Catalinbread Manx Loaghtan on the bench

We got a lot of great feedback and kind words from our Internet friends about our writeup on the Ibanez TS-9 Tube Screamer, so my brother and I figured we should get back at it and put another guitar pedal on the bench and see what we could learn. We had a request to do a fuzz pedal and we thought that would be fun, so this time we're looking at the Catalinbread Manx Loaghtan. My understanding is that this pedal is meant to be similar to the Ram's Head Big Muff, though with a Baxandall type tone control.

In the comments we received about our previous write-up, some mention was made of looking at even- versus odd-order harmonics and the relative contribution of each to the total distortion produced, with the idea that this relationship might help quantify some of the differences in timbre that we hear produced by different effects. Luckily, the DSA we're using has the Instrument BASIC option and I was able to write a small program, EVEN-ODD.BAS, for the instrument to compute these relative contributions to total harmonic distortion from the measured power spectra. I think this was a great write-up to have this new tool available for and I appreciate the timely advice!

As before, I started with swept-sine frequency response measurements to characterize the operation of the bass and treble controls. After that, I moved on to take distortion measurements with a fixed-sine input to characterize the sustain control. I noticed some unexpected interaction between the tone controls and the distortion product blend and that this interaction depended as well on the supply voltage. I thought this interaction was interesting, so I made some additional measurements to try to characterize it.

A Catalinbread Manx Loaghtan guitar effect pedal sits on an electronics
work bench connected to a dynamic signal analyzer. Also visible are a manual
for the DSA, a glass of water, keyboard, soldering iron, and some hand-written
notes.

Methodology

I was able to use essentially the same instrument setup for these measurements as I did for the Tube Screamer measurements, with the exception that for certain measurements the power supply voltage was adjusted as noted below. The DSA configurations for frequency response and distortion measurements are available as FREQRESP.STA and DISTORTN.STA.

Findings

I started out with a quick measurement of current draw, something I now wish I had measured on the TS-9 as well. I found that, at least when powered from the external power jack, the pedal draws 2.18mA at 9 Volts even when switched off. When switched on, the current increases to 2.91mA. I imagine most or all of the increase is accounted for by the LED lighting up. A typical alkaline 9V battery has a capacity of about 600mAh at low currents like this and so should be good for something like 20 hours of use, though the supplied voltage may be quite low (perhaps 5 Volts) at the end.

I suspected a true-bypass switching arrangement on a boutique pedal like this, but made a quick frequency response measurement with the pedal switched off to check. The measurement confirmed my suspicion: dead flat response across the measurement range. I saved this trace as FRBYPASS.DAT.

I switched the pedal on and took a baseline frequency response measurement with both the treble and bass controls at their middle positions. This showed a more or less flat response between 78 and 3,800Hz, though the bass response was a little higher and technically rolled off 3dB from its peak by 1.3kHz. I think with some knob tweaking, you could dial this in as flat as you please.

A frequency response trace labeled Baseline Frequency Response shows
fairly flat response and approximately no net gain or loss in the pass band.

Turning the bass knob all the way down rolled the bass off smoothly below 580Hz. Sorry I spaced out on matching scales in the print below; the response with this setting peaks at about the same level in the treble region as the baseline, it just looks higher because the top of the scale is down 20dB, as you can see in the scale labels.

A frequency response trace labeled Bass Cut Frequency Response shows
a distinct reduction of bass response and essentially unchanged treble
response superimposed over the previous baseline trace.

Turning the bass knob all the way up gave a less-pronounced but still respectable boost to the bass frequencies down to 66Hz.

A frequency response trace labeled Bass Boost Frequency Response shows
a bit of boost in the bass frequencies tapering into a close match through
the mids and into the treble region superimposed over the previous
baseline trace.

I tried the treble boost next. Turning that knob all the way up gives a more pronounced boost than the bass knob did, out to about 7.2kHz.

A frequency response trace labeled Treble Boost Frequency Response shows
a pronounced increase in treble frequencies and close match in bass and mid
frequencies superimposed over the previous baseline trace. The bump up in
treble frequencies along with the small peak in bass frequencies that this
trace has in common with the baseline begins to highlight a small dip in
the mid-range frequencies.

Turning the treble knob all the way down resulted in high frequencies rolling off from about 910Hz.

A frequency response trace labeled Treble Cut Frequency Response shows
a pronounced decrease in treble frequencies, a tiny boost in mids, and a close
match in bass frequencies superimposed over the previous baseline trace. The
small dip apparent in the treble boost trace is not apparent in this trace.

I took complete measurements at each combination of treble and bass at their minimum, middle, and maximum settings. I used the DSA's marker feature to measure the peak response and find the -3dB corner frequencies of the resulting pass-band. For some measurements, a little dip in the mid-range created two peaks in the pass-band. In these cases, I measured 3dB down from the highest peak and also 3dB down from the lower peak. The corner frequency of the lower peak is in parentheses. With both knobs in their maximum positions, the mid-scoop was pronounced and I broke out separate measurements for the bass peak, mid scoop, and treble peak on this setting.

Treble setting Bass setting Peak response Corner frequencies Data file
Minimum Minimum -5.4dB 360 to 1,800Hz FRMINMIN.DAT
Minimum Middle 2.3 81 to 910 FRMINMID.DAT
Minimum Maximum 7.8 68 to 425 FRMINMAX.DAT
Middle Minimum -2.4 580 to 4,400 FRMINMID.DAT
Middle Middle 1.7 78 to 1,300 (3,800) FRMIDMID.DAT
Middle Maximum 7.5 66 to 370 (2,900) FRMIDMAX.DAT
Maximum Minimum 6.6 1,400 to 6,700 FRMAXMIN.DAT
Maximum Middle 5.8 (74) 1,600 to 7,200 FRMAXMID.DAT
Maximum Maximum three segments, detailed below FRMAXMAX.DAT
Bass peak 7.2 65 to 360
Mid scoop -2.9 600 to 1,400
Treble Peak 5.7 1,800 to 7,300

Next, I turned my attention to distortion measurements. I wanted to investigate the effect adjusting the sustain knob would have on the overall output level and distortion products at various input levels. In each set of measurements, the output level stayed the same until the input level fell below a certain threshold, then it fell along with the input level. True to it's label, the sustain knob adjusted where this point fell: high output would be maintained at lower and lower input levels the greater the sustain setting.

I noticed that even-order distortion was most pronounced at the highest and lowest input levels, with the least amount of even-order distortion products near the input level above which the output level stopped changing. This showed what I thought was an interesting effect: though the output level stayed the same as the input signal fell off, the character of the distortion changed smoothly throughout the process.

Minimum sustain setting

Source Fundamental Distortion File
Even-order Odd-order Total
-70 dBVRMS -45 dbVRMS 10% 1% 10% DMIN70.DAT
-50 -18 26 6 27 DMIN50.DAT
-30 -7 9 28 29 DMIN30.DAT
-10 -4 10 38 39 DMIN10.DAT
-5 -4 12 39 41 DMIN5.DAT
0 -3 24 37 44 DMIN0.DAT

Middle sustain setting

Source Fundamental Distortion File
Even-order Odd-order Total
-70 dBVRMS -26 dBVRMS 29% 1% 29% DMID70.DAT
-50 -9 11 21 24 DMID50.DAT
-30 -5 8 36 37 DMID30.DAT
-10 -3 11 41 42 DMID10.DAT
-5 -3 13 41 43 DMID5.DAT
0 -3 25 37 44 DMID0.DAT

Maximum sustain setting

Source Fundamental Distortion File
Even-order Odd-order Total
-70 dBVRMS -9 dBVRMS 11% 17% 21% DMAX70.DAT
-50 -5 8 35 36 DMAX50.DAT
-30 -3 10 41 42 DMAX30.DAT
-10 -4 19 41 45 DMAX10.DAT
-5 -4 25 36 44 DMAX5.DAT
0 -5 50 34 61 DMAX0.DAT

In taking these distortion measurements, I noticed what I thought might be an unusual arrangement of distortion products in the frequency-domain plots. Particularly, I considered the irregular fall-off in odd-order distortion products and increase in even-order products suggestive of some asymmetry that might be neat to look at in the time domain.

For example, we can see in the frequency-domain plot below a significant proportion of even-order distortion products and some irregularity in their distribution.

A frequency-domain plot labeled to indicate it was taken with the power
supply set to nine volts, the treble knob set to the middle, and the bass knob
set to the middle. The plot shows a relatively smooth fall-off in odd-order
distortion products along with a relatively high proportion of even-order
distortion products that do not fall off smoothly; they fall off at first,
leveling out by the 16th harmonic, falling hard again around the 26th harmonic,
and actually rising again by the end of the plot.

And here's the corresponding time-domain trace. It's not too wild yet, but the assymetry is apparent even in this trace. In my notes, I marked this type of trace down as “square... ish”.

A time-domain trace labeled to indicate it was taken with the power
supply set to nine volts, the treble knob set to the middle, and the bass knob
set to the middle. The trace shows a relatively square shape in the
negative-going portion but the positive-going portion has a distinct upward
slope and a definitely rounded-off falling edge.

With the treble control at maximum, this frequency-domain plot shows the sort of thing that really got my interest. Even-order disortion products dominate for a significant portion of the plot.

A frequency-domain plot labeled to indicate it was taken with the power
supply set to nine volts, the treble knob at maximum, and the bass knob set to
the middle. The plot shows a smooth but steep fall-off in odd-order products up
to about the 25th harmonic where the contribution of odd-order products starts
to rise again. Meanwhile, the even-order products to not fall off as fast,
remaining strong throughout the plot and even dominating from the 18th through
the 30th harmonics.

As you might expect, the time-domain trace bears little resemblance to the input sine wave. In my notes I called this sort of trace “giraffes” due to the pronounced overshoot on the rising edge.

A time-domain trace labeled to indicate it was taken with the power
supply set to nine volts, the treble knob at maximum, and the bass knob set
to the middle. The trace shows a relatively square shape in the negative-going
portion, but the positive-going portion shows significant (100%) overshoot
on the rising edge, followed by a slump, a hump, and a significantly rounded-off
falling edge.

Doubling the supply voltage to 18V gave a significant increase in even-order distortion products for these same settings on the pedal, as can be seen in this frequency-domain plot (sorry I forgot to put the harmonic markers on this one).

A frequency-domain plot labeled to indicate it was taken with the power
supply set to eighteen volts, the treble knob set to maximum, and the bass
knob set to the middle. The plot shows a somewhat uneven fall-off in odd-order
distortion products, with a bit of a slump around the 21st harmonic. The
even-order distortion products pick up early and by the 6th harmonic or so
match or dominate the odd-order products.

In the time-domain, the trace looks similar to the one taken at 9V, with a bit less ringing on the negative-going side. I think this helps highlight the value of frequency-domain plots for distortion measurement. It can be hard to judge from a time-domain trace the relative contributions of different harmonics or the relative degree of distortion (except in the case of gross differences).

A time-domain trace labeled to indicate it was taken with the power
supply set to eighteen volts, the treble knob at maximum, and the bass knob
set to the middle. The trace shows a relatively square shape in the
negative-going portion, but the positive-going portion shows significant (120%)
overshoot on the rising edge, followed by a slump, a hump, and a significantly
rounded-off falling edge.

Lowering the supply voltage to 6V showed a significant reduction in even-order distortion products, particularly in the lower harmonics. They still dominate by the time we get to the right edge of the plot, as you can see below.

A frequency-domain plot labeled to indicate it was taken with the power
supply set to six volts, the treble knob at maximum, and the bass knob set to
the middle. The plot shows odd-order distortion products falling off
more-or-less smoothly while the even-order products hold steady through the
18th harmonic before starting to fall off a little (though by this point the
odd-order products are falling off faster and the even-order products
dominate).

Flipping to the time domain, we still see a pretty wild trace, but the positive- and negative-going portions are somewhat more symmetrical than when the power supply was set to 18V or 9V.

A time-domain trace labeled to indicate it was taken with the power
supply set to six volts, the treble knob at maximum, and the bass knob set to
the middle. The positive-going portion of the trace shows significant (100%)
overshoot on the rising edge, followed by a slump, a hump, and a rounded-off
falling edge. In the negative-going portion, the falling edge shows less (50%)
overshoot, but an otherwise similar (though inverted) shape to the
positive-going portion.

I performed a comprehensive survey of the effect of different tone settings on the distribution (even vs. odd) of distortion products and how this effect varied with supply voltage. These measurements were taken with the sustain control set to the middle and an input level of -30dBVRMS.

Power supply at 6V

Treble Bass Waveform Distortion Files
Even Odd Total
Minimum Minimum Symmetrical sawtooth 1% 24% 24% A6MINMIN.DAT
B6MINMIN.DAT
Minimum Middle Symmetric sawtooth 1 16 16 A6MINMID.DAT
B6MINMID.DAT
Minimum Maximum Sawtooth... ish 2 11 12 A6MINMAX.DAT
B6MINMAX.DAT
Middle Minimum Symmetric sails 3 45 45 A6MIDMIN.DAT
B6MIDMIN.DAT
Middle Middle Square... ish 1 33 33 A6MIDMID.DAT
B6MIDMID.DAT
Middle Maximum Symmetric sawtooth 2 26 27 A6MIDMAX.DAT
B6MIDMAX.DAT
Maximum Minimum Peaks 15 77 79 A6MAXMIN.DAT
B6MAXMIN.DAT
Maximum Middle Giraffes 8 64 65 A6MAXMID.DAT
B6MAXMID.DAT
Maximum Maximum Camels 4 62 62 A6MAXMAX.DAT
B6MAXMAX.DAT

Power supply at 9V

Treble Bass Waveform Distortion Files
Even Odd Total
Minimum Minimum Reverse sawtooth 18% 24% 30% A9MINMIN.DAT
B9MINMIN.DAT
Minimum Middle Sawtooth 10 19 21 A9MINMID.DAT
B9MINMID.DAT
Minimum Maximum Sawtooth 12 11 17 A9MINMAX.DAT
B9MINMAX.DAT
Middle Minimum Sails 28 36 45 A9MIDMIN.DAT
B9MIDMIN.DAT
Middle Middle Square... ish 8 35 36 A9MIDMID.DAT
B9MIDMID.DAT
Middle Maximum Sawtooth 13 24 27 A9MIDMAX.DAT
B9MIDMAX.DAT
Maximum Minimum Sails 43 71 83 A9MAXMIN.DAT
B9MAXMIN.DAT
Maximum Middle Giraffes 22 58 62 A9MAXMID.DAT
B9MAXMID.DAT
Maximum Maximum Camels 15 61 63 A9MAXMAX.DAT
B9MAXMAX.DAT

Power supply at 18V

Treble Bass Waveform Distortion Files
Even Odd Total
Minimum Minimum Reverse sawtooth 61% 24% 66% AHMINMIN.DAT
BHMINMIN.DAT
Minimum Middle Sawtooth 27 21 34 AHMINMID.DAT
BHMINMID.DAT
Minimum Maximum Sawtooth 30 16 34 AHMINMAX.DAT
BHMINMAX.DAT
Middle Minimum Sails 83 49 96 AHMIDMIN.DAT
BHMIDMIN.DAT
Middle Middle Square... ish 16 36 40 AHMIDMID.DAT
BHMIDMID.DAT
Middle Maximum Sawtooth 33 32 46 AHMIDMAX.DAT
BHMIDMAX.DAT
Maximum Minimum Sails 124 88 152 AHMAXMIN.DAT
BHMAXMIN.DAT
Maximum Middle Giraffes 41 64 75 AHMAXMID.DAT
BHMAXMID.DAT
Maximum Maximum Camels 40 63 75 AHMAXMAX.DAT
BHMAXMAX.DAT

The net of these measurements seems to be that the tone settings have a significant impact on harmonic production. That is to say, they don't seem to merely shape the output of the disortion circuit, but rather appear to interact with it directly. Anecdotally, I've heard that some folks find it tricky to dial in a pleasing tone with this effect. I wonder if this interaction may have something to do with that.

In summary, the Catalinbread Manx Loaghtan appears to be an interesting variation on a classic fuzz circuit. The Baxandall style tone-control provides broad tone shaping and shows some curious interaction with the distortion circuit. Furthermore, the effect seems to respond to changes in power supply voltage in a way that may be musically interesting.

I hope you found this write-up interesting and helpful. If you have any questions or comments, please let me know! If this is the kind of thing you're into, you may enjoy our other work.

Aaron D. Parks
Parks Digital LLC
4784 Pine Hill Drive, Potterville, Michigan
support@parksdigital.com