bent neck CRT pins...

[eccerr0r]

Re: bent neck CRT pins...

Oh cool... hmm... Disconnection experiments:
I disconnected all of the secondaries (all three HV, Cathode, Focus) ... still shorted - the 12V bulb is still bright(!)
I then disconnected the feedback circuitry (the op amp output that connects to the primary) (along with the CRT of course)... Short disappears..

I hooked up a 6V bulb to the filament wiring and connected the disconnected feedback winding to the low supply through a 5.6K ohm resistor (the existing circuit does this through a 2K resistor from the output of the 1458 op amp)... it started oscillating and the bulb connected to the filament output LIT UP! and I measured fairly high AC voltage on two of the secondary outputs...

This still makes no sense... the coil is weak and cannot oscillate at full power any more with the op amp hitting the rail perhaps? Though it is still strong enough to pass enough power to light the lamp if 'weakly driven'?

[stj]

Re: bent neck CRT pins...

i was going to suggest replacing any analog chips incase they are out of spec actually.
maybe the gain is wrong.
i have had to replace op-amps in stuff a lot.

[eccerr0r]

Re: bent neck CRT pins...

Wow... Yeah I was thinking... how could it be, I've never seen any op amp fail before (this would be a LM1458 and would be a first). But there's not really much circuitry here and why would Tek design it such that it would kill fuses if that op amp fails.

At least I think I have a bunch of these in my spare parts collection along with LM741s that I don't think I'll ever use...Some of them may even be brand new.

Now whether to add a socket or not... past experience is to never put sockets into a power supply...

Darn. Just checked my stash. Could find just two 1458's, a few 358's, a few lm6142's, and a boatload of TS512's and LM741's.

[stj]

Re: bent neck CRT pins...

i have had to replace 1458's made by JRC in Harmon-Kardon pc speakers a few times.

i use sockets - BUT ONLY TURNED-PIN / MACHINED-PIN TYPES.
non of those pressed contact crap - the chips slowly work loose in those!!

[eccerr0r]

Re: bent neck CRT pins...

Hmm.... doing some math. I suspect large spikes could cause damage to the op amp, but the way I calculate, it seems the cathode voltage needs to hit almost -5000 volts to stress the op amp negative feedback input. However don't know about how many spikes needed on the op amp output to do damage coming back inwards (flyback on the feedback coil). The resistors should dissipate most of the spikes however.

I still doubt this op amp failed...

BTW I saw one webpage that using an op amp as a comparator is "bad" and could damage op amps prematurely. I have my doubts in that but can anyone confirm/disprove that? Not going to link that page because then Google will start taking it as gospel.

[stj]

Re: bent neck CRT pins...

could be true, there are chips intended as comparaters for a reason.
stuff like the lm339

what concerns me about designs like yours is the potential EM field around the pack and if it can effect nearby semiconductors.

i remember a philips tv in the past that kept corrupting it's eeprom.
philips eventually found it was NOT a bad batch of chips but the EM field from the scancoils above it.
they issued a repair kit that was a tin shield with double-sided tape on it to stick it over the eeprom and mcu and it had an earth wire soldered to the corner!

[eccerr0r]

Re: bent neck CRT pins...

I always thought those comparators were such as they do not even try to be linear, otherwise they were the same as op amps. Weird. I still don't trust that web page that I saw yet, and still think that 1458 is still good...alas I'll swap anyway because I have the part and will never design with these crappy 1458's anyway (rather use 358 or the boatload of TS512's I have... )

[R_J]

Re: bent neck CRT pins...

I was thinking that if the op amp you think is bad is whats listed as U1956b Its possible R1951 or more likely C1951 are open, If that cap. is open I don't believe the transfomer will operate, seems similar to some of the Samsung power supplies where the return cap goes bad and the power supply will not give any voltage. and if the cap was going bad it may have placed too high a spike nocking out the op amp.
Just a thought.

[stj]

Re: bent neck CRT pins...

hmm polarised cap - is it electrolytic, or tantalum?

[eccerr0r]

Re: bent neck CRT pins...

Yes indeed C1951 could be bad and may be the reason for backdumping spikes. R1951 is still good (measured with ohmmeter) but C1951 does have ESR=~4 ohms, which I think is OK for a capacitor that small. It's currently a tantalum I think, then again I'm not sure. I think it's too small to be an aluminum electrolytic so I have to assume tantalum.

It's one of those silver colorred ones that's the size of a 1/4W resistor. In the earlier pictures it's that silver thing just to the left bottom of the "flyback". The 22 ohm resistor right next to it is R1951.

I just ripped out U1956 and found a machine pin socket for it. I hate desoldering double sided boards, even with a vacuum pump rework station (it still looks ok I think.)

I guess looking what I did: I pulled out R1971 and R1952 as a test and hooked up the + of C1951 to the collector of Q1981 through a 5.6K ohm resistor from my junk pile. Not a great negative supply but with that connection I was able to get it to oscillate and get AC output on all of the "flyback" windings.

And whoops I just realized with it hooked up low to -15V I may have reverse connected that capacitor. I didn't try simply shorting that pin to GND to see if it would turn on, then again I'm not sure if that's sufficient to get the oscillation going if it's tied to GND (though C1980 should be sufficient).

....


oh no. Maybe 4 ohms is not so good. Then again I don't know if I can compare to ceramic and polypropylene caps that I have in my spare collection (0.22uF) which are coming in at much less than 1 ohm.

[stj]

Re: bent neck CRT pins...

too bad you cant see the leakage.

[eccerr0r]

Re: bent neck CRT pins...

Desoldered one pin and used my Sencore LC53. Leakage is undetectable (<1 uA) at 15V. Not sure what the voltage rating of this cap is... markings are very hard to read.

[R_J]

Re: bent neck CRT pins...

in the other hv power supply diagram page 438, c4363 is the same Cap. listed an page 393 in the parts list as a Elect: 0.22uf 5% 35v 1khz

[eccerr0r]

Re: bent neck CRT pins...

Hmm. My homemade ESR meter operates about 60KHz. ~4 ohms. Looks like tantalum,
and after staring at the label:

+/- 10%
0.22J
35V

I remeasured leakage at 25V (crappy Sencore, nothing between 25V and 50V). Still undetectable (00.0 microamps at 25V)... Cap appears good
----
I just swapped the IC as is and... no dice, still drawing too much current

Back to the drawing board...

[eccerr0r]

Re: bent neck CRT pins...

Now trying to activate the HVPS out of the chassis with a current limited bench PSU set at 500mA...

With U1956 socketed, I can remove it easily. So with it removed and 15V supplied to the "joule thief" oscillator, I can get it to oscillate if I connect pin 6 of the "flyback" to -15V with a 5.6K resistor (which weakly reverse biases C1951, but so be it for now; then again the coil will affect voltage there too). Once again adding a 6.3V lamp to the heater output, it lights up. However one observation is that Q1981 gets quite hot, then again it could be biasing with that 5.6K resistor. About 100mA flows through my bench supply when I have nothing connected, 300mA with the light bulb hooked up (~0.25W lamp). This is "excessive" IMHO but I still haven't replicated the "dead short" that shows up when I have it in the system (plugged in and CRT disconnected)...

[R_J]

Re: bent neck CRT pins...

Are you supplying +15 volts at the fuse or -15 volts at the fuse? it should be negative 15 volts into the fuse U1956 seems to be used only for regulation and should work with out it. however output voltages may be to high, I would supply the circuit with less voltage for testing, use -10 volts should make it work not being loaded by the crt etc.

[eccerr0r]

Re: bent neck CRT pins...

Well, correctly biasing C1972 (which is the main filter cap on the power to the "flyback". So yes it would be -15V at the "input" relative to ground.

Hmm. While the PSU was "ON" I noticed that the -300V supply was really low (-120V). My next experiment is that I pulled R1953 that separates the stuff going into the DC restorer (WTF is a DC restorer? Seems like trying to shift an AC signal upwards so it never goes negative? but why?) -- This made the potential across C1886 go back all the way to -300V (not using regulation though... but it is much better than it was). So something is drawing current here though unsure if this is the problem...

[R_J]

Re: bent neck CRT pins...

dc restoration is basically used to improve the contrast of the image. In a crt tv you would never get blacks without dc restoration. do you get any voltages at the +87 volts and on Q1980 should be about +50 volts aprox so Q1980 can regulate out +42 volts

[eccerr0r]

Re: bent neck CRT pins...

Ahh thanks... that makes sense. I guess it's similar in usage as trying to force a MOSFET off by force discharging the gate, else the pulses will "blur" together... interesting.

OK. I reconnected that R1953 39K resistor and pulled R1941, the 200 ohm resistor that's at the interface of the multiplier. With that 200 ohm resistor gone, I get -300V on that -300V rail once more. Hmm...this would somewhat imply the multiplier unless the source impedance of this rail is not low enough. That multiplier pin has higher resistance than detectable with my multimeter, which it better be with that capacitor in series inside.

I have not tested the +87 supply on this board but the main low voltage PSU and the main board I do see +87 and +42 supplies... now the question is, if it's possible for these voltages (and +15V that I don't have hooked up while testing out of circuit) to make this circuit behave like a dead short on -15V...

[eccerr0r]

Re: bent neck CRT pins...

Hmm... I'm not sure what I did now:
I have the op amp removed, and tied the output of the op amp to -15V.
If I have the CRT anode disconnected, the -300V comes right up. I can feel some static along the high tension lead when it's off :o

When I connect the anode, boom -300V, is dead.

I put the new op amp in with the anode connected: dead short
Disconnect the anode: The neon bulbs turned on indicating the control grid and cathode voltages are getting too far from each other.

Blah...now...cracked crt is a distinct possibility? :-(

The anode still has capacitance. I left a gap at one point and let a spark jump across (can't see it as the connector's shrouded). I used a screwdriver and shorted the anode cap to the chassis and got a small spark.