Splicing 20KV CRT high voltage cable help

[Ian.M]

Looks pretty good.   Start with the 1mm handsaw kerf - if it doesn't clamp the cable tightly enough you can always sand down the jaw faces.  Consider putting the bolts through tapped holes from below so you have studs protruding to locate the jaws on then using wingnuts for easy clamping.

[OmegaHyperion]

i should have added that the bolts are held in the bottom plates via a thread, and a tightened nut (basically clamping the threaded rod in place). The top nuts are knurled steel nuts, not wing nuts. I assume they will develop enough force to clamp the cable, though id have to test it. If need be i can just replace them with something that i can apply more force to.

[OmegaHyperion]

Here is a cutaway of how i intend the clamp to be constructed. This might be able to clear things up slightly i hope.

[OmegaHyperion]

Here is an update on the splicing jig, i built the clamps now, and tested them with RG11 cable. They work perfectly. It took a bit longer, since my handsaws motor decided it no longer wanted to live. (the stator coils burned out)

Next step is to build the U frame, and finish the parts.

I intend to oil the wood clamps using boiled linseed oil, it should not be problem, as the cables will clamp well free of the splice its self. As well as linseed oil drying.

[OmegaHyperion]

It took a while to complete (life got in the way, as per usual), but i finally got the jig done, well sort of, the only thing its missing now is for me to oil the wood with boiled linseed oil, and pain the metal.



Now to practice splicing. I have also come up with a useful way of soldering the shielding. I will fold the shielding braid back onto its self, with a fiberglass sleeve between the inner and outer part. (alternatively i could fold it onto the outer insulation, also having the fiberglass sleeve there to protect the insulation). Then slide the new braid over the connection. Then fold the original braid onto the new one, and solder in place. The fiberglass should prevent the insulation from being damaged, as well as allow a full solder joint of all strands to be made. This should achieve the best shielding-

Any suggestions on streamlining this process are welcome, as usual.

[OmegaHyperion]

I have a new plan for the thermal insulation of the joint. I will use Fiberglass band to wrap the splicing tape, then fix it with kapton tape. After that i will put the braid over the joint, and place the braid of the original cables onto the new braid, and solder it into place. Then put glue heatshrink over that.

Any opinions on the joint idea?

I will test this on RG11 Coax before going to the CRT. I will also do a high pot test on the joint as well.

[OmegaHyperion]

I have made an experimental splice using some scrap RG11 cable

Here are the pictures of preparing the splice (the solder joint, and the splicing tape), as well as the post high pot test images, showing the spots it arced.


I only made the high voltage insulation part of the splice, and did not pay much attention to the shielding, it was connected and wrapped with wire to press it against the high voltage insulation to simulate the heat shrink.

The results of the high pot test where a arc voltage of 30KV (+-3 KV) Peak, or roughly 21KV (+- 2KV) AC 50Hz. This voltage was maintained for 1 min before a continuous arc was observed.

I will redesign my high voltage test bed to better do high pot. The above test was done as a sort of eyeballing test, to see if it even worked or not. The new test bed will have a more accurate high voltage volt meter permanently mounted, as well as a amp meter in the high voltage path (between ground and the transformer chassis). This should give me a more accurate reading. As well as a high voltage rectifier to test with both AC and DC (as needed), since i can only test with AC at the moment.

Also the scale bars are accurate, if anyone wants to do their own measurements.

[OmegaHyperion]

Conclusions from the test:

- The solder joint needs to be optimized, since it still has and allows spots to form where a corona discharge becomes more likely, as well as reduction of insulation thickness due to the soft splicing tape.
- Soldering with heatsinks is required to percent insulation damage due to heat from the center conduction (the center conductor was off center after soldering).
- 2 cm tapers should be increased to 3 cm tapers for more safety, as well as 1 cm of overlap of new and old insulation.


Possible solutions:

The use of crimp based joining of the high voltage conductor seems like a good idea at first, but would again create spots for corona to form. The best solution i can come up with at this moment is the use of heatsinks, and longer open conductor for soldering, as well as the use of a copper or brass tube filled with solder to mate the conductors, without forming sharp edges.  The new insulation should be wrapped to 1 1/2 the original diameter, half lapped.



Any input is welcome

[OmegaHyperion]

This is the redesigned solder joint. It eliminates all solder joint related problems found in the previous experiment.

It consists of a 2mm diameter 10mm long brass tube which was filled with solder. The conductors which will be mated where tinned and inserted under heat into the tube.

Due to a longer 15mm exposed cable, and this technique, no decentering of the central conductor was observed, due to the conductor temperature at the insulation interface staying below 100°C (melting point of PE)


This section of cable was a scrap from the previous test. It was not intended for a high pot test, it was only meant for a solder test.

[OmegaHyperion]

Im already using 3M high voltage splicing tape. Also in this particular CRT, there is no protection to trip, the high voltage power supply (not a flyback transformer) will just continue to pump power into the short, until i manually shut it down, or the fuse trips.

Thus i need to get this splice as close to perfect as i can get without spending thousands of euros in the process.

However, thank you for your comment none the less.

[OmegaHyperion]

Thus i need to get this splice as close to perfect as i can get without spending thousands of euros in the process.

For god sake. Just bodge it and wait it to fail. It will take considerably longer than you thought.
I won't be surprised if a street side repair shop will just bodge them in place and leave it to float in the air.

Well, good thing im not a street side repair shop then. I intend this splice to last at least as long as the piece of equipment the CRT is in, if not longer.

At the moment im looking into a better way to create the tapers on the high voltage insulation. So far i have used abrasion to grind the shape.

My new idea is to use a method similar to a pencil sharpener to taper it. Though i doubt a standard pencil sharpener will give me a long enough cone. I tested it on some RG11 scrap, and the cone on the 7.3mm insulation was roughly 10mm long, where as the ground tapers where 16mm long.

I will do some more testing and experimentations, until i have a way to make a splice which outperforms the cable its putting together.

[OmegaHyperion]

Well, good thing im not a street side repair shop then. I intend this splice to last at least as long as the piece of equipment the CRT is in, if not longer.

Okay, so it's not a cheap TV, it's actually a test gear. That makes sense.
If I were you, I would consider finding a plastic tube that's seamless and looks fine for 20kV (hint: gel ink pen refill), put the splice junction inside, and fill the thing with epoxy or silicone.

You are correct, its part of a piece of lab equipment im restoring.

I was considering the method you described, however previous members in this forum have discouraged me from perusing this rout. You can read the previous posts to see what this forum recommended and or came up with.

[andy2000]

It might be helpful to know exactly what this HV cable is used in (brand and model).  Are you certain the complete HV cable assembly is unavailable? 

[andy2000]

He mentioned that it's potted at one end (is the the entire HV power supply, just a HV multiplier, or something else?).  He said the other end is the anode cap, which is presumably removable from the CRT.  Is the shield grounded, or is it also carrying a high voltage?  More information might generate some alternatives to making a very difficult splice.  For all we know, it's just the coaxial HV cable required by older Sony Trinitron CRTs.  If it's something ultra exotic, it would be interesting to find out what it is.

It's hard to believe that no parts are available for something that appears to be very valuable. 

[OmegaHyperion]

Well, i might as well say what it is im working on. Its the photo CRT of a Scanning Electron Microscope (id say it dose not get more exotic then that). The shield is grounded inside the potted secondary high voltage tank (sadly no oil or SF6), thus i cant remove and replace the entire cable.

The other end of the cable is connected to a potted high voltage resistor, and then the anode cap of the photo CRT. A cap id rather not remove from the CRT.

As to available replacement parts, those have not been made since the early 90's, and finding one in unknown condition is hard enough, finding one in good condition is even harder. Thus id rather fix a known working component that was lightly damaged by an incompetent decommission team.

(they removed the photo CRT assembly to get the control unit threw a door where, even with the correctly installed CRT, a clearance of 30 cm on each side would have been present. (and instead of calling me they just decided to cut the cable...))

So, this is not your average TV im working on. Nor can i just simply move the unit to some repair shop. Thus im stuck repairing it in its installed location.

My previous splice test is capable of handling the voltage present in the high voltage cable, though id rather perfect the splice before i implement it in the microscope.  Just because something is hard to do, is no deterrent for me, i see it as a challenge.

edit:
I should probably add that the high voltage power supply that powers this CRT, also powers the other CRT's and other parts of the microscope. So having a short is a no go.

I did not say what it is im working on, since i thought that no one would reply if i stated i was working on an exotic CRT in an electron microscope. This way i got constructive replies, and am on the right track i think to fixing this cable.

[andy2000]

There's no substitute for unbroken insulation.  What about stripping back (and removing) several inches of shield on each side of the splice (2-3")?  Then insulate the splice with heat shrink tubing rated for at least 20kV plus a comfortable margin.  Finally, replace the missing shield with foil tape and one final layer of heat shrink over the entire splice. 

[Gyro]

Agreed, it's a simple matter of creepage distance and dielectric strength. The OP does seem to be making a bit of a meal of it (sorry).

There is no reason not to use continuous high dielectric heatshrink, the cable may have big things at each end but it is parted in the middle, so there is no difficulty in sliding on decent shrink ratio heatshrink before joining, (EDIT: probably 4 layers - two from each end).

From the creepage perspective, you can't assume that anything shrunk over the cable inner insulation is going to form a dielectric 'seal' (the same applies to the overlap between turns if insulation tape is used). The outer braid needs to be stripped back far enough to be able to resist 30-40kV on its own in open space.

The inner conductor joint area should be as small as possible (not long bare wire and splice) to minimize the airspace in which ionization can take place and degrade the insulation. coating the inner conductor joint with some silicone or dielectric grease would also help to fill the airspace. The heatshrink needs to overlap the inner conductor insulation at each end for a creepage distance capable of resisting the same 30-40kV.

Using copper foil tape over the splice area rather than braid is also a good idea. As the center conductor is +ve, the corona discharge will mostly come from the -ve charged screen near the inner core splice. That is probably what caused the insulation breakdown in the OP's wrapped tape insulation test - from the sharp ends of the braid overlaying it. Use of smooth copper tape will minimize any corona formation. It doesn't really matter how the braid is arranged outside that, the outer heatshrink will keep everything compressed and tidy.

With a little pre-planning it ought to be a reasonably easy joint to make.