capacitor shielding/grounding?

[Scenaria]

ok myself and keef discovered something that I have never seen before...

oscillation/distortion that occurs in an la2a... (using sowter trannys)

this non-symmetrical clipping occurs at a certain point as gain is increased on the compressor..

while poking around I noticed that if I wrapped my hand around the coupling cap and allowed my hand to touch the chassis the oscillation/clipping artifacts would go away and I could continue to increase gain until I had symmetrical clipping... at first the thought was some form of vibration? but even if I use double sided sticky tape to adhere the cap to the chassis it wont go away.. unless I touch both the caps case AND the chassis metal with my one hand...

any input on what this could be? is there a name for this problem?

by the way.. this does not occur when using the UTC transformer...

 

[jrmintz]

Shielding the cap? Would that help?

 

[Scenaria]

no clue...

could try it I suppose..

just seems like this shouldnt be required... as it seems to stem from the o/p tranny design or something?

 

[SSLtech]

It should be added that the point at which this occurs is immediately before clipping, which first occurs on the top half of the waveform.

As the signal gets within 5-10% of clipping, it sporadically 'blips' with the frequency of the blip increasing with nearness to clipping. That last half-decibel on a steady state tone produces a slight ring-modulator effect, with increasing and decreasing sidebands as the level is increased slightly.

the waveform is such that the 'blip' is a jump to the clipping point... almost as if the signal 'snaps' to the rail, resembling an attraction to the rail!

It's interesting enough, -I wish I had time to investigate it thoroughly, but right now things are a bit mashed... I'll be interested to know if Scenaria finds out anything else.

BTW, the linearity of the Sowter input tranny is vastly better than the UTC. I measure a -3dB point of about 55kHz, as opposed to around 25kHz with the UTC HA-100, the HA-100X or the A-10.

Nice work, Mr. Sowter!

Keith

 

[SSLtech]

[quote author="Scenaria"]no clue...

could try it I suppose..

just seems like this shouldnt be required... as it seems to stem from the o/p tranny design or something?[/quote]

Well, the UTC has another layer of shielding in the form of its can... but the LA-100 case should also be a decent shield... unless of course the paint on the adaptor plate is preventing decent contact... it is of course possible...

Keith

 

[Scenaria]

GOOD idea!

the paint! I'll scrape off some paint where the tranny is mounted... never toughts of that!

 

[Scenaria]

still distortion... only goes away when I ground the cap to the chassis with my hand...

funny thing is it can actually do +22dbu into 600 as long as I keep my hand on it before any clipping begins

at 100k it can do +28dbu

ok soo.. I essentially used the sticky tape on each end of the cap to apply pressure to the leads against the chassis.. this has essentially made the distortion just about dissapear... ofcourse its not a permanent fix. it can definately handle a pretty decent signal... almost +23dbu into 600 and the +28 into 100k

sweeps look sweet too... after trimming out the trim cap it has a very similar curve to the UTC... just higher frq resp.. same nice little bump and all


is this called parasitic oscillation?

yea... im reading and probably going in a completely wrong direction :/

lit

 

[PRR]

Put that cap in the other way around. It DOES matter.

Outside foil goes to lower-impedance or less-sensitive side of the circuit. Foil caps have (had? should still have?) a stripe on the end where the lead connects to the outer foil. Usually that goes to a Plate, so the inner foil goes to a Grid. You have the outer foil going to the Grid, a nice big pick-up antenna.

You may also have layout issues. How far have you departed from the well-considered original, both in general layout and in fine details like wire routing?

You want big oscillations? Put the Plate lead of an 8417 1/4" away and crossing a wire-wound Cathode resistor. This forms an RF transformer and thus an RF oscillator. Amazing amount of power flows. Around 80 Watts. Far more than an 8417 can stand. Just moving the plate lead cured it.

 

[Scenaria]

if your talking about physically rotating the cap.. I tried that with no improvements..

the cap is an electrolytic.. though I did try a poly cap with no additional success..

its definately a big antennae...

I wrapped some foil around the cap and tried grounding it.. this actually increased its reception :roll:

BUT when I wrap my fist around the cap it works perfect...

it has to be a combination of capacitance with the impedance of that particular tranny doing this...

what would my hand be doing to cause the elmination of this interference?

 

[Guest]

WARNING: I am probably full of crap.

You may be providing an additional capacitance to ground by placing your body in the signal path. I think the output combination of the coupling cap and transformer begin to resonate. Maybe a different value capacitor will change things.

 

[Scenaria]

ok...

tried other cap values... with no success

BUT with foil wrapped around the cap it does help...

one thing we realized was the hinged panel wasnt grounded well... until the paint was scraped off.. the box is now full shieled when closed but the interference is still present.. though the cap seems to require less "touch" to get the interference to go away... while before I had to enclose the cap in my hand now all I have to do is touch it...

I was thinking of swapping out the lytic and trying a poly or film..

 

[Gus]

PRR the stripe does not always show the outside foil maybe some are mismarked.

There is a good site with a writeup for finding the outside foil in film caps

http://www.aikenamps.com/TechInfo.htm

look in the advanced section at the bottom.

 

[CJ]

The LA2 is not the most stable layout in the first place. I had intermittant oscillation problems in mine which is why I switched to the 12AY7 for stage one. Remember that input signal level matters, and that the early LA2's had that pad thingy on the back. Try dropping the input signal down and see what happens. I notice mine went into oscillation more often when messing with the hinged cover. I believe the HA-100 has more leakage capacitance tio ground. I will have to go check my notes. This is probably due to the three mu cans and case.

By holding your hand around the cap, you are just adding picofarads to ground in the right place of the circuit.

Everything is kind of crammed together in the LA2, so there are not a lot of options for moving the cap around. It might not be a good idea to have a metal shield around a cap that is seeing B+ voltages at high frequencies.

I just measured my capacitance while holding my hand around a lytic can and the other end of the cap checker alligator. I got 14,000 pf or .014!
So you might want to check how much high end you are dropping when you shunt the signal to ground. The funny thing is, this is the output stage, not the more sensitive input sections.

You might try soldering a small cap to the lytic can lead and grounding it.

 

[Scenaria]

tried the cap on the lytic..

no progress. (though that was a good idea)

also tried a polyprop cap instead of a lytic... again no luck

the entire enclosure is fully grounded.. so theres no weak spots in that area (anymore) hehe

 

[CJ]

Did you try pouring some Apricot Ale inside the chassis? :razz:

 

[NewYorkDave]

Steve, I got your PM. Here are my thoughts...

Does the nature of the 'squeeging' change as you adjust the HF trim cap?

What happens if you open up the feedback loop? (Of course, you'll have to drop your input signal way down before you do this).

I had a similar problem in a circuit of mine that uses a 12BH7 White cathode follower to drive an output transformer. Just for laughs, try 1K grid stopper resistors on both grids of the 12BH7. Mount these right on the socket pins, in series between the grids and the 470K grid-leak resistors. These will not affect the normal operation of the circuit, but can make the output stage more stable.

 

[Scenaria]

welp.. I was about to try some of your suggestions but now my damn AP wont power up.. fuse blew (im hoping it was just week from being turned on and off) :shock:

anyhow.. I'll have to grab some fuses from the rat shack tomorrow

 

[NewYorkDave]

AP? Audio Precision?

Doesn't the squeeging show up on a plain ol' scope?

 

[Scenaria]

yea

but my audio source comes from the ap

no basic osc

:\

 

[CJ]

Just plug a CD player into the LA2.