Removing a base from a good octal valve. - UK Vintage Radio Repair and Restoration Discussion Forum

G4XWDJim · 17th Jul 2009, 1:25 pm

I have a KT66 with an intermittent problem. When it's working it's close to manufacturers spec so I'd like to save it.

On the valve tester gentle taps start or stop it working. I'm assuming that the cathode or screen pin has a bad connection to its internal lead. I've re-soldered the pins several times and each time I think I've fixed it only to be disappointed. There is no distress inside the electrode assembly during the fault condition so I doubt that it's the anode pin.

I've come to the conclusion that I need to remove the base and investigate more deeply. The question is how? It's very firmly fixed and I don't want to damage it by wrenching and ruin a potentially near new valve.

Does anyone have any experience of doing this with a tight and secure base. I'm less worried about unsoldering the pins all at the same time. I could probably make a little solder pot to heat all the pins at once but it's just the adhesive that holds the base on that seems difficult. As it's one of those slightly rounded brown bases I'd like to save the base for reuse if possible.

Could I boil it in water to soften the adhesive?

Any ideas Chaps, I'd be grateful for some thoughts.

Jim

AlanBeckett · 17th Jul 2009, 1:52 pm

Jim,
I had this sudden vision of your KT66 bobbing about in a pan of potatoes alongside the carrots and peas

The 'cement' that was used has always seemed to me to be very hard and brittle. It will be interesting to see what Members say because I want to try a couple of base swaps.
Alan

GrimJosef · 17th Jul 2009, 2:07 pm

I've had some experience of doing this with KT66s. I'm afraid it's always been a bit stressful !

You may be able to weaken the bond between the glue and the glass by dribbling solvent into the joint. I've used acetone (available very cheaply online, as it's sold in small bottles to nail salons for dissolving false-fingernail glue !). Be very careful of flash fires and of breathing too much of it in. If at all possible I'd recommend doing this outdoors on a breezy day. And it makes sense to desolder the wires (use a solder sucker and then prise them off the inside of the valve pin with a sharp, pointy thing) before you get acetone trapped inside the base !

You will probably still have to split the glue joint by rocking/twisting the base though. I must have done this ten times and I've only had an accident once. The problem is that the original glue is used in quite a large quantity and when the joint splits almost all of it stays stuck to the base. If you're unlucky it will have filled the bottom of the base and will be banked up around (or stuck to) the very thin tube through which the valve was originally evacuated. This tube is super-weak and if it's still glued to the base when the main joint splits then I'm afraid it will certainly snap off. There will be a short hissing noise, followed by prolonged wailing and gnashing of teeth

.

You may decide to have one more try at re-soldering the wires in the pins. The trick is to suck the solder out, detach the wire and then to try to clean both the wire itself and the inside of the pin using the thinnest rat's-tail file you can find, or some other very fine scraper. Frankly M-OV didn't seem to do a very good job of this and if it is the cause of the problem then even a very small amount of cleaning/scraping might sort it out. Otherwise I'm afraid it may be a snapped weld insde the valve itself. These can be very hard to see from the outside.

Cheers,

GJ

XTC · 17th Jul 2009, 2:19 pm

It's a long time since I've done this and it was with a KTW61.

I soaked the base overnight in a jar of meths, just to cover the base. I'm sure the cement they used, at least in the old days, was shellac and wood dust. That softens the cement. After that, the pins can be unsoldered with an ordinary iron, because they were only soldered at the tips of the pins. If you've put a lot of solder in there, you'd have to find a way to heat all the pins. Maybe a heat shield and hot air gun?

I recall I didn't bother to replace the cement in the base, but used tape or heat shrink which is maybe not appropriate for a KT66. Plastic padding or epoxy maybe, or even shellac and wood-dust paste.

They may not have used shellac with later production KT66s but I can't see that the meths would do any damage, if it failed.

Boiling it sounds like a bad idea. If the cement was likely to soften on boiling, wouldn't it soften when the valve was used?

I suppose you could cut round the base and use the bottom half of another base.

Sometimes you get intermittent heaters with a break inside the envelope.

Pete.

GrimJosef · 17th Jul 2009, 6:25 pm

Quote:

Originally Posted by XTC

I'm sure the cement they used, at least in the old days, was shellac and wood dust.

Yup, all the ones I've ever seen look like they're shellac-based.

Quote:

Originally Posted by XTC

Plastic padding or epoxy maybe, or even shellac and wood-dust paste.

A trick from Andrew Emmerson's book "Electronic Classics" which has always worked for me is as follows. Leave any shellac/sawdust which is still stuck to the glass in place (when the glass was new it was also chemically clean, and glue would stick to it, but after 40 years it's oxidised and now glue really struggles to stick to it). Put the base back on, having effected your repair, and resolder the wires into the pins, having cleaned and tinned the ends first. Run a couple of lengths of sticky tape round the bottom of the base and up over the top of the valve to hold the two tightly together. Dribble adhesive made of one part clear nail varnish to two parts acetone down into the join between the base and the glass (the acetone both allows the adhesive to run freely and also partially redissolves the shellac).The little brush in the nail varnish bottle is perfect for doing this. Don't worry about dribbling too much in - it doesn't seem to conduct electrically, so it doesn't matter if it runs down onto the inside of the pins. Leave it for 24 hours (on the couple of occasions I took the tape off early the joint cracked). If the joint gap was very "open" before you did this the joint may crack and you may need a second dose of adhesive. When everything's dry scrape any excess adhesive off the glass and wipe any dribbles off the base using clean acetone.

Quote:

Originally Posted by XTC

Sometimes you get intermittent heaters with a break inside the envelope.

Yes, I've seen that too. If the valve tester readings change quickly when you tap the valve then it probably isn't the heater. If they take seconds to respond then it might well be.

Cheers,

GJ

PJL · 17th Jul 2009, 7:50 pm

Not sure about KT66 but I was taught to cut the base off with a hacksaw and use a base from a faulty valve. The valves I have a repaired have usually had a ridge which makes removal very difficult. I am afraid it is most likely to be a weld problem inside the valve.

murphyv310 · 17th Jul 2009, 10:42 pm

Hi.
As PJL has said it's most likely a weld problem in the valve and unlikely a base problem.
If you really want to loosen the base from the glass, Cellulose thinners is the answer, leave the valve/base to soak for 24 hours, this will do the trick and then the wires can be de-soldered from the base enabling inspection.
Cheers
Trevor

G4XWDJim · 20th Jul 2009, 7:01 pm

Thanks everyone for your thoughts and ideas. I'm sorry to report that I have not achieved the success I hoped for although the valve still works after reassembly as well as it did before. At least I didn't wreck it.

I soaked it in cellulose thinners for a day and a half which softened the adhesive and allowed me to twist the base gently to the extremities of the connecting wires and loosen it. Incidentally does anyone know whether cellulose thinners is still made to the same spec or has it been "modified" to make it safer to use. It seems to smell nastier and less like the relatively pleasant smell of peardrops than it used to.

To remove the base I protected the spigot with a piece of thick shrink tubing and simply rotated the pins in a flame on the gas hob. A few seconds, a little pull and the base was off. I could see immediately that all the connecting wires were sound and well tinned. I had been hoping that one had broken or was at least really dirty.
Looking inside the pinch didn't reveal anything that looked amiss so at this stage I knew that I couldn't repair the fault and that it must be inside the envelope.
Putting it back together was not too difficult. I drilled out the pins with a 1.5mm drill and extended the wires with about 3 inches of wire. Just individual strands taken from a piece of flex. If I do it again I'll use 6 inches which will make it easier to fish them through without pulling previously inserted ones out again.
A smear of araldite around the base before full insertion of the glass was fairly easy and has resulted in a very solid fit.
Soldering the lot back together and testing it again revealed exactly the same fault. Anode current instantaneously disappears with a slight tap on the valve or indeed anywhere close on the valve tester. I have checked the octal base in the tester thinking some time ago that that's what the problem may be.
Later I set the screen volts high and the bias low to make it take a lot of current in the hope that carefully looking in a dark room might reveal a spark where one of the welds was duff but without success.
Experimenting with heater volts down to about 4 or so delayed the onset of the bad connection for quite a time. 7.5 volts brought the fault on very quickly so I suspect that there is a bad weld very close to the cathode itself so that it responds to heater temperature change very quickly.
My last chance is that I might be able to charge a big capacitor to about 1Kv or so and zap it somehow to make the bad weld re-weld itself but I don't hold out much hope.

Any further useful progress and I'll report.

Thanks all,

Jim

Barry Lloyd · 20th Jul 2009, 7:26 pm

Interesting thread, this. Reading the last message, I wonder whether it might be worth measuring the electrode currents under fault conditions - this should identify which connection is faulty (i.e if the anode or screen is o/c then the current on one or other will drop to zero, whereas if the cathode is disconnected all three currents will be zero).

Just a thought. It won't solve the problem but might enable yopu to work out whjich part of the valve needs zapping.

Barry

G4XWDJim · 20th Jul 2009, 7:57 pm

I've thought about doing this but I don't think it will work because they're all interrelated.
If the cathode opens no current will be read through any electrode. Like wise with the screen. If the anode opens the screen will act as an anode and heat up which doesn't happen. On a tetrode changing the anode volts makes little difference to the anode current. This is very easy to demonstrate on a valve tester.

I'm very happy to be convinced otherwise and really would like to find out where the fault is particularly if it lends itself to zapping.

If only it had been an SP61. I could have chucked it away without concern.

Jim

Barry Lloyd · 20th Jul 2009, 9:16 pm

I agree that zero screen volts will mean a big drop in anode current; but I was assuming that there would still be a small residual anode current, maybe only microamps, whereas an o/c cathode would mean absolutely no current at all. But I could be wrong. When I get the chance (i.e. not in the next week!) I will try measuring the anode current on a typical output valve to find out.

Barry

GrimJosef · 20th Jul 2009, 9:29 pm

Quote:

Originally Posted by G4XWDJim

My last chance is that I might be able to charge a big capacitor to about 1Kv or so and zap it somehow to make the bad weld re-weld itself but I don't hold out much hope.

I'm not sure zapping it with a capacitor would work, but if you can see the fault then maybe what you need is a laser ? I Googled "laser welding through glass" and I found these people http://www.appliedenergysystems.com/laser-welding.html Their third bullet point looks promising. Of course maybe the process would release too much gas into the vacuum. But maybe not ?

If you want to follow it up in the UK then perhaps these folks can help http://www.ailu.org.uk/ ?

Cheers,

GJ

PJL · 20th Jul 2009, 10:17 pm

The Mullard tester claims it tests for this kind of fault but I'm a bit dubious. It just disconnects the electrode moentarily and if you notice a change in anode current then the electrode connection is OK!

You say you are using a tester, what happens when in the fault condition?

G4XWDJim · 21st Jul 2009, 2:41 pm

Quote:

Originally Posted by PJL

You say you are using a tester, what happens when in the fault condition?

Monitoring the anode current when the fault is on shows an instantaneous drop to zero.
Disconnecting the various electrodes by means of the selector switch on the Avo valve tester simply drops the current to zero. Even dissing the anode doesn't cause any visible stress to the screen but maybe that's because the Avo tester uses AC for its tests. I'm on the point of giving up with it.
The laser idea sounds good if I could just see the problem but wouldn't it cost more than a few KT66s.

Jim

17th Jul 2009, 1:25 pm	#1
G4XWDJim Octode Join Date: Oct 2008 Location: Littlehampton, West Sussex, UK. Posts: 1,465	Removing a base from a good octal valve. I have a KT66 with an intermittent problem. When it's working it's close to manufacturers spec so I'd like to save it. On the valve tester gentle taps start or stop it working. I'm assuming that the cathode or screen pin has a bad connection to its internal lead. I've re-soldered the pins several times and each time I think I've fixed it only to be disappointed. There is no distress inside the electrode assembly during the fault condition so I doubt that it's the anode pin. I've come to the conclusion that I need to remove the base and investigate more deeply. The question is how? It's very firmly fixed and I don't want to damage it by wrenching and ruin a potentially near new valve. Does anyone have any experience of doing this with a tight and secure base. I'm less worried about unsoldering the pins all at the same time. I could probably make a little solder pot to heat all the pins at once but it's just the adhesive that holds the base on that seems difficult. As it's one of those slightly rounded brown bases I'd like to save the base for reuse if possible. Could I boil it in water to soften the adhesive? Any ideas Chaps, I'd be grateful for some thoughts. Jim

17th Jul 2009, 1:52 pm	#2
AlanBeckett Rest in Peace Join Date: Dec 2008 Location: Burton upon Trent, East Staffordshire, UK. Posts: 1,686	Re: Removing a base from a good octal valve. Jim, I had this sudden vision of your KT66 bobbing about in a pan of potatoes alongside the carrots and peas The 'cement' that was used has always seemed to me to be very hard and brittle. It will be interesting to see what Members say because I want to try a couple of base swaps. Alan

17th Jul 2009, 2:07 pm	#3
GrimJosef Dekatron Join Date: Sep 2007 Location: Oxfordshire, UK. Posts: 4,311	Re: Removing a base from a good octal valve. I've had some experience of doing this with KT66s. I'm afraid it's always been a bit stressful ! You may be able to weaken the bond between the glue and the glass by dribbling solvent into the joint. I've used acetone (available very cheaply online, as it's sold in small bottles to nail salons for dissolving false-fingernail glue !). Be very careful of flash fires and of breathing too much of it in. If at all possible I'd recommend doing this outdoors on a breezy day. And it makes sense to desolder the wires (use a solder sucker and then prise them off the inside of the valve pin with a sharp, pointy thing) before you get acetone trapped inside the base ! You will probably still have to split the glue joint by rocking/twisting the base though. I must have done this ten times and I've only had an accident once. The problem is that the original glue is used in quite a large quantity and when the joint splits almost all of it stays stuck to the base. If you're unlucky it will have filled the bottom of the base and will be banked up around (or stuck to) the very thin tube through which the valve was originally evacuated. This tube is super-weak and if it's still glued to the base when the main joint splits then I'm afraid it will certainly snap off. There will be a short hissing noise, followed by prolonged wailing and gnashing of teeth . You may decide to have one more try at re-soldering the wires in the pins. The trick is to suck the solder out, detach the wire and then to try to clean both the wire itself and the inside of the pin using the thinnest rat's-tail file you can find, or some other very fine scraper. Frankly M-OV didn't seem to do a very good job of this and if it is the cause of the problem then even a very small amount of cleaning/scraping might sort it out. Otherwise I'm afraid it may be a snapped weld insde the valve itself. These can be very hard to see from the outside. Cheers, GJ

17th Jul 2009, 2:19 pm	#4
XTC Retired Dormant Member Join Date: Oct 2004 Location: Bristol, UK. Posts: 837	Re: Removing a base from a good octal valve. It's a long time since I've done this and it was with a KTW61. I soaked the base overnight in a jar of meths, just to cover the base. I'm sure the cement they used, at least in the old days, was shellac and wood dust. That softens the cement. After that, the pins can be unsoldered with an ordinary iron, because they were only soldered at the tips of the pins. If you've put a lot of solder in there, you'd have to find a way to heat all the pins. Maybe a heat shield and hot air gun? I recall I didn't bother to replace the cement in the base, but used tape or heat shrink which is maybe not appropriate for a KT66. Plastic padding or epoxy maybe, or even shellac and wood-dust paste. They may not have used shellac with later production KT66s but I can't see that the meths would do any damage, if it failed. Boiling it sounds like a bad idea. If the cement was likely to soften on boiling, wouldn't it soften when the valve was used? I suppose you could cut round the base and use the bottom half of another base. Sometimes you get intermittent heaters with a break inside the envelope. Pete.

17th Jul 2009, 7:50 pm	#6
PJL Dekatron Join Date: Sep 2005 Location: Seaford, East Sussex, UK. Posts: 5,997	Re: Removing a base from a good octal valve. Not sure about KT66 but I was taught to cut the base off with a hacksaw and use a base from a faulty valve. The valves I have a repaired have usually had a ridge which makes removal very difficult. I am afraid it is most likely to be a weld problem inside the valve.

17th Jul 2009, 10:42 pm	#7
murphyv310 Dekatron Join Date: Sep 2006 Location: Kilmarnock, Ayrshire, UK. Posts: 5,422	Re: Removing a base from a good octal valve. Hi. As PJL has said it's most likely a weld problem in the valve and unlikely a base problem. If you really want to loosen the base from the glass, Cellulose thinners is the answer, leave the valve/base to soak for 24 hours, this will do the trick and then the wires can be de-soldered from the base enabling inspection. Cheers Trevor __________________ Cheers, Trevor. MM0KJJ. RSGB, GQRP, WACRAL, K&LARC. Member

20th Jul 2009, 7:01 pm	#8
G4XWDJim Octode Join Date: Oct 2008 Location: Littlehampton, West Sussex, UK. Posts: 1,465	Re: Removing a base from a good octal valve. Thanks everyone for your thoughts and ideas. I'm sorry to report that I have not achieved the success I hoped for although the valve still works after reassembly as well as it did before. At least I didn't wreck it. I soaked it in cellulose thinners for a day and a half which softened the adhesive and allowed me to twist the base gently to the extremities of the connecting wires and loosen it. Incidentally does anyone know whether cellulose thinners is still made to the same spec or has it been "modified" to make it safer to use. It seems to smell nastier and less like the relatively pleasant smell of peardrops than it used to. To remove the base I protected the spigot with a piece of thick shrink tubing and simply rotated the pins in a flame on the gas hob. A few seconds, a little pull and the base was off. I could see immediately that all the connecting wires were sound and well tinned. I had been hoping that one had broken or was at least really dirty. Looking inside the pinch didn't reveal anything that looked amiss so at this stage I knew that I couldn't repair the fault and that it must be inside the envelope. Putting it back together was not too difficult. I drilled out the pins with a 1.5mm drill and extended the wires with about 3 inches of wire. Just individual strands taken from a piece of flex. If I do it again I'll use 6 inches which will make it easier to fish them through without pulling previously inserted ones out again. A smear of araldite around the base before full insertion of the glass was fairly easy and has resulted in a very solid fit. Soldering the lot back together and testing it again revealed exactly the same fault. Anode current instantaneously disappears with a slight tap on the valve or indeed anywhere close on the valve tester. I have checked the octal base in the tester thinking some time ago that that's what the problem may be. Later I set the screen volts high and the bias low to make it take a lot of current in the hope that carefully looking in a dark room might reveal a spark where one of the welds was duff but without success. Experimenting with heater volts down to about 4 or so delayed the onset of the bad connection for quite a time. 7.5 volts brought the fault on very quickly so I suspect that there is a bad weld very close to the cathode itself so that it responds to heater temperature change very quickly. My last chance is that I might be able to charge a big capacitor to about 1Kv or so and zap it somehow to make the bad weld re-weld itself but I don't hold out much hope. Any further useful progress and I'll report. Thanks all, Jim

20th Jul 2009, 7:26 pm	#9
Barry Lloyd Retired Dormant Member Join Date: May 2006 Location: London, UK. Posts: 129	Re: Removing a base from a good octal valve. Interesting thread, this. Reading the last message, I wonder whether it might be worth measuring the electrode currents under fault conditions - this should identify which connection is faulty (i.e if the anode or screen is o/c then the current on one or other will drop to zero, whereas if the cathode is disconnected all three currents will be zero). Just a thought. It won't solve the problem but might enable yopu to work out whjich part of the valve needs zapping. Barry

20th Jul 2009, 7:57 pm	#10
G4XWDJim Octode Join Date: Oct 2008 Location: Littlehampton, West Sussex, UK. Posts: 1,465	Re: Removing a base from a good octal valve. I've thought about doing this but I don't think it will work because they're all interrelated. If the cathode opens no current will be read through any electrode. Like wise with the screen. If the anode opens the screen will act as an anode and heat up which doesn't happen. On a tetrode changing the anode volts makes little difference to the anode current. This is very easy to demonstrate on a valve tester. I'm very happy to be convinced otherwise and really would like to find out where the fault is particularly if it lends itself to zapping. If only it had been an SP61. I could have chucked it away without concern. Jim

20th Jul 2009, 9:16 pm	#11
Barry Lloyd Retired Dormant Member Join Date: May 2006 Location: London, UK. Posts: 129	Re: Removing a base from a good octal valve. I agree that zero screen volts will mean a big drop in anode current; but I was assuming that there would still be a small residual anode current, maybe only microamps, whereas an o/c cathode would mean absolutely no current at all. But I could be wrong. When I get the chance (i.e. not in the next week!) I will try measuring the anode current on a typical output valve to find out. Barry

20th Jul 2009, 10:17 pm	#13
PJL Dekatron Join Date: Sep 2005 Location: Seaford, East Sussex, UK. Posts: 5,997	Re: Removing a base from a good octal valve. The Mullard tester claims it tests for this kind of fault but I'm a bit dubious. It just disconnects the electrode moentarily and if you notice a change in anode current then the electrode connection is OK! You say you are using a tester, what happens when in the fault condition?