Green ringer problems ....FIXED !

[MartyMart]

Symptoms : Splatty and low output with some octave present
seems to work better with "hard struck" strings and HB guitar ...

Project ; Green Ringer from GGG
Schematic : http://www.generalguitargadgets.com/diagrams/gringer_sc.gif
Built on : vero
Substitutions: R9/R10 both 20k
R4 6k4
R2 150k
R3 20k

Voltages :
Battery  9.84 volts
Q1 C 7.72
     B 1.32
     E 0.81
Q2 C 1.20
     B 7.74
     E 8.41
Q3 C 9.84
     B 3.83
     E 3.34
Diodes:  cathodes 3.83v and anodes 3.91volts

Q2 looks reversed, but since there is a "mirror image" of 20k/68k/diodes/10k's to
Q3's base connection, it makes zero difference when reversed  !!

All values checked, off board is just jacks and battery snap, no cold/bad joints
Ideas .... ?
Thanks,
Marty.

[benfox]

Hi,
I had no problems building this. Check your diodes Maybe they are dead or not in the right place....
They should match too i mean they have to reach the same value.
Have a nice day

[Ge_Whiz]

Hi Marty

Sounds to me like not enough gain from Q1. My GR 'fizzles out' as the note decays, and only produces a distinct octave-up at higher input levels, i.e. after the note is hit it goes octave / distorted fundamental / fizzle. What diodes are you using? I got best results with plain vanilla silicon.

[MartyMart]

Thanks chap's ...
Diodes are 1N34a's and are in correctly, should I try a pair of 1N4148's ?
I dont remember "matching" them but the readings are equal ...
I could try a 2N5089 in Q1 or even MPSA18 for some more WUUMPH ! ?

Marty.

[Mark Hammer]

Just stick a 1-transistor booster of your choice on the front end.  The GR sounds juicier and delivers a more audible octave when pushed harder.

[petemoore]

 Â 1n34's tried, and they 'clamped' too much for my preference, which would be the 1n914 Si type diodes.
 Also, measureing the 'symmetry' [choosing from what you have there and finding 'pairs'] of resistances of the *22k and *68k resistors, as well as the diodes, may be of some 'mileage' improvement for octave...not sure...
 Could be the symmetry of the + and - swings going into the FWR section are less than symmetrical or something, so having perfect symmetry in the FWR...hopefully you're catching something about this symmetry/assymetry 'thing' because I'm failing to illustrate it perfectly as I type it...messing around with the actual values Of the FWR and what goes into the FWR may be worth messing with...I just matched the mentioned parts and hoped the signal fed to them was symmetrical...and still don't 'really know how much attention to symmetry is really necessary here.

[Mark Hammer]

The symmetry IS reasonably important, but at the end of the day there has to be enough signal amplitude that after the diodes have sucked out their voltage drop there is something you can derive an octave FROM.  The unit does not have to be pummeled, but the gain of the first transistor, and whatever you can do to supplement that, is helpful.  The supplemental level COULD be a compressor placed in front, or, as suggested, just a simple one-transistor booster stage with a gain of 2-5.

[petemoore]

  Yupp, this makes the GR 'Dependant'...sounds wierd on it's own, only you don't really get into octave and the really cool stuff when something's driving the input.
  I find this is really preferrable in use, simply because rare are the occasions when I want a transition to heavy octave from 'clean', no, most occasions call for Octave to be segwayed from Fuzz.
  So having the GR post Fuzz makes it possible to get Octave, nice octave by hitting only one switch.
  Octaves that octave with 'internal' fuzz, often sound oversaturated or somehow 'unfriendly when 'another' fuzz is left on before or after them, which tends to necessitate the hitting of 2 switches [one to turn the fuzzed octave effect on, another to turn whatever other fuzz off], or less than desirable octave wierdness.
  For some reason I find a FF does a good job driving the Green Ringer, and this lends itself to pretty nice and wide choices by simply adjusting the guitar volume control to get more or less Fuzz / 'Octa-ring' tones.

[MartyMart]

My symmetry "may" be a bit off, though i do have a pair of 20k's where 22k's are called for
the 68k's are 68's and the 18k is also a 20k.
It just sounds "wrong" and mis-biased at the 'mo
Perhaps i should use 22k's and some matched 1N4148's  ?
"Fuzz" seems to be missing here now ..... ?

Marty.

[Toney]

 Hiya Marty,
I built one of these recently too. Dont know if you went with Torchy's vero layout, but I had an early version that didn't show all the trace cuts.... shoulld be six in all. Symptoms were real similar until resolved.
Another point I read over at GGG was that Q2 needs to be a low gain 3906 (around 150 hfe) for a decent ring. That one's got me stumped at the moment. Closest I could find was 190.... tested 20 or so.

[MartyMart]

It was Torchy's and I do have six trace cuts.
I've tried several si PNP's wih no luck, I may try some lower gain Ge's
Tomorrow, I'll change the few values that i have wrong and go with some Si
diodes also.
Thanks,
Marty.

[johngreene]

The symmetry IS reasonably important, but at the end of the day there has to be enough signal amplitude that after the diodes have sucked out their voltage drop there is something you can derive an octave FROM.  The unit does not have to be pummeled, but the gain of the first transistor, and whatever you can do to supplement that, is helpful.  The supplemental level COULD be a compressor placed in front, or, as suggested, just a simple one-transistor booster stage with a gain of 2-5.

Except that the diodes should be already forward biased with DC and therefore shouldn't suck out any voltage drop from the signal. Notice how the CD on the anodes of the diodes is higher than the cathodes. The signal is just going to ride on top of the DC already present.

For optional operation I would think you would wan't 6.67 V at the collector of Q1. Which would cause Q2 to have 1/3 the supply dropped across the CE terminals this mid-point biasing it. This would give you maximum swing on both the inverted and non-inverted signals.

--john

[R.G.]

For optional operation I would think you would wan't 6.67 V at the collector of Q1. Which would cause Q2 to have 1/3 the supply dropped across the CE terminals this mid-point biasing it. This would give you maximum swing on both the inverted and non-inverted signals.

Actually you want half the supply across the transistor, and 1/4 across each of the resistors for max swing. The biggest swing happens when the transistor is either off and the full supply is across it, or saturated, and then there's half the supply across each resistor. Halfway between is 1/4 the supply on each resistor.

Other than that, you're dead correct, as usual.

That being the case, we look and find:

Substitutions: R9/R10 both 20k
R4 6k4
R2 150k
R3 20k
Voltages :
Battery  9.84 volts
Q1 C 7.72
     B 1.32
     E 0.81

We have some problems. Ohm's law does not hold for the voltages and resistor values you list.
For the input divider, you should have Q1's base at Vbase= 9.84*(150K/(150K+560K))= 2.07V, and you show 1.32. So either the resistors are not what you think, something else is sucking current, or the transistor's gain is way low.

The base-emitter voltage is 1.32-0.81 = 0.51V, which is reasonable.

The emitter resistor is 6.4K, so the emitter current is 0.81V/6.4K = 126uA. Almost that same current has to flow through the collector resistor, so that one should have a drop of 126uA * 20k = 2.53V, and should read 9.84-2.53 = 7.31V instead of 7.72. The voltage across it is down by about 16%, which is more than just resistor tolerance.

Q2 C 1.20
     B 7.74
     E 8.41

The emitter and collector voltages from their supplies don't match. The emitter is 1.43V from 9.84V and the collector is 1.2V up from ground. They ought to match more closely than that. Maybe Q2 is quite low gain, and is sucking a lot of base current that never shows up in the collector. The collector current is 1.2V/10K = 120uA. The emitter current is (9.84-7.74)/10K = 210uA. Since 120uA is going to the collector, the base current is 90uA, and the gain is 120/90 = 1.3. UGH. Of course, we don't know the exact resistor values, so this kind of evaluation can be off a bit, but it tells you where things are wrong.

In this case, the excess base current from a low gain device could well be pulling up the collector of Q1.

One situation that causes otherwise respectable transistors to work but with low gain is when they are swapped emitter for collector. They do work in this inverted mode, but the gain is low and other strange things happen. Did you by chance get the pinout backwards on Q2?

If so, that could account for the difficulty in getting any octave. The octave effect would not happen until you had a big enough signal on both collector and emitter to flip the diodes, and the emitter side has about a 2:1 advantage. It's one theory to check. (Ooops... I forgot, theoretical things are always wrong... ) What the heck, check it anyway.

Q3 C 9.84
     B 3.83
     E 3.34
Diodes:  cathodes 3.83v and anodes 3.91volts

Hmmm... with R9/R10 both 20K, their junction is 9.84/2 = 4.92V. The diode anodes are at 3.91V. That means that there's a volt across each 68K? Ohm tells us that there must be (4.92-3.91)/68K = 15uA flowing in each 68K resistor, and also the diodes. So Q3 is getting 30uA into its base. It's emitter is sitting at 3.34V, so the emitter current is 3.34V/10K, or 334uA. It's then exhibiting a gain of ten ((334-30)/30) which is 'way low.

Something is rotten, as you knew. The transistors are showing terribly low gains in practice. Some of the currents don't add up.

- check the pinouts
- check parts values
- check soldering

You're going to find something in there, I think.

[Tony Forestiere]

Thanks R.G. for an in-depth analysis of troubleshooting! I have bookmarked this thread for its educational content. Also, thanks to MartyMart for showing the PERFECT format of "What To Do If It Doesn't Work" symptom and readings.

R.G....sounds like you've written most of another "Technology of..." article

Thanks to both of you for showing how it is done!

[gaussmarkov]

marty,

i don't have R.G.'s analytical abilities, but i'd like to help.  so i ran the green ringer through superspice.  the ggg schem gives these voltages for a 9.84 volt battery:

Q1 C  5.48
     B  2.14
     E  1.51
Q2 C  3.66
     B  5.48
     E  6.16
Q3 C  9.84
     B  4.91
     E  4.27

diodes:  4.91

after making your substitutions, i get

Q1 C  5.46
     B  2.04
     E  1.41
Q2 C  3.68
     B  5.46
     E  6.14
Q3 C  9.84
     B  4.91
     E  4.27

diodes:  4.91

so not much change.    i hope these give some idea of what you ought to be getting.

[MartyMart]

Thank you RG/Gauss,
There is definately a problem.
As this was made some time ago, I now have many more parts and will
re-solder the new 22k's and 6k2, also new diodes, Si 4148's matched !
I have the pinouts spot on, so its another error somewhere, I'll also hfe
check my 5088's and 3906.
"It's out there.... somewhere"  !
Regards,
Marty.

[MartyMart]

Sorted  !!
Not exactly sure how, but I did the following :
Re -heated all solder points and cleaned the vero "gaps" again.
Replaced the 1N34a's with "matched" 1N4148's reading 0.728v each
Replaced Q1 with an MPSA18 - more gain here !
Replaced Q2 with a 557 fairchild PNP gain of 140 hfe
Done, it now produces a nice octave and some "fuzz" but is almost "clean"
perhaps this is how it's meant to be ?
Again, better with HB guitar and front pup/tone rolled out.
Strat with vol rolled back "does not work" !!
So I presume all the talk of "boosters" in front etc is due to this problem ...

Marty.

[johngreene]

Replaced Q1 with an MPSA18 - more gain here !

I'm not sure why this would make a difference. The 6.2K is surely going to dominate the gain over the HFE, the gain should be around 3 no matter what transistor you put in there. The input impedance would be higher .... But hey, if it worked.....

Replacing the 6.2K with a 5K pot in series with a 1.2K resistor and a large cap on the wiper might be an interesting mod. At least it would allow you to adjust the gain of the first stage for maximum effect.

--john