I've a question about LFO SPD=64 in the DT manual

In chapter 11.7LFO PAGE of the manual there’s a table where LFO SPD settings and the Multiplier are mapped onto musical note lengths. Seems pretty straight forward at first glance. But I noticed something weird. In the last row Elektron denotes for the SPD a value of 64, in spite the fact that the value 64 isn’t available on the device. The encoder range on my device goes only from -64 to 63 !! Where is the missing 64?

The reason why I’m asking is that when I had a long running single cycle wave form going across 2 consecutive patterns (using the amp EG on the amp page) the LFO (spd=63) that was modulating the filter freq, ran in and out the of the grid. It didn’t lock itself to the beat. Holy crap.

I guess either my unit must be at fault or I’m not getting it (or both).

nope - sounds like a normal split of teh available 128 values

if you can’t dial 64 via speed alone - try to ensure the factor of speed and multiplier is an integer factor of 64

so speed*mult is 32 / 64 / 128 etc - then you will get tempo sync if the multiplier is not the fixed-to-120bpm version


Instead of spd = 63 and depth = 63 I went with
spd = -64 and depth = -64

I think this could be used as a replacement for the missing 64 value. I think that sounds like a 1/16 pattern.

But my problem is still there. The lfo can’t hold up with a beat over a longer period. For instance the pattern I’m working on contains 32 steps. First step holds the sample. then the Amp EG is on HOLD and the lfo modulates the frequency at 1/16.
On the second half of the pattern I notice how the lfo looses its grip on the beat . There is a noticeable lag.

Perhaps I’m still missing something?

Thanks again!

i don’t know the DT, but normally for Elektron stuff the factor of Speed x Mult equal to 128 will cycle through one full cycle of an LFO shape or 16 steps of random s/h

if you have internal clock running and have a means to start the LFO phase where you want then i’d expect it would keep sync

but be sure that the LFO multiplier is not set to a fixed tempo multiplier (see the manual)

depth values above are not relevant to the problem, -ve speeds will just reverse a shape - try a speed of 32 and vary teh Mult to taste for whichever shape you use - depth can be anything

if the sample is a long loop that is more likely to be the source of drift relative to LFO

Do you have the LFO as free running? Then it can be drifting when you stop and play the sequencer. As it still running free in the background. Then you need to set it to trig and let it retrig at the first step of your sequence. See quote from me in the Unofficial guide. Perhaps that helps. It helped me.


Yes the table from the manual I was referring to .

It’s just misleading since SPD can never be positive 64. Maybe Elektron should update the manual (hahaha).

No, it’s set to trig mode (both on the TRIG page and the LFO), sorry should have mentioned that.

But in my case it won’t make a difference for the sync, because the pattern is long enough for me to notice the lag way before the pattern turns around for the next cycle and retriggers the LFO once again.

It could be In the sample like mentioned early. Does it happen on other samples aswell. Do you send midi to the DT?

Do you use trig locks on the frequency? Or a trig that also retriggers the LFO?

i’d like to see that last row’s actual values for 63 :smiley:

i just use 32 due to the lack of 64. The Mults usually have me covered.

good to know.

yes I probably I could set trigless trigs to keep it tight. But I don’t want to. I’d expect the LFO to be tight as f**

sure, LFO runs on project speed.

Sync is not a function of the depth parameter, but defines how the LFO modulates the signal.
I wanted to know whether sonically I could achieve the same result if I turned SPD to negative 64. So in order to have the same outcome I had to pull the depth value to the negative side as well. At least this is what I was thinking.

The sample is just a short (1/16) burst of a singly cycle wave. I’m sustaining the note by having the Amp EG set to HOLD (or a long decay time instead)

You could also use SPD 32 instead of 64 and double the mult. I never experienced drifting to be honest.


Nope, no trig locks on the freq or anything like that. Sample is very short (1/16), sustain is done via the Amp EG

By the way I think the table you mentioned might be wrong.

The values I’m using can be found on p .44 of the manual


True :slight_smile: But it would be much easier to memorise if we had +64.

Check out the table from the manual. For spd=64 bottom row the timing / note length is basically 1/MULT.
With spd = 32 you have to multiple the values for spd=64 by two.
And we still haven’t talked about dots and tripplets 1/12, 1/6, 1/24… :wink:

Wait a sec. Why are the note lengths (SPD,MULT) for

(64,16) and (64,32) both 1/16
(32,16) and (32,32) both 1/8
(16,16) and (16,32) both 1/4


Fault I guess

The table I created is the same. It is about steps. So SPD 64 and Mult 2 is 16 steps long. Ergo a bar or 1 note. But the table mentioned in the Digitakt Unofficial guide is wrong.

Q: does your LFO draft aswell when using SPD 32 and MLT 64?

1/2 is missing for three combos (and other values are duplicated) - that’s a typo

pinging @eangman

ps -ve speed reverses the wave shape, reversing the polarity of teh bipolar modulation depth won’t always get you a similar effect, some LFO shapes are monopolar

go with the speed default which is 16 or 32 probably and vary teh Mult - the LFO speed is expressed as Speed combined with Mult, but i’d suggest to @eangman that the phrase factor of should appear in that sentence to clarify things

Help me out please.

“my” table is from the official Elektron manual. I was wondering why you mentioned the “Digitakt Unofficial guide”. Am I wrong?

“your” table gives for (spd,mult)
(64,2k) = 1/64

However to the manual of the Dt that should be 1/1024.

Getting confused :wink:

As you can see in the picture of my table is that the value is in steps of your sequencer. I created this as I didn’t understand note values back then. So a half note didn’t mean something to me. So in my table a 1 is that the cycle plays trough one trig of your sequencer.

In the Unofficial guide there is in the top listing also a table of LFO. But that gives different values.

Ok, got it.

and 1/8 is missing too. Values on the diagonals should be the same.

checking now…

Ok, all good now. My bad, I was sending the LFO to the pitch of the sample, that ofc messed with the sample length which then gave the impression of the LFO running away. Now I have panning as the destination and everything is in sync exhales

@all : I really appreciate your help. Big thanks! (But we also found a typo in the manual so the thread wasn’t for nothing after all :slight_smile:)


And there goes my Wednesday evening :wink: haha. Good you have found the problem.

1 Like

That typo was fixed a while back. I think you are looking at an old version of the manual. Go to the Elektron page and download an updated version.
Hmm yes, will have look at that factor thing. It is mentioned in the MULT parameter description though.


I don’t see the manual, I don’t have/need that - I think we assumed the new user would have the latest manual

they linked to it


fwiw - it does seem that googling Elektron pdfs can get you to older versions - maybe those older ones should be displaying an additional message to find the later ones (if possible / practical etc)