Another Blast From The Past
Circuit Analysis And Re-design Criteria
Hands-on Experience
4049 vs. 4069
Reference

Another Blast From The Past

A contemporary version of The Spectra Distortion Generator by Charles R. (Chuck) Fischer (click on the pictures to load a larger image)  Charles R. Fischer was an inventor and writer for a musician oriented magazine called Electronic Musician. In the 80ies, there existed a lot of such magazines addressing the electronic musician who wanted to make his own gear, and indeed, needed to do so, because there was no stomp box culture as we enjoy it today. Those mags were probably the forerunners of today’s stomp box (a.k.a. “pedal”) makers forae. No standards had been established by then, and it its due to this fact that some design decisions appear awkward by today's light, and consequently are worth to be overhauled, while the basic circuits may be as fresh today as they were then. The pot settings shown are typical settings.

Amongst this league range overdrives using CMOS chips, as is the case with Fischer’s Spectra Distortion Generator, also termed EM distortion box and (misleading) EM fuzz. The documents are kindly hosted by GGG [FISCHER]. Fischer had invented it in his late days, and he was convinced that it will become his most popular build. We'll see.

As mentioned before, some of the decisions that were made are due to the circumstances of time, so I overhauled the whole schematic to meet today’s requirements and quasi standards.

Using a CMOS gate biased linearly as a distortion generator was en vogue at this time. Fischer’s design differs from the plethora of such circuits by the interesting filter section and his spectra control, which feeds the CMOS gates with a dedicated current source.

With all the amendments, certainly one of the more usable specimen of the kind.

Back To Index

---

Circuit Analysis And Re-design Criteria

To download the schematic of the re-design, click on the icon By doing so, you agree to the Terms of Use

---

Input Stage

Using the 5534 for a guitar interface is questionable. The 553x and all bipolar op-amps have a higher noise contribution with the given input impedance than j-fet op-amps, caused by their bias current. I suspect that the NE553x was new at that time and had the fame of being the ultimate op-amp for all designs. It is not for this position, however, in non-guitar applications (read: suitable impedance conditions), the 553x will win no doubt, as myriads of builds will proof.
Bipolars do not like high value bias resistors and tend to cause offsets on the output, which can lead to rumble upon sweeping direct coupled pots.
Fischer proposed a maximum gain of 100, which appears way to high considered that there is nothing to put boundaries to the signal except the supply rails. The input amp will clip in no time, and I am not keen to hear a 553x in this mode – unless that is what he wants. Several noteworthy designs rely on op-amp clipping, and the nod towards a MXR distortion+ would be indicative for that, but I doubt it. I settled for a gain range of 6-25.

I replaced the whole input arrangement with my trusted Boscorelli input stage [BOSCORELLI], albeit using an LF356, which promises the lowest noise level in this configuration (closely followed by a TL07x). I found that biasing the LF356 slightly below mid yields a bit more of clean headroom. The TL071 on the output will behave differently, but by then the situation has relaxed hopefully. Note the added “body” control on the input stage. I will explain it later.

---

CMOS Chip Choice

For a CMOS gate, 4049 and 4069 will work. Why not try both and hear the difference. There does not seem to be much of a difference in their construction, but people have reported a different tonality, which coincides with my own experience.

Populating the CMOS chip(s): (click on the pictures to load a larger image)  From left to right: empty slots / CD4049 / CD4069.Beware, they do not have the same pin out. Note pin1 orientation. On my schematic, both chips appear being in parallel, but they are not present at the same time. Only one socket is populated. I chose the gates so that the two sockets can be positioned inside each other, since only one ever is used. This way both chips can be evaluated, with only a little more PCB estate consumed. However, this is optional. In the picture, sockets are shown, which will interfere with each other. In reality, I used precision single-in-line pin-strips.

---

Bias Trim Pot

I added a 100k stop resistor to the “bias trim” pot to provide a minimum resistance. Otherwise, the user may inadvertently dial down the pot to a flat short circuit between input and output.

---

Rail Clipping Danger

The CMOS chip is powered by a constant current source. Later designs have used a resistor in that position, which is somewhat similar in effect. By using a dedicated current source, the voltage swings are unpredictable. They may well bring the output buffer into clipping. From past experience, I installed a 2:1 divider to counteract this beforehand, because half of the signal at this point is still a monstrous signal.
Care must be exercised because the filter section allows for both cuts and boosts. Under full supply current the gates will swing full range, and despite our divider, the signal can run into clipping.

---

Treble Filter

Since normally no treble boosts are needed for such circuitry, I have increased the stop resistor towards the boost side from 2.2k to 15k to mitigate the chance of exhausting headroom.
Also, with some settings, treble cut appears too weak. I also changed the other stop resistor from 2.2k to 560 Ohms for more effect.The tone cap has been changed to 15nF for more efficient treble control.

---

Power Supply

Fischer’s powering scheme is due to years bygone:

“This project requires a high-quality, regulated power supply for the audio portion”    - excerpt from the accompanying text

It does, which we will explain in a second, but he starts out from using a wall-wart AC supply. We can assume safely, that with the advent of today’s pedal culture we most certainly will meet a regulated 9Vdc power source.

The original circuit is powered by a regulated 12Vdc source, which gives a bit more clean headroom than 9V, but still cannot accommodate an input signal boosted 100 times.

I left away all the gew-gaw for bypass switching, so we can omit the entire 5V section along with the associated circuitry. This is just bloating the unit and is nowadays done differently anyway.

Nevertheless, due to the lack of power supply rejection capability, the discrete components need some filtering. A capacitor multiplier comes to mind. Like all regulators, this one is lossy, but the CMOS circuitry’s analog function works better on lower supplies anyway. A lower headroom is welcome in this instance.
The OPAs on the other hand will be powered directly by the full 9Vdc voltage for maximum headroom, and they have sufficient PSRR by themselves. However, as mentioned earlier, we can expect a fairly well behaved supply anyway.

---

Constant Current Powering

According to the datasheet, the LM334Z inherently introduces noise on the output, so a cap is recommended across the CMOS chip. Also, a look at the oscilloscope revealed that lower frequencies tend to modulate the output current, which appears as a frequency which is synchronous with the audio signal, riding on top of the DC supply of the CMOS gate. This may or may not be pleasant. A 1 µF cap stops this. Fact is that tone becomes feeble if there is no capacitor.

---

Spectra Control

Fischer uses a small stop resistor to each side of the SPECTRA pot. This probably is a remnant of an earlier design, because it makes no sense to split up that resistor. Series is series, no matter to which side the fixed resistor abides, so a single 150R resistor performs absolutely equivalent.
The author of the accompanying text apparently re-engineered the circuit after the demise of Charles R. Fischer. He might have been in error over that. He certainly made some minor errors in the description, which aren’t of any consequence.

Back To Index

---

Hands-on Experience

Not surprisingly, this unit again exhibits the CMOS-typical lack of dynamics. It turns into full compression very quickly. Because bass is pure poison for those and leads to quick congestion and splattering, I strongly recommend a bass cut feature before distortion like the one I used. It makes tone slimmer, more modern, hence the name body. Cutting bass ahead of distortion can make the meanest distortion unit palatable. Conversely, a bass cut afterwards cannot repair the damage already done.

The spectra control is less spect(r)acular than Mr. Fischer may have wished. There is some nice sounds to be had with low settings of “drive” and reduced bass, but nothing, we have not heard before.

The 4049 and 4069 behave differently in this application.
Both of them exhibit gross cross-over distortion symptoms with decreasing current. Naturally, the MOSFETs are running below their turn-on threshold. This sounds like a mis-biased Fuzz Face and is not appealing to my ears, but it might to some. Maybe this is what some guys deemed 70ies tone.

The treble control is surprisingly efficient, but even that is not enough to tame the overall brightness.
On low drive settings the unit appears dull, which is due to the large feedback cap. Yet increasing distortion even slightly increases the treble content rapidly and you have to throttle back on them. But overall the control’s tonal range is useful.

An input gain range from about 6 (safe value) to 25 appears more than enough for even the lamest single coil pickup. The CD4069UB sounds clearer and less splatty to me.

I tried the LM334 with and without a larger output cap, and it overall sounds better with it in place. In the stock version tone loses its body when spectra is dialed back. I have a small 100nF ceramic cap there permanently to filter out HF noise, but I now recommend 1µF.

From all the C-MOS dirt boxes I made, this is the most versatile up to now. It does not, however, end the quest for the ultimate, all-singing, all-dancing dirt box, as Fischer may have wished for.

Back To Index

---

4049 vs. 4069

In the light of recent events I tried different maker's chips. I found a N.O.S. MC14069UB, genuine Motorola, ceramic package, and this sounds much mellower and not splatty at all. How can that be? I looked deeper.

The subject seemingly is as old as the hills (it may however appear new to us, who never looked into that matter...). A forum entry from 2004 says:

“I think it may be difficult to generalize too much about the 4069 vs the 4049. the reason is that there is a ridiculous amount of variation among 4069s alone (and similarly the 4049) that I'd be hesitant to say one chip always behaves like x and the other like y.”        - Tim Escobedo, DIYstompboxes.com (post#7), CD4049UB v/s CD4069UB?

“As was pointed out, the internal circuit varies from maker to maker. All that you have to do to sell one as a 4049 (or 4069) is meet the input/output and logic specs. They will (and do) vary from maker to maker unless you use them as logic chips. ”        - R.G. (Keen), DIYstompboxes.com (post#11), CD4049UB v/s CD4069UB?

Remember, those chips are used outside the purpose they are specified for.

While I do not advocate the current trend for op-amp-rolling and buying prohibitively priced N.O.S. chips (why? because they are normally used within their designated purposes), I have a feeling that in this case indeed old ones perform better. I observed a steady increase in noise and gain in recently manufactured chips if used left field like in the given application.
Again, this may be totally inconsequential for their designated purpose.

The 4049 is frequently preferred because it yields a non-symmetric distortion due to its internal structure, as different to the 4069. This adds some even harmonics. Unfortunately this does not help if the chip is unsuitable from a start.

I have no means to prove this, but I suspect that CMOS units generate sub-harmonics as bi-products of intermodulation distortion [BLENCOWE], which would explain the bass overload. You can cut bass pre-distortion quite substantially, and bingo, there it is again on the output. Whatever the cause may be, bass tailoring each subsequent distortion stage (there is only one here) is mandatory.

---

Observations

The 4049  and 4069 have different behavior regarding gain and bandwidth.
With "drive" dimed, the 4049 sounds more distorted, but not necessarily better.

►

Applying the increased knowledge aquired with this project, I revisited an earlier CMOS unit I had made a while ago, which I had never been happy with. I found a SCL4049UB by Solid State Systems, an extinct company (and not to be confused with SCL = Semi Conductor Laboratory in India), in a ceramic case, and this sounds great. The unit is equipped with a bass cut pot similar to the one shown here, but I oversaw, that the second stage (it is a two-stage unit) again requires some bass  taloring.

Using the 4069, with "drive" at 3 o'clock and "Spectra" full, is very dull (and pretty low distortion too). The treble control has to be substantially turned towards brighter settings, and some of the body has to be removed. With drive maximized, treble has to be turned down far towards its minimum position. Body needs a correction towards less bass.

Lowering "Spectra" to 3 o'clock, yields a different, more grainy, but fluent tone, however body gets thin and has to be increased. In lowering it even lower the signal starts to quickly diminish and eventually look like cross-over distortion and finally no distortion. The usable range is from 2 o'clock to maximum with a 5k pot and 4069.

All controls are very interactive, but there is several useful tones with a somewhat different texture to be had.
Different maker's chips behave differently.

One would think, there is not much distortion, but looking at the scope, there is. However, the waveform consists of mostly rounded corners, hence not many aggressive harmonics. You will notice some heavy compression though.

---

Conclusion

Both chips are different, compared side-to-side, and supposedly equivalent ones (from a different manufacturer and / or era) are different despite the same name. People noticed differences even between the exact same batch. This is one of the rare cases where the brand does make a difference, because it is used at its (technically not described) limits, outside the purpose it is destined and specified for.

Try some brands and keep the one you like.

Back To Index

---

Reference

[FISCHER]
GGG kindly hosts the EM fuzz documents, https://generalguitargadgets.com/how-to-build-it/technical-help/articles/em-fuzz/

[BOSCORELLI]
Nicholas Boscorelli, The Stomp Box Cookbook, 2nd ed.,  pp.112ff,
https://pdfcoffee.com/qdownload/the-stompbox-cookbook-nicholas-boscorelli-1999pdf-pdf-free.html

[BLENCOWE]
Merlin Blencowe, Designing Tube Preamps for Guitar and Bass,  pp.17f,