Topics

454 HV regulation off

Jack Ohme
 

I created a thread a few days ago about my 454, and believe I have isolated
the issue, in its -1960v supply. The HV reg is totally off, which is
causing my grid voltage to be far too high, and as an effect, my screen
brightness as well. The LV regulation is spot-on, and cranking the HV
regulation control pot, R1401, mentioned on page 6-11 of the manual, to the
max and then dropping it to the minimum causes my screen and trace to drop
to an acceptably low level for a moment before returning to far too bright.
Any idea as to what could be causing this? Something in the actual HV
section, or something in the LV reg that controls the HV? Oh and to put a
number on my -1960v test point, its hovering around -1830 currently.

KB6NAX
 

You have a bunch of 3 megohm resistors in series in the feedback path from the CRT cathode to the oscillator control Q1414. Check the resistor values and sub the .001 cap. Yep, ceramics do go bad. -Arden

Albert Otten
 

Hi Jack,

It seems that you didn't read my message https://groups.io/g/TekScopes/message/162916 .

When the HV is "motorboating" strange effects can happen. But as long as the cathode HV is off spec because of a regulator fault but more or less stable, it won't cause your too bright trace. The grid HV will be off by about the same amount.

Albert

On Thu, Jan 9, 2020 at 01:51 AM, Jack Ohme wrote:


I created a thread a few days ago about my 454, and believe I have isolated
the issue, in its -1960v supply. The HV reg is totally off, which is
causing my grid voltage to be far too high, and as an effect, my screen
brightness as well. The LV regulation is spot-on, and cranking the HV
regulation control pot, R1401, mentioned on page 6-11 of the manual, to the
max and then dropping it to the minimum causes my screen and trace to drop
to an acceptably low level for a moment before returning to far too bright.
Any idea as to what could be causing this? Something in the actual HV
section, or something in the LV reg that controls the HV? Oh and to put a
number on my -1960v test point, its hovering around -1830 currently.

Roger Evans
 

The manual shows the voltage that should be present at the input to the HV regulator (R1414 - R1402, marked L on the board connector) as -11.6V. If this measures OK then the regulator is probably working as it should and the error is in the divider chain. You would need to check R1404, less likely R1403 and C1403, as well as the 3M resistors.

The cathode voltage is only off by about 7% and as Albert says the grid voltage will be off by a similar percentage so it shouldn't account for your symptoms. I have had to replace transistors in the Z axis amplifier of both my 454s and I would strongly suspect that the real culprit lies therein.

Regards,

Roger

Albert Otten
 

Hi Roger,

When the HV is working (almost) normally a Z-axis fault can never increase the positive voltage at TP1349 far enough to make the grid voltage over 150 V positive w.r.t. cathode. (150-180 V measured by Jack.)

Albert

On Thu, Jan 9, 2020 at 12:33 PM, Roger Evans wrote:


---
The cathode voltage is only off by about 7% and as Albert says the grid
voltage will be off by a similar percentage so it shouldn't account for your
symptoms. I have had to replace transistors in the Z axis amplifier of both
my 454s and I would strongly suspect that the real culprit lies therein.

KB6NAX
 

Pardon my ignorance but looking at the schematic the Z axis input to the CRT is capacitively coupled via C1479. Disconnect C1477 to isolate the Z axis input. There are three inputs to current summing junction R1403 - R1410. You won't see much voltage change when adjusting R1401. Better to look at TP1415 for correct operation of Q1414 and subsequently the voltages at Q1413 and Q1423. If all responds properly to adjustment of R1401 the problem is in the feed back string from the CRT cathode, assuming C1403, C1408 and C1469 are all OK, As mentioned previously high value carbon comp resistors can drift far out of tolerance. Open the resistor strings so that each resistor can be accurately ohmed. If any are more than 10% out of tolerance replacing all of them with new would be my first choice. Metal film or metal oxide film resistors would be fine replacements. If all other CRT element voltages are problem free the CRT is suspect. My 2 cents -Arden

KB6NAX
 

BTW, the mention of "Z axis fault" and "Z axis amplifier" has been referring to the unblanking circuit on the Z axis board. Tek calls the circuitry the "Z Axis Amplifier." That it ain't, IMO. It should be called the "Unblanking Amplifier." But what do I know......? -Arden

EB4APL
 

I think that this is not totally correct, if the Z Axis signal was AC coupled only, the intensity control (and the external Z Axis modulation) would not work.

As I understand the schematic correctly, the AC component of the Z Axis signal is coupled via C1449, but there is a DC component that is fed through the grid bias power supply, because this is not referred to ground but to the Z Axis amplifier output, so when Q1343 emitter changes its voltage, the CRT grid voltage varies accordingly.  The Q1343 output is formed by the combination of blanking signals, the intensity control and the external Z Axis input.

Regards,

Ignacio EB4APL

El 09/01/2020 a las 19:53, KB6NAX escribió:
Pardon my ignorance but looking at the schematic the Z axis input to the CRT is capacitively coupled via C1479. Disconnect C1477 to isolate the Z axis input. There are three inputs to current summing junction R1403 - R1410. You won't see much voltage change when adjusting R1401. Better to look at TP1415 for correct operation of Q1414 and subsequently the voltages at Q1413 and Q1423. If all responds properly to adjustment of R1401 the problem is in the feed back string from the CRT cathode, assuming C1403, C1408 and C1469 are all OK, As mentioned previously high value carbon comp resistors can drift far out of tolerance. Open the resistor strings so that each resistor can be accurately ohmed. If any are more than 10% out of tolerance replacing all of them with new would be my first choice. Metal film or metal oxide film resistors would be fine replacements. If all other CRT element voltages are problem free the CRT is suspect. My 2 cents -Arden

KB6NAX
 

To be precise, the circuits on the Z axis board provide intensity control and unblanking. There is no provision for intensity modulation such as TV video via that circuit. The rear panel AC coupled Z axis input to the CRT cathode provides intensity modulation from an external source. OK, picky, picky.... -Arden

Jack Ohme
 

Ok Roger, I've been meter checking around the transistors on the Z axis
board, like you said may be bad, as well as checking points mentioned by
Arden, and I've discovered something interesting. Z axis board output P
(connected to emitter of Q1423) is supposed to be at around -4.35v, when
really its at about -3.9v. It goes up to about -4.5v when I adjust the HV
adjustment pot, remember what I said earlier about the screen going to an
acceptable level when I tweaked that pot? But anyway, after a glorious
second, it returns back to around -3.9v. :(
The ohmmeter measures between point L (at the top of the chain of 3M
resistors), and TP1349 (the -1960v test point) at around 1.2Mohm, but
across C1408 its 10M so I'm fairly sure it ain't a leaky cap. Strangely
enough, there are 1.2Mohms worth of resistors between point L and +75...
but I think this is all beyond the point. Measuring large resistances in
such a complex circuit is worthless. I could pull the string and measure,
but frankly I'm not so sure that's the issue. I feel like the strange
behavior would be more appropriate of a bad cap or transistor. I will keep
looking for any bad transistors on the board, but so far, measurements seem
alright besides the bad one on Q1423. I could of course replace it, but I'd
like to be sure before suspending my work for 5 days for an eBay order,
haha.

On Thu, Jan 9, 2020 at 5:45 PM KB6NAX <gumbear@...> wrote:

To be precise, the circuits on the Z axis board provide intensity control
and unblanking. There is no provision for intensity modulation such as TV
video via that circuit. The rear panel AC coupled Z axis input to the CRT
cathode provides intensity modulation from an external source. OK, picky,
picky.... -Arden



Jack Ohme
 

I was going to inspect the resistor chain individually, but that would
involve desoldering of the entire HV section. As much as I love the rest of
this unit, hiding a bunch of stuff under a panel of soldered components
seems like a poor move on Tek's part.

On Thu, Jan 9, 2020 at 9:17 PM Jack Ohme via Groups.Io <machinamancerjack=
gmail.com@groups.io> wrote:

Ok Roger, I've been meter checking around the transistors on the Z axis
board, like you said may be bad, as well as checking points mentioned by
Arden, and I've discovered something interesting. Z axis board output P
(connected to emitter of Q1423) is supposed to be at around -4.35v, when
really its at about -3.9v. It goes up to about -4.5v when I adjust the HV
adjustment pot, remember what I said earlier about the screen going to an
acceptable level when I tweaked that pot? But anyway, after a glorious
second, it returns back to around -3.9v. :(
The ohmmeter measures between point L (at the top of the chain of 3M
resistors), and TP1349 (the -1960v test point) at around 1.2Mohm, but
across C1408 its 10M so I'm fairly sure it ain't a leaky cap. Strangely
enough, there are 1.2Mohms worth of resistors between point L and +75...
but I think this is all beyond the point. Measuring large resistances in
such a complex circuit is worthless. I could pull the string and measure,
but frankly I'm not so sure that's the issue. I feel like the strange
behavior would be more appropriate of a bad cap or transistor. I will keep
looking for any bad transistors on the board, but so far, measurements seem
alright besides the bad one on Q1423. I could of course replace it, but I'd
like to be sure before suspending my work for 5 days for an eBay order,
haha.

On Thu, Jan 9, 2020 at 5:45 PM KB6NAX <gumbear@...> wrote:

To be precise, the circuits on the Z axis board provide intensity control
and unblanking. There is no provision for intensity modulation such as
TV
video via that circuit. The rear panel AC coupled Z axis input to the
CRT
cathode provides intensity modulation from an external source. OK,
picky,
picky.... -Arden





Albert Otten
 

Hi Jack,

On Fri, Jan 10, 2020 at 06:17 AM, Jack Ohme wrote:


Ok Roger, I've been meter checking around the transistors on the Z axis
board, like you said may be bad, as well as checking points mentioned by
Arden, and I've discovered something interesting. Z axis board output P
(connected to emitter of Q1423) is supposed to be at around -4.35v, when
really its at about -3.9v. It goes up to about -4.5v when I adjust the HV
adjustment pot
This is all normal behavior. How else do you think the regulator could respond to changes in required HV voltage?

, remember what I said earlier about the screen going to an
acceptable level when I tweaked that pot? But anyway, after a glorious
second, it returns back to around -3.9v. :(
This is also normal behavior. Have a look at the RC-time in the grid HV circuit. When you quickly turn the HV down (less negative) with R1401, the cathode voltage nearly follows immediately, but the grid voltage follow slowly with a time constant about 1.5 s (ignoring grid current effects).

The ohmmeter measures between point L (at the top of the chain of 3M
resistors), and TP1349 (the -1960v test point)
you mean the Z-axis test point

at around 1.2Mohm, but
across C1408 its 10M so I'm fairly sure it ain't a leaky cap. Strangely
enough, there are 1.2Mohms worth of resistors between point L and +75...
but I think this is all beyond the point. Measuring large resistances in
such a complex circuit is worthless. I could pull the string and measure,
but frankly I'm not so sure that's the issue. I feel like the strange
behavior would be more appropriate of a bad cap or transistor.
For these resistor measurement +75V and -12V can be considered as grounded. So measured 1.2 M is perfectly OK.
The 10 M measured across C1408 is in the ballpark of the two parallel resistor strings for HV and focus.

I will keep
looking for any bad transistors on the board, but so far, measurements seem
alright besides the bad one on Q1423. I could of course replace it, but I'd
like to be sure before suspending my work for 5 days for an eBay order,
haha.
No indication yet that Q1423 would be bad.

May I ask you once more if you read my previous messages?

Albert

Jack Ohme
 

Albert, I've read all of your messages, I've just been attempting to rule
everything out before having to tear into the HV supply, though at this
point I suppose its necessary. Wonderful.

On Fri, Jan 10, 2020 at 8:05 AM Albert Otten <aodiversen@...> wrote:

Hi Jack,

On Fri, Jan 10, 2020 at 06:17 AM, Jack Ohme wrote:


Ok Roger, I've been meter checking around the transistors on the Z axis
board, like you said may be bad, as well as checking points mentioned by
Arden, and I've discovered something interesting. Z axis board output P
(connected to emitter of Q1423) is supposed to be at around -4.35v, when
really its at about -3.9v. It goes up to about -4.5v when I adjust the HV
adjustment pot
This is all normal behavior. How else do you think the regulator could
respond to changes in required HV voltage?

, remember what I said earlier about the screen going to an
acceptable level when I tweaked that pot? But anyway, after a glorious
second, it returns back to around -3.9v. :(
This is also normal behavior. Have a look at the RC-time in the grid HV
circuit. When you quickly turn the HV down (less negative) with R1401, the
cathode voltage nearly follows immediately, but the grid voltage follow
slowly with a time constant about 1.5 s (ignoring grid current effects).

The ohmmeter measures between point L (at the top of the chain of 3M
resistors), and TP1349 (the -1960v test point)
you mean the Z-axis test point

at around 1.2Mohm, but
across C1408 its 10M so I'm fairly sure it ain't a leaky cap. Strangely
enough, there are 1.2Mohms worth of resistors between point L and +75...
but I think this is all beyond the point. Measuring large resistances in
such a complex circuit is worthless. I could pull the string and measure,
but frankly I'm not so sure that's the issue. I feel like the strange
behavior would be more appropriate of a bad cap or transistor.
For these resistor measurement +75V and -12V can be considered as
grounded. So measured 1.2 M is perfectly OK.
The 10 M measured across C1408 is in the ballpark of the two parallel
resistor strings for HV and focus.

I will keep
looking for any bad transistors on the board, but so far, measurements
seem
alright besides the bad one on Q1423. I could of course replace it, but
I'd
like to be sure before suspending my work for 5 days for an eBay order,
haha.
No indication yet that Q1423 would be bad.

May I ask you once more if you read my previous messages?

Albert



EB4APL
 

The rear panel Z Axis input is specified as DC coupled, usable from DC to 50 MHz. So this could be used for TV video using suitable sweeps in horizontal and vertical amplifiers.

I believe that the extra AC coupling provided by C1449 is mainly as a means of obtaining fast unblanking, because the unblanking signal must have fast transitions or you will miss the beginning of the trace at fast sweeps. The stray capacitance of the grid HV power supply degrades the bandwidth so providing a direct path solves this, and also gives good bandwidth for the external Z Axis input.

In the context of the thread, I think that temporarily removing C1449 will not affect the Intensity signal continuity to the grid.

Ignacio, EB4APL

El 10/01/2020 a las 2:45, KB6NAX escribió:
To be precise, the circuits on the Z axis board provide intensity control and unblanking. There is no provision for intensity modulation such as TV video via that circuit. The rear panel AC coupled Z axis input to the CRT cathode provides intensity modulation from an external source. OK, picky, picky.... -Arden

Albert Otten
 

On Fri, Jan 10, 2020 at 05:31 PM, Jack Ohme wrote:


Albert, I've read all of your messages, I've just been attempting to rule
everything out before having to tear into the HV supply, though at this
point I suppose its necessary. Wonderful.
Hi Jack,

Sorry, I don't want to hurry you up, but I'm very curious to see if my suspect of a bad R1447 was right. Did you check that this pot is not interrupted (internally) at the C1440 end of the carbon trace? If the connection to C1440 end is broken then there is at least 7.5 M between the C1440 and the grid, reducing the grid voltage considerable in positive direction. The pot is only 2M so easy to check with a DMM.
I understood from your findings that the grid is positive w.r.t. the cathode, right?

Albert

Jack Ohme
 

Albert, the R1447 measures from 0 to 1.28M from wiper to the C1440 end, and
1.28M from end to end (Pot is supposed to be 2M according to schema). Since
I already cracked open the HV, I decided to measure the feedback resistors
as well. The entire string (3M x 8) measured 11M, but each resistor
measures within spec, so I can only assume that the 11M is just a path
somewhere else, probably normal. Apologies for my lack of knowledge and
thank y'all for bearing with me, this is the first scope I've ever worked
on, so I'm new to the theory and such. Anyway, could that lower resistance
be a bad R1447, and if so, could resistance loss in 1447 even cause my
problem? Or is this just parasitics/other resistances.

On Fri, Jan 10, 2020 at 10:43 AM Albert Otten <aodiversen@...>
wrote:

On Fri, Jan 10, 2020 at 05:31 PM, Jack Ohme wrote:


Albert, I've read all of your messages, I've just been attempting to rule
everything out before having to tear into the HV supply, though at this
point I suppose its necessary. Wonderful.
Hi Jack,

Sorry, I don't want to hurry you up, but I'm very curious to see if my
suspect of a bad R1447 was right. Did you check that this pot is not
interrupted (internally) at the C1440 end of the carbon trace? If the
connection to C1440 end is broken then there is at least 7.5 M between the
C1440 and the grid, reducing the grid voltage considerable in positive
direction. The pot is only 2M so easy to check with a DMM.
I understood from your findings that the grid is positive w.r.t. the
cathode, right?

Albert



Jack Ohme
 

Also HV diode D1440 measured 8M across with my meter and 10M in reverse. Is
this just because of the low voltage in my meter, or is this a bad diode?

On Fri, Jan 10, 2020 at 3:59 PM Jack Ohme via Groups.Io <machinamancerjack=
gmail.com@groups.io> wrote:

Albert, the R1447 measures from 0 to 1.28M from wiper to the C1440 end, and
1.28M from end to end (Pot is supposed to be 2M according to schema). Since
I already cracked open the HV, I decided to measure the feedback resistors
as well. The entire string (3M x 8) measured 11M, but each resistor
measures within spec, so I can only assume that the 11M is just a path
somewhere else, probably normal. Apologies for my lack of knowledge and
thank y'all for bearing with me, this is the first scope I've ever worked
on, so I'm new to the theory and such. Anyway, could that lower resistance
be a bad R1447, and if so, could resistance loss in 1447 even cause my
problem? Or is this just parasitics/other resistances.

On Fri, Jan 10, 2020 at 10:43 AM Albert Otten <aodiversen@...>
wrote:

On Fri, Jan 10, 2020 at 05:31 PM, Jack Ohme wrote:


Albert, I've read all of your messages, I've just been attempting to
rule
everything out before having to tear into the HV supply, though at this
point I suppose its necessary. Wonderful.
Hi Jack,

Sorry, I don't want to hurry you up, but I'm very curious to see if my
suspect of a bad R1447 was right. Did you check that this pot is not
interrupted (internally) at the C1440 end of the carbon trace? If the
connection to C1440 end is broken then there is at least 7.5 M between
the
C1440 and the grid, reducing the grid voltage considerable in positive
direction. The pot is only 2M so easy to check with a DMM.
I understood from your findings that the grid is positive w.r.t. the
cathode, right?

Albert





Jack Ohme
 

Yeah I'm an idiot, the diode just has a big voltage drop/fwd voltage. Also
as Roger mentioned, I measured R1404 to be about .85M, .15 away from its
spec of 1M.

On Fri, Jan 10, 2020 at 4:05 PM Jack Ohme via Groups.Io <machinamancerjack=
gmail.com@groups.io> wrote:

Also HV diode D1440 measured 8M across with my meter and 10M in reverse. Is
this just because of the low voltage in my meter, or is this a bad diode?

On Fri, Jan 10, 2020 at 3:59 PM Jack Ohme via Groups.Io <machinamancerjack=
gmail.com@groups.io> wrote:

Albert, the R1447 measures from 0 to 1.28M from wiper to the C1440 end,
and
1.28M from end to end (Pot is supposed to be 2M according to schema).
Since
I already cracked open the HV, I decided to measure the feedback
resistors
as well. The entire string (3M x 8) measured 11M, but each resistor
measures within spec, so I can only assume that the 11M is just a path
somewhere else, probably normal. Apologies for my lack of knowledge and
thank y'all for bearing with me, this is the first scope I've ever worked
on, so I'm new to the theory and such. Anyway, could that lower
resistance
be a bad R1447, and if so, could resistance loss in 1447 even cause my
problem? Or is this just parasitics/other resistances.

On Fri, Jan 10, 2020 at 10:43 AM Albert Otten <aodiversen@...>
wrote:

On Fri, Jan 10, 2020 at 05:31 PM, Jack Ohme wrote:


Albert, I've read all of your messages, I've just been attempting to
rule
everything out before having to tear into the HV supply, though at
this
point I suppose its necessary. Wonderful.
Hi Jack,

Sorry, I don't want to hurry you up, but I'm very curious to see if my
suspect of a bad R1447 was right. Did you check that this pot is not
interrupted (internally) at the C1440 end of the carbon trace? If the
connection to C1440 end is broken then there is at least 7.5 M between
the
C1440 and the grid, reducing the grid voltage considerable in positive
direction. The pot is only 2M so easy to check with a DMM.
I understood from your findings that the grid is positive w.r.t. the
cathode, right?

Albert







Harvey White
 

some HV diodes are made of stacks of lower PIV diodes.....

I'd expect the forward voltages to add.......

Harvey

On 1/10/2020 7:27 PM, Jack Ohme wrote:
Yeah I'm an idiot, the diode just has a big voltage drop/fwd voltage. Also
as Roger mentioned, I measured R1404 to be about .85M, .15 away from its
spec of 1M.

On Fri, Jan 10, 2020 at 4:05 PM Jack Ohme via Groups.Io <machinamancerjack=
gmail.com@groups.io> wrote:

Also HV diode D1440 measured 8M across with my meter and 10M in reverse. Is
this just because of the low voltage in my meter, or is this a bad diode?

On Fri, Jan 10, 2020 at 3:59 PM Jack Ohme via Groups.Io <machinamancerjack=
gmail.com@groups.io> wrote:

Albert, the R1447 measures from 0 to 1.28M from wiper to the C1440 end,
and
1.28M from end to end (Pot is supposed to be 2M according to schema).
Since
I already cracked open the HV, I decided to measure the feedback
resistors
as well. The entire string (3M x 8) measured 11M, but each resistor
measures within spec, so I can only assume that the 11M is just a path
somewhere else, probably normal. Apologies for my lack of knowledge and
thank y'all for bearing with me, this is the first scope I've ever worked
on, so I'm new to the theory and such. Anyway, could that lower
resistance
be a bad R1447, and if so, could resistance loss in 1447 even cause my
problem? Or is this just parasitics/other resistances.

On Fri, Jan 10, 2020 at 10:43 AM Albert Otten <aodiversen@...>
wrote:

On Fri, Jan 10, 2020 at 05:31 PM, Jack Ohme wrote:

Albert, I've read all of your messages, I've just been attempting to
rule
everything out before having to tear into the HV supply, though at
this
point I suppose its necessary. Wonderful.
Hi Jack,

Sorry, I don't want to hurry you up, but I'm very curious to see if my
suspect of a bad R1447 was right. Did you check that this pot is not
interrupted (internally) at the C1440 end of the carbon trace? If the
connection to C1440 end is broken then there is at least 7.5 M between
the
C1440 and the grid, reducing the grid voltage considerable in positive
direction. The pot is only 2M so easy to check with a DMM.
I understood from your findings that the grid is positive w.r.t. the
cathode, right?

Albert





Albert Otten
 

On Sat, Jan 11, 2020 at 12:58 AM, Jack Ohme wrote:


Albert, the R1447 measures from 0 to 1.28M from wiper to the C1440 end, and
1.28M from end to end (Pot is supposed to be 2M according to schema). Since
I already cracked open the HV, I decided to measure the feedback resistors
as well. The entire string (3M x 8) measured 11M, but each resistor
measures within spec, so I can only assume that the 11M is just a path
somewhere else, probably normal.
Apologies for my lack of knowledge and
thank y'all for bearing with me, this is the first scope I've ever worked
on, so I'm new to the theory and such. Anyway, could that lower resistance
be a bad R1447, and if so, could resistance loss in 1447 even cause my
problem? Or is this just parasitics/other resistances.
Hi Jack,

The 454 cabinet is so easy to open, but I forget that part of the HV circuits is buried in the HV-box. I saw that at least the anode lead has to be unsoldered; probably even more than that. (In the older 5xx scopes access usually is much simpler.) So that's what you mean with "cracked" I guess.

My idea of an interrupted carbon trace in R1447 can be rejected now. You results are strange though. Because of twice 1.28M the wiper seems to be in the extreme upper (in schematic) end. But the pot resistance normally never gets smaller and the parallel resistance is many M. Strange. What happens with the resistance when you turn the wiper from end to end? If the pot resistance really decreased for some reason then the grid would get more negative.

11 M is OK.

Also HV diode D1440 measured 8M across with my meter and 10M in reverse. Is
this just because of the low voltage in my meter, or is this a bad diode?
Probably low voltage (see Harvey's message). However, an open D1440 would be my next candidate to be the culprit. Hence as a guide I measured the diode types with my 576 curve tracer.
152-0192-00 visibly starts to conduct at 4.5 V, reaches 0.2 mA at 5 V and 0.8 mA at 6 V. (D1440 and D1452.) The spec is < 13 V at 50 mA.
152-0218-00 starts at 7.5 V, reaches 0.1 mA at 10 V. (Multiplier diodes.) Spec < 25 V at 20 mA.
Just feed D1440 from a 9V battery in series with a 22k resistor and measure the voltage across the diode; should be about 5 V or so. If 9 V the diode is probably open (confirm with larger series resistance and if you have higher voltage source).
When D1440 is open the grid will become positive w.r.t. the cathode. The voltage difference will be limited either by the grid current (through the high resistance chain) or the neon bulbs.

Albert