Different opamps in the OCD

[thermionix]

Not really a question, more musing I suppose.

I built an OCD clone on perf, v4 but without the Ge diode.  Or maybe it's a v3 with a log taper volume pot.  Whatever.  As I don't have the opamp yet, I put a socket in.  While I'm waiting for the TL082, I tried a few other duals I have around just for shiggles.  I tried TLC2272, 4558, OPA2134, and 1458 so far.

To my surprise, I like the 1458 the best.  Especially at medium gains, it sounds more, uh, "natural" with less high-end trash.  The TLC2272 did the higher gain near-metal sound the best.  The 4558 and OPA2134 were rather uninspiring.

I'm curious how the TL082 will compare to the 1458.  I've never played a real OCD.  I also have TL072s coming, as I'm out of those.  I have some NE5532s but haven't tried one yet, I'm guessing that one would be in between the 4558 and OPA2134.

Well, that's it really, just kinda thinking out loud.  If anybody has some other suggestions, good or bad experiences, I'd be interested to hear 'em.

[antonis]

Not quite suprisingly, Thermionix.. 

Both OCD & 1458 are of the same age (more or less..)

[Mark Hammer]

The 1458's strength is its weakness.  As an earlier generation op-amp, its ability to reproduce high frequencies at higher gains is quite limited.  When a pedal is designed to generate lots more harmonic content, being able to rein it in a bit is good.  A local fellow had read somewhere that the Timmy pedal sounded "better" with a 1458, and came by asking me to swap chips on his.  We spent the evening trying out about 8 different dual op-amps of various categories, and both agreed that the 1458 took the brittle edge off his Timmy nicely.

Lest this get out of control, the 1458 is not some superchip!  That said, there are various instances where "golden oldies" from the 1970's, like the LM741, LM308, or 1458, can help out a circuit a bit, via their limitations in that particular context.  The typical context will be one where some clipping element to ground is employed, however the Timmy uses diodes in the feedback loop, like a Tube Screamer.  So, I guess the moral of the story is that, when one or more op-amps are used in the clipping portion of a circuit, unless there are serious space requirements, it's a good idea to use an IC socket, because you never know when a swap might make a difference you can appreciate.

[duck_arse]

last time I did messing with opamp swaps, I was looking at the clean output levels. from memory, the 1458 had a greater clean swing than the TL0's at the same supply and bias voltages. don't quote me, tho.

[jsleep]

Another one to add to Marks list:  LM351 (dual version LM353).  I scored a couple older "Can" version of the LM351s and I love it in a DIST+ etc circuit.

JD

[iainpunk]

Hey, great that you chose to put an socket in, i have had goos luck with the CA3160. It is a single opamp, but i build it on a breadboard so i had the space. it feels to me like they sound more like they soft clip a tiny bit before the mosfets kick in to clip harder. Maybe if you could find them you could solder 2 of em on a board with headers, so it has the same pinout as a dual opamp. Godspeed and have fun building the pedal.
Iain

[thermionix]

don't quote me, tho.

[thermionix]

Another one to add to Marks list:  LM351 (dual version LM353).  I scored a couple older "Can" version of the LM351s and I love it in a DIST+ etc circuit.

JD

Cool, I'll look out for one of those, don't think I have one right now but the next parts order is always right around the corner it seems.  BTW, I've used 3 or 4 of your boards now and they've all worked beautifully.  The Phase 90 board is especially awesome.

[thermionix]

Both OCD & 1458 are of the same age (more or less..)

About 30 years apart, aren't they?

[Mark Hammer]

Another one to add to Marks list:  LM351 (dual version LM353).  I scored a couple older "Can" version of the LM351s and I love it in a DIST+ etc circuit.

JD

Long time no hear.  Hope all is well with you and the family...and the business.

The LF351 is all over the Electronotes binders I have.

[antonis]

About 30 years apart, aren't they?

Give or take.. 

If you also have got some LF411, give them a chance ...

[duck_arse]

hmmm, LF351 and the dual LF353.

[EBK]

I'm just imagining you obsessing over taking out op amps, putting them back in, taking them out, putting them back in, etc.  It's so meta! 

[teemuk]

It may seem obsessive but the OpAmps are not all equal and do behave differently in certain conditions. Which can explain why they sometimes introduce different "tones".

Besides few details (of which most matter very little in "LoFi" -effect design) OpAmp circuits operate pretty much "uniformingly" as long as their feedback loop remains operational (no pun intented). This is pretty much the sole purpose of an operational amplifier to begin with. This all, of course, changes if the OpAmp is overdriven outside its linear range. This is the region where circuit architecture differences of the OpAmps begin to display themselves.

- Different OpAmps have different limits in how close to supply rail voltage(s) their output signal can swing. It is common for these limits to have some degree of "asymmetry". e.g. positive half wave swings within 1V and the negative swings only within 2V.
- Some OpAmps "latch to supply rail" when overdriven to clipping. The poor recovery characteristic introduces much "nastier" harmonc content than plain brickwall clipping,
- Some OpAmps reverse phase of conduction when their common mode voltage is exceeded. This generates sort of an "octave effect"
- Some OpAmps may introduce "ringing" oscillation bursts at sharpest waveform corners
- Low slew rate can prevent the gneration of most obtrusive very high order harmonics
- Noise ratio is usually an issue in "high-gain" -design
- Current limiting circuits may begin to operate if the opamp's load becomes too low in impedance.
- "Rail-to-rail" -type opamps are - for obvious reasons - designed to have decent overdrive characteristics (they pretty much just brickwall clip without any "side-effects") but their low limits of operating voltage can pose practical designing challenges.

Ok, where do we encounter overdriven opamps...
- Stages configured for sole OpAmp overdrive should be pretty self-explanatory.
- "Shunt clipping" circuits driven by an opamp gain stage. (Isn't OCD in this category?) These can often be deceiving. The opamp stage WILL overdrive with high-enough input signal or gain levels (though typically this happens well above clipping threshold of the shunt circuit). So, with "high gain" settings dialed in overdrive of the opamp stage is very much a possibity. The shunt clipping circuit will limit signal peaks exceeding its clipping threshold as usual, but it can not prevent any of the "ill-effects" it's driver stage introduces when overdriven. These may contribute to total distortion in certaim conditions.
- Feedback loop clipping in non-inverting configuration. If these circuits are overdriven excessively the opamp stage begins to clip to rails.

Could be some other examples too.

[EBK]

It may seem obsessive but the OpAmps are not all equal and do behave differently in certain conditions. Which can explain why they sometimes introduce different "tones".

I agree.  It was just the notion of constantly fiddling around with a circuit named OCD that struck me as humorous. 

[jsleep]

Hey Mark!  Yeah, I am well, family is good, business is so-so which is probably why I have some time to check in on my favorite forum again     Hope you are doing well also.

I don' t know if Gus is still around, but I think he had said that the newer manufactured 351s weren't as good as the old ones, (old are probably "National Semiconductor" long before they got sucked into TI).  Assuming the "can" version I have is truly old and not a Chinese fake,  and that the 8-pin ones I have are new, I couldn't tell much, if any difference between the old and new.   I have very bad "echoic memory", so if I don't have an A/B set up to hear the two sounds immediately one right after the other, I can't tell much. Testing different ICs is definitely worth the time if you have a good ear or if you're are looking for certain sound characteristics.  Yeah, it might be OCD   

JD

[thermionix]

I ordered a DIP LF353 from The Bear as I needed a few other items anyway.  Don't know what brand it will be, but something else to try.

I think he had said that the newer manufactured 351s weren't as good as the old ones, (old are probably "National Semiconductor" long before they got sucked into TI).  Assuming the "can" version I have is truly old and not a Chinese fake,  and that the 8-pin ones I have are new, I couldn't tell much, if any difference between the old and new

Interesting.  I've built three Rat clones so far.  In the first two I used NOS Motorola LM308AH metal cans.  They sound cool, but not quite like I remember a real Rat pedal sounding, though I haven't played a real Rat since the 80s.  Then I got some new National Semiconductor metal can LM308AHs from Ebay, and I built the third with one of those.  It definitely sounds different.  It sounds more like I remember the real ProCo Rat sounding.  I would say the NOS Motorola builds have a bit more of a fuzz sound, the mids are more scooped, and the new NS has more of a JCM Marshall sound, more mids and presence.  Brash, but in a nice way.

The builds are otherwise the same, except for the obvious variations within tolerances.  Same PCBs and clipping diodes.  Given that the 80s Rats were built with DIP 308s and not metal cans, it makes me wonder if there was some actual difference in the old metal cans, where the new ones are just the same as the DIPs but in a different package.  OR, maybe the new NS cans aren't really 308s, but something else that works in its place, repackaged in metal cans and relabeled.

[thermionix]

I wasn't totally happy with the tonal response of this circuit, the lows and mids, so I tried a few simple mods.  First I changed the output cap from 10uF aluminum electro to a 1uF tant (there doesn't seem to be a consensus online as to what the stock value is), and at the same time changed the volume pot from 500k to 100k.  Might have been an slight improvement, but didn't get me where I wanted to be.  So I changed the 68n coming off pin 2 to 47n, and that was the ticket.  I think JHS does similar, I've read that they change 68n/2.2k to 100n/4.7k, for a reduction in gain and roughly the same tonal shift.  Either way, it's now voiced much better for my guitar/amp.  YMMV as always.