Spaceman Sputnik II

[digi2t]

Just remark for DCA55 users, that DCA55/75 not intended for testing in circuit. It can be damaged by DC stored in capacitors, or if AC/DC is connected.
digi2t, thank you for sharing you experience!

The AC/DC connected most certainly makes sense, and to date, I've never dreamed of testing with the power on. My brain sort of figured that doing that would result in "bad things" happening. Then again, it could give me an excuse to buy a 75... right? (Aw dang! It blew up... aww shucks.)

As for the caps, I must admit, it never dawned on me, so I guess I've been lucky that nothing's happened to my 55 yet. I suppose I could always drain the caps to ground before testing directly. Like I said, in some cases, removing components for direct testing isn't an option, especially with something of limited availability, not to mention Ge transistors that can be easily damaged.

Gonna start breadboarding this weekend. There's some component values that I don't have, so I'll fake them by chaining other values together while I'm waiting for my order to come in. I'll at least be able to see if I'm in the ball park. Stay tuned folks....

[digi2t]

Another update. Noticed that I had the DRIFT footswitch in the wrong position on the schematic, and I retook the voltages. Also managed to get a reading for the third Ge transistor on the DCA55. They're not accurate mind you, just a starting point. The voltages will take you the rest of the way. I also added the resistances of the trimmers as they are set. This should help in narrowing down the correct transistors for the first and third spots.

[digi2t]

Well, good news folks!

I have it on the breadboard right now, and it pretty well emulates what the original does. Instead of Russian transistors, I pulled out some Japanese 2SB172's instead. Call it the "Nippon" instead of the "Sputnik". Anywho, they're all 0.03mA or better leakage-wise, 72 in Q1, 82 in Q2, and 80 in Q3 for gain. The voltages are more or less aligned with the original, +/- 0.1 of a volt here or there. I'm attributing the slight variance to the fact that I didn't have any 475K resistors, so I have 470K's in there right now. Instead of two 10K's below the CALIBRATE pot, I used a single 20K. Also, on the high filtering, a 472 cap is filling in for the 502 while I wait. The 243 CALIBRATE cap is being temp'ed by a 273.

As for the unknown JFET, I tried a 2N5457, 2SK30Y, and a 2SK30GR. As for the Ge diode, I tried a 1N60, 1N270, and a Russian D9E. The combo of the 2SK30Y and D9E gave the best performance when it came to setting the edge of oscillation, and the best decay as well. Funny, I was singing the praises of the D9E earlier today, and once again, it didn't disappoint. My source and drain voltages are a bit higher than the original when the gate is set at 6.26v, but again, I'm attributing this to the missing 5K at the B-C of Q1. Once my order comes in, I'll be able to tweak it some more, and retest the voltages.

I'm almost positive that the BS170 that's just below the unknown JFET is used in conjunction with the Mill bypass to switch the JFET off. Otherwise, I think the oscillation would bleed into the audio in bypass. This is my theory, but it would be cool if some one with more brains on this could confirm.

All in all, pretty fuzzy, and when you kick in the DRIFT (oscillation), you can get all sorts of textures. The filtering is pretty cool as well. With the DRIFT off, at lower settings, it can be discreet. At higher settings, more effective. With the DRIFT on, freakin' awesome. Tone-wise, it's similar to the original, but the stand-in values, as well as different transistors, give it a different flavor. Say... a bit crunchier.

I'll audition some Russian transistors as well. I've got lots of them. The 2SB172's do sound pretty rad though.

More to come.... 

[stringsthings]

Interesting!

[digi2t]

Well, this is pretty much it. I've played and I've tweaked, and I've tweaked and I've played with it. This is what I have;



An interesting mod would be a diode polarity reversing switch for the Ge drift diode. I tried a BAT48 in here as well, and it works pretty well. A reversing switch changes the character of the drift effect by raising the oscillation onset point on the SCAN pot. While the initial oscillation attack is a bit more subdued with the diode the other way around, the decay is much nastier. I'll be adding a toggle for this on my build.

As I expected, the second MOSFET gets it's cue from the Mill. bypass, and kills the oscillation by grounding the JFET when you switch to bypass. The drift LED will still stay on if the footswitch is engaged, but you won't get any screaming oscillation bleeding into the audio. Of all the JFET's I've tried; 2N5457, 2N5485, BF245A, 2N3819, 2SK30A-Y, 2SK30A-GR, and J201, the J201 and 2SK30A-Y seem to perform the best. The 2N5457 is my second choice.

Going to try to clear up the bench (somewhat), and try to shoot a video tomorrow.

[stringsthings]

Looking forward to the vid!  I always enjoy reading about your circuit pursuits.
I need to build me one of them.  Maybe the spaceman!

[digi2t]

Well, I've decided to kick the bucket where Photobucket is concerned  . I've uploaded it to the Gallery now. Here's the schematic again;



Here's the video. First half deals with the original, second half with the breadboarded version. It's long winded, I know, I suck at these things, so feel free to shuttle through it.



Like I mention in the video, the hi and lo filter switches are more discreet when just on fuzz mode. Of course, a crappy camera mic doesn't help their cause either!  Anyway, what I have on the breadboard for the video, I combo'd some of the components to closer match some of the weird values of the original.

[stringsthings]

Very good video.  Lots of unusual sounds.

[digi2t]

Here ya go folks. Not build verified yet, but back trace to the schematic looks good. Let me know if you spot something that's out.



To help with the cuts, a reversed image;



Use the schematic to get your voltages dialed in. The gains are merely guidelines, depending on the transistor used, you can tweak it to your liking. If you can get the voltages close, you won't be disappointed. As for the weird ass component values, you can try more conventional values, or if your anal, hunt down the original values. Your call.

EDIT:

Images are updated. Forgot the transformer number, and the number of jumpers was incorrect. The "Diode reversing" switch is my own addition, after my breadboard testing. If you don't want it, then just connect from the vero direct to the Drift 5 lug. Also moved a few items around.

[PMowdes]

Here ya go folks. Not build verified yet, but back trace to the schematic looks good. Let me know if you spot something that's out.



To help with the cuts, a reversed image;



Use the schematic to get your voltages dialed in. The gains are merely guidelines, depending on the transistor used, you can tweak it to your liking. If you can get the voltages close, you won't be disappointed. As for the weird ass component values, you can try more conventional values, or if your anal, hunt down the original values. Your call.

EDIT:

Images are updated. Forgot the transformer number, and the number of jumpers was incorrect. The "Diode reversing" switch is my own addition, after my breadboard testing. If you don't want it, then just connect from the vero direct to the Drift 5 lug. Also moved a few items around.

Dino

A coupla things I've noticed.

You are missing D3 from the BOM, and I think that you may have D1 and D3 mixed up (unless they are the same)

You seem to have two 3.92K resistors but I can only find one on the schematic.  Is R20 the CLR for the stomp LED?  in which case does it need to be such an obscure value?

Can you update the Vero so that the transistor numbers are visible.  They are obscured by the pin orientation.

Thanks

Phil

[digi2t]

Gotcha Phil.

The CLR's are 3.6K, and there's only one 3.92K resistor.

There are two THREE diodes, and two LED's.

As for the obscure values, these are the original values. What's in there is what's represented. That said, there's no law that says that the builder cannot sub more common values. That's up to the builder. Personally, I build using the original values. This isn't my first dance with Spaceman builds, so I already have many of these weirdo values in stock.

I'll make the corrections later today.

[digi2t]

Vero updated.

Thanks Phil! 

[digi2t]

Another update. Couple of resistors were mislabeled, and the BS250 image orientation was wrong way around. Also corrected both filter switch connections. They now reflect what the original does when the toggles are up or down.





Progress so far on the vero...



The sockets are temporary for testing, I'll solder everything in permanent once I've settled on the right transistors. Initial voltage test puts me right in the wheelhouse of the original. 

[digi2t]

Well, that's it!! I'm stickin' a fork in this one, it's done. 








Works like the original, but I prefer the sound of the 2SB172's though, so I hucked the original into the trash. (Just kidding  )

I have made some slight changes to the vero. Firstly, I forgot the wire for the bypass LED - (dummy!), and I replaced the 22K (R17) resistor with a trimmer. Different diodes react a bit differently, so the trimmer allows for some adjustment where the detection point is concerned, especially when reversing the polarity. You can set it to 22K before you solder it in, record your initial voltage for reference, and then play with it to set your oscillation envelope.

Here are the verified veros, one with the switch, one without (as per the original). I've also added reversed images with the cuts highlighted.










I'll get a video of it up soon.

[stringsthings]

Awesome !!  I haven't made a vero project in a long while.  ( used to do it all the time )
This would make a great addition to my DIY pedal collection.  Many thanks for your work/efforts!

[digi2t]

Video is done.



The overall brightness of the 2SB172's is especially apparent in Drift (oscillation) mode.

What's interesting here is tonight, after staring at the schematic for a while, I suddenly hit me that I had seen that diode feedback oscillation scheme somewhere before.... the Schumann Two Face Fuzz.



The only difference here is that Schumann added a clamping cap to ground, but the premise seems the same. When I built the Two Face, I found that different clamping cap values dictated the oscillation tone when the signal was fed back. In the original Two Face, the cap values were quite small. In my clone, I found that 1800pF gave the oscillation a much better tonal range. Given that I used different transistors in the Two Face clone, I think that here the cap values may have been a help in getting the oscillation mode into a range that better matched the original.

So, it's kind of late now, but tomorrow, I'm going to try this angle in the Sputnik. Given that I'm not using the same transistors as the original, a clamping cap might be the ticket to adjust the oscillation tone range to better match the original. If so, it might be a decent add-on.

The fun continues...

[Cozybuilder]

Great job Dino. Another for the to-do list.

[digi2t]

Great job Dino. Another for the to-do list.

Thanks Russ! This one's been quite a bit fun to noodle with.

As I expected, a cap to ground on the anode side of the Ge diode was just what the doctor ordered. What I did was set the two units gain and tone settings to 9 o'clock, both filter switches to the middle position, and nudged up the oscillator voltage until they would self oscillate. My clone was oscillating at a much higher frequency, so I just auditioned caps until the clone was oscillating at the same frequency as the original. It ended up using a 0.1uF cap, which I soldered in. I could probably gone with a 0.12, or 0.15uF, but 0.1uF sounds good to me. The diode reversing switch still works as before, again, with a lower frequency range.

I used the frequency counter function on my DMM, measuring at the HIGH FILTER 1 side of the 39nF cap. Anywhere in the 400Hz to 600Hz zone gets you into the original's oscillation frequency zone.

The oscillation is now not only on par with the original, but also in a much better range high to low. The high end was almost into dog whistle territory! 

Belay last!!!

The cap to ground was NOT a good idea. After much noodling today, I found that it was a game of diminishing returns. Bigger cap, yes, does tame the oscillation, but it also darkens the original tone. By the time you get to sanity, the top end is too dark. So, the cap is out.

But, using a Ge diode with a low Vf helps. The D9E I used originally measured in at 0.323. Sorting through my bag, I found one at 0.272. I swapped it in, and there's a bit more variety in the jumping octaves, motorboating, and whatnot. At the end of the day, different transistors will yield different results, so socketing is highly recommended.

So that's it. Done.

[bluebunny]

So that's it. Done.

[digi2t]

So that's it. Done.

My feelings precisely... I can move on to something else now.