Grounding / shielding advice appreciated ("I accidentally built a radio")

[krabbencutter]

I was doing some measurements with my equalizer today and discovered that I have some nasty buzzing and pretty good radio reception
My "best" efforts to create a proper star ground obviously weren't good enough, so I'd be happy if someone could help me improve my power distribution & ground layout.
The PSU design is based on this article from Thomas Henry.

Picture 1:
This is my linear unregulated power brick. It has outgoing cables for +/- 12V, power ground and mains earth. Power ground and mains earth are separate wires and not connected.

Picture 2:
This is the connection to my EQ. I'm using 4 pin DIN jacks for no special reason.
1 = +12V Pin
2 = Power GND Pin
3 = -12V Pin
4 = Mains earth (shield)

Picture 3:
Purple mark: this is where the power brick connects to the circuit and power is routed to the 7812/7912 voltage regulator board.
Blue mark: After passing the regulators, power is routed to the power distribution board with 10pin ribbon cables. This spot is also my current star ground where power, chassis, mains earth, analog signals and digital signals run together.
Green mark: that's where I connect the shielding pins of my balanced TRS connectors to the chassis

Picture 4:
This is the bottom side of my power distribution board. I have separate ground rails for power, analog signals and digital signals. These ground rails stay separate throughout the rest of the unit.

To make my troubleshooting efforts easier, I removed all audio circuit boards except the balanced driver/receiver board. The I/O board has 100uF filtering caps at the power input and all opamps have 100nF bypass caps. The output of the balanced receiver and the input of the balanced driver are connected with a jumper. My suspicion is, that the issue is at least partly related to the 10pin ribbon connectors. Interference reduced noticably, when I removed the power distribution board and connected the I/O board directly to the regulator PCB. The length of the ribbon cable seemed to play a part as well. At the same time it didn't really matter if the case was open or closed. The case is assembled from aluminum parts that are connected with screws.

• Is there any obvious mistake in my grounding/shielding efforts? I'm especially unsure about the mains earth connection.
• Is the power distribution board the culprit and if so, could I improve the situation with additional capacitors?
• Are ribbon cables a bad choice? Should I use shielded & twisted cables to route power inside the unit?
• Could I improve the RF shielding by using a different material for the case?

 

[gyraf]

..we probably need schematic of the actual electronics to figure this out? I don't think grounding mistakes per-se makes RF messups

Perhaps try putting your "active" pcb in a small plastic bag and wrap it in aluminum foil, just to exclude local rf-problems
perhaps shielded cable to/froom inputs/outputs

/Jakob E.

 

[abbey road d enfer]

I agree with Jakob; it's hard to guess.
One thing you may try and would give us some idea, you could connect a 100nF capacitor between the 0V pin and the chassis at the power entry.

 

[Newmarket]

..we probably need schematic of the actual electronics to figure this out? I don't think grounding mistakes per-se makes RF messups

Perhaps try putting your "active" pcb in a small plastic bag and wrap it in aluminum foil, just to exclude local rf-problems
perhaps shielded cable to/froom inputs/outputs

/Jakob E.

Yes. But let's make that plastic bag anti-static, or use a cardboard box etc

 

[abbey road d enfer]

Picture 2:
This is the connection to my EQ. I'm using 4 pin DIN jacks for no special reason.
1 = +12V Pin
2 = Power GND Pin
3 = -12V Pin
4 = Mains earth (shield)

I don't see a wire for the earth there. Are you using the shell? DIN plugs are known to offer dubious contact via shells.

 

[rogs]

RF pick up often involves rectification by a early high gain active stage, rather than just a grounding problem.
This scan from National Semiconductors famous 1980 audio handbook has some suggestions... Notice the final bullet point at the bottom left of the page

 

[Bo Deadly]

What are the physics behind a bad solder joint rectifying AC? Is it because the crack with a little oxidation presents a high impedance with low voltage across it but with higher voltage the signal can somehow punch through the oxidation?

 

[gyraf]

..copper oxide rectifier,,

 

[MisterCMRR]

Almost any metal, including lead and tin, will form a semi-conductor when that metal and its oxide (a "corrosion" product) are in contact. This is a problem not only in audio, with connectors in particular, but in RF systems (where inter-modulation in the non-linearity of the "corrosion rectifier" creates new frequencies). Galena (lead sulfide, or PbS) was used in early crystal radios with a springy contact called a "cat whisker" because it was a pretty good rectifier/detector at small/weak signal levels (much like germanium that later replaced it). Corrosion at pressure contacts is definitely the enemy of good, clean audio!!

 

[Bo Deadly]

Eddie Van Halen once claimed that his Marshall Super Lead 100, the one used on the various early albums, did not sound the same after a tech changed the wiring. I thought that sounded like non-sense and he was known to tease people on occasion with false information about his setup. But I wonder what would happen if there were a bad solder joint somewhere in an amp like on a cathode resistor. Maybe that contributed to the unique sound of that amp? When it was re-wired, the bad joint was "fixed" and so was the unique sound of the amp?

 

[krabbencutter]

Thanks everyone for chiming in!

..we probably need schematic of the actual electronics to figure this out? I don't think grounding mistakes per-se makes RF messups

Perhaps try putting your "active" pcb in a small plastic bag and wrap it in aluminum foil, just to exclude local rf-problems
perhaps shielded cable to/froom inputs/outputs

/Jakob E.

The balanced transmitter/receiver are based on Project 87 from Rod Elliott:

1. The balanced input signal gets filtered https://sound-au.com/p87-f2.gif
2. the following receiver looks like this https://sound-au.com/p87-f1.gif
3. and here's the transmitter https://sound-au.com/p87-f3a.gif

I'll give the anti static bag and 100nF cap suggestions a try.

I don't see a wire for the earth there. Are you using the shell? DIN plugs are known to offer dubious contact via shells.

Yes, the mains earth wire is soldered to the DIN plug's shell. But I can try soldering it to one of the pins and see if it makes a difference.

 

[gyraf]

Unless you have the world's slowest opamps, you will need ~22pF capacitors across (parallel with) R3, R5 and R104 to ensure stability

And I miss a lowpass at the input of the transmitter, but it May work as-is most of the time..

Jakob E.

 

[JohnRoberts]

Eddie Van Halen once claimed that his Marshall Super Lead 100, the one used on the various early albums, did not sound the same after a tech changed the wiring. I thought that sounded like non-sense and he was known to tease people on occasion with false information about his setup. But I wonder what would happen if there were a bad solder joint somewhere in an amp like on a cathode resistor. Maybe that contributed to the unique sound of that amp? When it was re-wired, the bad joint was "fixed" and so was the unique sound of the amp?

As I shared already, inside a point to point wired tube guitar amp "lead dress" (how the wires are arranged inside the amp), can affect wire to wire crosstalk, especially if the wires are connected to high impedance nodes within a high gain tube guitar amp. I have no specific knowledge about Eddie's old marshal, but it seems plausible that changing how wires were arranged inside such an amp could make a subtle difference to a discriminating listener like Eddie.

Lead dress inside high gain point to point wired guitar amps is established by the original amp design engineer. Typically a hand wired prototype could be used by the factory as a model for correct wiring layout. I can imagine a repair technician making lead dress changes that looked neater, but sound different.

JR

 

[gyraf]

..it's been a couple of years since OP had the problem..

All components should be SMT.

Consider using a 4-6 layer PCB with one or two ground planes

etc, etc: BS

 

[Voyager10]

'Radio reception' sounding like a detuned AM radio is often a symptom of oscillation (i.e. instability) somewhere in the circuit.

 

[abbey road d enfer]

Gravedigging thread