Welcome, Guest. Please login or register.
Did you miss your activation email?

January 20, 2020, 07:09:19 PM

Login with username, password and session length

Recent Posts

Author Topic: Trem pulse LED or Why Won't This Work?  (Read 12576 times)

galaxiex

  • Legendary
  • ******
  • Posts: 519
  • Chip Points: 46
    • View Profile
Trem pulse LED or Why Won't This Work?
« on: September 14, 2014, 10:11:42 PM »
So in my circuit newbishness I built the Trill Tremolo from MODKitsDiy.

http://www.modkitsdiy.com/pedal/trill_tremolo

Turns out they have a Mod for the Mod kit....   :o  ;)  :cheesy:  ::)

Actually more than one mod...

They have...

DC power jack add... Mod

LED add... Mod

Intensity control... Mod

(Ok, enough already, they see the bad humor) .... Did I say that out loud?

I've already done the DC jack and the Intensity control.
Waiting for a 3PDT stomp switch to get here so I can add the LED... but wait...
Why have a boring plain old glowing-not pulsing LED when you could have....
....a pulsing LED.

There is already an oscillator circuit right there!
Just need to tap into it somehow...

Here's the schematic for the Trem...

http://www.modkitsdiy.com/sites/modkitsdiy.com/files/product_files/the_trill_tremolo_schematic_1.pdf

I tried to interface an LED with a 2N3904 BJT but could not get it to work.

EDIT; I hooked onto the collector of Q3 in the trem circuit.

No matter what I did it would either load down the circuit and kill the oscillation, or the LED was so dim you could barely see it.

Ok, new plan, try a JFET. PF5102

The schematic attached below is what I came up with.

Can anyone tell me why/if this..
A. is a bad idea
B. won't work properly
C. has potential trouble spots
D. has circuit design errors

Also, where to connect to the 3PDT switch (when it gets here) to kill the LED at the same time as the effect.

I was thinking to put the switch on the 9V line going to the LED circuit.

If I were to put the switch on R1 of the LED circuit, it would kill the pulse but the LED would stay lit.

I have bread-boarded the LED circuit and clipped it into the Trem circuit at power, ground and the collector of Q3
it seems to work...
But I did not actually plug a guitar and amp in to try it.
EDIT; actually I have now plugged in a guitar and amp and the Trem still works, so far...

Thoughts? Opinions? Help?  :)

Thanks!

EDIT 2; Here is their plain old boring, non-pulsing LED add schematic.

http://www.modkitsdiy.com/sites/modkitsdiy.com/files/product_files/led_add_schematic2_2.pdf
« Last Edit: September 14, 2014, 11:16:17 PM by galaxiex »
Fear leads to Anger, Anger leads to Hate, Hate leads to Suffering.

galaxiex

  • Legendary
  • ******
  • Posts: 519
  • Chip Points: 46
    • View Profile
Re: Trem pulse LED or Why Won't This Work?
« Reply #1 on: September 14, 2014, 11:55:24 PM »
Ok, to sum up that mess above ^^^

The LED circuit connects to the Trem circuit at +9V, ground and the collector of Q3.

It works... sorta..

With R1 LED connected to Q3 trem it adds a "thump thump thump" as the trem oscillates.
Also some added hiss in the signal.

I tried adding a low value cap between Q3 and R1. Doesn't work. LED won't blink.

Need some ideas here...

Thanks!  :)

Trem Circuit...

http://www.modkitsdiy.com/sites/modkitsdiy.com/files/product_files/the_trill_tremolo_schematic_1.pdf
« Last Edit: September 14, 2014, 11:56:57 PM by galaxiex »
Fear leads to Anger, Anger leads to Hate, Hate leads to Suffering.

Roly

  • Legendary
  • ******
  • Posts: 2184
  • Chip Points: 288
    • View Profile
    • Australian Valve Amps
Re: Trem pulse LED or Why Won't This Work?
« Reply #2 on: September 15, 2014, 05:09:54 AM »
The main problem here is that the oscillator, Q3, is fairly sensitive to loading on its collector, as you have discovered (and I have to say that I'm more than a little surprised that there isn't a limiting minimum resistor in series with the Depth control, but no matter).

It would be helpful to know what the available peak-to-peak signal swing is on Q3 collector, but the first thing I'd try is a Darlington arrangement.


Quote from: galaxiex
No matter what I did it would either load down the circuit and kill the oscillation, or the LED was so dim you could barely see it.

Ok, new plan, try a JFET. PF5102

Well that's half an answer in that it will have a very high input impedance, but how much current are you hoping for through the LED?  Will the FET carry that current?  Most of these small FETs saturate somewhere between 1 and 10mA.

A Darlington arrangement of two transistors gives a very high input current sensitivity allowing the use of a high input series resistor, say 1Meg.  You can check if the oscillator will tolerate this by simply tacking a 1Meg between Collector and ground and see if it still works.

The first circuit below is a straight forward Darlington emitter follower, so the voltage across the resistor+LED will be two BE drops below the Collector of Q3.  This may, or may not, be high enough to light the LED.

The second is a voltage controlled current sink.  After the input is higher than two BE drops the current in the LED is determined by the input voltage on the emitter resistor.  Depending on the available swing from the oscillator and the current you desire through the LED the value of the emitter resistor will be somewhere between 1k (1mA per Vin) and 100 ohms (10mA per Vin).

Even 'tho it has a lower input resistance than the emitter follower it's likely to still be quite high enough, so I'd be inclined to the second for smoother LED control.


...or you could add an LM324 quad op-amp.
If you say theory and practice don't agree you haven't applied enough theory.

galaxiex

  • Legendary
  • ******
  • Posts: 519
  • Chip Points: 46
    • View Profile
Re: Trem pulse LED or Why Won't This Work?
« Reply #3 on: September 15, 2014, 07:48:45 AM »
Thanks Roly!

Please see if I made a mistake on the following...

I am using a green led rated 3.6V @ 20mA. (on the package, I didn't measure it)

With 9V and 1K series R I get...

I=V/R   9/1000= .009A or 9mA .......

Or do I use the V drop of the LED?

9 - 3.6= 5.4V   5.4/1000=.0054  5.4mA?

Either way,
The FET seems ok at that current, does not get even warm.

The LED is nice and bright even at that low current.

PF5102 data sheet says Pd 625mW Max, data sheet does not tell me the "on" resistance.

Oh, maybe this? ...... Thermal Resistance, Junction to Ambient = Max 357  Units C/W

--------------------------

The P-P swing on Q3 collector is ~ 2 - 2.5V

Yes, Q3 and the depth control seem finicky.

I did the intensity mod..

http://www.modkitsdiy.com/sites/modkitsdiy.com/files/product_files/trill_intensity_modification.pdf

...but my 1M log pot is not here yet so it still has the 1M lin pot installed.  :o

I thought about a minimum series R for the depth/intensity pot cuz going from "0" to any increase in Depth there is a  "delay" for the effect to start while a cap somewhere, (probably the 10uF at Q3 collector) charges.

I'll try a few min R values and see...

Will try the 1M to ground test, and the darlington circuits also.

Something else I noticed after the pedal has been on for a while, the noise level seems to increase. (hiss)
LED circuit NOT hooked up.

Transistor going bad somewhere?

Thanks!
Fear leads to Anger, Anger leads to Hate, Hate leads to Suffering.

Roly

  • Legendary
  • ******
  • Posts: 2184
  • Chip Points: 288
    • View Profile
    • Australian Valve Amps
Re: Trem pulse LED or Why Won't This Work?
« Reply #4 on: September 15, 2014, 08:42:54 AM »
Quote from: galaxiex
Or do I use the V drop of the LED?

9 - 3.6= 5.4V   5.4/1000=.0054  5.4mA?

This is correct.  The LED turn-on voltage subtracts from the supply, so the voltage across the resistor is;

VR = Vbat-VLED

...and the resistance then comes from the current and that voltage;

R = E/I
R = VR/I(desired)
{K-ohms = Volts/mA}

At only 5mA you have obviously got lucky with your FET.

Quote from: galaxiex
Oh, maybe this? ...... Thermal Resistance, Junction to Ambient = Max 357  Units C/W

No, that's the thermal resistance of the case; "the chip will be 357ºC hotter than the case for every watt dissipated".  Obviously that's one hell of a thermal resistance and it isn't a power device.  Normally around power devices and heatsinks we are talking thermal resistances of between 1 and 10ºC per watt.  {you can do a thermal sim in LTSpice BTW because heat acts like electricity, thermal resistances, heat flow (current), temperature differences (voltages)}


Quote from: galaxiex
The P-P swing on Q3 collector is ~ 2 - 2.5V

Yes, Q3 and the depth control seem finicky.

I did the intensity mod..

http://www.modkitsdiy.com/sites/modkitsdiy.com/files/product_files/trill_intensity_modification.pdf

...but my 1M log pot is not here yet so it still has the 1M lin pot installed.  :o

Not nearly as much as I expected, more like rail to rail.  ...hummm...

That mod will kill the oscillator at some point as it is turned down.  I think it could be done better than that.

Pot taper can make a large difference to the operability of controls, e.g. touchy with a linear taper, smooth with a log.

Quote from: galaxiex
Depth there is a  "delay" for the effect to start while a cap somewhere, (probably the 10uF at Q3 collector) charges.

Having stopped the oscillator by loading, as you bring it up again it won't start until the loading is small enough and it may be marginal, taking several seconds to actually build up get going.

Quote from: galaxiex
Something else I noticed after the pedal has been on for a while, the noise level seems to increase. (hiss)
LED circuit NOT hooked up.

Transistor going bad somewhere?

Dunno.  Hard to say.  It's unlikely if they are new, but certainly not impossible.  Could also be salt sweat from your hands during assembly.  If you have some dry metho (i.e. fresh and kept sealed) then you could try giving your board a wash and scrub all over with a toothbrush, then bake it out in sunshine so it's bone dry and contaminant-free.  "Mud" board has a bad habit of picking up moisture and this can lead to "paper-rustling" and "popcorn" noises.  It's generally more important around high impedance sections such as FET Gate circuits.
If you say theory and practice don't agree you haven't applied enough theory.

galaxiex

  • Legendary
  • ******
  • Posts: 519
  • Chip Points: 46
    • View Profile
Re: Trem pulse LED or Why Won't This Work?
« Reply #5 on: September 15, 2014, 09:46:30 PM »
This is correct.  The LED turn-on voltage subtracts from the supply, so the voltage across the resistor is;

VR = Vbat-VLED

...and the resistance then comes from the current and that voltage;

R = E/I
R = VR/I(desired)
{K-ohms = Volts/mA}

At only 5mA you have obviously got lucky with your FET.

Got it, and yes, yes I did.
After you mentioned this (lucky) I tried a 2N3819 in the same circuit and it did not even light the LED.

Not nearly as much as I expected, more like rail to rail.  ...hummm...

That mod will kill the oscillator at some point as it is turned down.  I think it could be done better than that.

Pot taper can make a large difference to the operability of controls, e.g. touchy with a linear taper, smooth with a log.

Yes, I see what they did with the depth mod.
Before, the depth would not go very low, control at minimum and it still had quite a heavy pulse.
Original circuit... http://www.modkitsdiy.com/sites/modkitsdiy.com/files/product_files/the_trill_tremolo_schematic_1.pdf

So the mod allows you to ground the + side of the 10uF cap, killing the trem altogether.
Mod circuit... http://www.modkitsdiy.com/sites/modkitsdiy.com/files/product_files/trill_intensity_modification.pdf

A log taper pot would probably help, but with the modded depth circuit and a LIN pot it is VERY difficult (impossible) to get anything other than full depth no matter how carefully you turn up the depth control.

Me thinks this circuit was not well thought out...  ::)

So-far I have bought 3 of these MODKits pedals. This one, and..

The Verb Deluxe http://www.modkitsdiy.com/pedal/verb_deluxe

The Tea Philter http://www.modkitsdiy.com/pedal/tea_philter

Thus far I am only happy with the Verb Deluxe. GREAT reverb sound for a digital brick thing.

I had hoped the Tea Philter would give me that "cocked wah" sound, and it does, but not very effective.
A real wah (Cry Baby) with the pedal cocked you can get that honky mid-range sound.
The Tea Philter does that but it's weak at best and soon lost with any volume.
Kinda sad cuz they tout it as "BEST NEW Floor Effect of 2012!"
"Music & Sound Retailer Nominations for Best New Floor Effect of 2012"
Maybe there is something wrong with mine....

But I digress...

Having stopped the oscillator by loading, as you bring it up again it won't start until the loading is small enough and it may be marginal, taking several seconds to actually build up get going.

Yes, it does take several seconds to "come back on-line"
I don't like it and have now put the circuit back to "stock".
I tried putting a resistor in-line with the ground leg of the depth pot (modded circuit) of 100K as testing showed that value to "just" allow the circuit to oscillate with the depth pot at min.
Still didn't sound right. Too much pulse for a min setting.
Gonna play with it and see what I can blow up.... :o  :lmao:

BTW, I have to confess to being lost when it comes to oscillator circuits.
I mean, I know what they do, make a voltage flap up and down, but grasping how they *actually* function still eludes me.
I've read tons of stuff about them, but seems like another of those things I can't quite "get". (shrug)

Dunno.  Hard to say.  It's unlikely if they are new, but certainly not impossible.  Could also be salt sweat from your hands during assembly.  If you have some dry metho (i.e. fresh and kept sealed) then you could try giving your board a wash and scrub all over with a toothbrush, then bake it out in sunshine so it's bone dry and contaminant-free.  "Mud" board has a bad habit of picking up moisture and this can lead to "paper-rustling" and "popcorn" noises.  It's generally more important around high impedance sections such as FET Gate circuits.

Just a FYI, these kits are built using point to point on terminal strips.
No circuit board.
Another way they "cheaped-out" by not even bothering to do up a board for them.
I guess it gives the DIYer a sense of building from days gone by...  ::)

I think I'm gonna hafta start making my own circuits and boards. (or at least copying proven designs)

Long time ago I hand applied resist ink to "draw" simple circuits on blank copper boards.
Etched with ferric chloride. Then Drilled, Soldered.
Never tried the photo mask method. Thought it was out of my range...

Yesterday I watched a vid of how easy it is to do the photo method now. (I've got a good laser printer)
Gonna hafta try it sometime when I get to designing something.  :tu:

Sorry for the long wordy post!

Cheers!

EDIT; Oh, I forgot.. I tried a 1meg resistor from Q3 collector to ground and it does not (seem to) affect the circuit at all.
« Last Edit: September 15, 2014, 09:51:25 PM by galaxiex »
Fear leads to Anger, Anger leads to Hate, Hate leads to Suffering.

Roly

  • Legendary
  • ******
  • Posts: 2184
  • Chip Points: 288
    • View Profile
    • Australian Valve Amps
Re: Trem pulse LED or Why Won't This Work?
« Reply #6 on: September 16, 2014, 03:09:45 AM »
Quote from: galaxiex
I tried a 2N3819

This is essentially the same as an MPF102, and that quotes it Idss (max current Drain to Source, saturated) spread from 2 to 20mA, so it's about 50/50 that an individual device will be above or below your 5mA.


Quote from: galaxiex
(impossible) to get anything other than full depth no matter how carefully you turn up the depth control.

Me thinks this circuit was not well thought out...  ::)

There are an awful lot of circuits out there, and some of them even work!  It is inviting to think this may be an artifact of the Interweb generation but I can assure you that long before the Interweb there were examples of paper magazines publishing circuits that simply could not do what was claimed of them going back at least to the 60's.

In my youth I spent a lot of time trying to get a transistor amp circuit out of American Popular Electronics to produce power instead of smoke until I took it to an Electronics Engineer I knew who carefully explained to me why it wouldn't/couldn't possibly work.  It rather dented my faith that electronics magazine staff a) knew their topic and b)would filter out the rubbish - in the longer view both turned out to be wrong.

Quote from: galaxiex
"Music & Sound Retailer Nominations for Best New Floor Effect of 2012"
Maybe there is something wrong with mine....

...or maybe the opinions of "Music & Sound Retailers" on the topic aren't worth ...um... are worthless?

Okay, no beating around the bush; I don't think that these guys have much of an idea about what they are doing electronically.  10/10 for marketing, but close to zero for electronics design.  What I'm seeing with these mods on mods is cluelessness, people who can't get it right the second or even third time.  They don't even seem to have the acquired experience of going over a lot of other peoples pedals, reading R.G.Keen, etc.


Quote from: galaxiex
I have to confess to being lost when it comes to oscillator circuits.

An oscillator is only an amplifier yelling in its own ear.

I know people who have spent years tinkering with simple oscillator circuits because they can be fascinating.  The oscillator lives its live carefully balanced on a knife edge between oscillating too hard and not oscillating at all, between drifting up or down in frequency and pulling itself back into line (called phase noise or jitter).  Getting a pure sine wave out of an oscillator requires the exact satisfaction of the correct operating condition, something that is pretty well impossible; like a tight-rope walker not wobbling at all, but real oscillators are like real tight-rope walkers - they wobble a bit, always falling one way or the other then correcting.  Building an oscillator is easy.  Building a good (pure, undistorted) oscillator is hard.


Oscillation
First up, think of pushing somebody on a swing.  If you push at the right time, just after they have stopped at the peak of the swing, then each push will build the amplitude up.  Now try pushing just before they come to a stop naturally and the swinging oscillation will be reduced - so the timing of the feedback is important.

An oscillator consists of gain to overcome losses (an amplifier, which is normally inverting) and a delay network between the amplifier output and its input.  The Phase Shift Oscillator (PSO) in your trem is classic, a transistor for (inverting) gain, and a three-leg CR phase shift network (PSN) that delays its input by 180º of phase at one particular frequency.  The signal at the Collector that was 180º out of phase with the Base (negative feedback) is now in-phase (positive feedback) and is what is also called regenerative - more gets you more, and so the loop will oscillate trying to find a stable situation, all corrections will be over-corrections so they will build up until the amplifier clips against ground or supply.

If the loop gain is less than unity, i.e. the feedback phase shift network has more losses than the amplifier has gain, then any oscillations that start will quickly die out.  In a PSO it happens that the losses through the PSN are exactly 1/29, so the amplifier has to have a gain of 29 or more for the oscillation to be sustained.  If the gain is too high the oscillation will be distorted, clipped or even square.  The more excess gain the loop has the quicker the oscillation will start.


{mic feedback with a PA or feedback through your strings (good) or directly through the pickup (bad) have exactly the same set of conditions, a feedback path/loop, greater that unity gain around the loop, and a delay or phase shift that causes this feedback to be in-phase at some frequency, then ... scREEEECH!  The air does a pretty good job of the PSN by introducing a delay of 1mS per foot between speaker and mic (so in this case there may be quite a few whole cycles delay between speaker and mic).  The classic response to mic feedback, pointing the mic away from the speaker,  effectively reduces the loop gain below the oscillating level.  The mechanical escapements in clocks and watches are also classic oscillators that follow the same rules of loop gain and whole-cycle phase shift, the push at the right moment.}


With the Depth control arranged as it is to load down the collector of the transistor this will have the incidental effect of changing the amplifier gain.  When it is down the earthy end of the pot the loading will be so great that the gain will fall below 29 and the oscillator will die away and stop.  This is not "turning down the level", it's switching the oscillator off.

Two things determine how quickly an oscillator starts, excess loop gain, and circuit noise.

With perfect components an oscillator wouldn't start because it requires some microscopic discontinuity to be amplified around the loop to get things going - noise.  At some critical point on the Depth pot the amplifier will have a gain of exactly 29 and the circuit will be right on the cusp of oscillating as soon as some small disturbance causes the loop to start ringing, but because there is no excess gain it may take a very long time for the oscillation to build up to limiting, maybe seconds.

If you crank the Depth up high then the gain will also be high and the loop will kick off more quickly.

As it is wired the Depth pot should function more as an oscillator switch than as an output level control, which I think is what you are observing.  It is a worryingly "unsophisticated design" and looks very amateurish to my eye.


Quote from: galaxiex
I think I'm gonna hafta start making my own circuits and boards. (or at least copying proven designs)

A fresh waterproof fine tip felt marking pen is my favorite way of laying down resist by hand, effective, and washes off with metho.


Quote from: galaxiex
I tried a 1meg resistor from Q3 collector to ground and it does not (seem to) affect the circuit at all.

Okay, so the two LED driver circuits I put up can have a 1Meg input resistor without loading the oscillator - good.


Just looking at your LED buffer.  A prime reason for using a FET is because it has such a high input resistance at its Gate, which in turn allows you to use very high resistor values for its biasing so they don't load the source.  For 47k and 100k you could have used 4.7Meg and 10Meg.  In this case you could omit all the biasing and just connect the Gate to the oscillator Collector (however you have a few volts of G-S bias voltage to deal with).  The Darlington arrangement with transistors gives almost as high an input resistance, but with only two BE drops between the control and output.

As the oscillator will tolerate a loading of 1Meg, if you go with the voltage-controled current sink, the second circuit, then you have almost the full supply swing to smoothly control your LED brightness over the oscillator cycle (most of the other arrangements I expect to  be more or less like an on/off flash).
If you say theory and practice don't agree you haven't applied enough theory.

galaxiex

  • Legendary
  • ******
  • Posts: 519
  • Chip Points: 46
    • View Profile
Re: Trem pulse LED or Why Won't This Work?
« Reply #7 on: September 16, 2014, 12:08:17 PM »
Thanks! You have helped my understanding of oscillators by a factor of 100 at least.  :)  :dbtu:

So after doing a bit of research I think Fender originally got it right using a light source and LDR for the Tremolo effect.

I'm not stuck on this pedal circuit and will happily abandon the original for a opto project.
There are about a gazillion out there...

Just gotta pick one...  ::)

This is cool!  8| ... but I probably won't use it in this pedal.

http://makezine.com/projects/Optical-Tremolo-Box/

At least I have a good base to build from, the enclosure and associated bits.
I'm gonna scrap the original circuit (sows ear) and find something opto that I like. 
May even make a board for it.  ;)

Okay, no beating around the bush; I don't think that these guys have much of an idea about what they are doing electronically.  10/10 for marketing, but close to zero for electronics design.  What I'm seeing with these mods on mods is cluelessness, people who can't get it right the second or even third time.  They don't even seem to have the acquired experience of going over a lot of other peoples pedals, reading R.G.Keen, etc.

Ya, so no point to saving the original circuit except as an exercise to see what could be done with it, but that's a waste of time for trying to get something pleasingly functional.

EDIT; An interesting paper on Tremolo circuits...

http://www.strymon.net/tag/tremolo-circuit/

EDIT 2; just a side note.. I ordered the AMZ CD  <3)

http://www.muzique.com/amz-cd.htm

Figured the info was worth having for anticipated future projects...  ;)
« Last Edit: September 16, 2014, 07:55:51 PM by galaxiex »
Fear leads to Anger, Anger leads to Hate, Hate leads to Suffering.

Roly

  • Legendary
  • ******
  • Posts: 2184
  • Chip Points: 288
    • View Profile
    • Australian Valve Amps
Re: Trem pulse LED or Why Won't This Work?
« Reply #8 on: September 17, 2014, 01:41:02 AM »
Quote from: galaxiex
Thanks! You have helped my understanding of oscillators by a factor of 100 at least.  :)  :dbtu:

That is very pleasing.  :)  {One of the advantages of teaching face-to-face is that you can actually see the "WTF?" expression that might arise at some point, and know that you have to back up and get that point clear before moving on to the next.}


Quote from: galaxiex
So after doing a bit of research I think Fender originally got it right using a light source and LDR for the Tremolo effect.

I don't know that Fender actually invented the LDR method, but it has a couple of serous advantages over other methods, the main one being total DC isolation from the oscillator circuit - you just have this wobbling resistance you can apply wherever you fancy without having to worry about the Low Frequency Oscillator (LFO) injecting its LF signal into the signal chain.  In older Fender amps this often resulted in LF "pumping" of the speaker cone.


{ :grr = ON 
The Harmonic Trem is actually not a pure tremolo effect. It is really a dual-band filtering effect that alternately emphasizes low and high frequencies.

Well ... no.

Forgive me, but the mis-perception of the Fender "Harmonic Vibrato" as a simple amplitude/frequency crossover network is a personal hobby-horse.  This writer comes closer than most, but they are all so busy looking at the amplitude response they all miss the basic point about the signal phase; the Harmonic Vibrato is indeed a vibrato (frequency-changer, despite the claims made by Magnatone) because it is actually a phase modulator, the small amplitude modulation being incidental to its operation.  Leo Fender actually used two different forms of circuit for this effect, and by coincidence Tom Jennings was doing something very similar in the Vox AC-30, only using a slightly different and more complex circuit (which should give less amplitude modulation).

If you bother to actually investigate these networks you discover,as I did, that they are "all-pass" filters designed to primarily alter phase, not amplitude.

These circuits are strongly related to the "phasing" Single Sideband (SSB) generators used in radio communications.  The power tube bias-modulation tremolo is also used in radio communication and is known as a balanced mixer which produces Double Sideband Suppressed Carrier, (DSBSC).  The "suppressed carrier" in this case being the LFO signal.  The fact that this method is also prone to speaker pumping is mainly down to a lack of match between the two paths through the output stage, and everything has to be balanced for full carrier suppression, e.g. the OPT, not just the valves.
 :grr = OFF}



Jack Orman at muzique is one of the more credible circuit site around (at least I have found little I would argue with there, unlike some other sites).  To me it's important that a site has a name attached to it; it's about responsibility giving credibility.  Generally speaking the sites that are not talking rubbish have a name attached, R.G.Keen, Jack Orman, Steve Bench, Rod Elliot, Doug Self, Don Tillman, etc.

The second sign of a credible site is an easy way to contact the author if you want to ask questions or take issue with something.  Now look at the circuit library sites that have no name attached or any way of contacting the operators with corrections, etc.

To illustrate the "dark side", here is my favorite design abomination - you normally don't need trim pots to adjust FET stages, and when you do you don't put them in the Drain circuit.  This says to me "I don't know how to find the right value, in fact I'm not too clear on how FETs work, so I've made everything adjustable".  This sort of "design" should be hounded off the Interweb to protect the innocent.   :duh

Let's have a look at something that illustrates a better way of doing it than the circuit you are struggling with.  This was designed by Electronics Australia who had at least some idea of what they were doing.

Here is the un-modified original EA-tremolo;


(page 2), Page1Page 3.

As it happens they used an identical circuit (why reinvent the wheel?) in their Playmaster 125 S.S. guitar amp, and so this was in the preamp of my (much modified) Twin-50 for many years.

The signal amplifier (right) could be any kind of single-device gain stage, FET, BJT, triode, but the important point is what is happening with the Source/Emitter/Cathode resistor bypass cap.  It has a FET in series acting as variable resistance, so the amount of bypassing changes, and this changes the AC gain of the stage (hopefully without disturbing the DC conditions).  An LDR does much the same thing only controlled by light rather than a voltage, so you could put the LDR where the FET is (generally speaking, there are some details).

What is happening here is that the un-bypassed Emitter (etc) resistor give the stage local AC NFB, so reducing gain.  When the FET is low resistance the bypass cap comes into play, reducing the local AC NFB and thus increasing stage gain.

Now note how the LFO signal comes out via a very conventional level control, a "volume" for the LFO signal, the loading on the oscillator remaining minimal at all settings (because a FET Gate is effectively an open circuit).

Here is an "Improved" EA-Trem done by somebody who looks like they had some idea of what they were doing (Spanish);


http://www.pisotones.com/EA-Tremolo/MBC/EA-Tremolo_MBC.htm
Note LED D1 which should pulsate with the changing current through the oscillator transistor Q3.

...but from my experience the original works just fine. It calls for a MPF105 but any of the normal small signal FET suspects should work (the MPF105 has a slightly tighter spec spread than the MPF102; note the 560k in series with the Depth pot that allows trimming the Depth control range).
If you say theory and practice don't agree you haven't applied enough theory.

galaxiex

  • Legendary
  • ******
  • Posts: 519
  • Chip Points: 46
    • View Profile
Re: Trem pulse LED or Why Won't This Work?
« Reply #9 on: September 18, 2014, 12:38:09 AM »
Quote from: Roly
don't know that Fender actually invented the LDR method, but it has a couple of serous advantages over other methods, the main one being total DC isolation from the oscillator circuit

Didn't mean to imply Leo invented it, just that his use of it and the isolation factor makes sense.
Still seems like the "best" way to implement Tremolo.
Until I read that paper at the Strymon site, I didn't know that Fender used anything other than the "Classic" Tremolo like that used in the Black and Silverface amps.
BTW, I do take things I read on the interwebs with a grain of salt, no matter how convincingly and "authoritatively" written.
Strymon "seems" more credible than some, but anyone can make mistakes.
I am aware of the Magnatone amps Vibrato although I have never seen the circuit, or even heard one of those amps in person.
I've seen it described as "true pitch shifting vibrato" rather than amplitude modulated etc...
...but the Fender Silverface Tremolo (that Fender confusingly called "Vibrato") is the sound I grew up with and indeed, I own a near mint '78 Deluxe Reverb, (with the dreaded Volume "pull boost") all original  <3) except for the speaker  :( .
Are you aware the Magnatone amps are re-issued? Quite a story, I won't regurgitate it here...  ;)

Quote from: Roly
If you bother to actually investigate these networks you discover,as I did, that they are "all-pass" filters designed to primarily alter phase, not amplitude.

Interesting... I'm not going to investigate, I'll take your word for it.  ;)  :)
I DO find it interesting that they alter phase. The forerunners of Phaser pedals maybe?

All that is quite over my head right now, and for now will be happy with a decent Tremolo pedal that doesn't tick, or have other obvious artifacts from the LFO.
Just a nice clean volume modulation with decent speed range and depth control.  :) Oh, and a pulsing LED...  ;)

BTW, I do own a Boss TR2 and it's ok... has 3 different wave shapes and all... but it just doesn't have that "Vibe" ya know?
Seems too "commercial" somehow.

Jack Orman; I've been browsing his site quite a lot and wanted to acknowledge in some way,
so figured a purchase was a good way to do that.

runoffgroove; ya... I wondered at the use of trimmers in some (many?) of the circuits there...
It "sorta" made sense to me in that you can use the trimmer to optimize that particular value... in that particular circuit... but if you have the design skills? to make such a thing (circuit) in the first place, why wouldn't you publish the circuit "already optimized" ????
Component differences? I don't know... but I don't think I've ever seen ANY other site show a (finished) circuit with trimmers like that. Except maybe a critical timing circuit or some such, where you really need to trim something.
Maybe I don't get out enough....  ;)  :lmao:

A big THANK YOU! for that EA Tremolo pages.  :) The description of the circuit alone is VERY worth the time it takes to read it!
Your added description/points about the circuit are well taken too.  :)
THAT looks like something I can build (and tinker with) even tho it doesn't use a LDR, and even to my untrained eye, it just looks like a decent design. It's gotta be... it's lasted all these years, right?  8)

Some of those old magazine published circuits/article are great in that they sometimes give you a GREAT description of the circuit and how it works.
I've learned a huge amount from articles like that.
... but I've been burned too, much like you were with that amp from Popular Electronics.
I don't remember now what it was, or which magazine, but it was a simple circuit that had a schematic drawing error.
I managed to get lucky and saw the error (after 3 days of trying to get it working) and forever lost faith that ANY published circuit was "perfect" and error free.
If I build something out of a magazine and it doesn't work, the FIRST thing I check is the schematic, cuz I'm rather painstaking in my construction methods and rarely make a mistake in wiring and assembly.

(my schematic drawing is a different thing however...)  ;)  :lmao:
Fear leads to Anger, Anger leads to Hate, Hate leads to Suffering.

Roly

  • Legendary
  • ******
  • Posts: 2184
  • Chip Points: 288
    • View Profile
    • Australian Valve Amps
Re: Trem pulse LED or Why Won't This Work?
« Reply #10 on: September 18, 2014, 08:00:53 AM »
Well Leo could have been the first to use LDR's, I just don't know who was first.  From memory (so I could be mistaken) Fender amps have used several different modulation methods; wobbling the bias on the output pair would have been the first, I think an early one used modulation of the cathode current of a preamp, the two similar-but-different "Harmonic Vibrato" circuits, and of course the LDR, which I think makes five different methods.  I don't blame Strymon for repeating what even one (actually) authoritative book on Fender amps gets wrong.

The Magnatone claim is half right.  The Magnatone circuit is a phase modulator and if you modulate phase you are modulating frequency - it is simply impossible to do one without doing the other (since phase and frequency are different ways of looking at the same thing).  I can say 1kHz (one thousand cycles per second), but one cycle is 360º of phase so I can say with equal validity this is a "360,000º per second" signal.

The problem arises when Magnatone claim that their circuit is "the only true frequency shifter".  If they accept that a phase modulator like theirs shifts frequency then they have to accept that Leo Fender's "Harmonic Vibrato" and Tom Jennings very similar Vox AC-30 circuit are also frequency shifters.  Phase shifters have long been used in electronic organs and may even pre-date the Magnatone.


This will work fine, but it could be improved and simplified with a few FETs and LEDs.  Incidentally with three PSNs this gives three times the shift of the Magnatone, Fender, or Vox circuits.

Magnatone used a very different circuit using Voltage Dependent Resistors (VDR's) to get their result, and that is what the patent was granted for, the VDR method, not the phase/frequency result.


Quote from: galaxiex
I DO find it interesting that they alter phase.

Something to paste in your hat; frequency-selective networks also produce phase changes, lag for low pass and lead for high pass.

This is a classic Bode plot for a single low-pass CR network;


You will notice how the phase starts to rotate (lag) before the amplitude starts to drop off, being 45º at the "hinge" or corner frequency fc, and finally settling to -90º in the stopband.


Quote from: galaxiex
a decent Tremolo pedal that doesn't tick

"Ticking" can normally be traced to a serious discontinuity in the LFO signal (too much oscillator gain causing it to hard clip) or to cutting off the modulator, e.g. the LED going right out for part of the LFO cycle.


The small signal FETs we use have a very wide spec spread, Idss from 2 to 20mA for an MPF102 for example, but while this FET has a very wide datasheet specification it is rare these days to find individuals out near the limits.

If you look at the runoffgroove "Umble" preamp you will see that the Source 1k5 resistors are bypassed by 4.7uF caps.  If you change the Drain resistor (e.g. with the trimmer) you are changing both the DC conditions and the AC load, so the signal gain will also change.  If you vary the bypassed Source resistor however you are only changing the DC conditions because the AC signal is bypassed around the resistor, and that is what is actually required.  This is the right way to DC re-bias the FET without changing the AC gain (significantly).


Before Electronics Australia was killed off by Hi-Fi marketing types it was a very credible electronics mag.  This is not to say they didn’t sometimes have some real bloopers (a certain RIAA equalised phono preamp comes to mind) and they may have been a bit less than forthright about their mistakes, but they did issue errata to correct those mistakes (at least until the dying days when things seriously turned to $@#t).  The only mag I know of today that 'fesses up to its mistakes is Elektor, and they generally maintain a blistering standard of electronic engineering that others struggle to keep up with.

Errors are one thing, but outright misrepresentation is another.  I remember a ham telling me about a "100 watt" RF amplifier he built from an article in an American ham radio magazine and whatever he did he couldn't get more than a few watts out of it.  Later he traveled to the US and made a point of catching up with the author.  When he commented that he couldn't get it to make watts he was astonished to be told "Yeah, I couldn’t get more than that out of it either".  So how did it get to be published with the claim of "100 watts"?  Who knows, but it certainly undermines your faith in the magazine editorial committee.  Mistakes I forgive, deception I don't.
If you say theory and practice don't agree you haven't applied enough theory.

J M Fahey

  • SSGuitar Global Mod
  • Legendary
  • ****
  • Posts: 4142
  • Chip Points: 429
    • View Profile
Re: Trem pulse LED or Why Won't This Work?
« Reply #11 on: September 18, 2014, 02:53:17 PM »
Another dangerous site is Red Circuits, where all noobs seem to be attracted like the moth by the candle .... which will eventually burn it .

At least they pulled the 10000W and the 15000W amplifiers (no kidding) with 110V rails or so  :o  :duh .... yes, 220V inside a "supposedly safe" SS amp.

but they still post jewels like this, a thermally uncompensated switching Mos power amp:


Guaranteed to thermally rundown and self destroy as soon as you play it moderately loud.

galaxiex

  • Legendary
  • ******
  • Posts: 519
  • Chip Points: 46
    • View Profile
Re: Trem pulse LED or Why Won't This Work?
« Reply #12 on: September 18, 2014, 05:04:27 PM »
Heard about Red Circuits, never been there... might go now just to have a look.... always good to know what NOT to do... ;)
Fear leads to Anger, Anger leads to Hate, Hate leads to Suffering.

Roly

  • Legendary
  • ******
  • Posts: 2184
  • Chip Points: 288
    • View Profile
    • Australian Valve Amps
Re: Trem pulse LED or Why Won't This Work?
« Reply #13 on: September 19, 2014, 06:57:26 AM »
Well one thing you don't do is draft circuits like the one above without values or transistor types.   :grr
If you say theory and practice don't agree you haven't applied enough theory.

J M Fahey

  • SSGuitar Global Mod
  • Legendary
  • ****
  • Posts: 4142
  • Chip Points: 429
    • View Profile
Re: Trem pulse LED or Why Won't This Work?
« Reply #14 on: September 19, 2014, 07:35:40 AM »
All of them are drawn that way, except the odd one which is cut and pasted from somewhere else.
All of them with the same primitive software .... Paint?
Or something as crude.

Attracts beginners because they post incredibly simple projects, like this 30W power amp .... irresistible:

Just 3 transistors !!!!!!

He even publishes specs, such as 10 to 100KHz frequency response, 0.4% distortion, etc.
All either pulled out of thin air or, maybe, simulated.

Problem is, simulators imagine perfect heatsinks, parts always at 25 DegC , no ripple, stable supplies, etc.

In the real world, this uncompensated amp will overheat, thermal runaway and die.

And the poor beginner gives up in frustration and instead gets into some easier hobby, such as surfing without a board, parachuting with an umbrella or learning the Ramayana, in Sanskrit, by heart.