Univox U1226 Preamp Hiss

[PapaUkraine7]

Hi, I've been working on converting the preamp and EQ sections of my Univox U1226 into a pedal form. I've worked out most of the kinks and I love the way it sounds, but I'm still running into a few issues.



1. The biggest issue right now is that channel 1 (Q1) has a very annoying hiss. I could manage if it were subtle, but it's pretty prominent. The strange thing is that it only appears on the 1st channel, not the second. I also tried rebuilding the 1st channel only with new parts and a couple value changes, but the hiss is still there on the 2nd build. As the 2 channels are identical apart from values, and the hiss is present in two different builds, I have to assume the noise has something to do with the PCB layout. I also tried flipping the power filtering cap (which is near Q1) to the other side of the board, and it didn't make much of a change.

The hiss starts directly at the drain of Q1. Turning RV1 doesn't change the hiss until the last ~10% where it suddenly turns down to silence. The hiss does not appear on the gate or source of Q1.

2. RV3 & RV5 are almost useless. If I audio probe the center of each knob I can hear that they are lowering the volume of certain bands, but when I listen through the output they don't seem to make a change. RV4 has a very strong effect on the signal though, cutting most of the bass and lower mids.

Here are shots of the PCB layout:

[rankot]

Why do you run Q3 and Q4 in parallel?

[PapaUkraine7]

Why do you run Q3 and Q4 in parallel?

It's a (fairly) direct copy of everything before the phase inverter in a Univox U1226. This is the original schematic I copied it from.

[Rob Strand]

Channel 1 is the bright channel and Channel 2 has a small amount of treble cut.  This is a common scheme for a two channel amplifier and may explain why Channel 1 is perceived as being noisy.   

Both channel could in fact be noisy which you can prove by temporarily wiring C3 to Q2 and C4 to Q1.   The noise from Q2 should be same as Q1.  If not, there's something weird going on with Q1.

You are missing a resistor from the gate of Q3/Q4 to ground.   It should be between 200k (approx reverb full) and 500k (reverb off).    Adding that resistor might help lower the noise because it lowers the amount of treble boost on Ch1 and the amount of treble cut on Ch2, but only by a small amount.

[PapaUkraine7]

You are missing a resistor from the gate of Q3/Q4 to ground.   It should be between 200k (approx reverb full) and 500k (reverb off).    Adding that resistor might help lower the noise because it lowers the amount of treble boost on Ch1 and the amount of treble cut on Ch2, but only by a small amount.

I just tried this and the hum isn't gone, but it's certainly tamed a bunch. Good call!

I'm going to try finishing the 2nd build with swapped caps and see how that one sounds.

Thanks for the help!

[dschwartz]

Reduce R1...47K into the gate of the fet is just injecting thermal noise with no benefits...try 4.7k or 2.2k.

Also, you can model the circuit on LTspice and make a noise simulation..it will help you find the parts that are contributing to the total noise...recalculate the filters for higher capacitance/lower resistance to reduce thermal noise (hiss)

[PRR]

Reduce R1...47K into the gate of the fet ...

Obvious target, but..... 90% of tube amps have 34k here. 1.4dB difference; audible but not big. And a clean JFET will hiss a hair less than a 12AX7 with a series resistor.

[PapaUkraine7]

Reduce R1...47K into the gate of the fet is just injecting thermal noise with no benefits...try 4.7k or 2.2k.

Also, you can model the circuit on LTspice and make a noise simulation..it will help you find the parts that are contributing to the total noise...recalculate the filters for higher capacitance/lower resistance to reduce thermal noise (hiss)

This was it! I dropped the Q1 Gate resistor to 4.7k and it's nearly silent now. I honestly might raise it back up a hair to try to get the best of both worlds.

Also good call on LT Spice, the noise testing confirmed that the gate resistor was the cause of about 70% of the total noise.

[Rob Strand]

This was it! I dropped the Q1 Gate resistor to 4.7k and it's nearly silent now. I honestly might raise it back up a hair to try to get the best of both worlds.

Also good call on LT Spice, the noise testing confirmed that the gate resistor was the cause of about 70% of the total noise.

It will show up in LT spice because the 47k is the dominant source of noise.  In practice there is (thermal) noise from the pickups.  It will get worse if you back-off the volume control a bit.      (You can simulate the guitar impedance in LTspice as well.)   If all is well then you still should be able to hear small decrease in noise going from 47k to 4.7k.  That would normal.   If you get a large decrease then maybe there's something wrong with the JFETs.

One test is to plug a guitar directly into the preamp and play with the guitar volume.  See if you get *excessive* noise.

There's no big deal changing those resistors to 4.7k it's more about making sure the problem is gone.

[rankot]

You can simulate the guitar impedance in LTspice as well.

I've found this somewhere on the Internet: