Radio Shack Reverb Mods?

[Arn C.]

Someone mentioned this reverb on another thread and I was wondering if anyone had any info on modding one of these.  I don't have a schematic but I have the board out in front of me.  It has two jrc4558's , MN3207 and MN3102.

Thanks!
Arn C.

[Mark Hammer]

Most BBD-based echo boxes (and despite what Radio Shack calls it, it is an echo unit) lack appropriate filtering for mimicking the sound of natural reverberant spaces.  Ideally, what you want is some lowpass filtering in the regen loop to mimic the way that earlier reflections have more bandwidth than later reflections.  This can often be a simple matter of tacking a cap between input and ground on the regeneration amount pot.  The choice of value is up to you but .01uf is usually a good place to start.

To avoid runaway feedbacnk, many BBD-based devices use a combination of trimpot and chassis-pot to adjust regen.  The chassis-pot is maxed and the trimpot adjusted to the point where the regen stops oscillating.  Because the filtering of regen signal will reduce the apparent amplitude, you may be able to adjust the trimpot to permit more signal to pass through the regen path.

Note as well that while longer delays and more regen produces much poorer quality sound out of BBD's pushed to their limits (remember, each iteration acquires "error" due to leakage and other things intrinsic to BBD's; passing through a BBD 4 or 5 times is not a recipe for fidelity), chopping the bandwidth can make the signal *appear* to have not lost much fidelity.  So, to make it more concrete, if I have 4khz lowpass filtering of the delay signal post-BBD, but no filtering in the regen path, it will sound worse and worse (nd noisier and noisier as well) with each iteration through the regen path.  IF, however, I add 6db/oct of lowpass filtering in the regen path starting at 1khz, with each iteration I will tend to lose more and more of the crap that tends to accumulate with repeats.  In a sense, this acts more or less like a 1-pole filter the first time through, a 2-pole the second time, 3-pole the third time, etc., providing more treble attenuation for things that have already passed through the regen path, but less for content that is newly introduced.

With that cheap and cheezy "noise reduction", it may also be possible to up the capacitor value of the MN3102 and squeeze a little more delay time out of the MN3207, though not THAT much.

[Arn C.]

Thanks Mark for the tips!   I will play around with it a bit and see how it works out for me.    

Assuming this was built for stereo equipment and mic use, could I actually plug my guitar into the mic input or should I use the line in input and if so will I have to change any components?

Thanks Again!
Arn C.

[Mark Hammer]

Chances are pretty good the guitar signal is too hot for a mic input (few, if any, voice mics will pump out 100mv or more), and too cool for a line input.  On the other hand, any clean booster can rememdy the latter situation, and if I'm not mistaken the unit's user manual has a schematic which might help identify components that set gain on the mic input (assuming there is one).

[Vsat]

A very simple and effective mod is to replace the MN3207 with an MN3208 or MN3205 - get 2x or 4x the delay then. Requires some rewiring since the pinouts are different, but otherwise compatible. If you have a 2nd MN3207 kicking around, can piggyback-solder the extra MN3207 on top of the first one - the clock pins can be shared, as well as the power supply and bias pins. The two output pins and the input pin are bent-out to hook up a bias trimpot and coupling cap to the input, plus an clock null trim and associated resistors for the two output pins. The output of the first BBD goes into the input of the 2nd. Sounds more than 2x as good as the original when yu do this. Modded a unit to have three MN3207 piggybacked this way once... the MN3102 should be able to drive four of the MN3207. No compander chip though, so a null trim is mandatory to minimize clock-whine.
Cheers, Mike

[Dan N]

Somewhere in the archives someone modded one of these for guitar. I recall a schematic being uploaded, and a few parts changes. Also, recall the poster being pretty happy with the results!

SEARCH!

[noesis]

this is the schematic:
http://www.toothpastefordinner.com/realistic-reverb-schematic.gif

and this is a previous post on the subject:
http://www.diystompboxes.com/sboxforum/viewtopic.php?t=1198&highlight=realistic+electronic+reverb

I changed R53 on my reverb to get more output volume, but haven't tried any other mod.

[Peter Snowberg]

That was my very first "pedal".

Thanks for the schematic link noesis.

The first mod I would suggest is to change the bias resistors R1 & R2. Make them 2.2M or at least 1M.

2nd mod is to replace R52 & R53 with a single pot of 10K to 100K feeding the junction of those two resistors. There's your output volume control.

Congratulations, you are now the owner of a Realistic Guitar Fuzz-Echo.

I tried adding series resistance to the clock generator until you could hear the clock clearly in the signal..... a fine effect this is not....

[Mark Hammer]

There is something not quite right about that schematic.

When was the last time you saw a BBD circuit with EIGHT poles of lowpass filtering after the BBD but NONE before it.

Notice that IC1B doesn't really get a signal from anywhere?  Weird, and unlikely to be true.

I, for one, would be curious as to what happened to the bias voltage.  I can see Waldo in there, but I can't for the life of me spot the bias.

On the other hand, assuming the circuit around IC4-A is accurate, the combination of 1500pf and 18k yields a lowpass rolloff starting around 5.8khz, which is kind of high if the intent is to make it sound like reverb.  The  33k/470pf combo after the op-amp has a rolloff over 10khz, which doesn't help much.  Upping the value of C15 to 3300pf will reduce the rolloff so that it starts around 2.7khz.  If you bump up to 3900pf, that rolloff drops down to 2.2khz, and if you go up even more to 4700pf, it drops to 1.9khz.  If you accompany that 4700pf cap with a 22k resistor to bump the gain up a bit to offset the drop in perceived loudness because of treble loss, the rolloff becomes around 1.5khz which is probably about right for mimicking a true physical reverberant space.

[Rodgre]

By the way, I always loved this thing as a fuzz box for the guitar. At one point early on, I even ran the line out to a stereo speaker and cranked the level for a great fuzztone.

The funny thing is that I just got a PLUG-IN version of this for PROTOOLS!

Roger

[Top Top]

A very simple and effective mod is to replace the MN3207 with an MN3208 or MN3205 - get 2x or 4x the delay then. Requires some rewiring since the pinouts are different, but otherwise compatible. If you have a 2nd MN3207 kicking around, can piggyback-solder the extra MN3207 on top of the first one - the clock pins can be shared, as well as the power supply and bias pins. The two output pins and the input pin are bent-out to hook up a bias trimpot and coupling cap to the input, plus an clock null trim and associated resistors for the two output pins. The output of the first BBD goes into the input of the 2nd. Sounds more than 2x as good as the original when yu do this. Modded a unit to have three MN3207 piggybacked this way once... the MN3102 should be able to drive four of the MN3207. No compander chip though, so a null trim is mandatory to minimize clock-whine.
Cheers, Mike

Bumping an old topic here... old...

I just got one of these at a junk store for $3. I have a spare 3007 laying around (not 3207). They are both 1024 stage BBDs and have the same pinouts.

I realize the poster above could be long gone, but does anyone know if I could do as mentioned above with 3007 on top of 3207? I don't know enough about the different BBD chips to know the differences between the two chips.