Date   

Re: tants new vs old

Tom Gardner
 

On 18/07/18 09:29, Craig Sawyers wrote:
I have had axial tantalum capacitors leak and destroy nearby PCB traces, specifically
C6246 and C6341 in a Tek 1502 TDR.
Interesting - must check mine. Wonder why they used them only on the +25V rail, which is only used on
the high voltage board, to feed the switching transistors on the primary of the HV transformer, but
regular electrolytics on the other rails? Noted that these wet tants are 30V parts on a 25V rail
(rated = 1.2 rail voltage), so is somewhat close to the wire.
It has only happened in 20% of my 1502s. I seem to have a
fetish for the damn things :)

ISTR the 25V rails are not low current, which might increase
the stress.


Re: tants new vs old

Craig Sawyers <c.sawyers@...>
 

I have had axial tantalum capacitors leak and destroy nearby PCB traces, specifically
C6246 and C6341 in a Tek 1502 TDR.
Interesting - must check mine. Wonder why they used them only on the +25V rail, which is only used on
the high voltage board, to feed the switching transistors on the primary of the HV transformer, but
regular electrolytics on the other rails? Noted that these wet tants are 30V parts on a 25V rail
(rated = 1.2 rail voltage), so is somewhat close to the wire.

Craig


Re: tants new vs old

Craig Sawyers <c.sawyers@...>
 

Better yet replace them with higher voltage rated ones if they are available. 20V tantalum lasts
forever
on 5V rail as well as 35V one on 15V.
Well, yes. All I'm saying is that I have identical plugins (such as 7A26) in which at some point Tek
went from tant bead to aluminium electrolytic. So substitution is tried and tested.

And low ESR is not as good for tantalum beads as compared with modern low ESR aluminium
electrolytics designed for switched mode supplies. As an example an AVX 47uF 35V tantalum bead has
an ESR at 100kHz of 0.8 ohms. A Panasonic FR of the same capacitance, but 25V rated has an ESR of
0.3 ohms at 100kHz.

I can send you a handful :) They _ALWAYS_ leak sooner or later because of their very chemistry.
They
are filled with sulfuric acid and although silver sulfate has very low solubility it is still not
zero so sooner
or later that nice silver can turns to sieve. And the fact they are almost always used at their
rated
voltage or even above also doesn't make them last longer...
Operating any capacitor at or above its rated voltage is a short term recipe for capacitor death
regardless of chemistry. Even with wet tant, the current recommendation is that for long life the
rated voltage should be at least 1.6 times the operating voltage.

I have dozens of those leaked in 492BP/494AP SAs that I refurbish on a regular basis. The same is
true
for those HP 835xx microwave plugins that I revived quite a few -- I have never seen a single such
plugin that didn't have at least one leaked wet tantalum. The most spectacular was one 83572A
plugin
where sulfuric acid from one big wet tantalum ate off it leads completely, spilled on the
motherboard
and ate a hole through the entire motherboard PCB, through all of its copper and fiberglass
layers.

That is impressive! Were these rubber sealed versions?

Craig


Re: tants new vs old

Craig Sawyers <c.sawyers@...>
 

The only place I've ever seen a wet tantalum leak is in the 184 TMG; fortunately, corrosion was
limited
to the capacitor leads themselves. There must be environmental factors at work, because some
people see a lot of wet Ta failures. It's probably high ambient temperatures. -Dave
The only thing that can degrade a rubber-sealed wet tantalum, and also (particularly) axial
electrolytics, is the flux cleaning regime. As an example I bought a NOS P6046 differential probe, and
needed to change the power supply to 240V from 110V. Good job - the axial electrolytics were leaking
out of the rubber end seal and was chomping the circuit board traces. That was almost certainly a
result of defluxing solvent that was used, which penetrated the end seals causing long term
degradation. More affordable wet tantalum (which are still punishingly expensive) have/had rubber
seals which even when perfect can allow slow water loss, particularly over 100C.

Wet tantalum is used extensively in military and space applications, and is known for its long-term
reliability. Although an older paper this sets out the inherent reliability
downloads.hindawi.com/journals/apec/1976/123103.pdf .

Typical cost for a regular grade AVX wet tantalum, 400uF/100V is UKP176.

Also, current information from Vishay, particularly "Product which has been built and test supported
to 10 000 hours (per MIL Specification) has seen useable life well beyond twenty years. Some of the
tantalum products were the space probes launched in the 1970's which are still going."
https://www.vishay.com/docs/42096/he3faq.pdf

And https://www.vishay.com/docs/49584/49584_pt9033.pdf for glass hermetic sealed versions (which
resists water loss).

As a result, these hermetically sealed mil spec ones are seriously not cheap. But if you need long
life, and tiny leakage currents, and extended life, wet tant is the way to go.

Craig


Re: tants new vs old

Adrian
 

Hi Craig,

You would think that should be the case but as I commented on a recent thread, I've got a whole bunch of Tek 7000 series stuff working over the last couple of years, most recently the 7912AD and I guess I've replaced well over a dozen dead-short Tants in that time.

All bar one was blue, the other was 'brown with stripes' none have been yellow or orange. Yet, I would say that an (unscientifically derived) statistic is there are more yellow & orange Tants present in the equipment. Now, I've not looked to see if, for instance yellow/orange are always 5V and so on - just my observation - along with the comment I have worked on as much HP gear and have never had to replace a single Tant in that but don't even know if they used them though!

I assume the colour may be an indicator of manufacturer (the yellow is very 'AVX'?) I wonder if anyone else has noticed a colour bias in the failure rate?

...I now await a puff of (yellow/orange) smoke from the bench......

Adrian

On 7/18/2018 6:46 AM, Craig Sawyers wrote:
The only type you have to be concerned about are bead tantalum capacitors used for local power supply
rail decoupling. Colour is unimportant.


Re: tants new vs old

Dave Seiter
 

The only place I've ever seen a wet tantalum leak is in the 184 TMG; fortunately, corrosion was limited to the capacitor leads themselves.  There must be environmental factors at work, because some people see a lot of wet Ta failures.  It's probably high ambient temperatures. 
-Dave

From: Craig Sawyers <c.sawyers@...>
To: TekScopes@groups.io
Sent: Tuesday, July 17, 2018 10:47 PM
Subject: Re: [TekScopes] tants new vs old

The only type you have to be concerned about are bead tantalum capacitors used for local power supply
rail decoupling. Colour is unimportant.

They will be 6V rating on 5V rails and 16V rating on 15V rails. And bewilderingly 50V rating on 50V
rails. They randomly fail as a dead short. Replace  them with low ESR aluminium electrolytic with
three times the capacitance. That is exactly what Tek did in later serial numbers.

Anywhere else - as coupling capacitors between stages for example - are generally just fine.

I have never personally seen a wet tantalum leak. Which is just as well because replacements cost a
king's ransom.

Craig

Subject: [TekScopes] tants new vs old

I am guessing that this subject has been hashed over a number of times but as I am fairly new here i
will ask away.


Re: tants new vs old

Tom Gardner
 

On 18/07/18 06:46, Craig Sawyers wrote:
I have never personally seen a wet tantalum leak. Which is just as well because replacements cost a
king's ransom.
I have had axial tantalum capacitors leak and destroy nearby PCB traces, specifically
C6246 and C6341 in a Tek 1502 TDR.


Re: tants new vs old

Sergey Kubushyn
 

On Wed, 18 Jul 2018, Craig Sawyers wrote:

The only type you have to be concerned about are bead tantalum capacitors
used for local power supply rail decoupling. Colour is unimportant.

They will be 6V rating on 5V rails and 16V rating on 15V rails. And
bewilderingly 50V rating on 50V rails. They randomly fail as a dead short.
Replace them with low ESR aluminium electrolytic with three times the
capacitance. That is exactly what Tek did in later serial numbers.
Better yet replace them with higher voltage rated ones if they are
available. 20V tantalum lasts forever on 5V rail as well as 35V one on 15V.

Anywhere else - as coupling capacitors between stages for example - are
generally just fine.

I have never personally seen a wet tantalum leak. Which is just as well
because replacements cost a king's ransom.
I can send you a handful :) They _ALWAYS_ leak sooner or later because of
their very chemistry. They are filled with sulfuric acid and although silver
sulfate has very low solubility it is still not zero so sooner or later that
nice silver can turns to sieve. And the fact they are almost always used at
their rated voltage or even above also doesn't make them last longer...

I have dozens of those leaked in 492BP/494AP SAs that I refurbish on a
regular basis. The same is true for those HP 835xx microwave plugins that I
revived quite a few -- I have never seen a single such plugin that didn't
have at least one leaked wet tantalum. The most spectacular was one 83572A
plugin where sulfuric acid from one big wet tantalum ate off it leads
completely, spilled on the motherboard and ate a hole through the entire
motherboard PCB, through all of its copper and fiberglass layers.


Craig

Subject: [TekScopes] tants new vs old

I am guessing that this subject has been hashed over a number of times but as I am fairly new here i
will ask away.


---
******************************************************************
* KSI@home KOI8 Net < > The impossible we do immediately. *
* Las Vegas NV, USA < > Miracles require 24-hour notice. *
******************************************************************


Re: tants new vs old

Craig Sawyers <c.sawyers@...>
 

The only type you have to be concerned about are bead tantalum capacitors used for local power supply
rail decoupling. Colour is unimportant.

They will be 6V rating on 5V rails and 16V rating on 15V rails. And bewilderingly 50V rating on 50V
rails. They randomly fail as a dead short. Replace them with low ESR aluminium electrolytic with
three times the capacitance. That is exactly what Tek did in later serial numbers.

Anywhere else - as coupling capacitors between stages for example - are generally just fine.

I have never personally seen a wet tantalum leak. Which is just as well because replacements cost a
king's ransom.

Craig

Subject: [TekScopes] tants new vs old

I am guessing that this subject has been hashed over a number of times but as I am fairly new here i
will ask away.


Re: scanned: Tube aging and selection documents

tek_547
 

Same as for the Xformers Kurt, very valuable.
Thanx for that, René


Re: scanned: Tek Transformers

tek_547
 

Very interesting and good piece of information.
Thanx, René


tants new vs old

Jim Olson <v_12eng@...>
 

I am guessing that this subject has been hashed over a number of times but as I am fairly new here i will ask away.

I see on the different boards from the scopes I now have some from same unit and others same board different scopes that there is a vast array of colors of tants used. So my question are the solid blue ones better than the multi color ones or just a new better one.
Also what are the differences between the solid blue, solid orange and ones that are brown, green, blue?
Looking at Digikey and Mouser which are the best manufacturer of the new ones? There is a big difference in price.
On a couple of the main boards there are some leaking silver tants so best type replacement to use there or just replace with a appropriately adjusted electrolytic?
One board has a burned up tant and scorched board and other components will have to test them. I am thinking to replace all the multi colored ones with newer from one of the suppliers especially the wet axial ones they haven't leaked enough to damage the board but just barely. How does one dispose of the bad wet ones/

Jim


Re: 465B Square Wave problem

 

The termination is not required for the low and medium frequency
compensation which is just as well if a x10 probe is used because then
adjusting the x100 attenuator in a 465B requires a high level signal
for 5V/div. If a normalizer is used instead of a x10 probe, then this
is not a problem because the sensitivity will be 1V/div.

On Wed, 18 Jul 2018 00:25:19 +0000 (UTC), you wrote:

Timothy,
You do not mention the use of a 50 ohm termination at the scope end of the BNC cable.This termination is necessary in order to have a proper signal.Compensation adjustments should be done using a 10X probe, or a capacitance normalizer.The intent is for the input R/C time constant to be the same for all attenuator settings.
HankC, Boston, WA1HOS


Re: 465B Square Wave problem

 

Timothy,
You do not mention the use of a 50 ohm termination at the scope end of the BNC cable.This termination is necessary in order to have a proper signal.Compensation adjustments should be done using a 10X probe, or a capacitance normalizer.The intent is for the input R/C time constant to be the same for all attenuator settings.
HankC, Boston, WA1HOS


Re: 465 Stuck in Add Mode (solved)

Ray Burke
 

mailto:mustang_gt_y2k=yahoo.com@groups.io
I used to work on those old 465's and the diode switching circuits are fun to troubleshoot.  It was around 30 years ago, but you could check the diodes in circuit with the diode function of a DMM.  With the scope being that old
the switches maybe dirty?  Try some Deoxit in the back of all the verticle mode switches, and exercise them, about ten to twenty times.  ( You may need to remove verticle preamp board to get to the switches, or nozzle and
flashlight.  
Maybe some of the regular contributers  could chime in.
Later Ray Burke

On Monday, July 16, 2018 11:21 AM, Timothy via Groups.Io <mustang_gt_y2k=yahoo.com@groups.io> wrote:


I dont want to post bad info. I was up until 2am last night and I thought I had it licked before I went to bed.

I just powered up the scope and when I switch from channel one to channel two it is back in add mode. After about 20-30 sec this clears up and it works normally.

Could it be that it was in fact the capacitor that was at fault and the 1kv substitute is not responding as quickly as the 150v?

I want this scope to work perfectly. Its a very nice and clean scope.

Working on these scopes is like working on a classic car both rewarding and frustrating!


Re: 465B Square Wave problem

 

On Mon, 16 Jul 2018 11:28:00 -0700, you wrote:

Everything done in order as laid in the manual.
This is important. You can ignore the attenuator compensation to
start off with by working only with the 5mV/div range where all of the
attenuators are switched out. If you cannot get the transient
response of the 5mV/div range calibrated, then stop there are fix it.

You should be able to do the low and medium frequency transient
response adjustments with your 60mhz function/arbitrary wave generator
since you have another oscilloscope for comparison purposes however
the high frequency transient response adjustment will require a faster
clean source and a 50 ohm feedthrough attenuator.


Re: OT: Scopes and other electronics on "The Outer Limits"

 

On Mon, 16 Jul 2018 13:37:12 -0700, you wrote:

On Sun, Jul 15, 2018 at 08:45 AM, David Hess wrote:

They were using stories from actual science fiction writers. Larry
Niven wrote Inconsistent Moon which was episode 12 of season 2.
AAAAHHHH, I think I remember that one. the guy from Family Ties was on it and he was a scientist who figured the sun had gone nova, but it turned out to be a solar flare. He had the hots for this Woman and was a bit shy to approach her but when he thought the world was gong to end he went for her. She got a little bit miffed over that but got over it. the shockwave at the speed of sound hit them and the flare did kinda FUBAR the Earth but it did not kill everyone. he was kinda thinking when he wanted to cook all the food they had because the utilities were bound to go off.
For what it is worth, the woman was based on Marilyn "Fuzzy Pink"
Niven from the time that I assume he was dating her.

https://tvtropes.org/pmwiki/pmwiki.php/Main/WriteWhoYouKnow

It had a bit more of a human touch than I prefer in scifi but not too bad. Even with Star Trek, which I cut my teeth on, the ones that are like a soap opera are not my favorite to say the least. I don't really even watch DS9 because of that. not that it is bad, it is just not my taste.
There was a close link between Larry Niven and David Gerrold who wrote
many Star Trek - TOS episodes. I think Niven only wrote one which
ended up in TAS. I do not know about later Star Trek.

When I was a kid, 7 when it came out nobody was allowed to talk when it was on. We had the choice, shut up, outside or hell broke loose, maybe locked in a closet or something. My Parents were not really abusive at all. Tough, but then so was I. But the olman was allowed to talk and he made some jokes about it that literally had me on the floor rolling in laughter. I mean to the point where I missed part of the show. But they did reruns so all was not lost.

That series was great, as well as some that followed, but I have to say, the new Outer Limits actually beat them.
I found some of the old Twilight Zone and Outer Limits episodes to be
terrifying as a child. The Zanti Misfits comes to mind.

The Masters of Science Fiction and Masters of Horror anthologies which
ran around 2007 were pretty good. I think they drew on some of the
same source material as The Twilight Zone and The Outer Limits. The
more recent Black Mirror series is also worth checking out.

I recall seeing a couple instances of Tektronix oscilloscopes in those
series but would have to rewatch them to be sure. I have only watched
a couple Black Mirror episodes so far.


Re: 465B Square Wave problem

Fabio Trevisan
 

Hello Timothy,
You don't give much details about the conditions and scope settings you used...
What is the frequency of this signal you pictured or, alternatively, what is the Horizontal sweep (time / div)?
It would be useful you take a picture at a fastest sweep speed that you can, that can illustrate the first 2/3rds of the signal transition (and also report what sweep is ir)... There seem to be a distinct speed transition before the signal reaches 2/3rds of the amplitude, from the last 1/3rd and usually there's no single mis-adjustment that can cause this kind of waveform, but without more details of how much time this slow transition represents, and not knowing what is the transition speed of the fastest part, it's hard to tell.
To improve the diagnostic, there is a few tips / advice:
1.When you don't know anything about the step response of a scope (i.e. if there are defects and/or mis-adjustments), you want to eliminate as much possible problem causes in the beginning.
In that line of thinking, the first thing you want / need to eliminate are the input attenuators so, until you know that the pre-amplifier and output amplifier are working correctly, you need to feed a signal that's between 20 to 30mVpp and set the oscilloscope input to 5mV/div. That setting will by-pass all the attenuators and leave you only with the pre-amplifier and output amplifier.

2. To further eliminate possible sources of step-response problems (all the input circuitry and all the pre-amplifier), you can try to follow the instructions of the "Check/Adjust Low Frequency Compensation" which is the section 16, of the Vertical Calibration chapter, on page 4-32, steps g and h. (of the 465B User and Service manual (e.g. PN 070-2757-00).
This is a very handy procedure for folks like you (and me) who don't have the proper generators.
What this step g. instructs you to do is, essentially, using the scope's own channel alternation / chopping circuitry, as an internal square wave generator.
Here follows a brief description...Under normal operation when the scope switches the signal back and forth between CH1 to CH2 in alternate mode, you cannot see the transition (from one channel to another) because this transition is performed at the end of every sweep, when the trace is already blanked.
Following this procedure, by pulling the cable out from J4387, and injecting into this plug a TTL level signal (and use this same signal to trigger the scope), you actually cheat the scope into switching from CH1 to CH2 (or vv) in the middle of a sweep.
For that to work, you need to adjust the total horizontal sweep time (10 divs) to be >= 2x the period of the square wave you're feeding into J4387.
For this test you will keep both CH1 & CH2 inputs set to GND (no need to feed any signal to the CH1 and CH2 inputs), and the resulting square wave you're going to see on the screen is controlled solely by the position controls of CH1 and CH2 (CH1 position will control, say, the upper part of the square wave, and CH2 position will control the lower part, or v-v).
Adjust CH1 and CH2 position so that you get a 6 division p-to-p waveform on the screen.
If the rounded square wave corner still shows in this mode, you will have eliminated the pre-amplifiers and everything that is before the channel switching circuitry, as a possible cause.
If you see a relatively good square wave, without the rounded coners... then your problem will surely be on the input circuitry / attenuators / pre-amplifiers.

Hope this helps you advance some steps.

KRgrds,

Fabio

On Mon, Jul 16, 2018 at 02:28 PM, Timothy wrote:


I have a 465B that had a intermittent beam. 90% of the time the screen was
blank. I ended up changing the multiplier (although I am not sure that was the
problem) and the vertical output board and now my trace is back but there is a
problem that is the same on both channels. The problem is on all divisions to
a degree. I have changed some of the attenuator settings and even tried
adjusting the compensation from the preamp board but it has little effect.
Does anyone have any familiarity with this issue or can at least point me in
the direction I should look first?

https://groups.io/g/TekScopes/photo/56677/12?p=Name,,,20,1,0,0
https://groups.io/g/TekScopes/photo/56677/13?p=Name,,,20,1,0,0

Something strange I have noted is when I probe say test point 1 (the resistor
coming out of the attenuator box) or TP2 @ CR1193 on the vertical preamp the
trace on the broken scope blurs so much you cant even see it anymore, the
screen just becomes brighter and the test scope gets a blurred out trace, its
there, just fat. When I test the diode at point 9 I get a clean trace and the
damaged scope does not blur. Im not sure if this is related or if I am doing
something wrong. I was probing a 465 (not B) last night and did not have any
of these problems.

Another question about the manual. The arrows to the test points in the Figure
8-11 diagram dont seem to point to a definite spot. i have been looking at the
schematic to try and determine where exactly I am supposed to probe. Am I
doing this right?


Re: Homemade tunnel diodes

Leo Bodnar
 

Joseph,
I got it working at about 1MHz. This is not high by any standard but I was excited to make it oscillate at all.
Frequency depends on RC network time constant.
Leo


Re: scanned: Tube aging and selection documents

Alessandro Cattaneo
 

Hi Kurt,
thanks for these really interesting downloads.