Hummingbird - problem or normal?

[paul.creedy]

I recently finished building a Hummingbird clone, based on this layout:



It works nicely, but along with the beat of the note I'm getting an extra sound from time to time. It's more noticeable with lower notes and (not surprisingly) with the Depth turned up. With the rate quite high and the switch set to double speed it sounds like a crackle because it doesn't do it on every beat. I noticed it when swapping transistors (I went back to the 5089) to try and get a bit more volume as it barely reaches unity with the volume right up.

I tried lowering the input volume but no change.

I only bought one of each of the 2N2646 and 5457 as they were about £3 each, but I could always get more to try if that's likely to be the cause.

Any suggestions gratefully received.

[Elijah-Baley]

It seems a peculiar tremolo/pulser effect. I'm not sure I got what's your problem.
Did you compare it with this video? https://www.youtube.com/watch?v=HJYGpfgf6YY

[paul.creedy]

I'll see if I can record it tomorrow.

[paul.creedy]

Here's a quick recording which hopefully makes things clear.

You can hear that the first few notes sound ok, but the last two in particular the noise can be heard as the note dies away as a popping/crackling sort of sound.

http://www.arrowheadguitars.co.uk/pics/cracklebird.mp3

[paul.creedy]

I decided that without being able to try different 2N2646s and 2N5457s I could be attempting the impossible, so yesterday went in search of some.

When I bought the first two (one of each), they cost me £3.09 per component, £6.18 in total.

Yesterday I found 10 2N2646s for £10.99 and 50 (!) 2N5457s for £5.81  so hopefully I can get a better idea of what may be the cause of the noise.

[paul.creedy]

Regardless of the lack of responses I'm going to continue updating.

I tried a couple of new 2N2646s without much in the way of change, though wiggling them in their sockets did briefly make the switch noise happen repeatedly, which was better than it happening intermittently. Unfortunately when the pedal was returned to the board it was no different, so the search continues.

I'm still waiting on the 2N5457s to arrive, but in the meantime I'm considering swapping the volume control from the input side to the output side, mostly to see what happens

Looking at the layout, the first thing the signal from the level pot reaches is a 22k resistor, might a capacitor before (or after) it do anything? (clutching at straws or showing my lack of knowledge perhaps ? )

It's nice when mixed with some of my other builds, the Rattle Crow in particular, so I shall continue playing with it in the hope of some sort of resolution.

[duck_arse]

you should know by now how people complain when the circuit diagram is not included in thread. should. also when we have to ask to see build photos.

[paul.creedy]

you should know by now how people complain when the circuit diagram is not included in thread. should. also when we have to ask to see build photos.

Apologies - I thought the layout would be sufficient, I'll see if there's a circuit diagram available. From what I've read it's based on a Vox Repeater.

I'll take an interior photo when I've got a minute, but I'll reiterate that it works as expected apart from the issue of the slight glitch in the sound which (hopefully) can be heard clearly in the sound clip.

[nocentelli]

Regardless of the lack of responses I'm going to continue updating.

I'm familiar with the circuit (Vox repeat percussion but JFET rather than BJT audio stage) but missed this thread.

Schematic is here - http://effectslayouts.blogspot.com/2015/10/earthquaker-devices-hummingbird.html

Nice audio recording: I suspect the JFET spec could be the problem since a 22k drain resistor will only bias JFETs that fall within a range of physical parameters. Check the drain voltage out: If it is not around 4-6v, that might be the issue.

[paul.creedy]

Thanks for that

I'll take some measurements tomorrow.

[bluebunny]

Apologies - I thought the layout would be sufficient

Showing us only the layout is like showing us the cake and expecting us to know the recipe.   

[paul.creedy]

I have applied multimeter probes to various points (apologies if I've not taken a measurement at the right place)



Incoming from the power supply it was plugged into was 8.81v - the pedal was switched on but didn't have anything plugged in.

After R1 (100R) it was 8.66, after the diode it was 8.02 and after R5 (22K) it was 8.38v

I measured the voltage of the 2N2646 as follows:

B2-6.74
E-3.5
B1-0.06

Let me know if I need to check anything else.

[PRR]

> showing us the cake and expecting us to know the recipe.

OK, I thot there should be a meme for that. There isn't. And the closest hit I found reminds us all that not everybody reads "schematic" (or Urdu).

[paul.creedy]

A supplementary question (in English rather than Urdu)

Moosapotamus' Skippy Tremolo is also based on the Vox Repeater, and has an option to use a 2N6027 instead of the 2N2646 (with a resistor) - could I try the same with this ?

[paul.creedy]

One further piece of information, after running the pedal on its own through another amp, the problem is only apparent when the speed is doubled, with the switch set to the slower speed I'm not hearing any extra interference.

So it's when the signal passes through the 1u, not the 4u7 - which is polarised - is it worth trying a polarised 1uF?.

Either that or a different 1uF cap?

[paul.creedy]

. . . and here's a pic of the inside.

[duck_arse]

After R1 (100R) it was 8.66, after the diode it was 8.02 and after R5 (22K) it was 8.38v

yer circuit diagram show the 100R and the diode very firmly connected to each other, yet you report a voltage difference. something, circuit or writing, isn't quite right there.

and as for the 6027 sub - a qualified yes, will sub. I thought there was a recent discussion on the PUT's and UJT's operation and substitutions .......

and for your 1uF cap - if you have a not electro in there, test it for goodness, and then leave it.

[bluebunny]

yer circuit diagram show the 100R and the diode very firmly connected to each other, yet you report a voltage difference. something, circuit or writing, isn't quite right there.

The schematic and IvIark's layout don't match exactly in this area.  He has a series diode, not a shunt.  And this only feeds the 470R, not the 22K. 

[duck_arse]

yer circuit diagram show the 100R and the diode very firmly connected to each other, yet you report a voltage difference. something, circuit or writing, isn't quite right there.

The schematic and IvIark's layout don't match exactly in this area.  He has a series diode, not a shunt.  And this only feeds the 470R, not the 22K. 

excellent. see reply number six. carry on!

[kaycee]

I tried the hummingbird one and I don't think I got it to work properly. The skippy, however is a reliable build with the easily available BJTs instead of the jfet. The 2027 is readily available from bitsbox or on eBay. The 20k trim lets you tune in the variables of the oscillator and adjust the broad range of speeds.

On a tip from RG, run the ground of the speed capacitors direct to ground, not through the board to help limit ticking. Also add a second power filtering cap from the junction of the 470r and the 1k. For added fun an LED straddling the 33r makes a rate indicator.