SlowMo LFO?

You can but it will only be a square wave. And mostly a pulse rather than a LFO per se.
I don't see a way to slow down a LFO more than with the knob on the module itself

If you have grains, there’s a super slow LFO firmware for it

I have an ancient ICL8038 (literally just one, unfortunately), a VC-wave-generator chip, that I'm making into a six-parallel-output LFO function generator for my banana instrument.  It requires a minimum 10V power supply, so I haven't had the opportunity to build it into AEM; thus, it in particular is not a very helpful suggestion, but my main point (and I cite Lugia 's post, which follows in the same vein) is that you'll have to go outside of AEM to get this sort of signal.  It is readily available in other modular formats, as well as miscellaneous, external equipment.  

EDIT:


Of course, I forgot about the AEM duct tape, the patch-all, the Flex-Seal, GRAINS. Pretty sure there's a firmware out there to help fix this boat I sawed in half. Learn to code and the world is your oyster.

Uhm, you can do that now easily. Just connect the GRAINS via the USB cable to your PC, start up the Arduino IDE and then you can upload any of the existing firmwares with one button click. Sure you can't do it while playing a patch, but it's super easy and quick.

Now with the new version that has the Mozzie switch, programming has become even easier by using the really useful and convenient Mozzi Library. I'm currently working on a programming tutorial for this.

The wiki has all the links on available firmwares and how to install them: wiki.aemodular.com/pmwiki.php/AeManual/GRAINS

And then, if you have a gate output from the clock, send it to a SLEW/EDGE to form leading and trailing ramps. That should do it. You can even CV the rise/fall times.

Wouldn't the rate of the LFO be limited by the slew rate of SLEW/EDGE?  I'm not sure if it has the capacity (or capacitance, if you're cheeky) to stretch out a signal for a huuuuge length of time.  The rate of the "clock signal" (trigger pulse from converted MIDI, or wherever) has no effect on the slew rate of SLEW/EDGE, so the limiting factor is the SLEW module itself.  Unless the SLEW can stretch out to, for example, 30 seconds positive slew and 30 seconds negative slew, it's not possible to get a 1 cycle per minute LFO signal.  (up time + down time = full cycle time)  



Another interesting limitation of generating such a signal from the GRAINS module is DAC and bit resolution.  I'm not sure what the resolution is on GRAINS, but if we assume an 8-bit resolution, then 256 analog output states exist, so the output signal will very much resemble an analog wave but in reality be composed of differing time lengths of voltage levels from 0V, then (1/256)*5V, then (2/256)*5V, then (3/256)*5V, .... to 5V.  This is hardly an issue for most signals, as our human ears (and most of our senses) do a lot of guessing and approximating; we can't hear the 256 distinct voltage levels in an audio rate wave of say, 440Hz, because those 256 levels fly through our ears at a whopping 225,280 levels per second.  (steps in the one-half (0-5V) increasing part of a cycle + steps in the down part of the cycle * cycles per second)  Fast cycles mean smaller bit resolution isn't an issue, but let's say we're trying to make an LFO signal that completes one cycle per track, at say, (1/300)Hz, or one cycle completed in five minutes.  That same calculation yields 1.707 voltage levels every second.  A better way of looking at this number would be to invert it, and think about it as 0.58 seconds per voltage level.  This means that you'll be able to fairly distinctly hear a step in voltage, rather than an expected smooth transition, every 0.58 seconds.  In effect, a slow signal "stretches out" the limited number of "audio pixels" that GRAINS can output, and if one stretches them out too much, then one will be able to hear where one "pixel" stops and another starts.  It's still an LFO signal, but not a smooth one; rather, stepped, and observably so.  Just something to ponder.  

Now, I have no idea what the resolution on the GRAINS module is; I know it's built on an ATMEGA328, which has 14 digital I/O pins.  I'm unfamiliar with the circuit build to say how many of those digital pins are used in other functions, but the new GRAINS module has a toggle switch that controls program functionality, so I'll assume it's tied to one digital pin, leaving 13 left for DAC.  That's a strange number to work with in digital design, but it's a viable build.  I'd expect the resolution to range from 8-12 digits or so, if only the digital I/O pins are used.  The higher the resolution, the "smoother" the curve.  

This is not accounting for the PWM output capability of the ATMEGA328, which has six pins that can either be programmed to output a PWM-signal at I think 16kHz, or digital I/O pins.  If GRAINS uses the PWM capability to generate the output wave, then I simply have no idea as to its resolution and would have to reference the datasheet.  If GRAINS uses them as extra digital I/O, then we're at a possible 19 pins, which is more than enough for a more standard number of 16-bit information.  16 bit would mean 65,536 unique voltage states, compared to the 8-bit's 256.  That's 256x more voltage states-- equivalent to 256 different 8-bit microcontrollers all mashed together and acting as one, in terms of signal resolution.  (I'm sure there's a way to approximate 16-bit with two 8-bit microcontrollers communicating with each other, but 256x is much more illustrative a number of the practical differences, I think.)  With that many different voltage levels, you'd need to create one SLOOOOOOW signal (in the realm of hours-long cycle times) before noticing the "pixellation" in the wave.  


Just something to ponder.

Anybody know the resolution of GRAINS?  I couldn't find the info on the TW website.  

If I'm understanding what you mean, I think that's exactly it-- 10-bit color compared to 8-bit color will look so much smoother in transition, because the extra two bits allow for so many more colors "in between" the 8 bit color set.

Color is a good way of illustrating the phenomenon. Let's work with just black and white for this example.

1-bit BnW color means two color values: 0, for, let's say pure white, and 1 for pure black.
2-bit color means four possibilities exist: 00 pure white, 11, pure black, and everything in between those two binary numbers-- 01 would be a light gray, then, with 10 (the next binary number after 01) being a darker gray. So for 2-bit color, we have 00 white, 01 light gray, 10 dark gray, 11 black.
3-bit color: Well, with three digits, we have 8 colors-- 000 whitest white, 001 a touch gray, 010 a bit more gray, 011 for "middlemost" gray, then 011 for somewhat dark gray, 100 .... and so on, until 111 blackest black.

With each additional bit of resolution, more colors can be expressed. When we get to huge bit values, so many different colors exist that we have difficulty telling color 000011001001001 from the next-darkest color 000011001001010 (made-up numbers, of course). So for colors, we see a smoother transition of color and shade and brightness.

For analog signals, the same thought applies. Let's say we patch a super-slow, low-bit LFO into the frequency input of a VCO. In fact, a square wave LFO is a 1-bit signal, with states 0 for off (0V) and 1 for on (5V). You can easily hear the jump between different "states" (voltage levels) of the output wave by observing the VCO pitch. It would look something like this, if you wrote out the different states as they happened: 0 1 0 1 0 1 0 1 0 ... (maybe we should stretch the 0s and 1s out for clarity, say 0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 1 0 0 0 0 ... but it's all the same.)

Now let's take that 1-bit square wave LFO and make it 2-bit. It's not square anymore, but more like a staircase, with state 00 at 0V, 01 at 5/3V (which is one-third of the output range of 0-5V, as we have three voltage values greater than zero in our signal) , 10 at 10/3V, and 11 our full level of 15/3V = 5V. This signal might look something like this:
00 01 10 11 10 01 00 01 10 11 10 01 00 ... (These states are the same thing as shifting between color values or pitch, just represented as the binary states themselves rather than the analog value of voltage.)
This LFO would make our VCO span the same range of frequency that the 1-bit square wave did, except the difference is in the number of "stops" the pitch makes in between the lowest and highest state. So, we would get the lowest tone, then a higher, then higher again, then finally the highest tone. Then, if the waveform is like a triangle, it would reverse, going high to low in four steps.

With a 3-bit signal, the oscillator receives 8 total steps along the way from low to high. With 4, we hear 16 distinct tones. 5 gives us 32, and on and on and on...

As the resolution (the number of digits in binary that control the number of states) increases, the number of steps in our tone increases exponentially. Eventually we can make these steps so short that our ears kind of blend the steps together and they sound "analog" rather than stepped to us.

It's technically impossible for a digital circuit to perfectly replicate an analog wave; we get 99% (AKA "good enough") by using lots and lots of bits, and cramming all those different states together. If we don't linger for too long on any particular pitch (or "step" in our signal), then our ears are fooled and we're good to go.

BUT: if we slooooooooooow down the wave, then the length of time that we have to linger on those steps increases, and if the signal is "slow" enough, and the steps are spaced farther apart than we'd like (from lower resolution than needed), then our ears won't be fooled; we'll hear, by the example in my previous post, a distinct shift in pitch every time the next voltage level is sent to the VCO, which in that case was just over half a second for an 8-bit wave going from 0-5V and back again every 5 minutes.

Okay, I get it. Thems the breaks. So what we need is, like, trillions of bits, then set it up in Antarctica to keep it cool. Eventually we'd have so many bits in one space that it would collapse into a black hole, then we'd have a serious LFO!