Hi all,

I tore down a Marconi signal generator today. It's been awaiting my
attention for quite a while. Can't recall the model number off hand but
it does 10kHz to 5.4Ghz IIRC. I bought it from some chap who told me it
had a faulty smoothing cap in the PSU 'cos it was generating signals with
ripple on it. He told me he'd been quoted GBP280 ($387 in US dough as of
today's date) for a new replacement from Marconi and I bought it on that
understanding. Anyway, I tore it down today and located the said
capacitor. Here it is:

https://www.flickr.com/photos/128859...in/dateposted-
public/

This is the only pic that came out for some reason, but it's got most of
the important info on it. You can't quite see, but it has 5 terminals for
some reason, but on the board only 2 of them are connected. It's gone
seriously low-res internally, BTW, so *does* need replacing.

Questions: what makes this thing so special as to cost so much?
Why have the designers used such a huge capacity cap in this low current
drain application?
If I can source a generic electrolytic of the same spec or better for
30 quid, why should I not use that instead of the bespoke replacement??

Cursitor Doom wrote...

I tore down a Marconi signal generator today. ...
faulty smoothing cap in the PSU .. GBP280.

I have a 280-pound capacitor, four of them in fact.
Well, they must weigh something in that vicinity.
They cost $500 each, including pallet shipping.


--
Thanks,
- Win

On 06/08/2017 01:39 PM, Winfield Hill wrote:
Cursitor Doom wrote...

I tore down a Marconi signal generator today. ...
faulty smoothing cap in the PSU .. GBP280.

I have a 280-pound capacitor, four of them in fact.
Well, they must weigh something in that vicinity.
They cost $500 each, including pallet shipping.

The physically largest capacitor I ever saw in person was a PIO type
rated IIRC for a couple of uF at several kV; it weighed about as much as
a bowling ball and was about the same size

bitrex wrote:
On 06/08/2017 01:39 PM, Winfield Hill wrote:
Cursitor Doom wrote...

I tore down a Marconi signal generator today. ...
faulty smoothing cap in the PSU .. GBP280.

I have a 280-pound capacitor, four of them in fact.
Well, they must weigh something in that vicinity.
They cost $500 each, including pallet shipping.

The physically largest capacitor I ever saw in person was a PIO type
rated IIRC for a couple of uF at several kV; it weighed about as much as
a bowling ball and was about the same size


Our 170 pound energy discharge capacitors, each 70 uF at 12 kVDC:
http://capturedlightning.com/photos/...ps/MAXCAP3.JPG

"Bert Hickman" wrote in message
...
I have a 280-pound capacitor, four of them in fact.
Well, they must weigh something in that vicinity.
They cost $500 each, including pallet shipping.

The physically largest capacitor I ever saw in person was a PIO type
rated IIRC for a couple of uF at several kV; it weighed about as much as
a bowling ball and was about the same size


Our 170 pound energy discharge capacitors, each 70 uF at 12 kVDC:
http://capturedlightning.com/photos/...ps/MAXCAP3.JPG

I've worked with capacitors bigger than that, although I think they were in
sections so maybe it's not technically true to say "bigger capacitor"
(singular). :^) Ratings were around 100s uF, 2000V, lots of amps.

Tim

--
Seven Transistor Labs, LLC
Electrical Engineering Consultation and Contract Design
Website: http://seventransistorlabs.com

Tim Williams wrote:
"Bert Hickman" wrote in message
...
I have a 280-pound capacitor, four of them in fact.
Well, they must weigh something in that vicinity.
They cost $500 each, including pallet shipping.

The physically largest capacitor I ever saw in person was a PIO type
rated IIRC for a couple of uF at several kV; it weighed about as much as
a bowling ball and was about the same size


Our 170 pound energy discharge capacitors, each 70 uF at 12 kVDC:
http://capturedlightning.com/photos/...ps/MAXCAP3.JPG

I've worked with capacitors bigger than that, although I think they were in
sections so maybe it's not technically true to say "bigger capacitor"
(singular). :^) Ratings were around 100s uF, 2000V, lots of amps.

Tim

66 uF, 276 kV, 3000 A
but that was an aggregation of multiple cans.

Bert Hickman wrote...

bitrex wrote:
On 06/08/2017 01:39 PM, Winfield Hill wrote:
Cursitor Doom wrote...

I tore down a Marconi signal generator today. ...
faulty smoothing cap in the PSU .. GBP280.

I have a 280-pound capacitor, four of them in fact.
Well, they must weigh something in that vicinity.
They cost $500 each, including pallet shipping.

The physically largest capacitor I ever saw in person was a PIO type
rated IIRC for a couple of uF at several kV; it weighed about as much as
a bowling ball and was about the same size

Our 170 pound energy discharge capacitors, each 70 uF
at 12 kVDC:
http://capturedlightning.com/photos/...ps/MAXCAP3.JPG

Whoa, you have 11 of them! My caps look like those.


--
Thanks,
- Win

On Fri, 9 Jun 2017 08:05:20 -0400, bitrex
wrote:

On 06/08/2017 01:39 PM, Winfield Hill wrote:
Cursitor Doom wrote...

I tore down a Marconi signal generator today. ...
faulty smoothing cap in the PSU .. GBP280.

I have a 280-pound capacitor, four of them in fact.
Well, they must weigh something in that vicinity.
They cost $500 each, including pallet shipping.

The physically largest capacitor I ever saw in person was a PIO type
rated IIRC for a couple of uF at several kV; it weighed about as much as
a bowling ball and was about the same size

At 280 lbs, it would take several big men to move the thing. (Or a
forklift). Not the kind of thing you can just replace on your work
bench, because the bench would probably collapse.

On Friday, June 9, 2017 at 9:06:01 AM UTC-4, wrote:


At 280 lbs, it would take several big men to move the thing. (Or a
forklift). Not the kind of thing you can just replace on your work
bench, because the bench would probably collapse.


You realize the OP was referring to cost (280 pound sterling), not weight. If you're making a joke, the second poster beat you to it.

On a sunny day (Fri, 9 Jun 2017 08:05:20 -0400) it happened bitrex
wrote in :

On 06/08/2017 01:39 PM, Winfield Hill wrote:
Cursitor Doom wrote...

I tore down a Marconi signal generator today. ...
faulty smoothing cap in the PSU .. GBP280.

I have a 280-pound capacitor, four of them in fact.
Well, they must weigh something in that vicinity.
They cost $500 each, including pallet shipping.

The physically largest capacitor I ever saw in person was a PIO type
rated IIRC for a couple of uF at several kV; it weighed about as much as
a bowling ball and was about the same size

In the sixties I worked in a company that made HV transformers and equipment
for power stations, railways, etc, now the caps I have seen in the HV test room
were alsmost as big as me.
Soem of the transformers required a ladder to climb on those.
The caps looked a bit like these:
http://www.hvbright.com/products/hig...unt-capacitor/
Dangerous place...

In article ,
Cursitor Doom wrote:

Questions: what makes this thing so special as to cost so much?

At a guess - exact replacement parts might no longer being made,
the equipment manufacturer has a small remaining stock, there
may be no other source. Some owners of the equipment (e.g. military
and some businesses) may have an "exact replacement only" policy
for spare parts, to avoid the need to send equipment through a
formal requalification process.

So, Marconi can charge that much for a cap, because there are people
willing to pay it (rather than scrap the whole piece of equipment).

Why have the designers used such a huge capacity cap in this low current
drain application?

Might be "because they could". Or, possibly, some of the downstream
circuitry might have poor power-supply rejection, and having a truly
huge filter cap might be the only way to get ripple-related noise
and sidebands down low enough to meet the device's specs. They might
also have figured that this part might be prone to degrade over the
years (as it apparently has done?) and they installed one of larger-
than-initially-required capacity to stave off the effect of this
aging and degradation.

If I can source a generic electrolytic of the same spec or better for
30 quid, why should I not use that instead of the bespoke replacement??

The extra hold-down terminals might be needed in order for the device
to meet its reliability specifications, when installed under
conditions of high vibration and possible acceleration shock (e.g. in
military installs, on boats or airplanes). Without the additional
pins soldered to the board, vibration could result in the cap
shaking back and forth, with all of the stress placed on the two
solder joints (and the PCB traces) resulting in stress cracking.

A standard modern cap of the same capacity and voltage rating, and
equal or better temperature and lifetime specs, is likely to be a good
deal lighter than the original. If you can find one which fits the
connection terminals, and don't mind the fact that it might break
loose if you use the equipment in a bomber that's flying through
intense flak explosions for months on end, I suspect it'd work out
just as well for you.

In article ,
says...

In article ,
Cursitor Doom wrote:

Questions: what makes this thing so special as to cost so much?

At a guess - exact replacement parts might no longer being made,
the equipment manufacturer has a small remaining stock, there
may be no other source. Some owners of the equipment (e.g. military
and some businesses) may have an "exact replacement only" policy
for spare parts, to avoid the need to send equipment through a
formal requalification process.

So, Marconi can charge that much for a cap, because there are people
willing to pay it (rather than scrap the whole piece of equipment).

Sounds like the $ 100,000 diode for the military that could have been
replaced by a diode that cost less than one dollar except for the
military spec. Seems the military supply depot did not have any. The
company that made them did not have any,so they had to make one. Could
not make just one, had to do it in a large batch. Probably made 10 to
20 thousand of them.

On Thursday, 8 June 2017 19:56:29 UTC+1, Dave Platt wrote:
In article ,
Cursitor Doom wrote:

Questions: what makes this thing so special as to cost so much?

At a guess - exact replacement parts might no longer being made,
the equipment manufacturer has a small remaining stock, there
may be no other source. Some owners of the equipment (e.g. military
and some businesses) may have an "exact replacement only" policy
for spare parts, to avoid the need to send equipment through a
formal requalification process.

So, Marconi can charge that much for a cap, because there are people
willing to pay it (rather than scrap the whole piece of equipment).

Why have the designers used such a huge capacity cap in this low current
drain application?

Might be "because they could". Or, possibly, some of the downstream
circuitry might have poor power-supply rejection, and having a truly
huge filter cap might be the only way to get ripple-related noise
and sidebands down low enough to meet the device's specs. They might
also have figured that this part might be prone to degrade over the
years (as it apparently has done?) and they installed one of larger-
than-initially-required capacity to stave off the effect of this
aging and degradation.

If I can source a generic electrolytic of the same spec or better for
30 quid, why should I not use that instead of the bespoke replacement??

The extra hold-down terminals might be needed in order for the device
to meet its reliability specifications, when installed under
conditions of high vibration and possible acceleration shock (e.g. in
military installs, on boats or airplanes). Without the additional
pins soldered to the board, vibration could result in the cap
shaking back and forth, with all of the stress placed on the two
solder joints (and the PCB traces) resulting in stress cracking.

A standard modern cap of the same capacity and voltage rating, and
equal or better temperature and lifetime specs, is likely to be a good
deal lighter than the original. If you can find one which fits the
connection terminals, and don't mind the fact that it might break
loose if you use the equipment in a bomber that's flying through
intense flak explosions for months on end, I suspect it'd work out
just as well for you.

Marconi Instruments were hot on vibration tests since they're key to reliability in military use. Competitor equipment often failed their tests.

As well as what has been mentioned, a big cap would presumably help ride over an arcing mains connection, giving reliable service where a lesser device would cause malfunction.

As said if you're just using it on a bench you can put whatever cap you like there. It won't be a low ESR type on a 50Hz PSU. If you glue it down it will improve its shock/vibration resilience, but not to match the original marconi & mil specs.


NT

On a sunny day (Thu, 8 Jun 2017 17:28:02 -0000 (UTC)) it happened Cursitor
Doom wrote in :

Hi all,

I tore down a Marconi signal generator today. It's been awaiting my
attention for quite a while. Can't recall the model number off hand but
it does 10kHz to 5.4Ghz IIRC. I bought it from some chap who told me it
had a faulty smoothing cap in the PSU 'cos it was generating signals with
ripple on it. He told me he'd been quoted GBP280 ($387 in US dough as of
today's date) for a new replacement from Marconi and I bought it on that
understanding. Anyway, I tore it down today and located the said
capacitor. Here it is:

https://www.flickr.com/photos/128859...in/dateposted-
public/

This is the only pic that came out for some reason, but it's got most of
the important info on it. You can't quite see, but it has 5 terminals for
some reason, but on the board only 2 of them are connected. It's gone
seriously low-res internally, BTW, so *does* need replacing.

Questions: what makes this thing so special as to cost so much?
Why have the designers used such a huge capacity cap in this low current
drain application?

Nothing, it is a crappy old Philips, the contacts to the pins go wrong too.
Just replace with some caps with right capacitance / voltage and sintered wires,
not that crap.


If I can source a generic electrolytic of the same spec or better for
30 quid, why should I not use that instead of the bespoke replacement??

On Thu, 8 Jun 2017 17:28:02 -0000 (UTC), Cursitor Doom
wrote:

Hi all,

I tore down a Marconi signal generator today. It's been awaiting my
attention for quite a while. Can't recall the model number off hand but
it does 10kHz to 5.4Ghz IIRC. I bought it from some chap who told me it
had a faulty smoothing cap in the PSU 'cos it was generating signals with
ripple on it. He told me he'd been quoted GBP280 ($387 in US dough as of
today's date) for a new replacement from Marconi and I bought it on that
understanding. Anyway, I tore it down today and located the said
capacitor. Here it is:

https://www.flickr.com/photos/128859...in/dateposted-
public/

This is the only pic that came out for some reason, but it's got most of
the important info on it. You can't quite see, but it has 5 terminals for
some reason, but on the board only 2 of them are connected. It's gone
seriously low-res internally, BTW, so *does* need replacing.

Questions: what makes this thing so special as to cost so much?
Why have the designers used such a huge capacity cap in this low current
drain application?
If I can source a generic electrolytic of the same spec or better for
30 quid, why should I not use that instead of the bespoke replacement??

Insane ripoff. Good reason to never buy Marconi.

Looks like you ripped out the hole plating on one pin. With luck, it
will be one of the passive mounting pins.


--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

On Thu, 08 Jun 2017 12:19:54 -0700, John Larkin wrote:

Looks like you ripped out the hole plating on one pin. With luck, it
will be one of the passive mounting pins.

Fortunately it is. :-)

On Thursday, June 8, 2017 at 1:31:39 PM UTC-4, Cursitor Doom wrote:

If I can source a generic electrolytic of the same spec or better for
30 quid, why should I not use that instead of the bespoke replacement??

Many years ago, my Tek 7603 failed to start. I pulled the power supply out and was driven nuts by a simple DC voltage regulator problem in the power supply. A bypass electrolytic capacitor would have been the obvious solution, except this scope used at least a half dozen extra large Mallory built capacitors in parallel, and there's no way they all died together. Adding a bit of external capacitance though brought the voltage right back and the scope to life. Turns out those big caps were dropping out one by one over the years and gave no indication of anything going wrong as they did, until the very last one opened when the supply went out.

Why did I mention all of this? Because I just removed those big Mallorys and stuck in some standard electrolytics of maybe half the total value and taking up about a tenth of the physical area of the originals, and the scope still runs daily with a perfectly clean and stable trace.

In other words, I doubt you'll see any difference by doing what you instinct tells you. That cap may be very low ESR, have special impedance specs or ripple current specs, but I'd be stunned if it makes any real world difference with off the shelf caps. If it were mine, I'd use Panasonic FR series caps.

On 08/06/2017 21:39, wrote:
On Thursday, June 8, 2017 at 1:31:39 PM UTC-4, Cursitor Doom wrote:

If I can source a generic electrolytic of the same spec or better for
30 quid, why should I not use that instead of the bespoke replacement??

Many years ago, my Tek 7603 failed to start. I pulled the power supply out and was driven nuts by a simple DC voltage regulator problem in the power supply. A bypass electrolytic capacitor would have been the obvious solution, except this scope used at least a half dozen extra large Mallory built capacitors in parallel, and there's no way they all died together. Adding a bit of external capacitance though brought the voltage right back and the scope to life. Turns out those big caps were dropping out one by one over the years and gave no indication of anything going wrong as they did, until the very last one opened when the supply went out.

Why did I mention all of this? Because I just removed those big Mallorys and stuck in some standard electrolytics of maybe half the total value and taking up about a tenth of the physical area of the originals, and the scope still runs daily with a perfectly clean and stable trace.

In other words, I doubt you'll see any difference by doing what you instinct tells you. That cap may be very low ESR, have special impedance specs or ripple current specs, but I'd be stunned if it makes any real world difference with off the shelf caps. If it were mine, I'd use Panasonic FR series caps.


When Tektronix had a base in Guernsey, Channel Islands, thay adopted the
following spares procedure.
Each year, divide the stock by half, sell off that half at auction, then
double the price of what they kept in stock.

N_Cook wrote on 6/8/2017 4:46 PM:
On 08/06/2017 21:39, wrote:
On Thursday, June 8, 2017 at 1:31:39 PM UTC-4, Cursitor Doom wrote:

If I can source a generic electrolytic of the same spec or better for
30 quid, why should I not use that instead of the bespoke replacement??

Many years ago, my Tek 7603 failed to start. I pulled the power supply
out and was driven nuts by a simple DC voltage regulator problem in the
power supply. A bypass electrolytic capacitor would have been the obvious
solution, except this scope used at least a half dozen extra large Mallory
built capacitors in parallel, and there's no way they all died together.
Adding a bit of external capacitance though brought the voltage right back
and the scope to life. Turns out those big caps were dropping out one by
one over the years and gave no indication of anything going wrong as they
did, until the very last one opened when the supply went out.

Why did I mention all of this? Because I just removed those big Mallorys
and stuck in some standard electrolytics of maybe half the total value and
taking up about a tenth of the physical area of the originals, and the
scope still runs daily with a perfectly clean and stable trace.

In other words, I doubt you'll see any difference by doing what you
instinct tells you. That cap may be very low ESR, have special impedance
specs or ripple current specs, but I'd be stunned if it makes any real
world difference with off the shelf caps. If it were mine, I'd use
Panasonic FR series caps.


When Tektronix had a base in Guernsey, Channel Islands, thay adopted the
following spares procedure.
Each year, divide the stock by half, sell off that half at auction, then
double the price of what they kept in stock.

And this process led to their going out of business sale?

--

Rick C

On 06/08/2017 01:28 PM, Cursitor Doom wrote:

This is the only pic that came out for some reason, but it's got most of
the important info on it. You can't quite see, but it has 5 terminals for
some reason, but on the board only 2 of them are connected. It's gone
seriously low-res internally, BTW, so *does* need replacing.

Questions: what makes this thing so special as to cost so much?
Why have the designers used such a huge capacity cap in this low current
drain application?

Probably because they got a bunch of large value weird-ass caps cheap
and that's what they use in everything. Like a guy who asked me why they
used a 1N4002 in this one mass-produced rack effects box when a 1N4001
would've been fine from a ratings perspective and it's cuz "that's what
they use in everything"

If I can source a generic electrolytic of the same spec or better for
30 quid, why should I not use that instead of the bespoke replacement??

They're like 5 bucks:

http://www.mouser.com/ProductDetail/United-Chemi-Con/ESMH160VSN473MR50T/?qs=sGAEpiMZZMtZ1n0r9vR22dBjIkbB%252b54P4MErU9o8dM Q%3d

On Fri, 09 Jun 2017 08:00:30 -0400, bitrex wrote:

Probably because they got a bunch of large value weird-ass caps cheap
and that's what they use in everything. Like a guy who asked me why they
used a 1N4002 in this one mass-produced rack effects box when a 1N4001
would've been fine from a ratings perspective and it's cuz "that's what
they use in everything"

That, unlike most everything else you post, makes sense.

On 06/09/2017 02:43 PM, Cursitor Doom wrote:
On Fri, 09 Jun 2017 08:00:30 -0400, bitrex wrote:

Probably because they got a bunch of large value weird-ass caps cheap
and that's what they use in everything. Like a guy who asked me why they
used a 1N4002 in this one mass-produced rack effects box when a 1N4001
would've been fine from a ratings perspective and it's cuz "that's what
they use in everything"

That, unlike most everything else you post, makes sense.


I may be a "communist", but I do understand economies of scale. Build
everything out of LM324s, TL431s, and 555 timers if you can, so long as
it meets the spec.

The old '324 and 555 are sort of disparaged around here, but there's a
reason they're made in their billions each year and it's not because of
sales driven by hobbyists.

I think the reason a lot of software is so bad is because lines of code
are basically free. Hey! There's a library for that, don't "reinvent the
wheel." So what if it's 40,000 lines long and was written by God knows who

On Fri, 09 Jun 2017 18:32:35 -0400, bitrex wrote:

I think the reason a lot of software is so bad is because lines of code
are basically free. Hey! There's a library for that, don't "reinvent the
wheel." So what if it's 40,000 lines long and was written by God knows
who

Er, yes, good point but not sure what it's got to do with the subject
matter of this thread.

On Thu, 8 Jun 2017 17:28:02 -0000 (UTC), Cursitor Doom
wrote:

Hi all,

I tore down a Marconi signal generator today. It's been awaiting my
attention for quite a while. Can't recall the model number off hand but
it does 10kHz to 5.4Ghz IIRC. I bought it from some chap who told me it
had a faulty smoothing cap in the PSU 'cos it was generating signals with
ripple on it. He told me he'd been quoted GBP280 ($387 in US dough as of
today's date) for a new replacement from Marconi and I bought it on that
understanding. Anyway, I tore it down today and located the said
capacitor. Here it is:

https://www.flickr.com/photos/128859...in/dateposted-
public/

This is the only pic that came out for some reason, but it's got most of
the important info on it. You can't quite see, but it has 5 terminals for
some reason, but on the board only 2 of them are connected. It's gone
seriously low-res internally, BTW, so *does* need replacing.

Questions: what makes this thing so special as to cost so much?
Why have the designers used such a huge capacity cap in this low current
drain application?
If I can source a generic electrolytic of the same spec or better for
30 quid, why should I not use that instead of the bespoke replacement??

It will be a limited production component that is no longer made and
the remaining stock has a very high price. Replace it with an
electrolytic of the same capacitance and voltage rating. The extra
terminals are probably connections to internal parallel capacitors. I
once worked on a power supply that had a 600 uF capacitor but when it
went I discovered it was made of 8 x 100 uF in parallel all in the
same encapsulation with two terminals. I assume the 600 mark on the
case was a misprint.

Steve

--
Neural Network Software for Windows http://www.npsnn.com

On Thu, 8 Jun 2017 17:28:02 -0000 (UTC), Cursitor Doom
wrote:

https://www.flickr.com/photos/128859...posted-public/

47,000 uF 16v. You should be able to find that in a physically
smaller package.
http://www.ebay.com/sch/i.html?_nkw=47000+uf+16v

Carefully remove the base from the capacitor, preserving only the base
and the can. If you're really careful, you might be able to also save
the vinyl insulator. Tear out the guts and throw it away. Install
the replacement physically smaller capacitor inside the can,
connecting the capacitor leads to the base to match the original.
Solder it back onto the PCB and you're done.

If you don't care if it looks like the original, forget the
aforementioned process and just solder the replacement cap to the PCB
in place of the can in any manner that will fit.

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Fri, 09 Jun 2017 09:16:52 -0700, Jeff Liebermann
wrote:
(...)

A bit more on the cost of a replacement capacitor. A company that I
worked for in about 1977 was having a similar issue selling
replacement and repair components to its dealers. So, I calculated
what it cost the company to sell an empty box. That's a repair
component that costs zero dollars to purchase and requires no
manufacturing. However, it does carry all the overhead involved in
shipping a product, such as incoming inspection, QA inspection,
inventory control, warehousing, packing, order taking, boxing,
documentation, billing, handling, etc. I estimated $75 cost to
shipping (not including postage). My guess(tm) is that it would be
about 4 times that (due mostly to increased overhead and inflation)
today. That would be $300 to ship an empty box today, which is about
what Marconi is charging.

We "solved" the problem by offering the dealers almost any quantity of
the smaller parts involved for about the same price. Or, we would
throw in a handful of random floor sweepings with a little of
everything we thought might be useful. Either way, the minimum price
to ship anything remained at $75.

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Fri, 09 Jun 2017 09:16:52 -0700, Jeff Liebermann wrote:

Carefully remove the base from the capacitor, preserving only the base
and the can. If you're really careful, you might be able to also save
the vinyl insulator. Tear out the guts and throw it away. Install the
replacement physically smaller capacitor inside the can, connecting the
capacitor leads to the base to match the original. Solder it back onto
the PCB and you're done.

An old friend of mine who collected vintage broadcast radios would use
this technique when re-furbing them to keep up the appearance of
originality. I never considered doing this with non-classic gear before,
but it does make sense as the existing through holes can be used without
needing to accommodate the different lead spacings of the new component.
Thank you, Jeff.