Noise from DIY Amp

[Keryn O'Shea]

G'day gents,
An amp l threw together has always had a slight noise once both inputs are connected to the DSP, it sounds like a mosquito. With one input plugged in, its silent, with both, that's when you can hear it. Used with the DR280s, the noise is just noticeable, however l plugged SLAPs in yesterday and the noise is very noticeable.

After trying to diagnose, (insulating leads, lifting grounds, changing power points, leads, amp, ect) l found using a second power source eliminated the noise.

I'm aiming to use the SLAPs as DJ monitors in four weeks, and when you're that close, the noise is annoying! Apart from running a second external power source, is there something which l can try? I'm hoping a magical resistor in the right place does it!

[Tom Smit]

I wish I had an answer for you.

Side note: I got confused when I saw that the white speaker wires were on the positve terminal of the board. Here in NA, white goes to negative, and black to positive. To add some confusion, on NA commercial highway trailers, white is the ground, yet, on GM vehicles (back in the day) it was the black that was negative.

[Keryn O'Shea]

I wish I had an answer for you.

Thanks Tom
I'm wondering if the noise is more prominent with the SLAP as it doesn't have a woofer LP filter? I'll test with an OT8 tomorrow.

[Bill Fitzmaurice]

It could be a ground problem. When there's more than one path to ground on the inputs a ground loop can cause noise. Since the amps appear to be chassis grounded via the power supply the input wires should not be grounded at both ends of the input wires.

[Keryn O'Shea]

It could be a ground problem...

Right.. l removed the amp mounting screws which had the ground symbol

And lifted the ground on the power supply, which cut the noise in half, sweet!
It was still annoying me though, so l reconnected the grounds and checked the voltage out of the power supply, it was 47.8v which is surely acceptable..
Not being able to leave things alone, l tweaked the plastic voltage adjustment screw

The noise lowered it's pitch all the way to zero as it approached 48.0v , then a hiss developed on the way to 49v. So at 48.5v, there's a totally acceptable, quiet and normal amplifier noise
I'll leave it playing music for a couple of days & see if the amps continue to be happy with an extra 0.5v.

[Keryn O'Shea]

When there's more than one path to ground on the inputs a ground loop can cause noise. Since the amps appear to be chassis grounded via the power supply the input wires should not be grounded at both ends of the input wires.

Bill, would this apply to an amp with balanced input signals? Because l have another DIY amp with a similar problem, and it has balanced inputs, the first amp pictured is unbalanced.
If we could nail that amps noise, that would be excellent!

[Tom Smit]

It could be a ground problem. When there's more than one path to ground on the inputs a ground loop can cause noise. Since the amps appear to be chassis grounded via the power supply the input wires should not be grounded at both ends of the input wires.

Bill, are you saying that the input negative wires inside the amp case should be on pin 3 rather than pin 1? And, that the RCA terminals should be isolated from the case?
(I am using this as a reference.... https://www.ranecommercial.com/legacy/note110.html )

[Keryn O'Shea]

And, that the RCA terminals should be isolated from the case?

Luckily these sockets have insulating washers

[Bill Fitzmaurice]

What you want is only one connection from the amp signal inputs to the chassis ground. If one is present via the power connection don't add another via the input wires. This only applies to unbalanced inputs, not balanced.

[Tom Smit]

And, that the RCA terminals should be isolated from the case?

20251006_151142.jpg
Luckily these sockets have insulating washers

[Tom Smit]

What you want is only one connection from the amp signal inputs to the chassis ground. If one is present via the power connection don't add another via the input wires. This only applies to unbalanced inputs, not balanced.

Thanks.

[Keryn O'Shea]

Moving on to amp no.2... the first thing done was removing the mounting post screw, which indicated the earth from 1 amp module

And the noise is GONE! Just the normal, very slight and acceptable buzz remains.
Thanks very much Bill!

[Seth]

[Tom Smit]

[mtglass]

I'm sorry I didn't see this sooner, I've been doing something similar, and here are some considerations I've been using.

1) Multiple TPA3255s need to be synchronized, "For slave-mode operation, turn off the oscillator by pulling the FREQ_ADJ pin to DVDD. This configures the OSC_I/O pins as inputs to be slaved from an external differential clock. In a master/slave system inter-channel delay is automatically set up between the switching phases of the audio channels, which can be illustrated by no idle channels switching at the same time. This will not influence the audio output, but only the switch timing to minimize noise coupling between audio channels through the power supply. This will optimize audio performance and result in better operating conditions for the power supply."

2) Meanwell, I use them too, is notorious for being noisy. "360mVp-p" The ripple directly impacts the gain of the amplifier, so it gets passed straight through the TPA3255 to the output. It'd be nice to use the clock from Meanwell to synchronized the TPA3255, but I haven't tried that, and I'm pretty sure the Meanwell isn't clocked so it's switch frequency would have to be turned into a clock source... Based on the application, it might be a good idea to add some power supply ripple rejection, or reduce the ripple with external capacitors.

3) I have had luck with a few (4) TPA3116D2 running off one Meanwell power supply, but I think that's because the load is, relatively, so low there's not as much ripple from the power supply. I've been using that to run discrete amplifiers for monitors. 24V is a lot of SPL at monitor distance. Even though the TPA3116D2 is rated at 2x50W, and is more like 2x15W usable, I haven't had issues using it for monitors. In theory an 8 Ohm speaker would need a 72W amplifier at 24V. (I did this as an experiment to prove this forum's theory of Voltage matters way more than power. It's true, at some point running out of power means distortion because the drivers don't move as far as they need to. I fix that by using a crossover to roll off the bottom end of the monitor where they want all that power.) I've found that it is possible to over-size a power supply. Depending on their design, some switching power supplies have more ripple with lower load, and some with higher load: typically there's a sweet spot which Meanwell doesn't document.

Best of luck, a great experiment!