CPI Tube issues

[NECRAT]

So Thursday, one of my cabinets decided to "go south". Upon testing, have determined there is a potentially very serious issue with the IOT.

It's a CPI . Can run idle at around 28kV. Crowbars out around 31kV (which isn't good seeing that the tube is rated for 40kV and needs roughly 32kV).
Hi-Potted the lead sets, input cavity caps, etc., all good. Hi-Potted the tube directly, with the similar results. Gets to 31KV, trips the Hi-Potter. All indications are the tube is shot.
Unfortunately the manufacturer doesn't have a 24/7 support number. I worked with the field engineer from Comark on the tube today. He is also convinced, based on our testing, that the tube is shot.

One thing another in market engineer suggested was the tube may have "cathode poisoning". However, there is no ion current, and I would think if the cathode was poisoned, you'd see it in the ion current meter.

It's a bizzare issue to say the least. Wasn't fun explaining to the boss that we may need to replace the tube either.

I am calling CPI Monday. Has anyone else experienced this with a UHF IOT CPI tube before? What did you do to resolve it? (or were you not able to)

[Kelly]

Yes I had a couple of those in a Comark. I agree it sounds like the tube is toast, and doubt the issue is related to cathode poisoning if you have absolutely no ion current. Sounds like something collapsed inside the IOT, causing the arcing. And it was good you did an external hipot test too. That eliminates a problem with the crowbar. How many hours did you get out of this girl before she gave up the ghost?

Just be thankful it isn't one of those Thomson (French) IOT's. What a disaster those things were!

[NECRAT]

Yes I had a couple of those in a Comark. I agree it sounds like the tube is toast, and doubt the issue is related to cathode poisoning if you have absolutely no ion current. Sounds like something collapsed inside the IOT, causing the arcing. And it was good you did an external hipot test too. That eliminates a problem with the crowbar. How many hours did you get out of this girl before she gave up the ghost?

Just be thankful it isn't one of those Thomson (French) IOT's. What a disaster those things were!

Just shy of 17K hours. It's a 20kw tube running around 15kw. I had those Thomson IOTs in Rhode Island, and they were a pain. I had isolation issues with one, that kept on causing issues with RF getting into the focus supply. That wasn't fun. Even after the engineer from Thomson tubes came up from PA, and removed it, it still was cranky. We were able to resolve it using a very rudimentary spectrum analyzer and field antenna.

[Jason M]

Is it crowbarring at 31kV with no drive? DO you have square or round input cavities? Have you inspected for sins of arcing, corona and inspected the finger stock? Also may be a good idea to look at your coolant. That is the first thing CPI will blame.

[NECRAT]

Is it crowbarring at 31kV with no drive? DO you have square or round input cavities? Have you inspected for sins of arcing, corona and inspected the finger stock? Also may be a good idea to look at your coolant. That is the first thing CPI will blame.

It's all been inspected and checked. CPI actually didn't blame the coolant, but as of about a year ago, the pH was normal.

It is either going to be Cathode to Grid grounding (going to check the "frozen history" in the transmitter to inspect the grid current), or HV breakdown. Doing a slow nursing job on it, to see if she comes back. Chances are likely it's not going to, and the likelyhood that the tube is toast is very high.

[NECRAT]

Well the tube is now holding it's original beam voltage of 32KV without a crowbar event. This is a good sign, of idle. It wouldn't do this last time.
One of the other things I did when I had it out, was I cleaned the input cavity from top to bottom too, just in case some of the ozone dust was problematic.
(It's good to do this anyways)

The real bread & butter test comes tomorrow when I drip a little RF into the tube, to see how it behaves. Believe it or not, we had another tube do this about three years ago.
(Before I was there) Same exact cause and effect. Different socket, however.

(BTW, if you haven't been near a UHF transmitter running 32KV when it crowbars, it's a blast. Especially at 4am...)

[Kelly]

Yeah when the crowbar fires it really gets the adrenaline pumping, especially after midnight and you're working alone.

One of our stations had an Acrodyne UHF which started tripping the beam supply OL without firing the crowbar. The chief was convinced it was a bad tube, but I wasn't so sure the symptoms were consistent with a bad tube. I flew in the following day and we disconnected the beam supply from the cavity and tried turning on the HV. The beam supply still tripped! After some investigation, we discovered the high voltage lead was shorting against a bleeder resistor in the beam supply located outside the building. The manufacturer had used a sticky-tie to hold the HV wire to an internal cover. The sticky let go and the wire dropped against a hot resistor.

[NECRAT]

Yikes. We had an issue where we had a Crowbar fail, and upon inspecting it, one of the HV leads in the HV cabinet on the rig failed. We found this out, after bringing the rig back into beam. Of course, it failed without crowbar protection. As soon as I saw the small "fire" in the HV cabinet window, I hit the breaker as fast as I could, but it tripped the beam supply breaker. (And took out the CPU, the CPU P/S, and the metering board). Fortunately the lead only burned up on itself, and didn't take out anything else. (It was connected to the load stackpole, and I thought for sure, that would've been destroyed. Fortunately, it wasn't hurt.)

[davek]

The manufacturer had used a sticky-tie to hold the HV wire to an internal cover. The sticky let go and the wire dropped against a hot resistor.

UGH! I absolutely despise those adhesive cable tie clamps, they never last and end up falling off at some point. Some times I'll drive a screw through the center of them or replace it with a proper screw-down cable tie anchor.

[Dale H. Cook]

I absolutely despise those adhesive cable tie clamps, they never last and end up falling off at some point. Some times I'll drive a screw through the center of them or replace it with a proper screw-down cable tie anchor.

I use only the adhesive tie-downs that have a screw hole, and I only trust the adhesive to hold for the few seconds that it takes to screw it down.

[NECRAT]

Day 9.

Holding at 32KV with -10db of drive. I installed a step attenuator in it. 1 db at a time.

[NECRAT]

Day 10 on the "sad tube" update.

Increased the drive slightly today. -9db (yes one whole DB!)

Tube is humming along with 650ma of beam current and about 10% drive, with about -20ma of grid current. Wasting power into the Majic Tee reject, but it's not a lot.
The ultimate hope is to get 60-70% drive off the tube, to run it for 6-8 months until we get a new tube. The big issue we have is, this tube is the reference tube in the 3 cabinet transmitter. So I can't properly phase the forward / reject with this tube having low drive. We may end up wasting some power for a few months. I'll do the best I can with it, of course.

Next update in a couple of days.

[Dale H. Cook]

This thread makes me glad that my four years as a TV CE involved only metal-ceramic UHF tetrodes.

[NECRAT]

I should start a blog.

"The story of V2 at Fox25".

[BigRed]

Speaking of IOT problems and crowbar circuits I seem to remember a service bulletin from some transmitter company or the other (don't remember if it was Comark, or Thales, or Thomson or Grass Valley, or . . .never mind) about life issues, early failures with the thyratron tubes. (For some reason I never quite understood why they seemed to always use Thomson thyratrons . . .) Anyhow, I seem to remember that the response of the tube was slowing down at around 20,000 hours. They would still pass the front panel button test but wouldn't react fast enough to actually save the tube in the event of an arc-over. To test the crow-bar you actually have to do the "wire" test (which is a REAL pain to do, by the way . . .) And with those hours on it it might pass the test THIS time, but not the next. Best to just replace the thyratron at around 15,000-hours. I referenced the Thomson thyratron but it may have affected any brand in the DXC/IOX. Been a year or two back, just don't remember the details. And that SB may no longer be valid. But it might be worth having a look at that. Possible that an arc-over started and knocked some material loose in the tube before the thyratron completely fired? Just a thought . . . (and if they're relative large pieces the ion pump current might not spike).

And when a crowbar fires is a problem !? Heck, those of us in the back woods with crimson neck just think someone is taking a surprise picture of us! Now, for a REAL treat, be standing near a klystron when water from a leaky steam/water separator finally finds it way across the gun end, especially if the beam control cabinet is nearby too. Now THAT will put a stain in the old short REAL fast!

Oh, one other thing on the CPI tubes . . . had one once that actually had a burn-through in the collector, flooded the inside of the tube with coolant (if a couple of ounces can be considered a flood). The tube was fairly young when that happened and CPI graciously replaced it at no charge (probably because of the Earl effect). They were never able to determine what happened to cause that though. The tube was junk, instantly though, but that doesn't sound like what you're dealing with.