Squealing 8W SE

[marquimsp]

Hello, how are you all doing?
I have this mini amp, it's 8 watts (1xEL84 and 2x12ax7) with a 1x12 (eminence GB128) and I love how it sounds, but it emits a high-pitched squeal when it's getting close to max gain (with max MV).
I've tried getting a schematic from the former builder but he won't give me one (copyright). Maybe the components are too close to each other? I'd appreciate any suggestions on how to mitigate the squeal.
I've changed all the tubes already (tried different brands, the squeal persists, but the sound got much better with mullards and electroharmonix).

Here's a picture of the board:


And here is a video of what the squealing (unconnected to a guitar):


Connected to a guitar:


How it sounds:


A view from behind:

[xtian]

Any wire runs are suspicious when runaway oscillation is an issue. For example, I see kind of loopy wire runs from your input jack, and from the OT. Try moving these wires around with a chopstick or other non conductive tool, and see if it modifies the noise at all. Every loose wire in your build can act as an antenna and cause oscillation at high gain.

[marquimsp]

Hey, thanks for your reply!
I did what you said, moved those wires around with a chopstick while squeal was going on, nothing changed (intensity, pitch etc). But I found a VERY microphonic capacitor near the gain pot. Could that be it? Or is it normal?

Thanks again for your help!

[bepone]

i dont like design with this pcb. nobody is respecting any rules, like clearance for the start ?

try to see if the secondary is grounded ,if isnt then would squeal..
if gnd is ok , then too much gain with the certain layout can create oscillation, cure is to reduce gain between the stages, (change ecc83 for tube with lower gain, ecc81) and add grid stopper of 100k+ on the stage after the gain pot

[bepone]

and, from one El84 it is possible to get 5W not 8

[marquimsp]

i dont like design with this pcb. nobody is respecting any rules, like clearance for the start ?

try to see if the secondary is grounded ,if isnt then would squeal..
if grounded is ok , then too.much gain with the layout can create oscillation, cure is to reduce gain between the stages, (change ecc83 for tube with lower gain, ecc81) and add grid stopper of 100k+ on the stage after the gain pot

I don't know much about electronics, but I totally agree, it's way too much gain... I'll try the ecc81! If it's not too much to ask, how do I check if the secondary is grounded, and where exactly do I add a grid stopper? lol...
If you think it's way above my head it's ok, I'll just replace the tube.

Thanks for the very enlightening reply! I'm learning a lot.

[marquimsp]

and, from one El84 it is possible to get 5W not 8

Lol those lying bastards deserve to be out of business... Just kidding, 5W is loud enough!

[roberto]

Commercially they declare 8W because they just declare it at higher THD, where it is usually played: 5W at 5% THD and 8W at EG 15%.

[bepone]

this is poor EL84 not EL34. i think 8w is not possible not even with 100:% distortion / fuzz

[bepone]

I don't know much about electronics, but I totally agree, it's way too much gain... I'll try the ecc81! If it's not too much to ask, how do I check if the secondary is grounded, and where exactly do I add a grid stopper? lol...
If you think it's way above my head it's ok, I'll just replace the tube.

Thanks for the very enlightening reply! I'm learning a lot.

when the amp is off, check continuity between speaker wire and chassis,
for those simple amps, gain is relative, it is not best at full scale, after one moment starting to fall apart..raspy notes.. blocking the tone ..so to avoid this remove the unnecessary gain with lower gain tubes like ecc81/12AT7, you can find it more warm sounding (nos, not expensive, )

[marquimsp]

I don't know much about electronics, but I totally agree, it's way too much gain... I'll try the ecc81! If it's not too much to ask, how do I check if the secondary is grounded, and where exactly do I add a grid stopper? lol...
If you think it's way above my head it's ok, I'll just replace the tube.

Thanks for the very enlightening reply! I'm learning a lot.

when the amp is off, check continuity between speaker wire and chassis,
for those simple amps, gain is relative, it is not best at full scale, after one moment starting to fall apart..raspy notes.. blocking the tone ..so to avoid this remove the unnecessary gain with lower gain tubes like ecc81/12AT7, you can find it more warm sounding (nos, not expensive, )

Thanks again! I checked and there IS continuity between speaker cable and chassis phew! I'll order an ecc81 (not cheap here, around 24 USD)

[roberto]

this is poor EL84 not EL34. i think 8w is not possible not even with 100:% distortion / fuzz

Le'ts say it will reach 8W inbetween the 15% I told before (it will be more!) and 100%:
Considering 260-270V B+ on 5k Ra, swinging between 35-40V and 500V, you'll get around 5,5 W "quite" clean for an instrument amp (that's 163 Vrms squared and divided by the load).
Square waves will get 230 Vrms squared and divided by the load, that is something less than 11 W.
8 W is in between.

[bepone]

roberto, not even in theory, 5W is allready distorted.
but calculate from here.. let me know your result

/you can see i bias the tube on 100%, this is allready too much, but maybe you can extract 0.5W more /

EL84.png

[Stevem]

The first thing I would do would be to use shielded wire to go from the input jack to the board.

The shield should be grounded / connected only at the input jack.
On the other end it should be Heat shrinked off.

In the tight confines of the lunch box type Orange amps that I have kicked the gain up on for people I have had to make a card board and Aluminum foil shield to place around the whole input jack to get the amp to stop back talking at high volumes!

When a input jack / antenna and a speaker jack are within 5 inches of each other in a high gain amp, these are the things that can happen!

[roberto]

Hi Bepone,

290 V is not optimal for a 5k load with fixed bias as you can see, because it's strongly asymmetrical.
On top of that, you are setting the plate voltage at 290 V and the grids at 300 V.
Try with the data I've indicated before, screen voltage I'd say around 220 V.
As I says, it's "clean" for an instrument amp, it's not 0,000001% THD.