I've tried posting this info several times, but keep getting something like "server error 500."  So if the info below is posted like 5 times before this post, that's what happened.  (I'd apologize, but it's not my fault. )  OK, assuming the server didn't get the info below posted, here it is again:

Here's the BOSS-original SD-1 schematic.

Here's a link to R.G. Keen's 'The Technology of the Tube Screamer' as well, since the TS-808/TS9 use the same feedback loop configuration as the SD-1 and the proposed mod to the Tube 12a.

The proposed configuration is not due to a mistake on a schematic.  It's been used successfully in thousands/millions of Tube Screamer clones.

Quote from: phatt on November 17, 2009, 08:53:40 AM
Forgot,,
Re Bridge for clipping?? I breadboarded a few a lot of those clippers and the extra diode did little for me but heck try it out and make your own judgement.

Marshall used this idea in JCM 900's I think. The only way they could get more crunch from the amps as triodes alone can't do that trick.

The diode polarities on your diagram do not match mine.  On the two "left side" diodes, they need to "point" inwards towards each other (and the + side of the "middle" diode).  On the two "right side" diodes, they need to point outwards away from each other (and the - side of the "middle" diode.)

The diode arrangement in your diagram will block flow from the signal-side through the bridge rectifier to Vref/GND (except once reverse-breakdown voltage is met in the diodes, but I doubt you'd ever hit that voltage in this scenario. )

EDIT:  If you "rotate" the 4 bridged diodes 90-degrees to the left (and reposition the middle diode accordingly, but flip its polarity), your bridge rectifier clipper would work as expected.

Quote from: J M Fahey on November 16, 2009, 06:28:44 PM

Quote6) Remove and jumper the following: C3, C7, R5, R6, R14.

Jumpering : C3, C7, R5, R6 will effectively remove needed DC isolation,  considering that pins 1,2 and 3 of IC1A normally "live" at around +12V.

Could you please explain what this "DC isolation" is in the feedback loop?  The NJM4558 chips (and OPA2134PA) can handle a supply voltage of up to 30V, so 24V single supply should fall within spec.  Also, what purpose does this DC isolation serve?  (Please bear in mind that my experience is largely from the stompbox realm.  I have run SD-1/TS808 circuits at higher voltages than 9V (as have many others) and not observed any problematic behaviors.  However, I could be completely ignorant to the phenomenon you are describing.)

The chip can handle the +24V by spec, so I'm a little unclear on the rest of this assertion.  If you could please clarify, that would be greatly appreciated.

Quote from: J M Fahey on November 16, 2009, 06:28:44 PM
The B1Meg suggested as new VR1 will scratch heavily most of its travel, thumping loudly (and killing all sound) on either extreme, whether shorting Pin 1 to ground or placing around -11VDC (relative to Pin 3) on Pin 2, effectively hard switching the output to practically +24V.

If that's the case, then why doesn't a BOSS SD-1 "scratch heavily most of its travel, thumping loudly (and killing all sound) on either extreme...".  The 1M pot (and C3,C7,R5,R6 removals) make the IC feedback loop resemble that of an SD-1 (and TS-808/TS9 a little more distantly.)  A fixing resistor should keep from killing the sound at minimum resistance; at max resistance, well, the pot resistance takes care of that end too.

There are millions of SD-1 examples out there that bear testament that this arrangement does not produce the aforementioned artifacts when turning the pot.  However, just because I haven't personally seen a million dollars does not make me believe a million dollars does not exist...so I'll try to stay open-minded on this as well.

Quote from: Brymus on November 16, 2009, 12:58:56 AM
...
I think 1k5 cathode resistors would be an improvement like your OPA hifi opamps(which I want to try now after your glowing endorsment)
...

You can also stack (solder Pin 1 on Opamp A to Pin 1 of Opamp B, Pin 2 to Pin2 2, etc) two lo-fi op amps to get better response.  (Just make sure they can handle a supply voltage of +24V.)  NE5532 opamps are a popular one to stack, but I'd avoid them here at first due to their low input impedance.  I'd recommend trying any of the FET-input dual opamps (TL072, OPA2134PA, etc) first, then try non-FET input dual opamps and pick the one that makes your ears happiest.

Just for the record, the OPA2134PA isn't an end-all-and-be-all opamp for me.  I just have gotten into the habit of using it before others because I love the flat EQ and highly-detailed output I get from them.  Sure, you often have to tweak the pedal EQ to get a little warmth back in the final tone, but that's a small price to pay for the other positives the OPA2134PA brings.  YMMV, of course.

It's a classic bridge rectifier arrangement, with the exception of the addition of the 5th diode in place of the battery/mains differential load.  (The diode going between the two middle points is this "exception" diode.)  I first saw this clipping arrangement in the Marhsall JMP-1 preamp, then in the Radial Engineering ToneBone Classic.  I tried it in my DS-1 "Huevos Grandes" mod and liked it...and started using it in selected distortion circuits.

Attached is the diagram of the bridge rectifier clipper in question.  The dotted diode is there to show how the bridge rectifier replaces a hard-clipping diode in a distortion stompbox.  (In this case, replacing diode D5 in the BOSS DS-1.)

In this application, you're getting the clipping shoulder contour and proximity-to-clip of three 1N34A diodes in each direction, while only using 5 diodes (instead of 6).

FWIW, here are the mods I considered for the Tube 12a.  These mods would allow the gain stage to be active all the time.  These mods will also make the clipping boost phase into a soft-clipper (similar to the tubescreamer), as well as provide a "dist warp" control to control the amount of hard-clipping.

I doubt I'll do these mods myself, but they would likely improve the distortion on this little amp.

Mods:

1) Replace VR1 (A10KΩ) with B1MΩ pot + 33KΩ (fixed) in series.

2) Replace R4 with SW1A-controlled circuit for feedback loop clipping diodes (2 ->|-, 1 -|<-), use 1N4148 diodes from replacement of D3 thru D6 (below.)

3) Rewire SW1A to turn feedback loop clipping (R4) on/off.

4) Replace R7 (220KΩ) with 1µF film cap in series with -> 2.2KΩ (use 2.2KΩ pulled from R6, 2.2KΩ is closer to ground than 1µF cap.)

5) Replace C6 (150pF) with 51pF or 47pF.

6) Remove and jumper the following: C3, C7, R5, R6, R14.

7) Replace C8, C10, C16, C18 with 0.047µF film caps.

8) Replace with 1N34A Ge diodes and invert polarity of D3, D5.

9) Replace D4, D6 with 1N34A Ge diodes.

10) Install 1N34A Ge diode on solder-side of PCB between the union point of D4/D5 and D3/D6; Diode polarity points towards D3/D6.

11) Replace R15 (470KΩ) with 12mm B500KΩ pot; drill hole in front panel at 10:30 relative-position to GAIN KNOB; install DIST pot in hole.

12) Install 2200pF film cap on solder-side between R13 (union of R13/D5) and SW1B (side connected to union between R14/R15); disconnect and remove wire from union between R14/R15 and SW1B.

13) Replace and socket IC1 (4558) with an OPA2134PA or other stacked, dual ICs.

Quote from: Brymus on November 14, 2009, 06:34:16 PM
What kind of sound do the 4X4148s to ground give after the first triode?

Not that great.  I'd planned to mod them into a quasi-bridge rectifier arrangement to boost the proximity-to-clip a bit, but before I actually tried that out, I decided to design my own amp.  (I'll post that project here at SSGuitar once I've got it all done.)

Quote from: Brymus on November 14, 2009, 06:34:16 PM
I just tried something similiar with red LEDs and found 2 sounded great but 4 only added a little dirt not very noticable.

I usually like a proximity-to-clip around 1V on both sides of the waveform.  A bridge rectifier w/five 1N34A diodes is my current favorite.

Quote from: Brymus on November 14, 2009, 06:34:16 PM
How does the low plate voltage affect the clipping the make?
My triodes are closer to 250V- 300V plate voltage 80v seems low.

I don't think the valves are involved in the distortion much at all.  I think the majority of the distortion on this amp comes from the opposing 1N4148 diodes.

Quote from: Brymus on November 14, 2009, 06:34:16 PM
Also why are they biased with 2k2 instead of 1k5 or 820r?
Is it because of the low plate voltage?

While the tube in the Tube 12a appears to be more than a "marketing tube", but then again, at these moderately-low voltages, I doubt they factor much, if anything, to the distortion of this amp.  I've never tried to mod the tube parts mostly because of this belief.  Another thing is that they can't boost the signal too much before the LM1875, which they have to cut in 1/2 as it is to meet the max input voltage on the chip amp.

Quote from: Brymus on November 14, 2009, 06:34:16 PM
And finally have you tried added a bypass cap to either triode?
Seems like the first one could have a 4.7uf on a switch for a nice warm boost or even a .2 - .68 for some top boost .
Thanxs for sharing the schematic and pics,Its nice learning about new stuff.

Nope, never tried modding around the tubes.  To be honest, I've been a bit lazy with this amp, since it is pretty nice with just the upgraded op amp.  All the mods I was going to do I've put off in favor of building a backpack amp head instead.

I like the ideas you've mentioned and I'll look at them once I get my other amp project completed.  Luckily, this amp is so cheap, you can afford to have a couple to tinker with.

OK, here are the gut shots (Part 2) ...

OK, here are the gut shots (Part 1) ...

OK, I got to take things apart and get some pics.  First off, attached is the schematic corrections for the R40/R42 and D42/D45 neighorhoods.

The "dead short" between R40 (which is actually R41) and R42 is valid.  There is also a union between D42 and D45 (as was previously theorized.)  There is also an erroneous duplication of D45, but I've Red X'd the duplicate diode that is not present in the actual amp.

I'll post the gut shots next...

Quote from: J M Fahey on November 11, 2009, 08:54:26 AM
If someday you could post a couple of pictures, including some "guts", they will be welcome.

I think I'll have a chance to open up one of the amps tonight.  I'll try to get some shots posted here later on.

Quote from: phatt on November 11, 2009, 07:31:49 AM
BTW I think the schematic has a bobo next to *R40 and R42* looks like a dead short to me   as drawn that shorts out the secondary winding of the power transformer. Just thought I'd mention it.
Cheers, Phil.

Quote from: J M Fahey on November 11, 2009, 08:54:26 AM
Hi Phatt
Yes, you are right.
Another error is the undefined connection between D42 and D45.
I usually fiund many "official factory" schematics, with dumb unexplained errors.
I guess that they are there on purpose, so any technician will recognize them as such, and in a service situation, the dead amp is there, on the bench, and the actual wiring can be checked, but a non-tecnician, who can read a schematic and solder but not really understand design, will fail when trying to clone something, based only on an Internet downloaded schematic.
Even worse are "reverse engineered" schematics, when some component is not accurately read or some tracks get mixed.

When I'm taking pics tonight, I'll try to get some close ups on the R40/R42 and D42/D45 neighborhoods.

Thanks!

Quote from: J M Fahey on November 11, 2009, 08:54:26 AM
Hi 5thumbs (I'd *love* to hear you tapping a 5 string bass he he), thanks for sharing.

My handle is more about how my chops are suitable for silent movies and the hearing-impaired and not much else. ("I play like I have 5 thumbs." )  But that doesn't stop me from building/playing/enjoying it all just the same. 

Quote from: joecool85 on November 10, 2009, 04:27:28 PM
Whats the difference you get when you change that chip?  Better quality?  Better tonal reproduction?  Louder?

If you've used the OPA2134PA to replace *4558 chips in stompbox designs, the difference in this amp is similar to the qualitative changes seen there.

The *4558 chips, in my opinion, are rather lo-fi on the high end, as well as having that "mid hump" that it is famous for.  Once replaced with the OPA2134PA, you get better high-end clarity and lose that midrange mud of the 4558.  Since the Tube 12a is kind of mid-heavy already (due to the small 8" speaker), the addition of the 4558's mid hump is just a wrong move, IMO.  The flatter EQ and cleaner high end (as well as higher slew rate and all the other esoterica that you may or may not care about) of the OPA2134PA add some much-needed clarity to this little amp.

Finally, in all honesty, some of the additional top/bottom end performance changes and clarity might well come from the JFET input stage on the OPA2134PA.  This gives it a higher input impedance than the *4558 chips, which would likely result in more top and bottom end response. *shrugs*

I work on Kustom Tube 12A combos from time to time and have a couple of these amps kicking around my house at any given moment.  If you replace the 4558 IC with something hi-fi like an OPA2134PA, these $100USD amps are positively charming and useful as a practice, recording or workbench amp.

Since nobody else has previously posted this schem, I figured I'd share it with y'all.  Good luck!

Quote from: joecool85 on November 03, 2009, 02:50:00 PM
Wait a minute...1v into a chipamp?  I ran my guitar straight into my LM3886 and didn't get full volume, my pickups put out 2+ volts.  Then I hooked it up to my LM386 and ran that into it and got full volume, not sure how many volts it was putting out, but had to have been 5-6v.

I'm working on projects that are based upon the Tripath TA2020 and the LM1875 chip amps.  The TA2020 has a max input voltage of 6V.  (Yes, that's an absolute max rating...the recommended RMS input voltage wasn't specified on the datasheet, so it could be 1V or perhaps higher...unknown to me at this time.)  So you can feed 6V to a TA2020 chip amp and not fry it, but that doesn't mean it will not distort like crazy when fed that level.  (I don't know what voltage levels I can feed my TA2020 chips for clean operation, but I'll get the voltages "in the ballpark" in the design phase, then tweak to final values during the "prototype tweak" phase.)

The LM1875 is max rated for input voltage from -Vee to Vcc, which in my application is 48V (-24V -Vee to +24V Vcc.)  I own a couple of commercial hybrid amplifiers which use the LM1875.  The last valve section feeds +80V into in an inverting 4558 which effectively halves the signal before presenting it to the LM1875 chip amp.  So in this implementation, they have played exactly "by the book" and are halving the voltage to keep it within the spec'd input voltage range.  (In this particular amp, they use a single supply of +40V on the LM1875, but they are still within spec for that chip amp.)

So all of that is more or less reinforcement for what others have said about making sure your preamp output is within spec for the input voltage rating of your power amp.  After reading this thread and thinking it over, I'm going to add an inverting op amp with variable gain reduction just before the power amp outputs.  It's a trivial task to do at the design stage that will make tuning the first prototype that much easier.

Thanks to all on this thread for giving me some good food for thought as I work through my own design issues.