Echo base not working! Help

[Lost_soul]

Idk why i have problems with every build i make (how lucky !)

This time it's an echo base delay pedal Version 2. I put it on the breadboard without any mods. I used a 2N3904 instead of the NPN transistor and used a BC559 instead of the PNP.




So the problem is, when on bypass and the base of the NPN tranny is grounded there is a normal signal as you would expect and LED is OFF. When the base is not grounded (ie: effect engaged) there is still a bypass sound! No effect what so ever. But the LED is ON and is changing with the rate pot.

I tested it with an audio probe and at pin 14 or 13 (can't remember) of the PT2399, i had a 1khz sine wave.
The feedback, time, depth and rate were all working. As for the level i couldn't tell.

Here are the voltages:

[ElectricDruid]

I think you need to audio-probe it to see how far the input signal gets.

Here's some test points:

1) U3B pin 7, input buffer output. This should be ok, so this is just a check.
2) U4A pin 2 and both sides of the 100n cap connected to it. This is the input signal to the PT2399
3) Signal at the top of the Feedback pot. This is the output from the PT2399.
4) U4D pin 10, input to the output CMOS switch
5) Top of the level pot. This is the output from CMOS switch. We're expecting the signal to have died before now because if we had signal here, the thing would be working!

Somewhere along this path, the signal has to disappear, since it doesn't come out at the far end. Where that happens will tell us where to look.

Let us know how you get on. Good luck!

HTH

[Lost_soul]

I think you need to audio-probe it to see how far the input signal gets.

Here's some test points:

1) U3B pin 7, input buffer output. This should be ok, so this is just a check.
2) U4A pin 2 and both sides of the 100n cap connected to it. This is the input signal to the PT2399
3) Signal at the top of the Feedback pot. This is the output from the PT2399.
4) U4D pin 10, input to the output CMOS switch
5) Top of the level pot. This is the output from CMOS switch. We're expecting the signal to have died before now because if we had signal here, the thing would be working!

Somewhere along this path, the signal has to disappear, since it doesn't come out at the far end. Where that happens will tell us where to look.

Let us know how you get on. Good luck!

HTH

Well, thanks to you tom i got to know where the problem is!
It was that stupid 100n cap coming out of the feedback pot. I misplaced it by one hole so it wasn't connected to the pot at all hahaha.

But now i have another problem (oops). There is ticking when i turn on the Modulation (LFO). How can i get rid of it?
Also the feedback pot oscillates early, say like after the pot at 11-12 o'clock.
Here is a sound sample
so you can hear the ticking and also hear the delay to see if it's normal for the repeats to be a little dark or not.

First 40 seconds modulation is off.
Also i think the mod rate pot isn't doing anything? Maybe i am not hearing it but it just changes the ticking.

[Mark Hammer]

LFOs that employ a dual op-amp, like the one in the Echo Base, start with a square wave and then convert it to triangle. That initial square wave makes a sudden draw on current that produces a spike on the power line.  It is a problem frequently raised here in a variety of circuits. It is MOST problematic when a circuit is battery-operated, but still an audible nuisance even when an external power supply is used.

There are essentially two "cures".  One is to use a low-current op-amp, like a TL022 or LM358 and possibly a TL062.  If the current the chip requires to do what it does is minimal, then spikes won't be audible.  The second "cure" is to decouple the chip from the overall supply.  That would involve cutting any V+ trace to the chip (pin and inserting a low value (e.g., 100R) resistor between V+ and pin 8, with a medium-value electrolytic (e.g. 10uf) from V+ to ground.  This acts like a secondary "battery", holding enough current to supply the initial square pulse.  Think of it like "found money" in an old jacket that means you don't have to go to the bank to buy a meal or beer.

[ElectricDruid]

The dark repeats is definitely normal. The filter on the output of the PT2399 is set at 1.9KHz, so that's pretty dark. But if you open it up a lot more, you get more noise as well as more treble, so it's a design decision.

+1 What Mark said about using a low current op-amp might help a bit with the LFO ticking, and a big electrolytic cap across the power pins of that particular op-amp might also help. Giving that op-amp separate wires back to the power supply input rather than using the breadboard power rails might also help. Stopping LFO ticking is a dark art, really, but decent PCB layout helps a lot, and those ideas are equivalent to what you'd do on a PCB but for a breadboard.

You could moderate the feedback by increasing the value of the 20K resistor that's attached to the feedback pot wiper. Try 33K or 47K and see if that doesn't settle it down a bit. The diodes to ground are supposed to limit the feedback, so check they're in the right place and connected correctly.

HTH

[Lost_soul]

LFOs that employ a dual op-amp, like the one in the Echo Base, start with a square wave and then convert it to triangle. That initial square wave makes a sudden draw on current that produces a spike on the power line.  It is a problem frequently raised here in a variety of circuits. It is MOST problematic when a circuit is battery-operated, but still an audible nuisance even when an external power supply is used.

There are essentially two "cures".  One is to use a low-current op-amp, like a TL022 or LM358 and possibly a TL062.  If the current the chip requires to do what it does is minimal, then spikes won't be audible.  The second "cure" is to decouple the chip from the overall supply.  That would involve cutting any V+ trace to the chip (pin and inserting a low value (e.g., 100R) resistor between V+ and pin 8, with a medium-value electrolytic (e.g. 10uf) from V+ to ground.  This acts like a secondary "battery", holding enough current to supply the initial square pulse.  Think of it like "found money" in an old jacket that means you don't have to go to the bank to buy a meal or beer.

Thank you Mark for the great info about LFOs. It's my first time dealing with one so i gotta get into dark magic as tom mentioned 

I will try a TL062 tomorrow and report back how it goes. As for the second cure that you mentioned. Isn't it already applied in the schematic?


There is a 220ohm resistor between 9v and LFO+ And there is a 100u cap to ground from 9v.
Is this the way you mentioned? Or is it something else.

[Lost_soul]

The dark repeats is definitely normal. The filter on the output of the PT2399 is set at 1.9KHz, so that's pretty dark. But if you open it up a lot more, you get more noise as well as more treble, so it's a design decision.

+1 What Mark said about using a low current op-amp might help a bit with the LFO ticking, and a big electrolytic cap across the power pins of that particular op-amp might also help. Giving that op-amp separate wires back to the power supply input rather than using the breadboard power rails might also help. Stopping LFO ticking is a dark art, really, but decent PCB layout helps a lot, and those ideas are equivalent to what you'd do on a PCB but for a breadboard.

You could moderate the feedback by increasing the value of the 20K resistor that's attached to the feedback pot wiper. Try 33K or 47K and see if that doesn't settle it down a bit. The diodes to ground are supposed to limit the feedback, so check they're in the right place and connected correctly.

HTH

The diodes are connected as they supposed to be. I will try to increase that resistor and see how it goes. Did i mention i am using a Linear pot instead of Log:icon_lol:
Maybe that has something to do with it.

As for the ticking. I will replace the TL072 with a TL062 As this is what i have on hand and see how it goes.
But for the other methods you mentioned, i am not quire sure i understand them.

"and a big electrolytic cap across the power pins of that particular op-amp might also help"
Do you mean between LFO+ And ground?

"Giving that op-amp separate wires back to the power supply input rather than using the breadboard power rails might also help"
But doesn't it already have pin 8 connected to LFO+ which is 9v across a 220o resistor and it has pin 4 connected to 220ohm to ground.

As for the LFO+ It has it's separate rail but for the ground, the 220ohm is connected to common ground.

[Matthew Sanford]

For the cap, they mean right on the pins of the dual op amp with least amount of leads. With a LM358 LFO similar issue I'd had, a 100n ceramic soldered across them under the chip cured it. They suggest a bigger one, not wrong, puts a local reservoir of current right at the LFO chips power input

[ElectricDruid]

Did i mention i am using a Linear pot instead of Log:icon_lol:
Maybe that has something to do with it.

That would certainly make the point at which it starts to oscillate happen sooner on the pot rotation.

"and a big electrolytic cap across the power pins of that particular op-amp might also help"
Do you mean between LFO+ And ground?

Yes. I hadn't noticed that they've done a separate filter 220R/47u supply for the LFO. Make sure the 220R goes from the power input to the LFO's +V supply pin, and make sure that 47u is from the LFO's +V supply to the LFO's -V supply - not connected to something that's connected to those points, but actually right on top of the op-amp from one pin to the other.

"Giving that op-amp separate wires back to the power supply input rather than using the breadboard power rails might also help"
But doesn't it already have pin 8 connected to LFO+ which is 9v across a 220o resistor and it has pin 4 connected to 220ohm to ground.

Yes, I see that now. They've already implemented a few of the tricks we've mentioned.
You can try connecting the +9V end of the 220R back to the point where the power comes into the breadboard, rather than just into the power rail. Similarly with the 220R into ground - connect the Ground end of it to where the Ground wire connects to the breadboard, not just into the ground rail.
You need any currents that are flowing to or from that LFO section to go straight back to the power supply without going near *any* other parts of the circuit, so you don't want them flowing along the breadboard power rails that everything else uses.

[Lost_soul]

You can try connecting the +9V end of the 220R back to the point where the power comes into the breadboard, rather than just into the power rail. Similarly with the 220R into ground - connect the Ground end of it to where the Ground wire connects to the breadboard, not just into the ground rail.
You need any currents that are flowing to or from that LFO section to go straight back to the power supply without going near *any* other parts of the circuit, so you don't want them flowing along the breadboard power rails that everything else uses.

Well i did exactly what you said.

1st thing i connected the 47u cap from pin 8 to pin 4 going right above the IC. (However idk if you meant to connect the -ve side of the cap right at pin 4 or to the ground rail)

2nd thing i used a TL062

3rd thing i connected both the 220r resistors at where 9v and ground connects and not the rails of the breadboard.

In this image i was yet to connect the 220r resistors to the power supply points.





After all that the ticking is still there
Are there any other solutions.

[Lost_soul]

For the cap, they mean right on the pins of the dual op amp with least amount of leads. With a LM358 LFO similar issue I'd had, a 100n ceramic soldered across them under the chip cured it. They suggest a bigger one, not wrong, puts a local reservoir of current right at the LFO chips power input

There is a 47uf cap like that in the schematic. However it didn't fix the ticking unfortunatley.

[Matthew Sanford]

Odd thing to me is the uneven ticking. There is a bunch else on the LFO side going to ground, perhaps if all of that was through the same ground path then straight to where the power supply ground comes to the board then you could avoid mixing with the audio ground?

Plus a bit odd it doesn't seem to influence the npn for pitchy modulation, can you read the triangle voltage change all the way through the pot to the npn base?

[Lost_soul]

Odd thing to me is the uneven ticking. There is a bunch else on the LFO side going to ground, perhaps if all of that was through the same ground path then straight to where the power supply ground comes to the board then you could avoid mixing with the audio ground?

Plus a bit odd it doesn't seem to influence the npn for pitchy modulation, can you read the triangle voltage change all the way through the pot to the npn base?

I don't quite understand what you mean.
Do you mean in the audio sample i sent? The ticking changes at the end as i adjusted the mod speed pot.

Also could you elaborate the following: "can you read the triangle voltage change all the way through the pot to the npn base?"

[Matthew Sanford]

Odd thing to me is the uneven ticking. There is a bunch else on the LFO side going to ground, perhaps if all of that was through the same ground path then straight to where the power supply ground comes to the board then you could avoid mixing with the audio ground?

Plus a bit odd it doesn't seem to influence the npn for pitchy modulation, can you read the triangle voltage change all the way through the pot to the npn base?

I don't quite understand what you mean.
Do you mean in the audio sample i sent? The ticking changes at the end as i adjusted the mod speed pot.

Also could you elaborate the following: "can you read the triangle voltage change all the way through the pot to the npn base?"

Yes, that sample, the tick sort of swings the tempo instead of a straight rhythm.

On reading, I mean using the multimeter to read the voltage range on pin 1 of UA1 and pin 8 of 4066 (should be fine as LED shows it), then top of mod pot, and both sides of the 220R to make sure it's getting to the base of the mod npn, since it isn't modulating.

On ground I just meant a dedicated ground rail on the breadboard only for LFO grounds, then take it through a wire to where ground comes in from the power supply - I think that is akin to Druid advising to get power directly from source rather than power rails to separate both sides all the way to the power input. You already have a 220R to separate LFO+ (maybe a small 47R could help on the audio V+ side?), but the ground may mix through various ground returns prior to the ground in, possibly modulating the audio power to create ticks. I'm spitballing a bit, but it has to be mixing in somewhere.

[m4268588]

https://postimg.cc/hQqH7ZkJ
Will the LFO-tick be reduced when this capacitor is removed?

[ElectricDruid]

https://postimg.cc/hQqH7ZkJ
Will the LFO-tick be reduced when this capacitor is removed?

I'd guess it'd get worse, not better. That cap is stuck on the output of the schmitt trigger to try to slow down the switching speed. Whether that works or not isn't clear to me.

Taking it out is a simple thing to try, so it's maybe worth an experiment.

[Lost_soul]

Well, it was that 2.2uf cap. Once i removed it from the ground rail and connected it to the point that i made for only LFO ground coming directly from the PS the ticking was gone!

I wonder if there would be ticking again when i make it on veroboard based on sabrotone layout...

Also there are a couple things i don't understand.

First, the pots lugs numbering on the schematic for the time and mod rate. What i made on the breadboard for the time is connecting lugs 2&3 to pin 6 of pt2399 and lug 1 to the collector of the pnp. And it's working as it should. Time increases when adjusted clockwise.

As for the mod rate i connected lugs 2&3 to the 27k and lug 1 to pin 7 of the LFO chip. But this way, the mod rate increases when adjusted COUNTER clock wise.

But on sabro layout he did the opposite!

[Lost_soul]

On reading, I mean using the multimeter to read the voltage range on pin 1 of UA1 and pin 8 of 4066 (should be fine as LED shows it), then top of mod pot, and both sides of the 220R to make sure it's getting to the base of the mod npn, since it isn't modulating.

I measured the voltages on the points you said and they were all changing constantly. I fixed the ticking and all seem to be working rn. I hope it doesn't tick when building it on vero

Thank you Matthew for your help

[ElectricDruid]

Good work!

I think you'll be fine on vero, but the same rules apply, and if the vero layout doesn't follow them *change it* or the ticking will be back. You've got a good understanding of what the problems are now and how you work around them and that's about all any of us have.

For the pots, if the rotation goes the wrong way swap lugs 1 and 3, the outside ends. It's a very easy thing to get the wrong way around and it happens all the time, so I'm not surprised if you've found a layout where it's wrong. At least with wired pots it's very easy to fix. On a PCB layout, not so much! Another board revision required for a dumb mistake...been there, done that..

[Lost_soul]

Good work!

I think you'll be fine on vero, but the same rules apply, and if the vero layout doesn't follow them *change it* or the ticking will be back. You've got a good understanding of what the problems are now and how you work around them and that's about all any of us have.

For the pots, if the rotation goes the wrong way swap lugs 1 and 3, the outside ends. It's a very easy thing to get the wrong way around and it happens all the time, so I'm not surprised if you've found a layout where it's wrong. At least with wired pots it's very easy to fix. On a PCB layout, not so much! Another board revision required for a dumb mistake...been there, done that..

thank you tom for all the knowledge you share!
i will build the board and if there was ticking i would then try the solutions that i learned one by one and see which of them would make it go shhhhhh.

i can't even imagine making that mistake on a pcb. i would be angry all day long