Finished this off during the last few days, using Torchy's layout and the schem
from www.home-wrecker.com
He had all the pots lugs 3&1's reversed, so I'm glad I had the schem around also !
I only changed 1 part, 47ohm for the 51ohm that I didn't have.
OK, it all works fine and after some "fiddling" with knobs ( technical term for adjusting! )
I got a setting which almost works, sensitivity/attack and range in perfect harmony......
But .... I have a nasty "ripple" happening most of the time, even on long single chord
strokes !
I've searched and read a few other comments/suggestions, but there doesn't seem to
be a "fix" that I can see ?
I've used the 22uf cap, would a 22+10uf help for this kind of "loooong" filter type chord use?
If so, is it better to use a switch to add that in, or even go with a 47uf ?
Thanks to Torchy for the layout which works a treat :D
Marty.
Bump !
Anyone have suggestions for "smoothing" this out a bit ?
Could an LDR help ?
M.
That ripple is one of the shortcomings of the Dr. Q and its derivatives. A larger cap may reduce ripple, but the decay will also be increased. Try a 47u and see how that helps.
Also, you could try adding a 100n cap to ground between the LED and 51 ohm (47 in your build) resistor. That may smooth things out a little. I haven't tried that, but it could work.
Thank's for that, I'll certainly give it a try !
Marty.
Marty,
Is the amount of ripple reduced very much if you play just a single sustained note (even a loud sustained note)?
Becomes more obvious when more than one note is played at once, particularly lower notes?
Mike
Quote from: Vsat on October 16, 2005, 08:09:47 PM
Marty,
Is the amount of ripple reduced very much if you play just a single sustained note (even a loud sustained note)?
Becomes more obvious when more than one note is played at once, particularly lower notes?
Mike
Yes, there's almost no ripple on single notes and 5ths, but
anything more and its "wobbling" all over the place as it decays !
M.
Well perhaps I've been mis-reading some of the posts on the Quacky ?
This version with a TL072cn was thought to be a "fixed" version, but I've
just been trying both LM358 and LM1458 with much better results !!
all three controls are more responsive, perhaps the range of the sweep is
a little more controlled, but its a definate improvement, with more variation
available :D
I'll still try the small mods suggested, but its now much better, perhaps I'll
stick with the 1458 ...... Mark ?
Marty.
MartyMart - check this out. The chip AND the tranny are important to the overall sound of the pedal.
http://www.diystompboxes.com/smfforum/index.php?topic=35760.msg251999#msg251999
Quote from: spudulike on October 17, 2005, 05:56:06 AM
MartyMart - check this out. The chip AND the tranny are important to the overall sound of the pedal.
http://www.diystompboxes.com/smfforum/index.php?topic=35760.msg251999#msg251999
Great, thanks for this, which never came up in any search BTW !!
I'm happy with LM1458/2N5088 at the moment :D
seems to be a good "pairing"
EDIT: The suggested 100n cap to ground from the junction of the LED/51r is worth doing,
whilst it didn't solve things totally, the decay seems smoother and the "ripple" is slower and
also smoother sounding :D
Marty.
In general, the degree of satisfaction people get from their envelope-controlled wahs is a function of their playing style. If the notes come and go quickly, it's a bit like never noticing how truly bad the dessert is if you fill up on the bread and salad that precedes the steak. Admittedly, a great deal of what drives interest in such autowahs are fabulous bass-slapping sounds heard on record or in concert, which tend not to allow notes to linger. My own experience is that a great many autowah designs just don't sound that wonderful during the second half of the decay phase because the envelope followers tend to be bare-bones, or because the time constants are fixed in a manner that anticipates the note/chord being swept aside (pardon the pun) very quickly in lieu of the next note.
One thing to try is a larger cap or capacitance value, with a decay pot in parallel. So, a 33uf cap in tandem with a 500k trimpot. The idea is that somewhere in there is an optimal decay time that will smooth out the ripple enough without being too sluggish from a playing perspective. Note that a 33uf capacitance will also slow down your attack time, (so 47R for attack/charge-up/current-limiting resistor is a good choice).
The 1458 is not exactly a gift from the audio gods, but it works well enough in this instance. Jack Orman's biasing change that came in via the Dr. Quack should have made op-amp choice one driven purely by noise and current draw concerns, rather than sweep width concerns, however if a 1458 works for you, I say stick with it.
Ross comp in front...changes the way the EF works, might be worth messing around with, will certainly 'even things out' as far as peaky voltage swings hitting the input, may help out the decay steadiness...may be too much also...something to think about thinking about, anyway...lol.
NQ Sounded great when set up just so' for me, but change anything before it like guitar volume and it's needing reset up for max Quax.
Breadboarded an envelope follower on the weekend, fairly generic design. Checked out it's operation with the dual-trace scope.
Upper trace displayed the audio signal which was fed in, lower trace displayed the output from the averaging filter of the follower.
With strictly monophonic inputs, it works well... does a good job of replicating the envelope of the input signal. Play two or more notes simultaneously, and things are not so nice.... especially noticeable on sustained notes.
As a test, fed in these two signals - a steady 80 Hz sine plus a steady 85 Hz sine, equal peak-peak levels. Fed in individually, each produces a steady DC output is produced with very little ripple. Feed in both, and you get a huge amount of ripple at 5 Hz... the two signals are beating with each other. Looking at the input signal with the scope, the beating is clearly visible (no surprise there), it looks like a string of pearls, and the envelope follower is doing a wonderful job of replicating the beat envelope. Although the follower time constant is set up to filter out ripple at the expected range of frequencies produced by an instrument, it is not able to filter out the ripple from a polyphonic signal with all the low-freq beating taking place... unless a very long time constant is used, which will make the decay very sluggish.
This is probably the reason why some people report that full-wave rectification does not yield much better results than half-wave...the major source of ripple with these followers is low-freq beating between notes that are close together, ie. chords. Use it to control a filter and it sounds like the filter is modulated by an LFO.....an interesting problem.
Cheers, Mike
Useful observation there, Mr Pike. ;)
One of my guitars has a home-made fret job, and with the frets so tall (and its owner so damn cheap :icon_rolleyes: ), there is often wearing or other deformations of the strings at the most commonly-used frets. The sort of thing that when you pluck the string, you can see little red and blue bands resulting from the beats within a string.
I'm wondering if there is some virtue to highpass filtering a rectified signal, to remove any sub-audio components, or will that have much less impact than expected for the likely parts count?
How would the highpass scheme be implemented? Would there not be a risk of filtering out the slow-changing parts of the envelope along with the beating-ripple? The idea is to defy the laws of nature... the env flwr is doing it's job... just that this nuisance phenomena called beating takes place. Perhaps use a harmonizer to shift up 3 octaves so 5 Hz beat becomes 40 Hz and is more easily filtered? (without harmonizing the audio fed into the wah circuit).
Mike
Which brings us back nicely to my Space Filter envelope generator
http://www.diystompboxes.com/smfforum/index.php?topic=36221.msg255718#msg255718
Mike, I'd be interested to hear what you see on your scope with this one because the start frequency = strength of note played, then decays at sweep rate time and not per input.If matched to the right VCF I think this would sound good.
Stephen
Quote from: Vsat on October 17, 2005, 02:29:20 PM
How would the highpass scheme be implemented? Would there not be a risk of filtering out the slow-changing parts of the envelope along with the beating-ripple? The idea is to defy the laws of nature... the env flwr is doing it's job... just that this nuisance phenomena called beating takes place. Perhaps use a harmonizer to shift up 3 octaves so 5 Hz beat becomes 40 Hz and is more easily filtered? (without harmonizing the audio fed into the wah circuit).
Mike
Or do what Harry did - use a hex pickup and 6 envelope followers! :)
But didn't he use an element of fundamental extraction too, if I understood correctly what he recently posted here?
Stephen
I seem to remember this circuit not woking at all well in a thinned out "guitar" mode ...
It was left in full range Bass mode, to work well .....
Love tones filter pedal ( name escapes me "filteratron/wobulater thingy" ) sounds great ... but then it should for
600 bucks !
Marty.
Steve,
Could you please post the original Space Drum schematic to the list....(again).
Mine did not print out right, looks like it's sliced into 4 sections which are jumbled about in an unreadable manner.
I think I see what the env-sweep circuit in your drawing is trying to do, however it won't do it properly. The 13600 Iabc pin is always about two diode drops above the V(-) line (as it should be, since it is a current mirror input referenced to negative rail), but in the drawing it always has a current going through it, since the op amp which drives it has an output which sits at ground in the quiescent state. The Iabc pin is being used as a half-wave rectifier, and this pin should be deprived of current for alternate half-cycles to work properly. There should probably either be a Zener diode in series with the Iabc pin, or the pin coupled to the op amp output through an electrolytic capacitor with a reverse clamping diode between the Iabc pin and V(-).
Probably still behaves the same way wrt beating input signals.
Mike
how about massively lowpassing the control-path before rectifying?
(pre-de-emphasis)
How about something simple like a compression diode [like Joe Davissons Antiquity Fuzz], or even a clipping circuit might make less voltage ripple when two notes are played [if the clipping threshold were set low]...of course there is the trade off between amount of sweep and steadiness of sweep when compressing the voltage swing of the EF's input source, but maybe some mileage could be gotten using something like this ?
A series of delays [like reverb] into the EF, would smooth out 'two note ripple', you could still get good changes in amplitude [of course slowed by however many milliseconds the delay extends input to the EF]...I'think delays made short enough to allow the follower to still 'follow', and long enough to fill in the 'dips between the swells between interval harmonics [like when you try to harmonically tune a guitar with a buncha reverb or some echo on, you can't hear the pitch/notch/passover].
Because the echo need not be in the signal path, a relatively low grade [inexpensive] chip [I think there are premade ones around for fairly cheep?] could be used without any degredation of audio signal, and improvement of sweep steadiness...also sweep slowness of course...DPDT...
Am I rambling yet?
So here's a couple of envelope generators ( fairly complex ) ...
http://www.musicfromouterspace.com/analogsynth/ADSR.html
http://www.musicfromouterspace.com/analogsynth/adsr_new.html
Is there a way of generating a simpler but "stable" release portion for using
in a filter/auto wah type circuit ??
Do we need ADSR, or just "R" ?
would one 4066 and a few parts do the job here ....
Marty.
pete: the end-result of that kinda envelope would be in the same realm as the "quadrature" one mentioned on bottom of this page: http://www.diystompboxes.com/smfforum/index.php?topic=37679.60 (http://www.diystompboxes.com/smfforum/index.php?topic=37679.60)
(the thread that is combinable with this one...)
Mike - I'll repost the original Space Drum schematic this evening.
Stephen
Quote from: MartyMart on October 18, 2005, 04:08:19 AM
So...
Do we need ADSR, or just "R"Â ? ....
Marty.
or AD...
?
Looked at Steve's re-used EH sweep schematic again. Easy to confuse the symbols for V(-) and ground on that schematic! Re-draw with proper ground symbol is recommended...
The op amp (part of LM324) used for audio input has it's (+) input referred to V(-) through a one meg resistor.... not to ground! It is used a noninverting gain stage/buffer that clamps to V(-) on the negative input half-cycles, making it a sort of half-wave rectifier. The output is able to swing very close to the V(-) rail since the LM324 is used. The output would consist of "half-wave rectified" positive half-cycles. This is fed into a second LM324 non-inv stage which provides additional gain for driving the OTA Iabc pin directly. Should be OK since the current into the Iabc pin will drop to zero when the output of the op amp is less than two diode drops above the V(-) rail. The amplified positive half-cycles force current into the Iabc pin, the OTA is connected as a variable lag circuit, with the lag time becoming shorter as the audio input becomes larger and average Iabc increases. The "start" voltage is applied to the OTA (+) input, negative feedback forces the OTA to charge up the integrating capacitor to make the voltage at the (-) input equal to the "start" voltage, the larger the input signal the faster the OTA is able to achieve equilibrium. On the negative input half cycles the OTA output current becomes zero (the cap "holds" the most recent value, with slow discharge through the "release time" pot)...when the note finishes the integrating cap discharges fully.... not to ground, but to the "stop" voltage, which is present in buffered form at the output of the op amp connected to the lower side of the integrating capacitor. In summary, the circuit is an "attack-release" envelope generator which tries to produce an output equal to the start voltage at the beginning of each note, while decaying to stop voltage between notes. The louder the note, the faster the circuit is able to reach the start voltage, giving a dynamic response to playing style by providing a level-dependent attack time. A nice feature is that the start voltage can be set higher than the stop voltage, and vice versa, for normal and reverse sweeps.
Right, Steve? :D
Cheers, Mike
Absolutely Mike a brilliant synopsis. Just got back from a swim at a public pool on my way home from work. Funnily enough, in the public gallery was a young lady pouring over what looked like a huge - I mean building plan size schematic as I swam past. So I just had to get out and take a look, but it was in fact a building plan, and she said that she was doing a little work whilst her kid was having a swimming lesson, and I had not been the first to comment on her huge piece of paper. The original Space Drum schematic with a CA3094 is here:
http://img.photobucket.com/albums/v317/StephenGiles/EHspacedrum.jpg
Upward & onward
Stephen
Thanks Steve! The schemo is a bit different but the analysis is the same.
An aside: if you find a trashed EH SD with bad crystal mic, rubber-cementing one of those flat- circular brass piezo beeper plates from an old watch works even better than the original (and doesn't turn to powder on hot days, unlike the Rochelle salt slab in the crystal mic). And if you see a Marshall Time Modulator schemo while swimming, be sure to get a copy and post to the list!
Cheers, Mike
Mike I used to have a couple of SDs which I bought for £2 each and sold for £100 each about 12 years ago!
Incidentally, for all you staunch Canadians, it was on this day in 1929 that women ceased to be"Chattels" and became "Persons" - heard it on the radio today.
Stephen
Also should look at the original context of this circuit... ;)
Used in a drum synth, where the output voltage was used to sweep an oscillator up/down or reverse.
The input signal is a short impulse provided by a crystal mic attached to a drum pad. A sharp tap on the pad would provide a nearly instantaneous sweep up to the start voltage, followed by decay to the stop voltage. A light tap would provide a noticeably slower ramp up to the start voltage, which would then decay to the stop voltage. A very light tap would provide an even slower ramp up to an intermediate voltage.
Mike