I'm thinkin this one is beyond the realms of DIY but here goes.
I'm a guitar repairman by trade. Effects come up in conversation now and again (I had some fun building the Omnidrive, the only effect I've so far found the time to build. Thank you Mr. Hollis and Mr. Keen it was fun. Don't worry though, I haven't been selling anybody elses designs as my own. Though I must say quite a lot of my customers would pay good money for custom built effects. I'm starting to ramble here....). Anyway, a number of customers have expressed an interest in tap controlled delay (for delay time that is, tapping foot along to the beat) pedals. As far as I'm aware there are one or two commercially available.
Now my own take on it is that this would probably be fairly difficult (if not downright absurd) to build. I'd imagine it would need to be digital with some awkward detection. My own thoughts were that it might be easier to have an analog delay with delay time controlled with a rocker pedal (I reckon I have the necessary mechanical skills and equipment to build one of those). However, this got shot down by complaints of "it'd be too awkward to tune to the correct delay on stage" and "the control would have to be too course to give sufficient range, making it too inaccurate to be useful". Though one did say that a rocker pedal to control feedback with lateral movement to control sweep would be nice.
In any case, it's a nice thought, and for the entrepreneurial a market. Feasible? Or perhaps my ignorance is a little over-evident. Cheers for taking the time.
This has been talked about before. The general conclusion is that it can be done but you would need a microprocessor to do it properly. This is beyond the capacity of the average DIYer.
Andrew
It's possible but not easy. There are PLENTY in-production of tap tempo delays out there:
Digitech Digidelay
Line 6 Echo Park
DLS Echotap
Line 6 DL4 (well, almost)
DOD DFX4 (I think that's the name, this one might be OOP)
Akai Headrush (The original is OOP, there is a new one released)
And there's more, too.
-Colin
BEGIN
(setup all the counters, registers and other junk to satisfy this particular
chip's operating environment; initialize registers, blink the lights, whatever
fire drills you want to do before getting down to work)
WHILE (power is on) DO
READ Switch()
IF (switch is different now)
THEN
GOSUB (debounce the switch)
(calculate how long it's been since the last time the switch was
pressed)
IF (it's been a reasonable tap tempo time)
THEN
DO
(calculate whether this interval satisfies making a new tap
interval based on the last few tap intervals
END
IF (it's a new tap interval according to the averaging criteria)
THEN GOSUB (write the new interval into whatever does
the interval)
ELSE (just remember this interval because it may be the
earlier ones of a valid average)
ENDIF
ELSE
DO
(reset things for intervals were too long; this could be the
first switch make for a new interval determining set)
END
ENDIF
ENDIF
ENDWHILE
END
There are also production analog delays which feature expression pedal control of both repeats and delay time.
I'm thinking of one in particular......
Ed R.
Quote from: anonymousexperimentalist...Line 6 DL4 (well, almost)...
Why "almost"?
Quote from: Ed RemboldThere are also production analog delays which feature expression pedal control of both repeats and delay time.
I'm thinking of one in particular......
Ed R.
Ahhh, yes, that one's a beaut...... ;) :)
Mike
Quote from: R.G
BEGIN
(setup all the counters, registers and other junk to satisfy this particular
chip's operating environment; initialize registers, blink the lights, whatever
fire drills you want to do before getting down to work)
WHILE (power is on) DO
READ Switch()
IF (switch is different now)
THEN
GOSUB (debounce the switch...
...
...
To control an analog delay with BBD's
Use an microprocessor to create the clock do a little math and..
Here you go Tap to delay time algorithm: :) :)
Tap'ed time/Delay time = 51.2ms (Short but just an example)
BBD stages = 1024
Clock: 0.5 * (1 / (0.0512 / 1024)) = 10kHz
Claus H
Hello Claus, you here too!! :D
I have this Headrush that I use with a Skrydtstrup buffer in font , but just for the feeling , does anybody know how this pedal handles the signal when the effect is turned off?
Btw : I heard that some people found them noisy and opened to find a very messy stressed korean work, but I opened mine and it is very clean inside.
Quote from: vhollundHello Claus, you here too!! :D
I have this Headrush that I use with a Skrydtstrup buffer in font , but just for the feeling , does anybody know how this pedal handles the signal when the effect is turned off?
Btw : I heard that some people found them noisy and opened to find a very messy stressed korean work, but I opened mine and it is very clean inside.
Jep I am here... I have been trying to keep it a secret But I seen now the jig is up :) :)
Buffers and tap tempo stuff is still in pre production :) I got a new amp/guitar tech job and I am very busy like 45+ hours a week.. Hope to get some time to do a buffer run soon..
See you later
Claus H
QuoteTap'ed time/Delay time = 51.2ms (Short but just an example)
BBD stages = 1024
Clock: 0.5 * (1 / (0.0512 / 1024)) = 10kHz
Getting a 51.2ms delay between foot taps is faster than I can move, but I think the math may be a bit flawed.
The clock is going to be 20kHz, I think, even with a two phase clock, because two phase clocks are simultaneous out-of-phase waveforms, not half-speed waveforms.
A 10kHz clock will give you 102.4mS of delay, and require a signal limited to under 5khz to avoid aliasing.
I think that my pseudocode does the whole task, and the actual timing is subsumed in the pseudoinstruction
QuoteTHEN GOSUB (write the new interval into whatever does
the interval)
Quote from: R.G.QuoteTap'ed time/Delay time = 51.2ms (Short but just an example)
BBD stages = 1024
Clock: 0.5 * (1 / (0.0512 / 1024)) = 10kHz
Getting a 51.2ms delay between foot taps is faster than I can move, but I think the math may be a bit flawed.
The clock is going to be 20kHz, I think, even with a two phase clock, because two phase clocks are simultaneous out-of-phase waveforms, not half-speed waveforms.
A 10kHz clock will give you 102.4mS of delay, and require a signal limited to under 5khz to avoid aliasing.
I think that my pseudocode does the whole task, and the actual timing is subsumed in the pseudoinstruction
QuoteTHEN GOSUB (write the new interval into whatever does
the interval)
I have to disagree with you.. But let's try it !! I got some 1024 delay lines laying around (or just use a chorus pedal ;) ) If you read the data sheet for a 1024 stage BBD line the max delay time is 51.2ms the min clock is 10Khz this fits nicely with the math :) The 51.2ms is way faster than anyone can move ther feet I completely understand "This delay time is only use as an example" (It also says so in the text ! ) I also understand that this does not do it as a pseudoinstruction set.. I just wanted to share the math with all you.. It took me some time to figure out how the data sheet's delay time related to the clock.. Or might be completely off THEN GOTO END
Claus H
Quote from: The Tone GodThis has been talked about before. The general conclusion is that it can be done but you would need a microprocessor to do it properly. This is beyond the capacity of the average DIYer.
Andrew
For any average DIY-er with a little knowledge of uP's this is a faily sioimple thing to program. Info on programming is widely available. Check the sites of microchip (PIC's) or Atmel (AVR's, my favourite since the're more 8051 compatible). These guys work hard at making the step to uP circuits as low as possible in order to sell chips. And: those chips are cheap as well.
Boy this is a fairly old thread to be dug up.
Quote from: Dirk_HendrikFor any average DIY-er with a little knowledge of uP's this is a faily sioimple thing to program.
A "DIY-er with a little knowledge of uP's" is not an "average DIY-er". Most DIYers here do not have a background in uC. Only a handful of people who visit do and yes for them I agree this is a simple thing.
For the record I already have done a uC based tap tempo. Been there, done that, got the prize. :)
Andrew
Proof that my math is right :)
http://factaee.elfak.ni.ac.yu/fu2k01/fu01.pdf
See Page 4 and 6 here 3 cascaded mn3005 (4096 stage) BBD's are used to make 256 ms of delay with a clock of 24khz this means they each produce 0.256 / 3 = 85 1/3 ms of delay
My math is
0.5 * (1 / ((0.256/3) / 4096)) = 24kHz
No harm intended
Claus H
Quote from: R.G.
Quote from: Claus H
Tap'ed time/Delay time = 51.2ms (Short but just an example)
BBD stages = 1024
Clock: 0.5 * (1 / (0.0512 / 1024)) = 10kHz
Getting a 51.2ms delay between foot taps is faster than I can move, but I think the math may be a bit flawed.
The clock is going to be 20kHz, I think, even with a two phase clock, because two phase clocks are simultaneous out-of-phase waveforms, not half-speed waveforms.
A 10kHz clock will give you 102.4mS of delay, and require a signal limited to under 5khz to avoid aliasing.
I think that my pseudocode does the whole task, and the actual timing is subsumed in the pseudoinstruction
Quote:
THEN GOSUB (write the new interval into whatever does
the interval)
I did a little research to day I plugged in a Carl Martin Delayla and measured the clock at max delay time.. The Delayla contains 6 cascaded MN3007 (1024 stage) BBD's and it's maximum delay time is 500ms
0.5 * (1 / (0.5 / 6144)) = 6.144 kHz
My measurement of the clock at max delay time (500ms) was oddly enough 6.149 khz :) :) :) So my math is all right..
Well no reason to gloat :)
later
Claus H
nosy...