Gemini III dual fuzz dissected

[digi2t]

You are the Spaceman source! Please PM me when you start the WOW thread!

Tony

No problemo. For those who are curious, there's finally a video out...

Jimi... does this one tickle your ears brother? 

[digi2t]

OK... clone is DONE!!!

I've done some surgery in my time, but I think this one really tested my wiring/soldering skills. It was tighter than a frogs a$$ in there, and that's watertight!













My first Millenium Bypass too. I'm quite pleased with it. Customary video will be forth coming soon.

[dorrisant]

Nice... It looks so vintage!

I love the look of those old Russian cans... I'm waiting for mine.

Tony

[digi2t]

Update on the Millenium bypass.

R28 should be upped from 1.21K to 4.7K. This is the current limiting resistor for the LED. I believe there's too much current flowing through the diode, and it gets hot. My LED (super bright red) was going on and off. Probably the heat/cool cycle of the diode. Also, I was getting a pretty bad pop when I was switching the effect. With a 4.7K, both problems disappeared. I suppose that the original uses a 1.21K in relation to the type of LED that they are using. If you're using super brights, I recommend a 4.7K resistor.

I've updated the vero in the Gallery.

Video will be here on Monday.

Cheers,
Dino

[digi2t]

OK folks, video is here. But before we get to it, I have to lay down some more info.

First off, I was having a problem with the sustain on the silicon side. Actually, I narrowed it down to Q4. Although it's a 2N2222A in the original, an I had even tested a few different BC type transistors, the sustain wasn't anything to write home about. Voltage wise, I wasn't getting close to the voltages that the original was giving me on Q4, coming up short on the collector by more than 1/2 a volt. After a lot of head scratching, and sifting through a ton of different transistors, I finally found a correlation between the gain of the transistor, and Q4 collector voltage. With a 2N2369A (see Axis Face) in this slot, with a gain in the mid-60's on the Peak meter, I managed to not only meet, but surpass the collector voltage by .2 volts. So, high gain here, no good. Higher gain was killing my collector voltage. I figure if I had one in the low 70's, I'd be spot on. I did keep the BC109B in Q5, and the 2N3700 in Q6. I also ditched the 120nF cap after Q6, for the 100nF that I had in there during testing. This was a personal issue, my ears said "better".

Secondly, setting up the voltages of the Ge side. I had to do a bit of reading on the Harmonic Percolator. Especially where the HP's Q1 emitter to ground, and Q2 collector to +9 resistors are concerned. It dawned on me then why they used the particular trimmer values (25K and 100K) in the Gemini. I decided to set my two trimmers, as if I was dealing with an HP, to 20K and 91K. Plugging in the unit, I found two things; A) My Q2 and Q3 voltages were WAY off from the original, and B) with the Ge gain switch on high, I was getting oscillation. With the DMM on Q2 emitter, I started dialing the 25K trimmer in small increments, reducing the voltage. This would change the pitch of the oscillation, but not eliminate it. Slowly dialing the 100K trimmer in the same direction, I could then kill the oscillation. Dial the 25K a bit, oscillation comes back, kill it with the 100K, and so on. I kept on doing this back and forth ballet, in small increments, until I got Q2 emitter to .5v below Q1 emitter (as per the original), and had no oscillation. At this point, with the Russian trannies, I'm about +/- 10% within the voltages I get from the original unit.

That's it. A bit of fanagling, but with some CDF (Common Dog F*ck), we got 'er licked. On to the vid. Enjoy.

[zetamkiii]

Would it be possible for someone to post the back of the lower board? Trying to fit it all in a regular size enclosure
and a straight copy of the pcb is the only way. This one is far more daunting than anything I've ever worked on.
Any attempt I would make to rearrange the circuit would end in disaster.

[digi2t]

Sorry, the best I can do is this;



Since the pots and switches are soldered directly to the board, pictures of the trace side were quite useless. That's why I drew it by hand. Also, the lower board is double sided, the red lines indicate traces on the component side of the board. Your best bet would be to redraw it on the computer, using the pin spacing of the pots and switches to scale things. FWIW, the lower board is the same width as the upper board. Using the spacing of the stand-offs (1 through 7) to scale the lower board should help as well.

[zetamkiii]

^Thanks . I'm about to go transistor shopping. Can you recall what were the gain values (Ballpark figures) of the original transistors and of the ones replaced them with? I like what you achieved with the silicon side, The germanium side I will test as many possible candidates for high gain as I can while still retaining good cleanup.

I have heard that the Gemini III does not clean up well so perhaps it is a factor of the design rather than the transistors. I have a Mercury III and an Aphelion but sadly no original Gemini III that I can compare with.

[digi2t]

^Thanks . I'm about to go transistor shopping. Can you recall what were the gain values (Ballpark figures) of the original transistors and of the ones replaced them with? I like what you achieved with the silicon side, The germanium side I will test as many possible candidates for high gain as I can while still retaining good cleanup.

I have heard that the Gemini III does not clean up well so perhaps it is a factor of the design rather than the transistors. I have a Mercury III and an Aphelion but sadly no original Gemini III that I can compare with.

Gain values of the original transistors?

BWAAAAAAHAAAAAHHAAAAAHHAAAAAHHAAAAAA!!!!

Sorry. There was no way in hell that I was going to pull the transistors out of the original unit. Your best bet is to study the Harmonic Percolator and company, and follow recommendations for transistor gains in those circuits. I'll see if I wrote down the gains I used for my unit, but I don't think I did.

As far as cleaning up is concerned, using the silicons that I did, on full silicon, it cleans up really well. Better than the Ge side. I believe I demo'd this in my video.

I'll see if I can find those gains for you.

[zetamkiii]

Just an estimate would do.

[digi2t]

Just an estimate would do.

Your in luck. Look here; http://www.diystompboxes.com/smfforum/index.php?topic=97493.msg856308#msg856308

That's my vero layout, and the Ge transistor gains that I used are on the left hand side of the layout. These values are in my clone, and my voltages are almost spot on to the original. Mine sounds better though  .

I knew I had put them down somewhere. Just had to dust off the ole thinking cap.

[zetamkiii]

Any clue about the silicon side values? Doesn't need an NPN at the end of it too does it? I'm just trying to find an acceptable
range of the first, second and third transistors for these parts while I gather everything for a breadboard.
For Q5 you say mid 200's. A pretty big difference in gain from the GE side.
The impression that I got from comments and demos was that:
Ge1 is the lowest gain stage,
Ge2 and Si1 were a similar gain,
Ge3 and Si2 were also similar gain,
and Si3 was a gain stage above Ge3.
So I was going to base my hfe values according to that basic idea.
Thankfully this one isn't as finicky with the hfe as the wow signal.
Edit:Ah the silicon side is all npn...

[weadapt]

sorry to bump such an ancient thread, but i need one of these in my life.

mouser has part# 42TL019-RC backordered until july and i dont want to wait that long.

all i have here is # 42TL018-R and i am certain that wont work. is there a workaround available that any of you has seen? or any for sale anywhere that isn't mouser?

[UKToecutter]

Their first delivery date is 3rd July, that's only about 5 weeks away

:-)

[weadapt]

sorry to bring this thread up again, but i've looked all over and cant tell the difference between a DPDT on-off-on and a DP3T on-off-on. it says DPDT on-off-on in the BOM...

[chuckd666]

I am also exhuming this thread/this is my first diystompboxes post so apologies for that.

Anyway, I have completed my build on vero - thanks to all involved in the discussion and tracing because it all came in handy!

The enclosure has since been painted hammer finish copper:



and a blurry shot of the inside:



Germanium transistors are Sanyo 2SD187 for the NPN, and MP20Bs for Q1 and Q2.

Silicon transistors are as stated on the vero layout.

My main issue is with the germanium side output - it's significantly quieter than the silicon side. It is also 'looser' than my real Gemini III. Any suggestions on tweaking it to be louder on the Germ side? I'm not sure where to start.

[digi2t]

What are your voltages for the Ge side? It might only mean that you need to tweak the bias trimmers a bit to raise the volume. See here for adjusting the Ge bias; http://www.diystompboxes.com/smfforum/index.php?topic=97493.msg868878#msg868878

I found that adjusting the Ge bias trimmers generally helps even out the volume difference, unless the transistors are too leaky. In my unit (if memory serves me) I believe the Ge side is slightly louder, and more wooly, whereas the Si side is a tiny bit lower, but the fuzz is more aggressive, along the lines of a BMP.

[chuckd666]

Thanks digi2t!

I will report back with my previous voltages and then tweak as suggested. So set it like the HP, then work backwards from there... very good.

Appreciate the guidance.

charles

[chuckd666]

Well, your suggestion seemed to work! My biasing for the Germ side was just totally off. I am still kinda out of the ballpark with voltages on Q3, but it sounds good. Q2 is within 10% of the original voltages. Happy! Getting a highish pitched whine on the silicon side and it's a bit noisy, but that's probably because I didn't shield the wires and the lid is currently off heh.

Thanks man! YA RULE.

[digi2t]

Well, your suggestion seemed to work! My biasing for the Germ side was just totally off. I am still kinda out of the ballpark with voltages on Q3, but it sounds good. Q2 is within 10% of the original voltages. Happy! Getting a highish pitched whine on the silicon side and it's a bit noisy, but that's probably because I didn't shield the wires and the lid is currently off heh.

Thanks man! YA RULE.

What transistors are you using?