BB Preamp (Madbean Glitterrattii) oscillates at high gain

[bendeane]

Built a stripboard version and sounded great. Had a PCB fabricated using the Madbean glitterrattii schematic (this the schematic that the stripboard version I used is said to be based off of) and got it populated. It squeals something fierce at high gain settings. Never got that out of the veroboard. I noticed one resistor change from the schematic that is different from the stripboard version I used. R16 is 47k in the schematic and 16k on my stripboard. Would this alleviate the squeal?

It was recommended on tagboard effects to increase the clipping amp's 51pf cap to 220pf. I don't have any on hand so I wired a 47pf and a 151pf in series to get me close (I used a 47pf in both the stripboard and my schematic versions).

I also don't seem to be getting the same volume out of the PCB version (the PCB version and SCHEMATIC version I referenced are the same thing; I might've flipped those terms throughout the post). I've also run out of 5088 trannies and subbed in a couple 4401 transistors in my schematic version. The hfe out of the 4401's is only slightly lower, but assumed since they were being used as input and output buffers that this wouldn't make much of an audible difference.

Anyway, just thought I'd try to get a little input. It sounds great aside from the lower volume and the squeal at max treble and gain. I put a switch in to go between the standard silicon clipping diodes to some germanium ones and then none in the middle position. I have a thing for the harsh OPAMP clipping with no diodes but the switch between the GEs and the silicon diodes gives this thing 3 distinct flavors.

Any advice?

[mth5044]

Caps in series aren't additive like resistors, you will want to put them in parallel to make that work.

Sounds like a problem with the layout. How are your layout skills?

[bendeane]

Correction, I did wire the caps in parallel. Mistyped that. Layout skills aren't awful, but still pretty green.

[Rob Strand]

If the squeal is affect̀ed by the output volume, try keeping the input and output wires away from each other.  Also I would be expecting your foot switch wiring has input and output wires close together.

If the squeal is affect by the treble control then it could related to the above case.  If the treble control has an effect but the output volume does not then I would be looking at your tone control pot wiring.  Keep it away from the input wiring and the gain pot wiring.

Beyond this you should check your DC voltages on the opamps and transistors.   If these are wrong you will need to starting looking at point out and wiring problems on the PCB.

[bendeane]

Squeal is present when some combo of gain, treble and volume get too high. Gain and volume at max only produces squeal when treble gets too high. Volume and treble maxed only produces squeal when gain starts to get too high, etc.

My in and outs are on opposite ends of the board in all cases and pots are all at the top of the board but I'll check the the proximity of the treble and gain pots as well as my routes.

Any thoughts on the reduced volume issue? The volume doesn't seem to be as high as my stripboard version. As a clean boost (gain at minimum) I get unity at about 1 o'clock. Was getting unity at about 11 o'clock on the stripboard.

[Rob Strand]

Re gain:

- A larger R16 value should give you more gain.  If I understand you correctly this is opposite to what you are seeing.

- The pot tapers may be different on the two builds.

- If the circuit is oscillating supersonically it may be reducing the gain. ie. the circuit is oscillating but you cant hear it.

Re oscillation:
One thing worth trying is to add a cap from the tone control opamp output to the tone control opamp -ve input.
Maybe start at 47pF and increase in steps 100p, 220p, 470p.

If the circuit works at least you might be able to tweak the value to remove the oscillation but not affect the sound.

[bendeane]

Okay...thanks a ton for the advice...I think my PCB layout needs tweaked, though...

I may try to hack my way thru the 3 I have and tweak some values, try the cap on the tone control OPAMP, but I have my gain pot pads too close to my treble pot pads, I think. 

My pot tapers are all the same from the veroboard to the PCB version.  I used Logarithmic for the Gain and Volume and Linear for the Bass and Treble on both builds.  The only component difference is the R16 (47K on my PCB and 16K on my stripboard) and the two trannies (4401 on my PCB and 5088 on my veroboard).

As a side note, I'm so grateful for the input I get from all of you...learning so much thru that and a little trial and error...a year ago, I was having trouble getting some of the simpler stripboard layouts I was finding to work.  So much fun and potential with this...

[R.G.]

There are some standard things to check when you have oscillation.
1. Power supply decoupling; assume that you need one 0.01 to 0.1uF ceramic cap per opamp until proved differently. Some layouts get away without it, some don't. You could try soldering a 0.1uF between the +V and -V pins of your opamps on the bottom/pins side. This helps with but does not completely make up for problems with power supply and ground routing.
2. Reference voltage decoupling; any noise that does get onto your reference is shuttled to every opamp's + input. Controls floated on Vref send current directly into Vref. Sometimes this is OK, sometimes it bites you.
3. Wire routing; as noted, keep inputs and outputs far away from each other, and especially keep input wires going to high impedance inputs (like the + input of a JFET opamp) well away from other wires to eliminate capacitive pickup.
4. Ground wiring; don't mix up reference grounds, power grounds, and sewer grounds. Generally you get away with ignoring this on small, low power boards, but sometimes it bites you.

These are design issues. Implementation issues like poor soldering, wrong component value or orientation, mistaken wiring, and so on are always with you.

[bendeane]

I feel honored and humbled and in awe to get a response from the revered R.G. Keen.   

[PRR]

> honored and humbled

Never mind that.

Do you understand that ANY high-gain circuit WILL oscillate if the output signal sneaks-back to the input?

And that build-errors happen despite our best efforts?

(I just put an LED headlight on the snowblower, and have to find my latest build-error.)

[bendeane]

>

Never mind that.

Do you understand that ANY high-gain circuit WILL oscillate if the output signal sneaks-back to the input?

And that build-errors happen despite our best efforts?

(I just put an LED headlight on the snowblower, and have to find my latest build-error.)

Yeah, I've had oscillation issues before, albeit few cause I'm still pretty new to all of this and learning new things with each build. Just grateful for all the kind input and advice around here from everyone.

I used to be in the business of large format printing and signage. When I first got my feet wet and was seeking guidance from a forum, many (not all) of the members were not so kind to welcome newcomers. I find the pedal building community to be a very different place than that.

[bendeane]

Ok, so spent some time with both versions. I was wrong, the veroboard is part for part with the PCB version I laid out. Definitely layout related but I was able to mess with R16 and bring it down enough to get rid of oscillation. It definitely affects output volume but it's usable.

Put a switch on the diodes for LED vs silicon (it's an ON-OFF-ON so middle position is no diodes). Loving the LEDs for clipping. Might try mosfet. The germanium diodes didn't do much for me.

Gonna study other layouts and be aware of all of your guys' comments and advice with my next PCB layout. Thanks for the help. 

[bendeane]

So after dumping the PCB and building a veroboard version of this (that sounds spectacular, BTW), I was reevaluating my PCB and noticed something...C5 (220n going to ground from the output of the first clipping stage) was not routed correctly.  It wasn't going to ground.  It isn't going anywhere.  Would this be the source of oscillation?

I essentially removed the capacitor altogether from the circuit by doing this and the RC low-pass filter isn't an RC filter at all.  If I understand R.G.'s article correctly, if I get that C5 grounded correctly, I'd get a 26db decrease in signal around 14KHz.  20db decrease in signal around 7230Hz.  If I get some time tonight, I'll test these assumptions, but if anybody that has time to chime in before that, it would be great!

[Mattnezz]

With C5 open the circuit might sound a bit trebly and harsh too
But yes, I think grounding C5 might solve your oscillation problems since it indeed cuts a lot of high frequencies.