HP8596E Spectrum Analyzer repair

[SilverSolder]

Hard to say what is an artifact of the photography (it is not easy to capture a CRT image accurately) and what is actual blur.

However, it is clear that the CRT is very "tired" looking with ghost text burned into the phosphor (black lettering) from many hours of use.  (Is the text actually from this spectrum analyser?  It looks like the ghost text was burned in a different position on the screen than the current text being displayed.)

As a CRT gets up in the hours used, the emissivity of the cathode falls, and the tube becomes dimmer.  This can be compensated to some extent by cranking up the brightness, (and cranking up the filament voltage), but it will only go so far and isn't really a permanent fix.

[KCT_99]

Well, I did more measurements.

Rails      with CRT connected         with CRT disconnected
+600V              +568V                               +570V
-200V               -138V                                -139V
+60V                +70V                                 +70V

Then just for good measure I replaced R80 also, now 200 Ohm.

And I managed to take video of the event. Please note, during these flareups focus sharpens right up for brief period of time as seen in the lower left on the screen in the video. Also, for some reason, when I do connect the remote brightness pot, these flareups seem to occur more often and more pronounced, it seems. This video was taken with the remote pot connected.

Video download link: (Will be deleted on 14 October, 2019)
https://wetransfer.com/downloads/6dd972e01237a7c9e5999fe48b1be59020191007164902/e16a280e1942552d349837e0de905c7f20191007164902/1661fe

[KCT_99]

Here another attempt to take a reasonable picture of the screen image. I used a magnifier and cell phone camera.
Yes, this analyzer was used substantially, but over all it does seem to be in a very good shape. The burned in image seems 'normal' as I have seen this on other spectrum analyzer CRTs of this HP type.

[james_s]

Do you know anybody with a CRT tester? If you have any arcade collecting groups in your area that might be a good place to start, they're also pretty cheap on ebay as they are a niche thing these days however they tend to be rather bulky so shipping can be expensive.

Can you try monitoring the voltage on the G1 and G2 pins during one of the flares? It's quite possible that you have an internal arc in the electron gun or the flyback causing that. You might try gently tapping on the CRT neck with something like a plastic screwdriver handle and see if you can get it to act up.

The low -200V rail certainly suggests a fault in the circuit somewhere, the other rails are close enough to look pretty reasonable to me but -200V is way low.

[KCT_99]

So I opened up the CRT socket, found nothing indicating trouble. I removed each contact re-soldered it and gently pushed the contact forks to have more bite on the CRT pins when engaging. This seems to have helped noticeably with image sharpness. However the flareups still persist. I had the scope probe 10:1 connected to Pin 2 CRT to view the video signal during a flareup. This signal remains absolutely stable. as seen in the image below. Then I had the scope probe 10:1 connected to Pin 1 CRT, there I can observe an substantial instability during these flareup events, as seen in the video linked below. Average DC is rising erratically.

Videos Will be deleted on 14 October, 2019

https://wetransfer.com/downloads/3d511248784e4a874eb1fcb7fd5974a620191007214850/d4d6f38e229dc06bb70a7342661add3320191007214850/80b18e

Yeah, i tried already gentle tapping the boards, connectors, transformer etc. did not seem to provoke this problem. I tried cold spray as well. It does not seem to impact this instability in any way.

Initially, the -200V rail been only -140V seemed the obvious problem, but at this point it is still inconclusive to me. Alto been quite low this voltage remains stable during these flareup episodes.

[SilverSolder]

The pictures are good.  The video is excellent, that really got the issue across well.  Thankfully, CRTs are their own oscilloscopes!

First, notice how the "glowing shadows" of the text point towards the center of the screen from top left and top right:



These glowing shadows are caused by the electron beam no longer being perfectly round, and it is rotating as it moves across the screen (the shadow points the opposite direction on opposite sides of the screen).  This means we can completely rule out the video amplifier and concentrate on why the CRT is defocused, weak, and performing flare-up tricks for us.


Let's review how this CRT probably works in basic terms to help the investigation:



1.  The filament (3,4) heats the cathode (2) which causes a cloud of electrons to boil off its surface. 

2.  The control grid (1) is shaped like a cup with a pinhole in the bottom facing the screen - the control grid is more negative than the cathode which has the effect of repelling the electrons into a little ball around the center of the cup near the surface of the cathode.  The control grid voltage controls the overall brightness of the screen (beam current).

3. The accelerator grid (6) is a hoop shape that attracts the electrons out through the pinhole with a high positive voltage.  By the time the electrons reach the hoop, they are going so fast that they continue straight on.  Very few electrons actually hit the grid so we don't expect to see much current coming out of pin 6.

4. The electrostatic focus lens consists of three short cylinders:  the pre-accelerator (connected to EHT), the focus control (7), and another accelerator also connected to EHT.  The beam will be precisely focused and moving at very high speed when it comes out the other side.

5. The beam hits the screen, which is conductive and lets the electrons be sucked out of the tube by the EHT connection, which return the electrons to ground to start the cycle all over again. 

6. The tube is coated inside and outside with a conductive graphite material called Aquadag.  The inner coating is at EHT voltage, the outer is at ground.  The glass between the coats acts as the plate of a giant capacitor.  This capacitor smooths the EHT and retains charge for days/weeks after the CRT has been switched off and can give the unwary technician quite a surprise.  This capacitance has to be discharged in order to work on the EHT section safely.


With reference to the explanation above, what could be wrong with the unit?  Well, we can see from the pictures and the video that we have both a focus and a brightness problem simultaneously.  Since the electrostatic focus lens on this particular tube is primarily driven by the EHT,  and since the brightness of the tube is also dependent on EHT, the common denominator is - the EHT supply.  It seems prudent to find a way to measure (or even trace on an oscilloscope) the level of the EHT in order to be able to either confirm or eliminate EHT instability as the root cause.

[tautech]

@SilverSolder
I don't mean to nitpick only offer further clarity to your great post.

Typically in older instruments there are several voltage rails and we refer to those to 60V or so as LV while those say of the CRT plate drivers of ~120+V are the HV.
When we mention EHT it's normally referring to the cathode voltage applied to the CRT and is always negative and often ~-2kV and more.
The beam acceleration voltage is usually derived from the EHT via some uneven stage # multiplier (3 or 5 is most common) but is positive and referred to as PDA (Post Deflection Acceleration) voltage.

This large voltage potential difference, commonly 6, 8, 10 or more kV accelerates the electron beam at speeds required for the particular CRT display refresh rate/BW of the instrument and/or the phosphors used.

[james_s]

It sounds like we are getting terms commonly used for electrostatic deflection CRTs mixed up with those used with magnetic deflection CRTs as this one is. A magnetic deflection CRT has no plates and the cathode operates near ground potential, the EHT is a positive voltage feeding the final anode, in the case of a tube of this sort it will generally be around 6-9kV which almost universally comes directly from a single HV winding on the flyback transformer, half wave rectified by a HV diode which is often internal to the flyback and then filtered by the capacitor formed by the CRT itself. I have not seen the schematic for this particular monitor but the focus and G2 voltages are typically derived from a winding separate from the main EHT winding on the flyback, it's possible that the flyback is arcing internally or it could be a fault in one of the rectifier diodes or other components involved in conditioning these voltages. While the EHT could be fluctuating as I said earlier that *usually* results in an obvious "breathing" effect where the picture changes size at the same time as the brightness is also fluctuating.

Apologies if this has already been mentioned but do any of the flyback-derived voltage rails fluctuate when the picture is doing its thing? I still think the fault could go in one of two directions, one is a circuit fault causing the focus and G2 voltages to fluctuate which causes the effects noted in the picture. The other possibility is an internal fault in the electron gun in the CRT that is dragging these voltages around. A third possibility slightly different than the first is electrical leakage somewhere, I forgot to ask if there is any glue or other contamination on the board? I used to see fairly often a white glue used to reinforce parts on PCBs would turn brown and become slightly conductive. The focus and G2 voltages are very high impedance sources intended to deliver essentially zero current so it doesn't take much leakage to bog them down.

[SilverSolder]

I have probably spent far too much time poking around inside old television sets, my side line as a college student! -  so I am probably using TV terminology.

[SilverSolder]

[...] I have not seen the schematic for this particular monitor but the focus and G2 voltages are typically derived from a winding separate from the main EHT winding on the flyback, [...]

The illustration in my previous post is from the diagram (which is posted earlier in this thread).   This CRT appears to have a somewhat unusual focus and accelerator grid arrangement driven straight from the EHT.

[james_s]

I have probably spent far too much time poking around inside old television sets, my side line as a college student! -  so I am probably using TV terminology.

Your terminology is correct as far as I noticed, this is a magnetic deflection raster monitor so it is essentially a TV minus the tuner and audio stuff. IIRC these even use bog standard NTSC format composite video.

[james_s]

Ok now I found the manual linked earlier with the schematic, having that is very useful.

The G2 and focus are both fed by a 600V rail derived from the primary side of the flyback, this is essentially a boost converter and the HV secondary can be ignored for a moment. This looks very standard to me, it's pretty much the same circuit used by most small raster monitors and portable TV sets that are powered by low voltage DC. So I suspect that 600V rail is fluctuating, this could be the flyback itself, or it could be either of those diodes or the capacitor C23, or it could be something external loading it down. Or it could be an issue with the 200V line, if something is loading that down it will cause other voltages to be wrong. A HV probe would be handy here just as a sanity check on the EHT. Lacking a real one I would probably just make one with a bunch of 1M resistors in a plastic tube, the circuit is simple enough. One need not do any live probing, just slip the end under the anode cup and ground the other end of the divider, hook it up to a multimeter and then power up the monitor while not touching any of that stuff while it's live.

[KCT_99]

Thanks for your help.

Unfortunately, I don't know anybody with a CRT tester. We used to have a TV repair shop in town, but they are long gone.

I had measured the rail voltages during the flareups, the +600V, +60V and -200V rails are stable.

The only rail not yet verified is the HV. For that, as suggested, I will have to make up this large voltage divider.
I see if I can manage to do that tomorrow, got to go resister shopping first. Then I will attempt to make a safe measurement.

I could also disconnect the -200V rail from the rest and see if it recovers.

Also, I will thoroughly clean the board of flux residue. They used white silicone to glue some electrolytic to nearby coil. I guess this is to give the coil mechanically stability. The silicone is only applied on the component side. I see if I can remove the silicone for now.
It appears the silicone was used around L1, C17,C38 and possibly R25.  Also applied to L2, C19 and heat-sink for Q2.

[james_s]

It's sounding to me like it may be a fault in the CRT, if the 600V line is stable and the brightness and focus are surging that doesn't leave much else. Do check the HV though and let's see what that's doing.

If you can find one of those cheap 5" portable B&W TVs you could temporarily swap the tube from that into the monitor. The pinout on small tubes like that is pretty standardized and they are electrically similar enough to work. B&W monitors have no convergence or purity to muck with so swapping the yoke onto the TV tube is not difficult. Conversely you could test the tube you have with the electronics of the TV.

[SilverSolder]

Don't forget to discharge the CRT capacitance...  and remember that it can recover quite a bit of charge after the first discharge, as the Aquadag has high resistance and acts like a big RC filter.

How to discharge:

[KCT_99]

Thanks guys, I will be careful and keeping my left hand in the pocket.

[KCT_99]

Alright, just got back from resistor shopping, but I have had a second though on this while creeping along in my car. So before I go though this HV exercise I do something else.

We know, G1 looks very suspicious as I have measured there the instability while elsewhere there is little to none instability that stands out in any way. And here we are on to something I believe. For fun and gigles, I put 220nF 250V capacitors to ground on both sides of R50. This seems to tame these flareups right down and gives almost a constant brightness but without much control over it. With the Sharpness I could kind of live with altho I would want the image to be sharp.

Please bear with me here for a second as I explain my theory. I checked on ebay, and found two such CRT monitors offered, both seem to have noticeable screen burn-in's as bad as mine. Sure this could just mean operators cranked up the brightness with little regard to CRT life, but I suspect a design flaw here. I believe many of these CRT of this type suffer from the same issue. Someone suggested to look for contamination that could offer a creep path for high voltage I looked the CRT socket as suggested and as many have found problems there, at least on some TV. So at the CRT socket life is good, but what about on the other end of the connection ?

The thing that kept me up at last night is that this thing flares up in brightness not down, this means G1 gets more positive. Confirmed as seen in the video I took of G1. Everything is stable at some DC on G1 and then suddenly I see the G1 rise erratically just to settle back down to about the same DC for a while and then this spectacle repeats.

I suspect intermittent creep currents but not inside the CRT tube, but outside of it.

Looking at the PCB diagram, G1 (GRN) is located right next to G2 (RED) on this PCB, perhaps separated by about 4mm or even less than that. Other CRT connection are clustered in very close proximity there as well, except Yellow, the video signal on the Cathode. This signal is far away on the opposite side of the board and well clear of anything nearby.
So G2 (RED) sitting neatly at a healthy 567V just 4 mm away from G1 (GRN)....
Hmmmm....

I am not sure what the Layout guy was thinking when squishing these signals this tightly together and on top of that running a jumper J1 (G1) half across the board when he could have conveniently have G1 (GRN) located at the other end of J1. Saving himself the trouble of putting the jumper, using up board space while easing up the CRT connection congestion.
Hmmmm....
Must have been a bad Monday, I guess....

My theory is that, I might have creep right there between GRN and RED, either on the component side or on the bottom. I suspect on the component side, its very crowded an the wires leading to the CRT socket have a little contact crimped to them which is then soldered into the board.

To confirm this theory I will dismantle this entire assembly to get good access to clean this board thoroughly, I will then apply something to add insulation to these connections, I am thinking using my wife's finger nail lacquer or perhaps some silicone. Then after reassembly I have a strong feeling the flareups will be gone.

The more I look at this exceptionally terrible layout design, the more I lean towards my theory.

What do you guys think ?

Does this sound reasonable for a theory ?

[SilverSolder]

The theory sounds relatively easy to check.  Maybe lift the jumper and solder the wire to the end of it?

In an earlier post, you said:  " I found that during a flareup the voltage drop over the 1k R59 increases from about 50mV to about 350mV. I had a meter directly on the CRT pin 1 and there the voltage changes from about -50V to +30V during a flareup."

If R59 is being driven from the side away from the CRT (rather than the CRT driving it), your theory begins to look very good!

[tautech]

I've had CRO's with unstable CRT performance and generally replacing all high value (1M+) resistors and HV caps fixes it. The resistors drift high with age and caps leak and do crazy things with HV on them.
Very occasionally a HV zener or HV diode can be a problem too.

Don't lose sight that much of the higher voltage components are under some stresses and BTW 4mm clearance for 600V is plenty.

[james_s]

CROs have electrostatic deflection tubes using vector deflection, this is a magnetic deflection tube with raster deflection. The circuit designs typically used differ enough that I rarely see the same sort of problems with one as I do with the other. Most raster monitor problems I've encountered are failed electrolytic capacitors, bad solder joints, occasionally shorted semiconductors, and internal faults in the CRT. Flybacks occasionally fail but most of the time when that happens it pops the horizontal output transistor and the monitor goes dead, or it visibly arcs or has no HV output at all. It's been very rare that I've found any faulty resistors in a raster monitor but obviously it's worthwhile to check.

Screen burn is normal in this sort of instrument, they have a lot of static text and graphics on the display whenever they are powered up, and they have typically spent many hours powered up, it's the nature of the beast. Really, really heavily burned tubes are usually pretty tired but for the more moderate ones it's a crap shoot.

[KCT_99]

Well, in short, my theory did not work out. The flareups still persists. 

I took this thing apart again and throughly cleaned the PCBs, noticing they were not properly cleaned when manufactured, virtually every solder joint had some flux residue even on the component side.

I rewired G1 and G2 such that they are separated well from each other, I also put nail lacquer on the higher voltage part leads and solder joints.

So now I am back to square one. I guess now I have no further excuse to avoid messing with the HV. Altho I don't think this is a HV problem.

I was wondering why would they do a 270k R50 with a 1k R59 in series, while the 1k is a fat one. It does not make sense to me at all at this point. For the time been, I put a 2k and another 2k in series, so I have now 270k + 2k + 2k. In between the 2k's I put a 220nF to ground. This tames the flareups a little bit and slows things down altho it does nothing about the absolute brightness variation. Brightness is all over the place, from very dim to somewhat normal and then super ultra bright so much so I worry I could burn something out. For brief periods the image is quite pleasant and acceptable.

I tried to locate a CRT but no luck. It could be the company is long gone.

BWL 779009 MTI Monitronics

Coming to the realization there is not much more I could do, sucks.

Next thing to do, might be taking some money and sniping a CRT assembly of ebay. 

.

[james_s]

I suspect you've got an intermittent fault in the CRT itself, but I wouldn't be too quick to condemn it. Can't you find one of those cheap 5" portable B&W TVs? I've picked up quite a few of them at garage sales and thrift stores for a few dollars. I use them as monitors for various things and convert them to vector monitors but they would be handy for testing tubes like yours too.

Might not even be too hard to find a tube that is compatible with your monitor and the correct size. CRTs are very easy to sub in many cases, there are two common neck sizes, beyond that most of the differences are in the physical mounting details. There are three common phosphors used in small monochrome tubes but for your application it doesn't really matter, a white or amber phosphor would probably look just as good.

If it does turn out to be a tube fault, there is a reasonable chance you can blow out the short, my CRT tester uses a charged capacitor across various pins to do this.

[SilverSolder]

Once you eliminate the impossible, whatever remains, no matter how improbable, must be the truth.
-- Arthur Conan Doyle

Going back to first principles, no significant current should flow into (or out of) any of the grids of a CRT, it is a voltage controlled device.

The fact that you measure such a high current for the control grid during the flare-up is a red flag and could point to some kind of leakage issue inside the CRT.

Still, it would be irritating if it turned out to be a bad HV rectifier?  The last non-eliminated possibility...

In your last picture, the "glowing shadows" are pointing the opposite way from the earlier one -  indicating a vastly different focus voltage or other major change of the tube operating point...

[SilverSolder]

This article on all the ways a flyback transformer can fail (including causing blinking power) may be of interest:

http://www.electronicrepairguide.com/testing-flyback-transformer.html

[KCT_99]

Thanks for your help.

Well, at this point, I am not sure what to do further. Yes, I could test the HV on the Transformer, but I realize if I do need a new transformer or a new CRT they are equally hard to impossible to get. So what are my options at this point ?

If there is something funny going on in this tube, I suspect then between G1 and G2.  As last ditch attempt, could I use the 600V rail to attempt brutally blowing out a short ? I think I have some HV capacitors laying around somewhere.

If I understand correctly, I would unplug the CRT socket and temporarily connect G1 to ground and then hit G2 with 600V and a couple micro farads in hopes the short blow open and not welding short short ?
What about the resulting metal vapor inside the tube that might be generated in this attempt, would that not make things for creep much worse ?

I think, perhaps, before attempting that, I should measure on G1 and G2 again. If I measure G1 to G2, I would expect some voltage around 500V and during a flareup this voltage should drop. Also, measuring simultaneously on G1 again this voltage should rise.
Perhaps adding a third meter on G2 which should also slightly drop during the flare up event.

Would this be enough proof for a faulty CRT and for the blow out attempt ?

If all fails I might have to eat it and buy a well used spare display with potentially the same issue, this sucks.