DIY Audio Projects Forum

Dear Audio people,

I am working on a project to build a loop switcher for Guitar effects. I have a crazy problem I couldn't understand and need help to find a solution. Let me explain:

I use a LM7805 regulator with the suggested circuit from the data sheet. I have a 10uF and a 100nF cpacitor on the input side (9V supply voltage) and the same on the output side (regulated 5V). With the regulated voltage I power a PIC microcontroller which switches some relais (TQ2-5V). The audio signal is running through the switches of the relais and should therefore be isolated from the rest of the circuit. All digital stuff works really well.

My problem is, that the audio signal is crazy noisy. You clearly hear the typical 50Hz hum from the power source. The LM7805 is powered by a regulated and isolated 9V voltage from a multi-Output effect pedal supply. If I power the board with the 5V of a USB port (behind the LM7805 - so this is not active) , the signal is nearly clean. The audio ground is not connected to the board. It goes straight from the input to the output jack. The audio signal itself is connected to the relais and so should also be isolated from the digital side of the board. I hear the sound of my guitar so all should be connected correctly. I can't understand why I hear this crazy hum, because the audio signal should have no relation to the main circuit. Even if this would be the case, I guess the LM7805 should do its job quite well and produce a clean supply voltage. The hum is not a digital noise which is usually higher in the frequency range. If I power the digital parts with the usb supply voltage all works quite well, too. So my knowledge does not help me anymore.

I want to ask, if you may have any suggestions where I should find the solution of this problem? It seems to be the DC regulator or one of its components. Can a bad capacitor cause this kind of problem? Is it possibly an EMI problem? I would try to use another voltage regulator, but have to buy one.

Thank you for your help...

Hi, My guess is that the input and output capacitors are not large enough for proper filtering. I suspect that the DC being fed to the 7805 is not all that clean and shows up in the coils of the relays. I typically use capacitors in the 1000uf range in those locations. They are cheap and rather small physically.

Good listening
Bruce

_________________
Some of my DIY Tube Amplifier Projects:

• PoddWatt: 5751 / EL84 Push-Pull DIY Tube Amplifier Project
• OddBlocks: KT88 Push-Pull DIY Tube Amplifiers (Mono Blocks)
• ForeWatt: ECC802S (12AU7 / ECC82) DIY Tube Preamplifier Project
• GrooveWatt: Tube (Valve) RIAA Phono Preamplifier Project

Yes I agree with Bruce and would add that most of the "wall-wart" dc supplies are
sadly lacking in filter caps as larger caps, eg;4700 or 10,000uF are just too large
physically to fit into the little black box. So add your own especially before the 7805 to
clean the DC up before it goes into the 7805.
Also make sure the "wall-wart" has enough current capacity to supply to your cct.
It is ok to have 2amp supply and only draw 500mA. But if you are supplying 500mA
and drawing out more the voltage going into the 7805 will be low and you will not be
getting the advantage of it's regulation. In short, for voltage, you need a "wall-wart"
outputting 5v + 2.7v = 7.7v, (7.5v ok), going into 7805 and for current just make sure
it is rated for more than your total cct will be drawing. And of coarse add the 4700uF
or even 10,000uF before the 7805. After the 7805 add 1,000uF. The smaller caps are
for RF noise suppression.

Hello out there,

thanks for your answers. Over the weekend I found a solution for this problem, but I didn't understand why.

@Bruce and Laurie54:
The capacitor size was a thing I thought of. I tried to use other sizes but without any major improvements.The wall wart is an isolated power supply I paid a lot of money for (85€/~100$) and whick delivers 9V on the output. This should be a serious thing. The noise problem occurs even if all relais are switched off. The controller circuit itself should take a max of 30mA, so the 7805 should easily handle this (1,5A version).

A solution I found was to connect the analog ground from the audio signal to the digital ground of the microcontroller. With that connection all works well. The problem is, that now the audio singal isn't completely isolated anymore. As long as it works, I'm happy with that. But I didn't understand the real reason for this behavior. Based on my limited knowledge, it should work fine even the other way.

So thank you for your messages. I will investigate a little longer in different directions. Maybe I will find a better solution. For now, I'm quite happy...

Are the relays switching BOTH audio signal and audio ground? Think of the audio gnd path the same way you think of the signal path. Yo be really isolated the relay COIL gnd must be only connected to MCU gnd and not to audio pwr or cct gnd. A 5v V+ from the MCU is only 5v with respect to it's own gnd. Sending 5v from a MCU to a relay which has it's return gnd connected to the audio gnd will give you this problem if the gnds are not connected. Connecting the gnds together eliminates any difference between the 5v MCU gnd and the audio gnd.

When designing you must treat the/any relay as two different cct's. One is the contacts, the other is the relay coil itself.

If you insist on having the MCU gnd and audio lines isolated I would use opto isolation. This way the mcu and it's pwr spply will only be turning on an led (in the opto). The audio cct and relay including their gnds would then be powered by the audio power supply and the relays are switched by the trans in the opto. In practice though it is rarely necessary to go to such extra cct complexity.