JFET mu amp cell phone interference

Started by JFace, October 12, 2024, 12:23:47 PM

Previous topic - Next topic

JFace

I have a pretty standard JFET mu amp that is amplifying cell phone interference. Specifically, when a text is sent or received, there is a sound like a morse code coming through the audio.



The circuit is inside a metal box with jacks making contact, so everything is shielded. Here are my findings so far:

Grounding the input has no effect on the noise coming through
Placing a low value ceramic cap on output to ground has no effect
Placing a ferrite bead in series with the power supply has no effect
Placing ceramic 100n caps on VCC and VCC/2 to ground has no effect

I'm wondering if anyone has any other suggestions on how to mitigate the interference.

Thanks in advance.

stallik

I'm curious. Are you saying that the circuit itself is picking up the digital phone signal. Or is it amplifying the phone signal picked up by your pickups?
Insanity: doing the same thing over and over again and expecting different results. Albert Einstein

Rob Strand

Quote from: JFace on October 12, 2024, 12:23:47 PM'm wondering if anyone has any other suggestions on how to mitigate the interference.
Perhaps it's getting into the pots or the output leads.

Try adding a 1k in series with the output terminal.  Try a the socket end and the PCB/ckt end.

You can also add a small cap to ground on the output lead, either or both sides of the 1k resistor.

The same pattern on the input could help too.

There's no easy way to find the problem.  You have to do small experiments to try to put a box around the cause then slowly focus in.
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

JFace

Quote from: stallik on October 12, 2024, 06:43:51 PMI'm curious. Are you saying that the circuit itself is picking up the digital phone signal. Or is it amplifying the phone signal picked up by your pickups?
The circuit itself is picking up the phone signal. Even with the input grounded, the interference comes through.

Quote from: Rob Strand on October 12, 2024, 07:21:58 PM
Quote from: JFace on October 12, 2024, 12:23:47 PM'm wondering if anyone has any other suggestions on how to mitigate the interference.
Perhaps it's getting into the pots or the output leads.

Try adding a 1k in series with the output terminal.  Try a the socket end and the PCB/ckt end.

You can also add a small cap to ground on the output lead, either or both sides of the 1k resistor.

The same pattern on the input could help too.

There's no easy way to find the problem.  You have to do small experiments to try to put a box around the cause then slowly focus in.

Thanks for the suggestions. Definitely a tricky scenario and I'm trying to isolate where the it is being picked up.

I tried the cap to ground on the output, no change. The build out resistor doesn't make a difference either. The input doesn't seem to matter, because even grounding the input I get the interference.

I suspect it is due to the upper JFET, but not sure what I can do to filter it out. I'm attempting different caps after the 1M resistor at the gate, but not having good results.

Rob Strand

Quote from: JFace on October 12, 2024, 07:59:59 PM
Quote from: stallik on October 12, 2024, 06:43:51 PMI'm curious. Are you saying that the circuit itself is picking up the digital phone signal. Or is it amplifying the phone signal picked up by your pickups?
The circuit itself is picking up the phone signal. Even with the input grounded, the interference comes through.

Quote from: Rob Strand on October 12, 2024, 07:21:58 PM
Quote from: JFace on October 12, 2024, 12:23:47 PM'm wondering if anyone has any other suggestions on how to mitigate the interference.
Perhaps it's getting into the pots or the output leads.

Try adding a 1k in series with the output terminal.  Try a the socket end and the PCB/ckt end.

You can also add a small cap to ground on the output lead, either or both sides of the 1k resistor.

The same pattern on the input could help too.

There's no easy way to find the problem.  You have to do small experiments to try to put a box around the cause then slowly focus in.

Thanks for the suggestions. Definitely a tricky scenario and I'm trying to isolate where the it is being picked up.

I tried the cap to ground on the output, no change. The build out resistor doesn't make a difference either. The input doesn't seem to matter, because even grounding the input I get the interference.

I suspect it is due to the upper JFET, but not sure what I can do to filter it out. I'm attempting different caps after the 1M resistor at the gate, but not having good results.

Perhaps go extreme.  Try a short at the output socket then a short at the output lead at the PCB.
If it's still getting in it basically excludes the whole circuit!
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

Clint Eastwood

#5
Try a 10k or so series resistor close to the gate of Q1, after C3/R31

amptramp

Make sure the circuit is shielded against radiated susceptibility by being in a conductive box.  Make sure the power supply you are using is also shielded against radiated and conducted susceptibility.  If things are coming in over the air, you may have long enough output leads to pick up a signal regardless of the circuit.  Shielded wire is your friend.

R.G.

At cell phone frequencies, you're far into the realm of black magic. Wires act like inductors, resonators, and antennae even if they're "grounded".

My first question would be whether this happens with only one cell phone or many cell phones - that is, is it possible that your phone is blasting something out at a lower IF like 10.7MHz or so, which is a common-ish one? It's unusual to get discrete and non-SMD circuits to work up at a few Ghz. The pulse nature is what makes me think "IF" because you would otherwise be getting some other digital bleed over from voices, I suspect.

You can force the JFETs to not respond to anything over audio by combining a gate-souce or gate-source+gate-drain capacitor and a series resistor to the gate. Gate-source would probably do it, but adding a gate-drain bypasses the whole channel - it's not just the source lead in FET operation. The series resistor to the gate both makes this into a low pass filter, not just a funny resonator/oscillator by inserting damping; a bead may not help as much and can in rare cases make a resonance or oscillation worse.

Good RF practice is to shunt RF to the metal box right at the input jack. A 10pF to 100pF ceramic cap at the signal input with a solid electrical connection to the chassis is a good idea, as is a series resistor to the circuit input. Beads are good, but may rarely cause resonance issues without resistor damping. It's also important that the box is more or less seamless. Holes and slots in the box can let RF through when the RF wavelength is smaller than four times the dimension of the hole/slot. This can usually be ignored for audio effects, but if you're hearing artifacts from SMS only, the bang-bang nature of SMS stuff might be forcing an IF through where audio isn't. It would have to be up in the black magic realm if this is the case; 1Gz has a wavelength of about 299mm, so a quarter wave slot would be ~ 99mm/4". I don't know enough of the inner RF structure of cell phones to know if this is reasonable.

If it were me, I'd try some of the tips already suggested, and if that doesn't help, go to soldering tiny caps right on the JFETs gate-source and introducing series gate resistors in an attempt to kill their RF response. Keep your wires SHORT. And do test other cell phones.

R.G.

In response to the questions in the forum - PCB Layout for Musical Effects is available from The Book Patch. Search "PCB Layout" and it ought to appear.

duck_arse

Quote from: JFace on October 12, 2024, 12:23:47 PM



there is more stuff connected to the lower fet gate, isn't there.
"If you've got a blacklist, I want to be on it" - Billy Bragg
"You're on the list!" - The Death of Stalin

Jim Hagerman

Add an RC low pass filter before gate of input JFET. Something like 1k + 100pF. Cap to ground, of course...

fryingpan

#10
Quote from: Jim Hagerman on October 16, 2024, 12:34:58 AMAdd an RC low pass filter before gate of input JFET. Something like 1k + 100pF. Cap to ground, of course...
It might make sense to add a 1k gate stopper too, or wire that capacitor (or a smaller one) across drain and gate of the lower JFET.

amz-fx

Best Buy installed a radio in my old truck and it chirped every time my cell phone pinged a tower. I took it back for them to fix and they tried, but basically they said you have to live with it. So then I put a filter on the power connection and the problem went away.  :)

The power supply connection is an input too. Lots of rf junk can sneak in there. Just putting a cap on the power rail isn't enough, You need some resistance or inductance in series before the cap to complete the filter.

Try a 100 ohm resistor in series with V+ and then 0.1 uF cap to ground.

Best regards, Jack