Ampex 601 extremely weird start up problem

[substitute]

Hello.  I'm refurbing an Ampex 601, it's basically done and everything is sounding good.  However I'm having an extremely strange problem powering up the unit.  When I turn it on, it warms up, I can hear a little noise from the output but it doesn't really pass signal, I have to turn it way up to get anything and even that signal is very clipped.  Now here's where it gets weird, if I short pin 7 of V102b to ground for just a second there's a little pop and then everything works fine. 

I've tried different tubes, no change.  This one is so odd to me that I'm not even sure where to start.

Thanks!

 

[substitute]

Oh, here's the schematic...

 

[mjrippe]

Check coupling cap C104

 

[substitute]

That didn't do it.

To simplify/eliminate variables I disconnected the line input and two 270k mixing resistors, soldered a new cap directly from pin 1 to pin 7.  The powering up issue persists and a quick short to ground still fixes it. 

However, it sounds better without that extra stuff in there, so I've got that going for me.

Regardless, thanks MJ

 

[PRR]

You need a grid-to-ground resistor. 1Meg is popular.

As built it should be grid-leaked via Line Rec Level pot. If you have been removing stuff, that would explain the floating-grid symptom.

 

[substitute]

Yup, that did it.  As always, thanks PRR.

 

[PRR]

One "MUST DO" in any vacuum-tube or JFET design or project, which seems to have faded from many people's memory, is -- Where is the grid bias??

When people post sketches, that's one of the first things I look for. Does every grid have an appropriate path to a bias source? (Often circuit-common for simple stages.) This tidbit is frequently forgotten.

(Somehow it rarely happens on BJT designs??)

In this case we knew the Ampex diagram you posted -worked-, and without a screwdriver jump-start. I started thinking of bad tube and joints. But then I noted "without that extra stuff", and wondered how much stuff removed was non-extra. In this case the Line In pot was both a signal path and a bias path.

 

[substitute]

A gentleman and a scholar.