Boss DS-1 ciruit question

[WGTP]

It looks like to me, that the boost stage in front of the op amp is producing an asymmetrical signal as a result of the small 22 ohm resistor that is at the emitter to ground (r9) of the transistor.  The signal can then be "clipped" by the single diode (d8) (assuming it is "facing" the large side of the asymmetrical signal) before going on to the buffer (1st op amp stage) and then the second op amp stage for additional gain. 

If this is correct, then the actual clipping diodes are "seeing" an asymmetrical signal from the second op amp that is symmetrically clipped by the stock 1N4818 diodes, or asymmetrically clipped if modded.  If modded to be asymmetrical, and receiving an asymmetrical signal, wouldn't the direction of the diodes (for example a stock SI and an LED) matter, in terms of the harmonic content???

http://freestompboxes.org/members/soulsonic/schematic/DS-1_Distortion.pdf

[5thumbs]

It looks like to me, that the boost stage in front of the op amp is producing an asymmetrical signal as a result of the small 22 ohm resistor that is at the emitter to ground (r9) of the transistor.  The signal can then be "clipped" by the single diode (d8) (assuming it is "facing" the large side of the asymmetrical signal) before going on to the buffer (1st op amp stage) and then the second op amp stage for additional gain.

Please excuse me if this is an ignorant question, but how can the resistor from emitter to ground (R9) make the signal asymmetrical on the virtue of its resistance value?  In Jack's 'Fat Mod for DS-1', as well as my 'Vintage Ripper' and 'Huevos Grandes' mods, the Q2 resistor network (of which R9 is part) is rebiased to change the boost (and also EQing, via increased input impedance improving bass response, etc).  Part of the Q2 rebias is changing R9.  In all three mods, R9 is increased from its stock value of 22Ω to much larger values.  These changes do not cause an asymmetrical boost or clip of the waveform that I am able to discern, so hopefully you could help clear up my confusion.  (And no, this isn't an attempt for me to be politely self-effacing and sarcastic at the same time.  I'm not smart enough to pull that off with the skill that some do around here.  )

If this is correct, then the actual clipping diodes are "seeing" an asymmetrical signal from the second op amp that is symmetrically clipped by the stock 1N4818 diodes, or asymmetrically clipped if modded.  If modded to be asymmetrical, and receiving an asymmetrical signal, wouldn't the direction of the diodes (for example a stock SI and an LED) matter, in terms of the harmonic content???

http://....org/members/soulsonic/schematic/DS-1_Distortion.pdf

That's an old version of my doc.  You should pick up the newest version from the link in my sig.  Thanks!

[cctsim]

Increasing R9 should reduce the overall gain and improve linearity (reduce distortion). The concept is called negative feedback or current feedback in this case.
The stock value of 22 Ohm is too small to have a big effect on the distortion and is used for stability purposes (in my view).

[5thumbs]

Increasing R9 should reduce the overall gain and improve linearity (reduce distortion). The concept is called negative feedback or current feedback in this case.

Exactly.  I believe this is one of the reasons why Jack Orman suggested it [Q2 rebias] in his DS-1 Fat Mod.  It's also the reason I tweaked Q2...to get less treble boost and a more flat-spectrum tone going into the op amp stages.

Once again, I could be ignorant here, but I don't believe D8 is used as any sort of asymmetrical clipper (which was part of the original topic of discussion.)  If it is, could somebody please explain the science/engineering behind why it would work as such?  (Thanks.)

FWIW, the doc linked in the original post is an old version.  If you click on the link in my sig line, you'll download the latest version.

[cctsim]

As far as I understand, D8 is an opamp input protection diode similar to D1/D3 in BD-2.  It doesn't contribute much to clipping, at least in the simulations I did.

With regard to re-biasing Q2, I think is totally unnecessary. You can achieve exactly the same result by controlling the AC gain with the feedback capacitor C4.

Here is a SPICE simulation I did to investigate the effect of C4 on the gain of the CE stage:

[WGTP]

Sorry about the link (great stuff). Thanks for your explanations. I'm just trying to figure it out myself. The older version doesn't have the buffer in between.  I wonder if it was added to "smooth" the distortion of the op amp. It appears that I don't understand the operation of that gain stage. Isn't it the same as the stages in a BMP?  I was thinking the 22 ohm resistor would limit how far the signal could swing in one direction before distortion occurs. Maybe it doesn't have enough gain for this to occur.  

IIRC the diode is not set up in the feedback loop like the BD-2, but it may function in the same way.  

[5thumbs]

As far as I understand, D8 is an opamp input protection diode similar to D1/D3 in BD-2.  It doesn't contribute much to clipping, at least in the simulations I did.

With regard to re-biasing Q2, I think is totally unnecessary. You can achieve exactly the same result by controlling the AC gain with the feedback capacitor C4.

Very interesting info.  Thanks for posting it.

As for the need to rebias Q2, the stock biasing of Q2 on both the original and post-1994 DS-1 pedals gives a larger gain increase than Jack's or my Q2 mods, but also remember that Jack rebiased Q2 and the primary results were decreased treble boost (the stock misbiasing of Q2 caused a large boost in treble frequencies, not a consistent, linear gain increase in all frequencies) and increased bass (due to increased input impedance).  These are frequency-dependent effects that are not necessarily connected to the C4 value as they are the entire biasing of Q2 and/or its interop with the rest of the DS-1 circuit.

I rebiased Q2 to back away from some of the bass increase in Jack's mod values, plus I wanted to cut the treble as well.  I definitely want to play with C4 in future mods related to Q2, but at the start, I'm wondering if changing its value will have frequency-dependent effects as I've described previously.

Oh, and finally, WGTP, there is a Q2 booster in the original M.I.J. DS-1 as well.  It's listed on the old service sheet for the M.I.J., pre-1994 DS-1.

[WGTP]

Not the booster, the buffer in between isn't in the older design.

I better stick to the op amp, clippers, tone stack etc. until I get the front end figured out...

[cctsim]

I did some simulation re-biasing vs. changing C4  and it appears that results are very similar, e.g. in terms of the frequency response at least.




In both cases, the treble boost seems to be constant in both cases.

[5thumbs]

Not the booster, the buffer in between isn't in the older design.

I better stick to the op amp, clippers, tone stack etc. until I get the front end figured out...

Yup, you're right.  The TA7136P op amp is a single, decompensated op amp with an odd 7-pin pinout.  When Toshiba discontinued the TA7136P, BOSS switched over to in-line, dual-741-style op amps in the DS-1 in 1994.  I'm assuming the first op amp half was put into a buffer configuration to minimize the changes from the previous design.

What part of the front end are you still figuring out?  Perhaps we can all collectively scratch our heads together and further all of our knowledge on the DS-1 front end.

CCTSIM, once again, that's cool info, but it makes me wonder then if the treble-frequency-boost occuring with the stock biasing is a result of the transistor itself (2SC2240) rather than its biasing.  If it's the former, then the treble-boost might not show up in a SPICE simulation.  (FWIW, I'm trying to get what I've heard when rebiasing the Q2 tranny to jibe with with the sim data you've shared with us.)

[WGTP]

I don't understand why there is so much treble boost.  What is reducing the bass?   

[ayayay!]

Like everyone said above, it's in the Jack Orman article about Q2 w/ the 22ohm on the emitter  I personally like to just bump it to 100 ~ 120 and call it a day.  Works great.  More than that and you may have to start swapping other parts. 

[5thumbs]

I don't understand why there is so much treble boost.  What is reducing the bass?   

Jono's right about Q2 acting as a treble booster in the stock design.  R17 is another way to increase the bass in the DS-1 circuit.  Increasing R17 will hold more bass frequency in the mix.  15KΩ is a popular replacement value for R17 (which is 6K8Ω by default.)

[harrisxr650]

hey guys sorry if this is in a wrong section. i did jack orman's "ds-1 q2 gain mod" along with keeley mods and i liked the tone i got especially from Jacks mod the prb is i lost lots of distortion which would be nice if i could get it back while keeping the same tone ..is that possible? any tricky advice?