ROG Tri-Vibe Hiss/Whoosh

[Dasher]

In 2013, I built a Runoff Groove Tri-Vibe on a home-etched PCB using the Andre Schaap / John Lyons layout.

Schematic: http://www.runoffgroove.com/tri-vibe.png
Layout: http://www.runoffgroove.com/tv-pcb.pdf

The vibrato works as expected and I've enjoyed using it the past 3 years. I currently have some free time on my hands, so I decided to look into the one issue with the build that has always slightly annoyed me. There is a noticeable level of hiss generated from the effect, even at relatively low bedroom volume levels. The hiss is in time with the LFO, giving a wind-like, whooshing sound. For testing purposes, I'm running the Tri-Vibe alone directly into the input of my amp: Guitar -> Tri-Vibe -> Amp.

Below is a list of what I've tried so far:
1. I've checked components values and copper traces several times and haven't noticed anything wrong. I could have missed something, of course!
2. The hiss is there when using a battery or power supply.
3. I've tried replacing all ICs with no improvement in noise level.
4. I've used an audio probe to determine where the hiss originates and have found that it starts immediately following the 1n5 cap in the vibrato signal path (between the two LM13600 stages).
5. My IC voltages are:

NE5532
1 – 4.47
2 – 4.46
3 – 4.06
4 – 0
5 – 4.71
6 – 4.71
7 – 4.72
8 – 9.44

TL064
1 – 4.71
2 – 4.72
3 – 4.69
4 – 9.44
5 – 4.70 oscillates
6 – 4.71 oscillates
7 – 3.21 oscillates
8 – 4.70 oscillates
9 – 4.71
10 – 4.71
11 – 0
12 – 4.71 oscillates
13 – 4.66
14 – 4.70 oscillates

LM13600
1 – 1.26 oscillates
2 – 5.60
3 – 4.94
4 – 4.94
5 – 5.72 oscillates
6 – 0
7 – 5.72 oscillates
8 – 4.53
9 – 5.37
10 – 6.56 oscillates
11 – 9.44
12 – 6.55 oscillates
13 – 4.94
14 – 4.94
15 – 5.60
16 – 1.26 oscillates

Does anyone else who has built a Tri-Vibe using the Andre Schaap / John Lyons layout experience this type of noise? Google/forum searches have brought up varying reports on the noise level of the effect. Some say there is hiss, while others say it's essentially noiseless. If the level of hiss I'm experiencing sounds unusual, any ideas as to what can be done to reduce it?

[GibsonGM]

I haven't built it, but can say that if you look at the schem, you see what appears to be an internal Darlington pair at the output of the 13600.  This configuration offers a huge amount of gain.  This may be greatly amplifying noise that is already existing prior to the output!  The 1n5 cap you mean is the one coming from the emitter, correct?   

So, in the interest of covering a base you might be missing...good layout and design practices are important. That means using short leads, shielded input and output lines grounded at one end only, star grounding...do all your grounds lead to one spot on the PCB, and from there to battery "-"?   And leads running over the chips, or crossing each other (including component leads)?  How are your pot leads laid out?

How about trying a 1n cap from input to ground, right at the jack (on the off chance it's RF causing some kind of oscillation -doubtful it's that tho)?

Can you post a pic of the top & bottom of your board?

[Dasher]

Thank you for the reply.

To clarify where I first get hiss when using an audio probe, I've attached a modified image of the schematic. I renumbered the LM13600 pins to account for the fact that the chip is reversed in the layout I used.



Yes, the 1n5 cap I'm referring to is the one coming from the emitter of the first Darlington pair (Pin 9 on the modified schematic). The other end of the cap connects to pins 5 and 7. Perhaps using shielded input wires would help, but if the hiss I'm getting were a result of existing noise being amplified by the first LM13600 stage, I would also hear the noise when audio probing Pin 9 (on my modified schematic). Correct? Instead, I don't hear the hiss until pins 5 and 7. Either way, I may end up replacing the input with shielded wire to eliminate that possibility. 

I don't think the hiss is a result of oscillation. It sounds the same as thermal noise that we can get from resistors. But again, I'll try a 1nf cap from input to ground tonight to cover all possibilities.

I've attached images of the top and bottom of my board. It was a bit hard to capture everything in detail, but hopefully they give an idea of what I'm working with.

[Pedalhead]

Dunno if this will be of any help, but I recently built a Tri-Vibe and found it a bit trebly and glassy sounding. I upped the input cap to 100n and put a cap to ground on the output. 10n was ok but still a bit trebly. I next tried a 56n which instantly gave me a ton of loud hiss. No idea why (?). Next tried a 15n and no hiss for a couple of seconds then it slowly crept up. Suffice to say I went back to 10n. Peculiar anyways.

[blackieNYC]

I had mine on the breadboard until a week ago. Two in fact. Stock values. Except the audio path is all TL072s. I didn't get any hiss. So I think something's wrong with your circuit.  Wish I had better help to offer.

[Dasher]

Thanks to everyone for their assistance and comments so far. I was able to try a few suggested ideas tonight, unfortunately without any success.

First, I placed a 1nf cap from effect input to ground - this had no effect on the hiss. Next, I tried a 10nf cap from effect output to ground - this also had no effect on the hiss.

My gut feeling agrees with blackie; there's something wrong with the circuit as I've built it. I'll go over all my copper traces and check component values another night.

[DrAlx]

The two OTAs are configured as all-pass filters and the relative order of them should not matter. Therefore it should make no difference which one is using the 22nF and which is using the 1n5.  I would swap the 22nF with the 1n5.  i.e. make first OTA use 1n5 and second one use 22nF, in order to see if the noise moves from pins 5,7 to pins 10,12.  If it does then it suggests its just part of the design and that nothing is actually wrong. If the noise stays on pins 5,7 then it suggests  either a problematic IC or a layout/build problem. I have a trivibe of my own layout and I don't consider it especially noisy.  There is some swish noise but I'd expect that.  See if you can post soundclips.

EDIT: Just audio probed my trivibe. I definitely hear modulation on the first OTA using 22nF (which in your build corresponds to pins 10,12.)  So I am wondering if there other problems in your build.

ANOTHER EDIT: I'm not sure its sensible to use the audio probe on pins 5,7,10,12.  For this circuit those pins are effectively acting as inputs to the OTA, and the act of probing could inject noise.  I think it's safer to just measure at pins 8 and 9 since these are effectively buffered points.

I just audio probed my build again. I get the same thing as you. First OTA with 22nF is quiet but second with 1n5 is hissy. However the hiss is gone if I probe the final circuit output. I think the deemphasis employed on the final output opamp kills the noise. Check solder connection on the 4n7 and 15k on the output opamp. You may have bad connection.

[DrAlx]

Looking at that resistor in the photo its hard to tell if its 15k or something lower.  What should be a red band looks quite dark to me. Maybe just the photo?

[Dasher]

Thanks for breaking out the audio probe, DrAlx!

EDIT: Just audio probed my trivibe. I definitely hear modulation on the first OTA using 22nF (which in your build corresponds to pins 10,12.)  So I am wondering if there other problems in your build.

I re-audio probed Pins 9, 10 and 12 in my build and I do hear modulation. My mistake. I've updated the schematic posted above to reflect this.

I just audio probed my build again. I get the same thing as you. First OTA with 22nF is quiet but second with 1n5 is hissy. However the hiss is gone if I probe the final circuit output. I think the deemphasis employed on the final output opamp kills the noise. Check solder connection on the 4n7 and 15k on the output opamp. You may have bad connection.

I also re-audio probed Pin 8 of the LM13600 and compared the level of hiss here to the circuit output. The circuit output is definitely quieter. Seems the deemphasis is working... but just to be sure, I'll recheck all components and connections in that area.

Looking at that resistor in the photo its hard to tell if its 15k or something lower.  What should be a red band looks quite dark to me. Maybe just the photo?

Good eye, but just the photo. It is a 15k resistor, indeed.

Perhaps I'm just being too critical about the level of noise coming out of the effect? It's certainly usable and the hiss isn't noticeable when playing. Unfortunately, I don't have a way of uploading video or audio. One observation I've made since removing the PCB from the enclosure for debugging is that the hiss isn't as prominent now as it's being masked by all sorts of other noises. BlackieNYC, maybe that's why you didn't notice any hiss while it was on the breadboard?

[DrAlx]

My guess is you might be being too critical. OTAs are known to be relatively noisy components. The trivibe was the first modulation i built. Prior to that i only had distortions/fuzz/overdrives which all had noticeably lower noise levels.

[antonis]

I see carbon resistors around the 13600 chip. If one of them is the 10k which connects pin 1 with the LFO, maybe your problem is here..

[Dasher]

I rechecked components last night and everything looks correct.

I've built a few modulation pedals now (small clone, little angle, magnavibe and tri-vibe) and the tri-vibe is certainly the noisiest of the bunch. Unless somebody mentions that their Tri-Vibe is especially quiet over the next few days, I think I'll box the effect back up and continue to live with the hiss.

Is there any chance that noise is created by having the TL064 set both the LFO and VREF, rather than having two TL062s as in the schematic?

I see carbon resistors around the 13600 chip. If one of them is the 10k which connects pin 1 with the LFO, maybe your problem is here..

Thanks for the suggestion antonis. The 10k resistors from LFO to Pins 1 and 16 are both metal film in my build.

[duck_arse]

the tl064 is not in the audio path, so should not be contributing to hiss. do you have a more modern opamp you could try for the 5532?

[Dasher]

the tl064 is not in the audio path, so should not be contributing to hiss. do you have a more modern opamp you could try for the 5532?

I tried a tl072, tlc2272 and 4558. I'm not sure if any of those are more modern. Either way, no change in the hiss. As was mentioned earlier, it's probably coming from the lm13600/Ota which are known to be relatively noisy.

[Mark Hammer]

One possible strategy is to split the 47k input resistor going to pin 6 on the output mixer from the OTA stages.  Find some combination of resistance values that add up to 47k (or close enough) and run a suitable cap to ground from their junction.  So the wet/dry balance will be preserved, but the wet signal will have the hiss trimmed back.

For example, replace the 47k with 22k and 24k in series, and run a 1500pf cap from their junction to ground.  That will provide a 6db/oct rolloff starting around 4.8khz, which should reduce audible hiss somewhat.

If you want to get really nuts about it, run a pair of 10k in series, followed by a 27k, and run 3300pf caps to ground from the two junctions for a 12db/oct rolloff at the same frequency.

[Dasher]

Thanks Mark. I tried a 22k/24k series pair with a 1.5nf cap from the junction to ground. I didn't notice a change in the level of the noise. I think I've been mistakingly referring to the noise as hiss. I don't have a scope (or a particularly good ear apparently!) but the noise is probably more prominent in the midrange - low treble frequencies.

I'm going to box the vibe back up and continue to live with it as is. I've got another project I want to get going on. Thanks to everyone for their help!

[roseblood11]

Did you solve the problem? I get the same hiss, and used the same layout

[EBK]

Did you solve the problem? I get the same hiss, and used the same layout

May seem strange, but I would suggest breaking that shared ground between pin 6 of the 13600 and the 4k7 resistor and regrounding pin 6 directly.  Also, where are the decoupling caps?

[skumberg]

Had the same noise experience with a fuzzdog pcb. Used jrc13600 NE5532 and tl062 as recommended. I lent it to a friend and he loved the sound and didn't mind the swoosh. It sounds a bit like a dish washer as it swoosh with the lfo rate.
Also used it before a fuzz face first and a huge constant "white noise" was produced. It was ok after the fuzz face, a klon or tube screamer didn't mind the effect order.