Trip Hazard Optical Disruptor - Assymetric Optical Compressor/Low-Freq Saturator

General documentation, gut shot, schematic links, ongoing circuit tracing, deep thoughts ... all about boutique stompboxes.
Post Reply
User avatar
bow_and_error
Breadboard Brother
Information
Posts: 56
Joined: 13 Jun 2019, 23:13
Has thanked: 95 times
Been thanked: 41 times
Contact:

Post by bow_and_error »

So I recently built the Bajaman Optical Limiter and loved it. It's warm & responsive on bass/baritone guitar, and is better than my Xotic SP at adding that "glue" later in signal chain. Because I had to build a bunch of LED/LDR optocouplers to test in that, I decided to find some other optical circuits to play around with.

One of the things I found was the XQP 545 Optical Disruptor/DIYRE Colouruptor, and was super taken by the low-freqency saturation, bordering on waveform modulation at times. According to XQP, the Disruptor started from him messing with their Optical De-esser, which is a feed-forward compressor (sidechain signal is taken from input of GR circuit rather than output) with a highpass filter in the sidechain. He found some interesting behavior after removing the sidechain high-pass filter: smooth, predominately low-frequency, distortion that is asymmetric (compresses top of AC signal but not bottom).

I've been following DIY Recording equipment (who have a great podcast & YT channel), but was reluctant to buy a 500-series rack & not yet confident enought to build a bipolar +/-15V PSU. But surprise, there is a pedal version! It's under a different name, Trip Hazard Effects, and has some changes (more on that later), but they were kind enough to post a schematic, like they do for all of their products.

My Build:
The schematic is a bit difficult to work with, so I made a simplified version without the bypass circuitry & with some more common resistor values. I also didn't have the TLE2426 Rail Splitter IC, so I used the Precision 1/2V Supply from the Bajaman Optical Limiter/Holy Roller. This change is also why the numbering for the resistors/capacitors doesn't start at 1.
Image

I had a bunch of unlabelled LDRs that I've roughly tested & organized based on their dark/light resistance. I had the best luck with a:

- 3mm Red LED (mine were translucent red)
- LDR w/>20M dark resistance & ~2.5K light resistance (tested with 4.7K CLR off a 9V supply). If you're going to test your LDRs, make sure you get the LED perpendicular to the LDR face so that it's not shining off-axis. High dark resistance LDRs are a must, otherwise you will get gain reduction even when there is no input signal.

I also experimented with a trimpot in place of R12, which can help set the upper limit for the amount of Disruption and may be useful for dialing in different LED/LDR combos. To my ear, this sounds closest to the demos (XQP demo/Trip Hazard demo).

Image
I built this all up on a breadboard, and am working on a vero layout, but it turns out I'm @#$%ing really bad at that. If anybody wants to do a perf/vero layout, I would be glad to build & verify it!

Circuit Analysis:
Audio Signal Path - I'm pretty new to this, so please correct my assumptions:
1.) Input feeds 1st TL084 gain stage
2.) Sidechain splits off of the main signal path
3.) Main signal path continues through the Gain Reduction section (LDR shunted to ground)
4.) Post GR signal through 2nd TL084 gain stage
5.) 2nd gain stage ouputs through pot for level control to Output

Sidechain:
1.) Sidechain split feeds 3rd TL084 gain stage with open loop gain?
2.) Disrupt pot attenuates signal to 4th TL084 gain stage
3.) 4th gain stage output feeds cathode of optocoupler LED, which is in its feedback loop
4.) Optocoupler LED controls GR before 2nd gain stage

As I've spent quite a bit of time with this circuit, I have some questions that are beyond my level:
1.) I read that the sidechain LED is driven by unrectified signal from a transconductance amplifier (which is a V-to-I converter), allowing input voltage to control LED brightness & giving GR more soft-knee character. All of the gain stages, except the open loop one in the sidechain, look like what I've seen referenced as transconductance amps, is that what is going on here?
2.) Comparing the sidechain LED driver to Bajaman's Optical Limiter, the LED is in a different place. Most optical comp/limiters that I've found route the output of the final sidechain gain stage into a current-limiting resistor before hitting the LED anode and then ground. The Disruptor has the LED in the feedback loop, with the gain stage output feeding a CLR before hitting the CATHODE of the LED before splitting to resistor/ground & the (-) opamp input. Is this where the compression gets it's assymmetric character from? I understand that ideally an opamp wants to keep the difference in voltage between its +/- inputs at 0V, and that the output is an amplified version of the difference between them, but visualizing how this effects the sound is definitely beyond my understanding.
3.) The 500-series version of the XQP Optical Disruptor uses 2x identical vactrols for gain reduction, which I might try. I guess this works as a voltage divider across the 2 LDRs, possibly allowing for differences in dark/light resistance value?

If you have any circuit improvements or corrections, please let me know. I'll record some sound clips this week and hopefully finish a reasonable vero layout (or find one of you fine folks that would be kind enough to do it).

User avatar
bow_and_error
Breadboard Brother
Information
Posts: 56
Joined: 13 Jun 2019, 23:13
Has thanked: 95 times
Been thanked: 41 times
Contact:

Post by bow_and_error »

It's been a while since I originally posted this, and since then I've made a few revisions + verified my first perfboard layout.

The biggest changes were:
-Swapping the polarity of the LED in the vactrol. The original orientation ended up kinda sputtery, but when I flipped it around it sounded much smoother on the low frequencies
-Changing the Disrupt pot taper to linear to keep the control from bunching up at one end.

Updated Schematic:
Image

I'm hella new to creating layouts, so I found VeroRoute to really helpful & easy to use. The first perf layout worked great, but the component spacing was really tight. This version is an improvement, but still requires some vertical resistors:
Image

Image

Image

You can see the jumbled components above pin 1 of the TL084, but it sounded good!
Image

I'm uploading some sample clips to Soundcloud, so I'll post those later today.

User avatar
bow_and_error
Breadboard Brother
Information
Posts: 56
Joined: 13 Jun 2019, 23:13
Has thanked: 95 times
Been thanked: 41 times
Contact:

Post by bow_and_error »

I uploaded audio clips here. Each setting plays for 8 bars (2x 4-bar loops):
1.) Bypassed
2.) Density 10%
3.) Density 30%
4.) Density 100%


Higher settings are noisy (as expected), and I am getting some popping when I turn the pedal on. Any ideas as to why that might be happening? I still have the perfboard in my non-shielded test rig, so I'll see if boxing it up helps.

User avatar
bow_and_error
Breadboard Brother
Information
Posts: 56
Joined: 13 Jun 2019, 23:13
Has thanked: 95 times
Been thanked: 41 times
Contact:

Post by bow_and_error »

So I solved the switch popping on this circuit, but haven't had time to update the schematic:
- Any ground references before C2 or after C10 (so input GND/output GND/R2/R11) need to reference actual peadl ground - i.e. DC jack center pin & input/output sleeve)
- Any ground references inside those 2 caps are all referenced to Vb from the buffered bias.

That should remove any popping from the switch.

After lots of experimentation, I also found that the VTL5C1 is pretty crucial to getting the right sound in this circuit. As you can see from this VTL5Cx-series datasheet, it is the fastest vactrol they offer, and still has a rather high (>50M) off resistance:

Image

A VTL5C9 might work, but it has a slope around half that of the VTL5C1, so you'd have to try it out to be sure.

User avatar
bow_and_error
Breadboard Brother
Information
Posts: 56
Joined: 13 Jun 2019, 23:13
Has thanked: 95 times
Been thanked: 41 times
Contact:

Post by bow_and_error »

Here is an updated schematic with more clear delineation between chassis/IO ground & the Vb that the internal circuit uses:
Image

I also updated R9 to 7.5k, which provides more makeup gain after the saturation/compression section.

User avatar
roseblood11
Tube Twister
Information
Posts: 1887
Joined: 23 Aug 2008, 14:21
Has thanked: 366 times
Been thanked: 313 times

Post by roseblood11 »

The Trip Hazard website is down.
Here's the archived link, including the pdf's:
https://web.archive.org/web/20161108062 ... ducts.html

User avatar
bmxguitarsbmx
Cap Cooler
Information
Posts: 510
Joined: 25 Mar 2010, 21:15
Has thanked: 704 times
Been thanked: 216 times

Post by bmxguitarsbmx »

I think it sounds great. Thanks for your hard work documenting it all!

User avatar
bow_and_error
Breadboard Brother
Information
Posts: 56
Joined: 13 Jun 2019, 23:13
Has thanked: 95 times
Been thanked: 41 times
Contact:

Post by bow_and_error »

If anyone is looking for a vero layout, Dirtbox Layouts did one for the Optical Disruptor: http://dirtboxlayouts.blogspot.com/2021 ... pedal.html

User avatar
Zokk
Resistor Ronker
Information
Posts: 287
Joined: 05 Jul 2010, 20:00
Location: tantalum shortcut area
Has thanked: 211 times
Been thanked: 68 times

Post by Zokk »

Hello I'm always amazed to see a TL084 driving a VTL5C1... this opamp is ok for audio, but not really as LED driver :scratch:
It should not be able to provide enough current to the VTL LED, so I don't think the min/max chart posted earlier means something in this case. Remember those VTLs need 40mA to achieve the lower resistance as per the datasheet.

First thing would be to take a measure of the LED current in the original pedal if possible (it really uses a TL084?), it would help to define the real usable range in this circuit. So you can draw a chart mA/resistance based on the real thing. A bare resistance measure is meaningless without the LED current.
Second thing would be to swap IC2D for a "stronger" opamp (TS912INs are my prefered those days).

Do be fooled by those VTL, Perkin Elmer uses old design LEDs running at high current to achieve the linear curves. They are not suited for pedals at all... but you see them in Buchla modules... not the same beasts and not the same power supply.

Cheers.

User avatar
bow_and_error
Breadboard Brother
Information
Posts: 56
Joined: 13 Jun 2019, 23:13
Has thanked: 95 times
Been thanked: 41 times
Contact:

Post by bow_and_error »

That's an interesting observation, I was thinking it had to with the TL084 having a lower DC offset than the TL074, but the datasheets don't seem to reflect that:

- TL084 Input Offset Voltage: 3mV Typ, 15mV MAX
- TL074 Input Offset Voltage: 3mV Typ 10mV MAX

I wonder if maybe the idea is to NOT drive the LED completely, in order to keep the LDR in a lower-resistance range. Unlike an optical compressor, the idea is to modulate the dry signal with the AC sidechain signal to cause distortion, rather than to completely attenuate the signal. I'd imagine reaching complete attenuation would probably cause a dropout of that distortion effect as well.

Although with that said, I do note that the Response time of the VTL5C1 is different depending on the input current: the Turn-Off time is similar, but more linear, and the Turn-On time looks somewhat slower - see page 3 of the datasheet. I'd imagine faster response time to be desirable for proper AC modulation.

Either way, the Trip Hazard Disruptor pedal (schematic) is based on XQP's Optical Disruptor 500-Series module, which uses an LF351 to drive the LED instead of a TL084. The 500-series version also uses 2x vactrols in parallel, which may have driven the decision to use that opamp - see Schematic here. Would the LF351 be more capable of driving an LED? I think the [urlhttp://pdf.datasheetcatalog.com/datasheet2/f/0cxhzq2yxy7uwfkwj8aec59lu2ky.pdf]LF347[/url] is the quad version of that opamp, so maybe I'll have to order some and drop one in.

User avatar
Zokk
Resistor Ronker
Information
Posts: 287
Joined: 05 Jul 2010, 20:00
Location: tantalum shortcut area
Has thanked: 211 times
Been thanked: 68 times

Post by Zokk »

Hi you may be right, if the lower resistance on the LDR is not the goal there's no need to be able to supply 40mA from the opamp.
That's a good trick to have a faster time response from the LDR and achieve FM-like modulations in a narrow range.
In this case the LF351 should behave the same as a TL074 or TL084, they are all "low offset" opamps. But if you need to attain lower resistance meaning a higher current in the LED, the LF351 seem more capable to handle the load.
Regarding the offset itself, at worst 10/15mV is a very small voltage drift, not important in this circuit I think.

I don't understand the use of a dual vactrol in the 500 module... both in parallel :scratch:

Could you post some values: LDR min/max resistance in your build?

Post Reply