Power Supply Dilemma

[Todd, KA1KAQ]

It's almost like a play on the old saying 'when God closes a door, he opens a window'.

"When Murphy opens a window, he abruptly slams it down onto your fingers."

Like other big transmitters, mine has several power supplies ranging from 150v to 2500v. One is a 300v supply that powers several things from the pilot and dial lights to the audio driver section. Lately it's been consuming 3A fuses on start up like candy.

And here's the annoying part: it only happens on power up, usually after a few seconds, then after shutting down the transmitter and replacing the fuse - it's FINE for as long as the transmitter is left on.

Initially I was thinking it was some kind of a cap issue related to initial charging at start up. But the entire deck was out and recapped a year ago with new manufacture Sprague Atom electrolytics, AES twist-lock cans, and SBE ODs in place of all wax and paper caps. And it worked FB until recently.

I suspected a tube issue, so I changed out the phase inverter and 1st audio amp just because one was a bit weak and the other was an International re-brand. No shorts, nothing to indicate a problem otherwise. No real reason to change them or for them to be causing the problem.

Also checked the pair of 5V4 slow-heating rectifiers, thinking an issue there would pull current faster and overheat the fuse. Both check fine, though I should replace them with a known-good pair to be sure.

The supply is pretty simple: transformer, pair of 5V4s, dual section 40 uf electrolytic, 12 hy filter reactor, 100k resistor, 3A fuse on transformer primary. Schematic can be seen midway down the document at:

http://www.collinsradio.org/archives/manuals/KW-1-(01-52)Sec7,ant-ref.pdf

Sure - one of the new caps could have puked, but it doesn't appear to be a condition that remains in place and they've worked fine for over a year. Means nothing I know, but seems less likely. If the iron was shorting, either to ground or between turns, wouldn't it remain so? This doesn't seem to be thermal in nature since the transmitter isn't warmed up when it happens. Instead, the fuse will pop/lights go out, turn transmitter off, replace fuse, turn on and use for hours with no more issues.

The symptoms exist for only a few seconds, so potentially I could go through a lot more fuses trying to get initial readings at start up. I plan to swap out remaining tubes with known good examples, but none of them appear to be overtly bad, meaning shorted, open, etc.

Intermittents are always a pain, this seems to be a semi-intermittent: there at start up, then gone. Any of the related stages could be causing it, but all have been recapped and have good tubes in place. The fact that it blows a fuse, then works fine once it's replaced really has me confused. It would make more sense to keep blowing fuses if something is indeed bad.

Have you ever heard of such a thing?

[WD8BIL]

Try sticking a 25 ohm varistor in the primary of the transformer and see what happens.

Are the cap values the same as original or did you up'em a bit when you recapped?

[Carl WA1KPD]

I had an old funky power switch once that caused that. Never dissected it to determine why but I assume something was heating up and then going to ground inside the switch. I uncovered it becuase it did not happen when I was feeding the tranny directly during trouble shootig.

[WA1GFZ]

Could be a transformer flashing over. I would look at the 5 volt windings and hv.  All it takes is two turns flashing over to draw a lot of current.

[Todd, KA1KAQ]

That's a good question, Bud. I'll check tonight. Some of them were indeed bumped up a bit for convenience as well as keeping the same package in place, others remain close or exact to original values. Voltage ratings were increased.

Even if they are higher value, why would it work fine for over a year and start acting up now? I would've expected it to show up early on, like the first time the caps charged. It's as if the transmitter sensed the rewound mod iron was coming home, and decided to create another issue requiring attention.

It's possessed! 

Frank: I thought about that, but wouldn't it flash over again as soon as I replaced the fuse? I'm most concerned about the iron because it's such a pita to replace/repair. This one is a sealed unit, too. One thing I noted, there was a 3A slo blo fuse in the panel but the manual only states '3A'. Perhaps this has existed for a while and only recently revealed itself when I installed fast-acting fuses.

Carl, the switch works fine after the fuse blows once, so I'm guessing it's okay. I'd sure prefer that be the cause though - easy to access and I have a few spares kicking around.

Manual says 300v supply should read 280v on the multi meter setting for that supply. It has always read a bit high, 340 I think. My line voltage runs right around 120, which is what the autotransformer for the rig is set up for.

[w3jn]

I have some negative temp coefficient thermisters I can send ya.  Put one in series with the xformer leg and yer good to go.

[WA1GFZ]

Todd, pull out the rectifier tubes and see if you can make the fuse blow. This will isolate the transformer from the rest of the rig. All the windings jump when the power is first turned on. Maybe sometimes two turns rub and short taking out the fuze. When the rectifier heaters are cold there should be no current into the filter caps. Also look for caps across the line side of the transformer. Sealed transformers are not hard to get apart but a bit of work. fc

[Todd, KA1KAQ]

Y'know Frank, I meant to try that last night when I had the tubes out for testing, and completely forgot. That'll be test #1 for tonight. In the event that this locates the problem, I'm guessing the choices are:

 - install another 3A slo-blo fuse to get past that initial kick

 - try new rectifiers

 - install a Thermistor

 - get iron rewound (blech!)

Just to be clear: the fuse doesn't blow immediately on power up. It's usually 2-5 seconds, sometimes a bit more.

Johnny, I got one of those Thermistor kits off ebay with assorted values in a plasitc box for $9.99 a while ago. So it would just be a matter of properly sizing the Thermistor to the load, eh? They have gibberish numbers on them as I recall but the bins are marked well. Maybe if I put my glasses on it would make sense.

Adding these to the gear used regularly was the intent behind buying them in the first place.

 

[w3jn]

Yeah, you can get the data sheet from the mfgr, that's where I got mine.

Pick one that's a 3 amp job.

It sounds to me like the fuse ain't blowing until the rectumfriers heat up.  So something's drawing fairly heavy current from the gitgo.   Question:  what does this 300V PS run?  Is there anything that it supplies that has an associate bi-ass supply with it?  I'm wondering if the 300V rectumfriers and the biass rectumfriers aren't warming up out of sync - ie you're getting your 300V on some toobs before they're getting biass, so they're drawing some heavy current for a bit.  Perhaps those 6B4s???  THose'll sink a bit of current!

Might wanna check the biass rectumfrier to see if it isn't weak or something.

[Todd, KA1KAQ]

Yep, -150v biass supply also runs a 5V4 rectifier that I hactually replaced with a NOS Raytheon last year. I did have a problem with one of the 6B4Gs having a grid-to-cathode short which screwed up my biass world last year, made some tar leak outta the driver tranny. Posted about it, thought a cap had died so I yanked and rebuilt the entire audio/ps deck to prevent future issues. Shortly after recapping and fixing it all up nice, the mod iron went south. Now the mod iron is fixed but supply was acting up before the swap.

You, HUZ, and myself were talking about this supply a week or so back on the air. I suspected the rectifiers but Stevieboy thought not, said they would pass less current, not more. 5V4s are slow-heating though, that's the key. To prevent applying voltages affore yer spoda.

But still - nothing is in key-down/transmit, no audio is present, so there shouldn't be any load on the toobs beyond filament, si? I'm missing something obvious again, aren't I? Toobs aren't biassed off, therefore passing big current until biass supply wakes up (newer tube, so probably slower warm up)? WAKE UP, BRAIN!!

[W1VD]

Might try wiring a suitable current measuring device in series with the output from that supply. That'll show if the load is the problem...or not.

[WA1GFZ]

Jay has a point take a dead tube and a tube socket and wire up an extender cable with a milliamp meter in series so you can see the current rise on turn on. Any cap hanging off the supply could be suspect. It could still be the transformer. fc

[w3jn]

If the B-puss comes on before there's any biass, the 6B4s are gonna draw some current!!

This will happen if the 300V rectumfrier warms up before the biass rectumfrier.

[WA1GFZ]

Gee John,
maybe this is the reason the current is taking off. Maybe Todd can solid state the bias rectifier. Then add the right series resistor on each diode to get the bias voltage back to match the tube drop. You could build it in an old tube base so Joe doesn't disown you. Power up the bias would be right there.
Seems to me the bias divider is not simple so any leaking cap on the bias rail would cause high plate current. 

[Opcom]

How about temporarily using a current transformer to measure the surge or whatever it is?

[k4kyv]

I had an intermittent fuse blowing episode with my MV power supply.  I JS'ed several quarter-amp fuses at various spots in the power supply - plate leads from HV transformer to rectifier tubes, in series with filter cap, filter choke and output lead, and shorted out the main fuse.  Just waited for the intermittent to show up, to see which one would blow first. It was one between the HV winding and a rectifier plate.  Turned out to be bad contact at rectifier filament transformer not heating up the 866's enough to keep them from flashing over.  Fixed the problem, then power supply became intermittent without blowing fuse.  That one turned out to be intermittent filter choke.  I opened the choke (an old UTC LS series choke) and found a wire lead burnt in two just below the tar level.  Fixed that, reassembled the choke, and the supply has never given any more trouble.  Don't know if there was any connection between the two problems or if it was a classic case of Murphy's law.

[Todd, KA1KAQ]

Okay, here's the results of last night's test:

 - Pulled the pair of 5V4Gs from the 300v supply, fired up, sure enough - no issue. Tested and moved around 5V4s to get the slowest-heating (according to response time on tube tester) tubes into 300v supply and faster heating into -150 biass supply. Fired up xmitter, 4-5 seconds, *snick* out went the lights.

The 5V4s are meant to heat slowly and prevent B+ from being applied before everything is warmed up and ready, or at least that's how the manual explains it. I'm still confused as to why it will blow one fuse, then work fine after a replacement is installed. It acts like a charging issue, with caps being charged up already the second time around. Of course, I got a phone call and forgot to go back and check the replacement can specific to that supply installed during the recap last year.

It's annoying, partly because of the nature and partly because of my inability to figure it out. Thinking I might bypass the fuse holder temporarily and install one of those resettable CBs to at least save on 3AG fuses while I blunder through it. The old tube base idea sounds promising, plenty of ma meters kicking around.

So, no real progress beyond pulling the rectifiers and confirming the fuse doesn't blow then. Grumble, grumble. Possibly have a bad 'new' cap? You must remember the bias issue from last year, Frank, and how they chose the bias arrangement as a place to cut corners.

[w4bfs]

Hi Todd ... when I run into a proplem like this I use a ballast box composed of several various wattage 120V incadescent bulbs with individual shorting switches ...this way I can bring up power available to equipment under test without blowing fuses ...very handy and highly recommended ....helps with electro-mech type problems...73...John

[AB2EZ]

Todd

I agree with the suggestion that the bias supply is turning on too slowly.

Try temporarily rewiring the primary of the bias supply transformer (which includes the filament winding for the bias supply rectifier) so that you can turn on the bias supply before you turn on the rest of the transmitter.

If that helps, then you can decide whether to solid-state the bias supply.

It is also conceivable that one of the tubes that is connected to the 300 volt supply has a short that goes away when that tube heats up.

To test this... take out the 3A fuse... turn the transmitter on... wait 90 seconds... and then (carefully) insert the 3A fuse without turning the transmitter off. [The 300 volt output will come up slowly because the 300 volt rectifier tube filaments are off when the 3A fuse is removed].

Stu

[W1VD]

Wonder if you tried removing V-410, the 5R4GY 500-volt supply rectifier? Easy test...

[WA1GFZ]

Todd,
I have your schematic at home but it would be easy to cob in a pair of 1N4007 diodes to see if solid stating the bias would guarantee it come up first.
Without series resistors the Bias would be a bit higher than normal but it wouldn't hurt anything. Just grab an old tube base and install the diodes so you can just plug it in. Plate is the diode anode fc

[Todd, KA1KAQ]

Here's a schematic online, midway down:

http://www.collinsradio.org/archives/manuals/KW-1-(01-52)Sec7,ant-ref.pdf

Bias for the drivers is adjustable anyhow, Frank. I'm still trying to figure out what has changed to make this an issue. Haven't been able to scare up spare 5V4s to swap in beyond the ones I have, all are pulls but one. If the larger capacitance on some electrolytic is causing it due to longer charging time, why now and not last year when I recapped it? Moving and substituting the rectifiers produced no change. If the transformer had shorted windings, why would installing a second fuse work? Same thing if one of the new caps went bad. I know I'm not the brightest bulb in the string, but it just makes no sense. 4-5 seconds isn't a lot of warm-up time.

Wonder if you tried removing V-410, the 5R4GY 500-volt supply rectifier? Easy test...

To see if the 5v winding is causing it? The 500v supply has a separate 5v winding in the same transformer according to the schematic, but I'm not seeing anything else. I did install a new 5R4GY Monday just for shitzngiggles, though. No change. I'll yank it tonight if I get a chance.

[AB2EZ]

Todd

As Frank suggests... its easy to make up a plug-in solid state rectifier to test the slow-rising bias voltage theory.

Alternatively, if you remove the 1A bias voltage power supply fuse, that will isolate the primary of the bias power supply from the rest of the power supplies. As a result, it will be easy to use a clip lead apply 120 VAC to the primary of the bias supply transformer ... thus turning on the bias supply before turning anything else on.

Why would this happen now, and not previously? Good question. Perhaps the changes you made to the 300 V supply made that supply capable of delivering more current. As a result, the current that is flowing into tubes like the 6B4Gs... prior to the full turn on of the bias supply... could be larger now. This larger current might be enough to blow the 3A fuse (which didn't blow before you "improved" the 300 volt power supply).

Stu 

[AB2EZ]

All

There is something puzzling about the schematic that Todd pointed to in his posting (maybe a mistake on the schematic)...

Note:

The center tap of the secondary of the 300 volt supply transformer is not connected to ground (either directly, or through a resistor). Instead, it connects to the center tap of another filament winding of the transformer (not the winding that powers the rectifier tubes of that supply). That other filament winding is connected directly to the filaments of tubes: V401, V402, V403, V406, and the filaments of the 6B4G's. None of these filaments is shown as having either side grounded, Several of these tubes have indirectly heated cathodes with no connection to the filament; and, furthermore, other tubes which feed from the 300 volt supply also have no path back to the center tap of the secondary of the 300 volt supply's transformer

There must be a path to ground that is missing from the schematic. As shown, only the 6B4G's (which have directly heated cathodes) would be able to draw current from the 300 volt supply.

If I had to guess, there should be a relatively low value (maybe a few hundred ohms) resistor from the center tap of the secondary of the 300 volt supply's transformer to ground... to provide a return path for everything that is drawing current from the 300 volt supply (except the 6B4Gs, which are already connected via their filaments)

Stu

[Todd, KA1KAQ]

Hey Stu -

Heard you talking with Herb and Roger couple nights back, but had the transmitter torn apart. Completely forgot that the 32V is online now, could've worked you there and asked your thoughts on the matter.

"Improved": Actually, that's the odd part - All I did was recap the entire deck in Feb 07, shortly after one of the 6B4Gs got a grid-to-cathode short and made the bias go haywire. Nothing was done to the 300v supply beyond a new dual section can, tested/replaced tubes as needed on the entire deck. It worked fine since, until the day before NEAR-Fest when some of the guys were over. Went to turn it on, no dial lights. Replaced the fuse and all was well. Or so I thought.

It makes sense that the drivers could be pulling more current without bias, but the manual clearly states that the bias supply is supposed to warm up slowly to prevent K-403 (send/rec) from being energized. This makes me believe that even if the 300v supply comes online first, it shouldn't be able to nail the audio driver stage since it also uses 5V4s and because of the built-in time delay requirement. Meaning, the initial design should preclude this happening unless something is not working, which doesn't seem to be the case.

So something has changed somewhere recently, in a way that doesn't show up beyond blowing the 3A fuse on start up, then running fine once it's replaced. A shorted winding or cap would stay shorted, I'd think?

Just out of curiosity Stu - the next time you fire up the your transmitter, would you please switch your multi-meter to the 300v supply and tell me what yours reads? The book says it should read 280v, mine reads 340 and has for as long as I can remember. Slightly-elevated line voltage shouldn't account for that much increase.

UPDATE: There are a few errors in the online (older) manual. The photo of the deck also doesn't show C-417, although the schematic does. That's the cap from the driver CT to ground. Wouldn't surprise me if it was just an omission.