[Philip Pemberton]

Hi,
I'm trying to wire up a DG7-32 CRT. I've managed to get a visible spot,
which is more-or-less focussable down to a 0.5mm dot. Maybe not as good as my
oscilloscope, but pretty good for a 1950s CRT.

The tube itself is a Tungsram of unknown age, and was bought new-old-stock.
I'm using the Philips DG7-32 datasheet because I can't find a copy of the
Tungsram datasheet anywhere. Given that it's a standard tube, I doubt the
specs vary that much between manufacturers.

Now, the problem is, I can't get the spot to extinguish completely. Here's
the circuit I'm using to derive the G1, cathode and G3 voltages:

======

100k 120k 68k 220k 680k
-500V <<<---/\/\/---/\/\/---*---/\/\/----/\/\/---/\/\/----------------> 0V
^ | ^
(actually | | |
-530V) | | |
| | |
Vg1 (inten) Vk (cathode) Vg3 (focus)
-522V to -465V -396V -366V to -276V

Anode (Vg2,g4,a) voltage: +243V

Voltage across tube (Va-Vk): 639V.

======

First point: The deflection plate pairs (D1/D1' and D2/D2') are wired
together to disable them, but the pairs are left floating. The deflection
amps aren't wired up yet; I wanted to get the core of the CRT circuitry
working, then move on to the deflection control circuitry.

Now, if I set Vg1 to -522V, I get a large round spot on the CRT face which
can't be focussed down at all. Stepping Vg1 down to around -500 gives a
mostly round spot that can be focussed down fine, but it's far too bright.
This is the problem I want to solve. This CRT is going to be used as an X/Y
display for both a scope clock and an I/V curve tracer, so I NEED to be able
to blank the spot (for the scope clock at least).

Vg3 seems to work OK, but doesn't have much in the way of range - the CRT
spot focusses almost all the way to one end of the 220k potentiometer's
travel. Fiddling with the balance of the 68k and 680k might be enough to
shift the center point closer to the middle of the pot's travel though; it's
ranging from K+30V to K+120V and the datasheet specs the range limits as K+0V
and K+200V. This is more of an ergonomic issue than anything else, though.

At one point I also had a 47k potentiometer wired between +240 and 0V,
with the CRT anode connected to the wiper. This has since been removed to
simplify part of the circuitry, and the anode is now wired direct to the
+240V rail.

The voltages are derived from a 6VA 9-0-9/230V mains transformer run in
reverse, with a half-wave rectifier, 10uF capacitor and Zener circuit
generating the +240, and an unregulated voltage doubler generating the -500V
(which is actually supposed to be -600, but is actually closer -530).

Thanks.
--
Phil. | Acorn RiscPC600 SA220 64MB+6GB 100baseT
[email protected] | Athlon64 3200+ A8VDeluxe R2 512MB+100GB
http://www.philpem.me.uk/ | Panasonic CF-25 Mk.2 Toughbook
No software patents! <http://www.eff.org/> / <http://www.ffii.org/>
.... If at first you don't succeed, destroy all evidence that you tried.

 

[Fred Abse]

On Sat, 22 Oct 2005 12:31:40 +0100, Philip Pemberton wrote:

First point: The deflection plate pairs (D1/D1' and D2/D2') are wired
together to disable them, but the pairs are left floating. The deflection
amps aren't wired up yet; I wanted to get the core of the CRT circuitry
working, then move on to the deflection control circuitry.

You need to have a DC path from the deflection plates to anode, otherwise
you get stray charge building up on the plates, causing all sorts of
unexpected effects. _Never_ leave deflection plates floating.

Now, if I set Vg1 to -522V, I get a large round spot on the CRT face
which
can't be focussed down at all. Stepping Vg1 down to around -500 gives a
mostly round spot that can be focussed down fine, but it's far too
bright. This is the problem I want to solve. This CRT is going to be
used as an X/Y display for both a scope clock and an I/V curve tracer,
so I NEED to be able to blank the spot (for the scope clock at least).

According to the data I have, beam cutoff should occur with g1 between 50
and 100V negative wrt cathode, with 500 volts on the anode. g1 voltage
*does* affect focus. Google for "Wehnelt cylinder" for an explanation of
why. According to what you say, you can get good focus, so that would
appear to rule out an O/C g1 electrode.

Vg3 seems to work OK, but doesn't have much in the way of range - the
CRT
spot focusses almost all the way to one end of the 220k potentiometer's
travel.

ISTR these tubes were always soggy on g3 voltage. As long as the spot
focuses, don't worry too much.

Fiddling with the balance of the 68k and 680k might be enough to
shift the center point closer to the middle of the pot's travel though;
it's ranging from K+30V to K+120V and the datasheet specs the range
limits as K+0V and K+200V. This is more of an ergonomic issue than
anything else, though.

At one point I also had a 47k potentiometer wired between +240 and 0V,
with the CRT anode connected to the wiper. This has since been removed
to simplify part of the circuitry, and the anode is now wired direct to
the +240V rail.

That's how you implement an astigmatism control with this type of CRT.
Setting the anode voltage equal to the mean voltage on the plates. You'll
need this when you drive the plates from amplifiers, and they're
somewhere between 0V and B+ voltage.

Get those plates connected to anode via 100K resistors for a start, and
see if some of your problems go away.

The voltages are derived from a 6VA 9-0-9/230V mains transformer run
in
reverse, with a half-wave rectifier, 10uF capacitor and Zener circuit
generating the +240, and an unregulated voltage doubler generating the
-500V (which is actually supposed to be -600, but is actually closer
-530)

Any particular reason for this? I'd try to get hold of a B+ mains
transformer and voltage double the output of that.

Do you have a DC path from heater to cathode? Don't rely on cathode
insulation (only rated at 125V), and remember that your heater transformer
insulation needs to withstand cathode voltage. Heater windings going leaky
to frame used to be quite common in the days when 'scopes used mains
transformer fed heaters.

 

[Philip Pemberton]

In message <[email protected]>

You need to have a DC path from the deflection plates to anode, otherwise
you get stray charge building up on the plates, causing all sorts of
unexpected effects. _Never_ leave deflection plates floating.

Well, that explains that. I've redesigned the biasing circuitry and it now
works fine. The focussed spot is less than 0.5mm in size.

I've built up some deflection amps based on Sascha Ittner's scope clock
design. Problem is, the damn thing doesn't seem to have enough of a span in
the "X pos" and "Y pos" controls to get the spot moved to the correct
position on the screen :-/
The same basic circuit is used in the "OZ2CPU" CRT clock, though with
different component values.

The big problem now is getting the deflection amplifiers going. I'm having
problems where the Y deflection is fine, but the X deflection won't go to
the edge of the screen.
What I want to do is tweak the gain of the amplifier so that it'll run off 5V
and the HT, and take an input of either 0-5V or 0-2.5V, where 0V=full left,
5V (or 2.5V) = full right, 0.5V = spot centered.

Any particular reason for this? I'd try to get hold of a B+ mains
transformer and voltage double the output of that.

I wanted to run it off a 12V supply.

Do you have a DC path from heater to cathode?

Yes. One end of the heater is wired directly to the cathode.

Don't rely on cathode
insulation (only rated at 125V), and remember that your heater transformer
insulation needs to withstand cathode voltage. Heater windings going leaky
to frame used to be quite common in the days when 'scopes used mains
transformer fed heaters.

Let's see.. The transistor consists of:
Plastic bobbin - good to a few kV
Primary windings
A layer of Kapton tape
Two layers of Kapton-coated tinfoil, grounded
A layer of Kapton tape
Secondary winding

I'm not sure what its isolation voltage is, but it should be good enough for
a CRT.

Later.
--
Phil. | Acorn RiscPC600 SA220 64MB+6GB 100baseT
[email protected] | Athlon64 3200+ A8VDeluxe R2 512MB+100GB
http://www.philpem.me.uk/ | Panasonic CF-25 Mk.2 Toughbook
No software patents! <http://www.eff.org/> / <http://www.ffii.org/>
.... Apple: "I know! Let's call it the Raincoat."