Prototype set is here

[miniman82]

Here is the series of pictures from the convention, where we mounted the proto CRT into Bob D's CT-100 donor cabinet and attempted to run it with a CT-100 chassis.


Begin with a spare CT-100 cabinet.




Remove factory CRT mask.



Prefabbed proto CRT 'conversion' mask, plexiglass hasn't been installed yet.



Place conversion mask in cabinet.



Install plexiglass into front of conversion mask, secure with self tapping wood screws.



View from the rear of the cabinet.

[miniman82]

Install proto CRT into custom mask. This required the use of a CT-100 support ring, held in place by support rods that were fabbed on site because the factory ones were far too short. These were Bob Galanter's idea, made from threaded rod that we cut to size. We then heated the ends with Bob's mapp gas torch to bend J-hooks into the end so they would fit into the U-hooks in the CT-100 cabinet.



Front view, with CRT in place.



Rear view, with CRT installed.




CRT being powered by the CT-100 at the ETF, which had unreasonably low HV (first attempt, on Friday).



Attempt #2 on Saturday, with CRT being powered by the Model 5 chassis. The Model 5 worked a lot better than the CT-100, but horizontal frequency tended to drift out of lock and we weren't able to get it back. Various things were tried in hopes of getting the Model 5 chassis repaired including replacement of caps and the horizontal frequency transformer, but were unsuccessful.



On Sunday we tried Bob's CT-100 chassis, but in the end we discovered that there is an issue with how much current the convergence electrode on the proto CRT is drawing out of the HV divider string in the chassis, which plays hell with the HV regulation. I will soon be modifying my CT-100 chassis so that the HV pot is no longer at the end of the divider string, but rather on the B+Boost line. This will ensure that HV remains steady, so convergence voltage is not chasing anode voltage anymore. Between this and some different value divider resistors (I bought a pair of 100 and 25 meg to try), I am confident I will at last be able to converge the proto CRT and get some decent screen shots. By the end of the day on Sunday we were all pretty tired...

[tubesrule]

Nik,
I meant to ask you at the convention, but to determine if this is a fault with the proto crt or by design, how is the convergence voltage derived on the proto chassis?

Darryl

[miniman82]

Same way it's done in the production chassis, though the values in the bleeder string are double what the CT-100 has. That's why I'm going to experiment with those parts, I have some different ones on the way right now. What I'm not sure about is if the convergence grid needs more current, or needs to be current limited by more series resistance. It was getting closer to being good with more voltage on it, but it also drew more current at higher settings. We tried to lower the value of the first resistor in the string (before the convergence pot), but it got very hot as it was burning off about 9000 volts. I'm still convinced that most of the problem was the ultor voltage drifting up as convergence voltage also went up; in other words, convergence always lagged ultor voltage no matter what we did. If I can get the ultor voltage to sit still while pushing convergence a tad higher, I think that will be a step in the right direction. That's the reason for tying the HV pot to B+Boost, rather than having it at the end of the bleeder string where its voltage supply (and hence shunt regulator operating point) can be influenced by current draw from the convergence grid.

[Tom S]

I'll wait to see a picture on the tube. Very nice snag there. Make it glow.

[Steve McVoy]

Current limiting by more series resistance can only lower the voltage on the convergence grid. If a higher voltage produced better convergence, then you need less series resistance.

I think the problem with the test you did is that the convergence grid draws too much current when convergence is right, and that current is too much for the 19.5 kv supply to handle. I don't think returning the bottom end of the convergence pot will help. I think the only solution is a beefier HV supply.

[miniman82]

If it were drawing an amount of current that was too much to handle, anode voltage would have sagged- instead it got higher. Only testing with the modified chassis will prove anything, Ed says the metal tubes were electrically identical to the glass ones...

[jeyurkon]

If I understand the shunt regulator circuit, if the convergence grid draws more current that will shift the grid on the shunt regulator more negative allowing the HV to rise as you observed.

To minimize the effect the standing current in the divider would need to be larger, meaning a beefier supply. Otherwise you need to limit the convergence grid current, but then you won't have sufficient voltage for convergence.

John

Originally Posted by miniman82

If it were drawing an amount of current that was too much to handle, anode voltage would have sagged- instead it got higher. Only testing with the modified chassis will prove anything, Ed says the metal tubes were electrically identical to the glass ones...

[jeyurkon]

If the convergence grid needs a much higher voltage, then the other possibility would be to move the pot further up the divider. Keep the total the same by reducing the resistance above it and increasing it below the pot by the same amount. The voltage drop due to the current drawn by the convergence grid will be smaller then, causing the anode HV to rise less.

John

[Penthode]

Convergence is going to be affected by both the electrostatic convergence plates dimiension and positioning, the dimensions of the tube and the 2nd anode HV potential. The assumption is that the HV is to be 20kV.

Perhaps the HV in this prototype was only 15kV? The HV capacitor voltage rating suggested a lower voltage.

It would be interesting to know at least if static convergence can be achieved with the lower convergence DC potential of the CT-100 with a HV of 15kV.

[Steve McVoy]

Interesting point. I have always thought that even if the doorknob cap was rated at 20 kv and the negative end tied to the focus potential (2 kv) that there wasn't much margin if the HV was 20 kv.

However, the schematic we have of a RCA model 1, which uses the same tube, shows 20 kv as the HV. It doesn't make sense that they would reduce the voltage for this set.

[John Folsom]

The electrical specs for the earlier portotype metal tube is virtually identical to the later 15GP22.

[jeyurkon]

Originally Posted by John Folsom

The electrical specs for the earlier portotype metal tube is virtually identical to the later 15GP22.

I think Ed told Nick the same thing. If true, then it should just work.

I hesitate to suggest it, but perhaps there's a problem with the CRT.

[miniman82]

That is one possibility, but another is the strength of the static convergence magnets on the proto tubes. The stick magnets on a CT-100 are very weak, and the prototypes had an entirely different system with 'wands' that could move in many different axis. If those magnets were stronger, it could account for the issues we're having.

[Steve McVoy]

Nick, I don't see how the permanent magnets could be a factor. If the prototype tube has specs very similar to the 15G, then it should work with 15G voltages, and apparently it doesn't. Since a 15G converges without the permanent magnets, using only the convergence control, the prototype tube should too.