Electrostatic Deflection "Bunching" - Solved

[Tom Albrecht]

For years, I've had problems with sets that use electrostatic deflection with a CRT like the 7JP4. At high brightness/contrast, the vertical deflection would tend to show "bunching" of scan lines in bright parts of the image. To get rid of it completely, I would often have to run at a lower level of brightness and contrast than the CRT seemed to be well capable of.

I've noticed this problem particularly badly on a Sentinel 400TV and also my latest set on the bench -- the Airline version of the same set (basically identical; manufactured by Sentinel).

In the past, I put quite a bit of effort into trying to make sure the HV was stable, the voltage supply to the vertical deflection was stable, etc., by experimenting with adding extra HV filter capacitors. All to no avail. Today, I figured out how to solve it definitively.

In the schematic below, the vertical deflection coupling capacitors are circled:



Increasing the capacitance of C73 and C74 from .0047 uF to .01 uF solves the problem. Now the brightness and contrast can be run much higher, without picture distortion.

I suspect there is some difference in properties between the modern ASC 6000 V tubular capacitors and the original 6000 V paper capacitors that were used. I'm not sure exactly what is happening when the vertical deflection gets distorted, but it must have something to do with charge from the electron beam building up on the vertical deflection plates in bright parts of the scene. Adding more capacitance in the vertical deflection capacitors either increases the leakage (very slightly) or simply makes it so that it takes more charge to distort the scan.

An added bonus is that on this set, the picture was trapezoid shaped (top and bottom parallel, left and right sides sloping inward toward top) until I increased the capacitors. That also went away with the change.

And one more bonus -- the amount of available vertical deflection is drastically increased by the change. With the .0047 uF caps, the vertical deflection was not quite sufficient to fill the mask. Now there is way more than enough vertical deflection available, and I can adjust the linearity much better.

So I guess my advice would be to intentionally use larger vertical coupling capacitors than the set originally had, and you'll be much happier!

I'll post some pics later to show the effect.

[TV'S&MORE]

I have the serious trapezoid issue in a Firestone set I was restoring several months ago it drove me crazy so I put it back together and it's sitting on the shelf maybe it's time to bring it back on the bench.

[ChrisW6ATV]

Tom, that is a nice, easy fix for the problem. Thank you for posting your experiments and results. It makes me curious to see if those coupling capacitor values have been consistent in most or all of the 7JP4 sets.

[bandersen]

Interesting, I'll have to give that a try. I have an unrestored very, very early production Motorola VT-71 with giant 0.03 uF vertical caps. They later switched to 0.005.

[Tom Albrecht]

Most of them use .005 uF for the verticals and .001 uF for horizontal. One exception in my collection is my Hallicrafters 505, which has huge .05 uF 6000 V caps! The originals were huge. Looking at various Hallicrafters schematics, it looks like they reduced them to .03 uF in later production, as I'm sure the .05's were expensive.

The vast majority of electrostatic sets, however, use .005 uF. If you see the bunching problem, I'd recommend giving .01 uF a try. I'm amazed at how much better this set is now. It went from hardly watchable to basically perfect, with a picture bright enough to watch in a well lighted room.

[Tom Albrecht]

Here's what I mean when I refer to "bunching" of the vertical deflection. I put the .005 uF caps back in the set to demonstrate.

1. A moderate case of bunching. Note also the trapezoidal (or "keystone") image. This is all the vertical size I can get with the .005 uF caps.



2. A more severe case:



3. If you push the contrast up too far, it will go completely nuts:



Now, with the .01 uF caps in there, there is plenty of vertical deflection (the vertical size control has been backed off a great deal, and it still produces a full raster). The trapezoidal picture is gone. Most importantly, the "bunching" is gone, even in a picture like this, where the bright spot would normally have bunching:



I'll be trying this on my other identical set (a Sentinel) very soon. It has the same problem. I've seen this effect to various levels on other electrostatic sets as well, so it will be interesting to see if it fixes the problem in all cases.

[kvflyer]

Thinking out loud now.

When these sets were new, capacitors were huge and no doubt expensive when they had a WVDC of 6000 volts. So, could it be that they used the absolute smallest value possible? So, maybe it would be a good idea to make them larger. Or would it upset the circuit?

Seems to work in this case. I know that you don't want to go crazy in a power supply (first cap after the rectifier) because it will stress the rectifier. Don't know what it would do here...

[TV-collector]

Wow, thats the news of the year!!!!!!!!
I have a lot of new 1000 PF/12.000 Volts caps which I connected to about 5000 PF.
Now I could get a problem because they need their space.

TV-collector

[tvdude1]

Thanks Tom my set has the same problem.Will try it.

[dieseljeep]

Originally Posted by tvdude1

Thanks Tom my set has the same problem.Will try it.

There's been a lot of concern about the compatibility of the newer issue capacitors, in the deflection circuits.
Some restorers didn't have any success, using ceramic disk caps. I thought, using the high quality tubular caps, were the answer to all the problems.
The original caps, were nasty, old oil caps.
BTW, how did National get away with using door-knob caps, as coupling caps.

[RDusel]

Tom,
I've been fighting that exact bunching on my Admiral. I was suspicious of a replacement vertical output transformer.
I'll be trying the .01's shortly. Thanks a lot for being persistent and solving this!
Rob

[cameronflyer]

I wonder if this would give a brighter picture on a TS-18 Motorola 9VT1?
I don't have the bunching or trapezoid but do have the vert. height problem and a bit of a dim pic.

[Tom Albrecht]

It doesn't directly improve the brightness. However, if you have the bunching problem, and are turning down the brightness and contrast to mitigate it, then doing this fix will allow you to run at higher brightness and contrast and still have a very nice picture.

The reason this took me several years to figure out is because you would think that if the vertical deflection is working fine at low brightness, then the caps must be doing their job OK. Since the problem was occuring only at high brightness, I was assuming that somehow the bright spots in the picture were causing HV collapse or something like that. I don't really understand why this fix works, but it sure does. I can't get over how good this set looks now.

Another thing that prevents a more direct troubleshooting of this problem is that one would normally want to probe the deflection signals right at the CRT to see if there is anything funny about the signal on a scope. But you can't measure those point with a scope because of the HV. If you probe on the low voltage side of the caps, you don't see anything really amiss.

We've known for many years that ceramic caps don't work in this application. I wonder if I cut open one of the latest .005 uF ASC tubular caps, I'll find a little ceramic cap inside? Just kidding, but it does seem to work a little like a ceramic!

[old_tv_nut]

So, if I'm following correctly, the problem occurs even with NEW .005's? Weird. Hard to imagine how the vertical could modulate the horizontal (keystone) unless the .005's are leaky (?) The interaction with video seems a mystery also.

Would it be possible to make a decoupling test probe with a new .01 and a, say, 22 Megohm to ground, so you have a ground-reference waveform to scope as you poke around? Maybe you could discover which point is going wonky in the first place, as you switch the vert coupling from .005 to .01. Another thing is to look for the keystone on the plates of the horizontal outputs.

[Tom Albrecht]

It does indeed happen with brand new ASC brand .0047 uF 6000 V tubular capacitors.

If you probe the plates of the vertical and horizontal outputs (which I have many times), you don't see anything until the problem is extremely severe -- then you can see the slightest wiggle on the plates of the vertical output. No where near enough to explain the effect. The keystone waveform is slightly visible on the horizontal output plates. But surpsingly, when I add a huge filter cap (more than 10 uF) on the 800 V line feeding the horizontal, the the waveform cleans up, but the keystone on the CRT is still there.

The main effect has to be occurring beyond the capacitor at the CRT deflection plates, where you cannot easily probe. I may try what you suggested to make a probe using another HV cap. However, the behavior in that cap might be similar and obscure the effect. I won't bother to change this particular set back to the bad caps, but I have another one (Sentinel 400 TV) with the same problem that I will try within the next few days and see if I can learn more about what is happening.