LND1500 loop issues

[studiodunn]

I have built 2 of these and both have the same issue with not passing a good signal through.
The signal is thin, weak and obviously off from the bypassed signal. There is also signal bleeding through when it’s bypassed

The return takes a signal fine, but the send side has issues. The input has 310v and all my components have been double checked against a few schematics.

This is the schematic I built from and I’m not seeing anything drastically different from other versions of this loop I’ve seen.

What am I missing here?

5EB1EE27-393A-4218-A603-EDB4ECE7118D.png

[martin manning]

Looks like the send stage has a LOT of feedback, and it drives the loop from the drain, where most send buffers are source followers.

Where did that come from? Are you sure it is correct?

[frankdrebin]

Schematic is the metro amp zero loss,and it works as intended.
If it doesn't work it's either a faulty MOSFET,wrong component or bad execution.

[pdf64]

It’s worth confirming that the 10 ohm resistor on the send is wired as a 0V lift, rather than a load
How about taking a voltage survey, and adding the readings to the schematic?

The ‘bleed through on bypass’ issue description seems kinda cryptic, really needs a heap more detail adding.
On the face of it, it may support the ‘bad execution’ hypothesis.

[jml77]

How has the bypassing been achieved, perhaps a pic of the wiring might help debug the circuit, or rather the installation of it.

It's a very neat design, using so few components in the send stage, but achieving all that's needed. I noticed that the gate of the first stage hasn't got any protection diodes, I think the feedback might help keep the gate well within spec, but could that be a problem. I doubt that it would be. Just a thought.

[studiodunn]

Some pics and voltages. I’m not sure what’s happening at R1-100k but I’m guessing the issue lies there.

6CC4B292-5F59-4D4C-9BAE-FD196A92292A.jpeg

95ACD352-D0C0-48F9-9CD5-5BECB6419A03.jpeg

83DF6CA7-0E6D-4694-B009-36FDC63FD631.jpeg

[attachment

IMG_7546.jpg

=1]46C36FC0-E667-46E1-8D55-AE580CB6B056.jpeg[/attachment]

[pdf64]

Some pics and voltages. I’m not sure what’s happening at R1-100k but I’m guessing the issue lies there...

No voltages visible?
The circuit is in very close proximity to other circuits, specifically the 2 tube sockets underneath; maybe there's some unintended coupling going on, have you scoped for oscillation etc, have you checked to see if the board works if it's elsewhere, ie away from that place in the chassis?

[roberto]

That schematic cam be quite improved imho.

I would use a voltage divider at the input of the schematic, then use a stronger mosfet for it (IRF820?) with more current flowing through it but less feedback, then add feedback on the return stage.

[studiodunn]

Some pics and voltages. I’m not sure what’s happening at R1-100k but I’m guessing the issue lies there...

No voltages visible?
The circuit is in very close proximity to other circuits, specifically the 2 tube sockets underneath; maybe there's some unintended coupling going on, have you scoped for oscillation etc, have you checked to see if the board works if it's elsewhere, ie away from that place in the chassis?

Sorry, they are up now.

And yes, I moved the board and got the same results.

Why I'm going from 330v to 1.41v after the 100k resistor will surely shed some light on the problem.

Schematic is the metro amp zero loss,and it works as intended.
If it doesn't work it's either a faulty MOSFET,wrong component or bad execution.

I have bulit 3 Metro loops and everyone went in and worked without issue. I'm not saying I couldn't or didn't make a mistake, but for both loops to do the same thing( and not in the same build) has me confused.
I also replaced the MOSFETs in the loop shown to eliminate that being the culprit....no change.

[martin manning]

I think voltages on the driver stage should be something like this:

[studiodunn]

I think voltages on the driver stage should be something like this:

I wish they were

[jml77]

Those green mylar polyester caps, I've only ever seen them rated at 100v.

Q1 is hard on, RDSon looks to be about right, so what are the driving voltages. What amp is it, and where is the loop being inserted.

[martin manning]

...suggesting that C2 is leaking DC.

[studiodunn]

Those green mylar polyester caps, I've only ever seen them rated at 100v.

Q1 is hard on, RDSon looks to be about right, so what are the driving voltages. What amp is it, and where is the loop being inserted.

330v off the B+ in a Wizard Mtl 100w build. The Loop is fed directly from the Tone stack, out to the PI

[Ten Over]

The Q1 LND150 is functioning properly for a bias of +0.64V. The question is why there is 1.3V on the Gate creating that +0.64V bias. I agree with martin about C2 being a prime suspect and I would lift C2 and then measure the voltages.

The expected bias of -0.6V from the diode on the Source of Q1 would cause an idle current of 0.5mA according to the LND150 literature. This would give you a Drain voltage of 280V if the supply voltage remained at 330V, but it probably won't. As it sits, Q1 is taking down 3.3mA and Q2 is taking down around 0.8mA causing a 45V drop across what looks like an 11K dropping resistor (R8). If Q1 starts running at the expected 0.5mA, the drop across R8 will only be 14V and the Q1 Drain voltage could be 311V.