G. Forrest Cook Reverb

[yodude]

Looks cool.  I'm going to jump in and build one.  Link roundup:
Cook's page: http://www.solorb.com/elect/musiccirc/reverb2/
Gaussmarkov's page: http://gaussmarkov.net/wordpress/circuits/spring-reverb/

Trying to figure out which transformer to use.

I have an unmarked Accutronics tank that I pulled from an old Fender M-80.  According to google, it's an 8EB2C1B which means it has an input impedance of 800R (is there a way to measure it?).

The gaussmarkov standard BOM lists:
- an 8AB2A1B tank, which has an input impedance of 10R.
- a 42TM018 transformer, which is 10K:10K.
But the gaussmarkov commentors seem to think a 42TU013 transformer is better, which is 1K:8R.  The secondary of 8R seems to make sense with a 10R tank, but why the big difference in the primaries?

Cook's project page says:

The transformer matches the impedance of the driver amplifer to the reverb driver coil and allows a dual phase driving signal to power a reverb coil with one grounded side. The transformer is a standard "70 volt" audio line transformer that is often found on PA systems. One reader reported having good results using a Mouser 42TU013 (1K to 8 ohm) transformer. If you can find a reverb tank with a high impedance driver coil, the transformer may be eliminated, the driver coil will require isolation from ground.

Does anyone have a part number for a "70 volt" audio line transformer?

Does my 800R tank have a high enough impedance to eliminate the transformer?

Thanks for any input.

EDIT

Was looking at Cook's page again and realized I missed his transformer call out:
Speco 7010 70 Volt
Can't find a datasheet, but a seller says: Speco's 70 volt 10 watt line transformer has wattage taps. It has 10, 5, 2.5, 1.25, and 0.625 watt taps.

[yodude]

this here's a good old fashioned bump

[yodude]

The circuit has 3 ICs: 1458, LM833N, 4011N.

Smallbear has LM1458 and LM833.  Will those sub in for the 1458 and LM833N?

Can somebody point me to a Smallbear-available sub for the 4011N or even sell me one?

Looking at the schematic, the 1458 and 4011 seem to be involved exclusively with clipping detection.  Would the circuit be ok if I just left that whole section out?

Thanks!

ps As a small gift to future builders, here's the circuit BOM as condensed from gaussmarkov's parts list:

qty   desc   desc
1   1uF_NP   gm-caps-film-box
2   10uF   gm-caps-electro-pol
2   100uF   gm-caps-electro-pol
8   100nF   gm-caps-film-box
2   20pF   gm-caps-ceramic
2   1uF   gm-caps-film-box
1   1000uF   gm-caps-electro-pol
2   220uF   gm-caps-electro-pol
1   1458   gm-ics
1   LM833N   gm-ics (NE5532 on solorb)
1   4011N   gm-ics
2   LED5MM   gm-diodes
2   2N3906-   gm-trans
1   2N3904   gm-trans
1   27   gm-resistors
3   1K   gm-resistors
10   10K   gm-resistors
3   100K   gm-resistors
2   1.2K   gm-resistors
3   1M   gm-resistors
2   680   gm-resistors
1   47K   gm-resistors
3   100   gm-resistors
1   10   gm-resistors
1   47   gm-resistors
1   220K   gm-resistors
1   5K   gm-resistors
3   10K-A   gm-pots
1   100K-A   gm-pots
2   20K   gm-pots (trimmers)
1   SPST   gm-switches

No guarantees though!

[petemoore]

800R (is there a way to measure it?).
  I'm sure there is but I trusted the cross-code numbers that I got at Accutronics, most of them have relevant markings, I have a rare one here without ID marks.
  The gaussmarkov standard BOM lists:
- an 8AB2A1B tank, which has an input impedance of 10R.
  That's a bit off from the 10% rule, 10/800th's isn't a good impedance match.
- a 42TM018 transformer, which is 10K:10K.
  I'm not sure what the impdedance ratio is on the 42Tmo18, nor whether that will provide good transducer control. 

[yodude]

Ok all wired up.  I'm now the proud owner of a perfectly functioning volume cut.

I built it per the plan outlined above.  No clipping detection circuit and no transformer (the tank impedance is 800R).

The signal splits after C4 (using Gaussmarkov's version of the schematic).  Part of it goes through R5 to the output mixer amp.  The rest goes to the drive pot, the reverb driver amp, the transformer, the tank, and a recovery amp before heading to the output mixer amp.

The audio probe shows the R5 signal making it all the way to the output, but the signal through the drive pot drops to inaudible after R14.  Jumpering R14 has no effect on the output.  R14 was reflowed and inspected with a magnifying glass.

Any thoughts?

[PRR]

> the signal through the drive pot drops to inaudible after R14.

That is correct. Study "inverting op-amp". If IC2A had infinite gain, the signal voltage on the "-" pin would be zero. IC2A gain is not really infinite, but very-very-very large, so the "-" pin signal voltage is very-very-very teeny.

Current in R14 is not absorbed by the "-" pin, the op-amp is connected to force a compensating current through R15. Since R15 is twice the size of R14, the voltage on the right end of R15 is twice that at the left end of R14. Gain of "2" from drive pot wiper to IC2A output pin.

So.... did you get past R14? To IC2A output?

Did you measure the _DC_ voltages at each IC2 pin? If something is screwed-up, half the time it will show as odd-ball DC voltage readings. Aside from the power pins, all pins on IC2 should read half the supply voltage as seen at the R18 R19 divider.

Even if the spring driver is screwt-up, the reverb recovery should work. Smack the tank.

[PRR]

> no transformer (the tank impedance is 800R).

The driver requires a load where neither side is grounded. Some reverb tanks have one side of the drive coil grounded. That's not mandatory. You may be able to snip a jumper, or remove both coil-leads from the RCA jack and extend them to IC2,

[edvard]

The 70 volt transformer with wattage taps sound suspiciously like a 70-volt intercom transformer I picked up at Radio Shack years ago, hoping to put it to use as a low-voltage tube output tran.



Radio Shack still sells them for around $15 USD, but you may be able to find one cheaper elsewhere.
Just search for "70-volt line matching transformer", you'll find it.

[yodude]

Thanks guys.  I appreciate the help.  Regarding the transformer, I'm hanging my hat on Cook's statement that the transformer isn't necessary if the tank has a high impedance.  My tank is spec'd at 800R.  I don't yet know if that's high enough, but it's a lot higher than the 10R tank assumed in design.

The "IN" side of the tank isn't grounded.  Banging on the tank does not produce any sound at the output.  But when I plugged my ipod directly into the circuit at the tank return, I did get sound at the output.

With the ipod back at the input jack, the audio probe didn't find signal on any of the IC pins.

I measured some voltages.  First, I measured the input voltage.  On the gaussmarkov layout, there's a "+12V" in the upper left.  Weirdly, I measure 9.17V at this point.  The power supply (PP2+) measures 12.44V before I plug it into the circuit.  I'm sure this makes sense, but it's still weird.

Next I measured the IC:
pin  V
1  0.39
2  0.90
3  0.87
4  0.00
5  0.87
6  0.53
7  0.39
8  1.26

I also measured the transistors.  For Q1 and Q2, Cook says to adjust the relevant trimmer until the collector voltage is 6V.  These are NPNs, so I think he might mean the emitter should be 6V.  Both emitters are 1.25V and both are completely insensitive to the trimmers.  With the trimmers centered:
   Q1    Q2    Q3
E  1.25  1.25  0.01
B  0.50  0.50  0.68
C  0.03  0.03  1.20

Checked the markings on the IC and Qs: all match the BOM.
Checked the pinouts on the IC and Qs: ok.
Checked the orientation of electros: ok.

hmmm...

[yodude]

... which is what happens when you wire the power backwards.

... which is what happens when you swap the red and black probes on your DMM.

Sounds pretty good now but need to wait for daylight to turn it up a bit. I still have a little distortion. Hopefully just need to twiddle the trimmers a bit.

Thanks edvard, petemoore, and PRR for the help. Thanks couple of beers for the insight.

[PRR]

> 70-volt intercom transformer

The most-common Fender Reverb Tank is 5 or 10 ohms input (really 2 ohms bass 20 ohms treble).

The opamps work best at 500-2K ohm load.

So with the Fender tank you want a ~~1K:8 ratio. 70V at 5 Watts is just about 1K, and this IS a readily-available part. (It's a bit over-kill but what the heck.)

> My tank is spec'd at 800R.

The opamps can drive that directly (no transformer). Technically each opamp "sees" 450 ohms, which is low for TL072 but '833 or 5532 are rated 600 ohms and don't stink in 450 ohms (and it's "only reverb"; tank coloration overwhelms any small strain in the driver).

> wire the power backwards.

In this racket, you GOTTA keep your head, and your leads, on straight.

In my first job, I beat my head because after a "minor" fix I got the power connections backward and didn't realize it for several MONTHS. I've repeated the trick on a 8o88 PC. Both the mixer and the PC worked a whole lot better right-way-round.

[Earthscum]

So with the Fender tank you want a ~~1K:8 ratio. 70V at 5 Watts is just about 1K, and this IS a readily-available part. (It's a bit over-kill but what the heck.)

Actually, 1k:8r would be the same RS tranny ya use as a cheapie inductor. Only $4 and a lot smaller. Those 70V trannies are actually pretty huge in comparison. Also, you can use those 1k:8r trannies, a BJT, a 555, and a buttload of diodes and caps (and a few resistors) to make  a 20kV "FU up if ya don't STFU" zapper... cattle prod, stun gun...

[PRR]

> the same RS tranny ya use as a cheapie inductor.

Yes. It would probably work. The bass will distort. That's a point: the traditional Fender rig has a steep 500Hz bass-cut, to put more spring-action in the ear-tickling midrange and less where the springs just slap and sound flabby. You might try C7 at 10nFd or 5nFd (with ot without any transformer).

[yodude]

Just closing the loop on this one ...

I built it with the tank described above with no transformer.  I also omitted the child board.  Works great, sounds great.  Highly recommended!

Thanks for the help, forumites!  And a special thanks to G. Forrest Cook (design), Gaussmarkov (pcb layout), and John Lyons (pcb etch)!