#314

Thanks, @coreyr and @mdoudoroff. That makes sense. Much appreciated.

1 Like
#315

hyperlocal cold mac patching circle whenever yr ready

2 Likes
#316

Cool! Plus yoga and sushi and daycare.

3 Likes
#317

Anyone wants to trade a recent Cold Mac with a RARE first run one? :grinning:

7 Likes
#318

So mine is “vintage”! :star_struck:

4 Likes
#319

So a few of us have been nerding out about rate-of-change patches in the Hardware Physical Modelling thread the past few days, and I’m hoping one of you might indulge me and test an idea out on your Cold Mac.

One way of patching up a sort of RoC detector involves sending a CV and a delayed or slewed copy of it into a Min/Max, and then mixing the Max signal with the inverted Min signal, giving you the difference between both signals at any point in time.

So I’m thinking you could send a signal to Survey (or just use Survey as the signal generator), take the slewed signal out of Location and patch it into the OR input, and then mix together the OR output and inverted AND to get the rate of change. With the added benefit of all the other outputs being related to your original signal.

2 Likes
#320

Isn’t Location doing this already without any patching ?

#321

I was thinking this on my first read too. But if both your signals are -ve for example, won’t Location just continue to -5v?

#322

Not quite, Location is acting more like a slew. If you sweep an incoming voltage from, say, 0V to 3V, location will travel from 0V to 3V but slower and then stay there. What the patch above will do is only put out a voltage when the incoming voltage is actually in motion. So if you sweep from 0V to 3V, the final output will rise as the voltage is travelling but then drop back down to 0 once it stops moving. The faster that sweep occurs, the greater the difference between the original voltage and Location voltage will become, and the stronger signal you end up with at the end. So if you used it to open a VCA it would be like bowing a violin string, you have to be moving in order to generate sound.

4 Likes
#323

To be clear, this is not quite right. It does not stop at 3v. It will continue on rising until it hits 5 volts. The input sets the rate of change of the output, not the target value. The greater the input, the faster the rate of change. Negative voltages will make the output travel in the other direction. The location output only stops when the input is 0 (no rate of change) or it hits the boundaries (-5V,+5V). The output is actually integrating the input signal.

3 Likes
#324

Well… except that from my testing, location integration generally falls far short of -5v/+5v. This is a major source of confusion. In practice, the description @smbols offers (at 3 volts) is more accurate than the theoretical behavior.

Arguably, the location and follow circuits of Cold Mac are a bit undercooked.

2 Likes
#325

Have you tested this on multiple modules? It should definitely be getting closer to +/- 5V than that! It gets to +/- 4.9V or so on my CM at home… and if it was only getting to +/-3V, that would be an issue with the module for sure…

#326

My very early Cold Mac behaved similarly, but I haven’t tried any recent specimens other than the one in the video. Does this actually vary across specimens?

#327

Here’s a sampling I just performed on mine. Voltages shown are within ±0.15v.

-10v : -4.5v *requires external voltage source
-7v : -4.3v *requires external voltage source

-5v : -4.2v
-3v : -4v
-1v : -3.6v

1v : 3.7v
3v : 3.9v
5v: 4.2v

7v: 4.3v *requires external voltage source
10v: 4.5v *requires external voltage source

1 Like
#328

Ahhhh okay.

So, if there is any positive voltage coming into Survey or Slope, whether it is traveling or stationary, Location will travel upwards towards its limit at a rate related to how quickly the source voltage increased?

@mdoudoroff - have you tested the speed limits of the circuit at all? In your video manual it looked like the circuit moves quite slowly even if you make a quick sweep of Survey. What about if you send it an instantaneous signal from a gate, for example?

1 Like
#329

Yes, although in my experience, the upper (and lower) limits vary (see my table two messages above). We need some others to test their Cold Macs to verify.

I haven’t used a stopwatch, but there’s certainly an upper velocity limit in the vague vicinity of 1 second. Maximum velocity is if you pump 10v or -10v into the circuit externally. An instantaneous signal is going to have very little effect. As the gate length increases, you should be able to nudge the voltage upwards toward… some upper limit. Probably worth exploring.

2 Likes
#330

Like @voidstar said, LOCATION gives you the integral of the input signal, not just a slewed version (or the derivative, which is the goal of the rate of change patches you all have been discussing in the modeling thread).

I’ve been watching that discussion and wanted to try something out using FOLLOW – that’s the Cold Mac output that behaves more like a slew. It rectifies the input first, so in order to get the rate of change, you need to subtract the FOLLOW and rectified outputs. And the FOLLOW output’s gain is apparently higher than that of the rectified output, so FOLLOW needs to be scaled down before subtraction. All that is to say, I needed a lot of cables, and an inverting mixer (Shades in this case).

CM SURVEY knob/input as control
CM rectified output to CM FADE input
CM FOLLOW output to Shades ch1 (with gain about halfway)
CM RIGHT output (inverted rectified output) to Shades ch2 (at full gain)
Shades ch2 output (FOLLOW/2 + RIGHT) to OR input 1
…and via stackable to Shades ch3 (gain at -1)
Shades ch3 output to OR input 2
Tweak Shades ch1 gain so that ch2 output is roughly 0 when Survey knob isn’t moving.

For a vaguely breath/bow-ish voice (oscillator -> filter -> VCA):
Shades ch2 or ch3 output to filter cutoff (these two Shades channels output the un-rectified rate of change, so they will momentarily open or close the filter depending on which direction the SURVEY knob is moving)
CM OR output (rectified rate of change) to VCA gain

In the patch above, any time SURVEY crosses 0, the VCA closes (or nearly closes). I thought I’d take advantage of that and clock a sequencer when that happens.
CM LEFT output to Maths ch1 input
Maths EOR to sequencer clock

I have a feeling Maths alone could do the RoC thing a little more efficiently, since it’s a slew, inverting mixer, and slope detector all in one. I’ll post in the other thread when I’ve tried that out.

7 Likes
#331

I’m relatively new to the rodeo, but I’ve had great results patching an LFO into the SURVEY input, creating multiple variances of the LFO on the output jacks (thanks to @mdoudoroff’s great doc!). Then, sending 6 of the Cold Mac outputs to individual TRIG inputs on Just Friends in sound/cycle. Monitoring just the MIX output of Just Friends.

Sweeping the SURVEY knob slowly from fully CW to CCW was endlessly entertaining. I guess CM is changing the timing & voltage relationship of all outputs, and varying the V sent to JF TRIG inputs based on the SURVEY knob? Thus triggering oscillators based on LFO timing & voltage amounts.

The UI with just 1 knob really begets experimentation. I’m still learning.

18 Likes
#332

Nice! I don’t know how I’ve never thought to use CM in this way. So simple really. Not that I have JF but there is definitely still something I can get out of this.

#333

and i think it’s a great use of CM and is one of the modules strengths. you can get complex, reactive results without complex patching. i always start my CM with something simple like this and then mess it up along the way.

2 Likes