sell (on the cheap) a PCB for FXLoop
http://www.valvewizard.co.uk/fxlooppcb.html (http://www.valvewizard.co.uk/fxlooppcb.html)
(http://www.valvewizard.co.uk/fxlooppcb1.jpg)
(http://www.valvewizard.co.uk/fxlooppcb2.jpg)
Universal Amp PCBs
Building-block PCBs for guitar amps. 100 x 63mm.
£3.50 each
Franco
-
. . . a LND150 loop that's supposedly the Mojotone FX loop.
Does it look like this? The older version was slightly different.
-
Yeah that LND150 loop schem is very similar to the one I found.
Merlins FX loop looks good.
-
Latest version. Using Merlins FX loop schematic. The input divider is taken from some advice Merlin gave someone over on Amp Garage - recovery stage has gain of approx 60, so the input divider should be the inverse of this or 1/60th.
-
Are you missing a 100K plate resistor?
-
Also to me seems so
(https://i.imgur.com/qcSClIE.jpg)
(http://www.valvewizard.co.uk/fxlooppcb2.jpg)
Franco
-
If you aren't going to get a little color from the third stage before the FX loop, then what PRR said in Reply #4 makes a lot of sense. If you also don't have a bypass switch and you want the third stage to be driven about the same as it is in a Princeton, then the attached is one idea.
-
If you aren't concerned about adhering to the Princeton levels and you don't care about the difference between cathode and plate signals, then the whole thing can really be simplified.
-
Are you missing a 100K plate resistor?
Yes. Yes I am. How did I miss that?
-
If you aren't concerned about adhering to the Princeton levels and you don't care about the difference between cathode and plate signals, then the whole thing can really be simplified.
The whole idea is to preserve as much of the Princeton Reverb sound as possible. To me that means as much of the circuit design as possible, and therefore the FX loop shouldn't colour the signal. Of course, I may be wrong about preserving the circuit.
-
The whole idea is to preserve as much of the Princeton Reverb sound as possible. To me that means as much of the circuit design as possible, and therefore the FX loop shouldn't colour the signal. Of course, I may be wrong about preserving the circuit.
We're not going to preserve the circuit if an active loop is inserted, but we can preserve the relative signal levels.
As an example, let's say that we have a 50V signal at the plate of V1b. The stock Princeton has about a 500K/3.3M voltage divider, so the grid of V2a gets about 6.6V.
Your Merlin loop has an 18K/1M voltage divider, so the grid of V3a gets about 0.88V. If V3a has a gain of 35 (plausible), it will have 31V on the plate and 0.31V on the cathode. We now need a gain of 21 to 22 from V3b to get the signal close to the 6.6V that a stock Princeton would have. An unbypassed 1K cathode resistor and a 47K plate resistor on V3b is going to give a gain that is pretty darn close to what we want.
-
I had a minor eureka moment after reading your post. I began remembering a bunch of stuff I'd learned 20 or so years ago - theory and practice began to merge. I always knew I was a slow learner... But 20 years? Well, at least I know in another 20 years I'll be almost half as knowledgeable as you fine gents. :laugh:
I will need to measure-test-adjust on the amp once it's built. I've adjusted the 18k/1M divider to 27k/1M (I'm assuming gain of 60 for V3a and V3b). Once built and measured, it should be simple to adjust either input divider or recovery gain to sort out the signal levels.
-
I've adjusted the 18k/1M divider to 27k/1M (I'm assuming gain of 60 for V3a and V3b). Once built and measured, it should be simple to adjust either input divider or recovery gain to sort out the signal levels.
The 18K/1M divider does a real good job of matching V1b's output to V3a, so I wouldn't change that.
I figure a stock Princeton with 50V from V1b will deliver 6.2V to 6.6V to the grid of V2a depending on where the 100K Reverb pot is set. Attached is a suggested loop with predicted signal levels that should get you pretty close when no effect device is connected.
-
I was looking at your improved bias and I don't think it will give you the required bias voltage. I have attached a bias circuit with different values.