Jump to content

Capacitors, the GOOD, the BAD & the UGLY


Recommended Posts

Keep and eye on the this thread. Mark has graciously provided unlimited thread edits and thus I'll be able to initiate what I can share now and, continually update in the future with new or revised capacitor information.

Link to comment
Share on other sites

Keep and eye on the this thread. Mark has graciously provided unlimited thread edits and thus I'll be able to initiate what I can share now and, continually update in the future with new or revised capacitor information.

Hi Carl

When I saw the topic name I started humming that movies theme song. LOL

I just went to the Dynaco section and looked up a write-up that I started in 2007.

It appears that I can still edit it after all these years.

A few years ago I was timing out trying to write and write.

I mentioned to Mark then that I kept timing out.

He did something to extend that writing time, then.

A year or so ago I once again commented to Mark that I was able to go back and edit it a day or so later.

This means that you could read something one day and the next it may read quite differently or not at all.

For an ongoing and updating topic this is great, it allows one to just adding to the mix and not having to

go over and ditto each write-up.

Link to comment
Share on other sites

  • 4 months later...

Hi Carl

After over fifty years there is still no definitive advice as to which brand or type of crossover capacitors to use.

I've seen one person say Solen is the best and another saying not so.

I've never seen a general agreement yet.

The conversations have gone from a few dollar cap to the more expensive $100.00+ capacitors.

Even an agreement within a capacitor brand does not appear, a non-agreement as too which voltage rating even is better.

It would seem, at least on the surface, that replacing a crossover cap in the Dynaco A-25, for just one example, a $60.00 speaker system, with a $100.00+ capacitor would not seem appropriate.

Maybe in a blind test it would be a noticeable improvement over a common stock Jensen OEM capacitor.

I went to a website, I don't remember the name though, a few years ago.

The chap had tested several dozen capacitors of varying brands and reported his results on his website.

If I remember correctly there was a variance in his testing procedure.

I was terribly confused about what or who was the winner after reading all his data results.

Just a thought for today.

Link to comment
Share on other sites

  • 4 months later...

Keep and eye on the this thread. Mark has graciously provided unlimited thread edits and thus I'll be able to initiate what I can share now and, continually update in the future with new or revised capacitor information.

My apologies for the delay in getting this thread off the ground. I've been quite busy, but have accumulated some interesting data on vintage caps and new caps I've actually experienced in some restorations and xover upgrades. Things have slowed down a bit and I've had a chance to pull some of the test data together. But first, readers will need to understand what these complex looking test reports look like and what they mean.....

Okay, here we go. Attached are two documents that explain how I test capacitors and show examples of tests with some added text to explain what the data shows. I'm starting with this so readers can understand the test results I will be publishing in future posts.

CAPACITOR TESTS EXPLanation text.pdf

BAD CAP SCAN EXAMPLE.pdf

Link to comment
Share on other sites

I thought I'd start the data sharing with a cap from a very popular loudspeaker.

Here are two tests of 6.0 uF capacitors. One is of an original AR3a Industrial Condenser Corp. paper/wax cap and the other is of an ERSE 400 V Pulse-X capacitor. You can see from the data table how the ESR of the original cap starts out at about 1/4 ohm at 20,000 hz and rises from there to over 280 ohms at 10 hz. Also note how the capacitance varies from 6.83 uF @ 20 kHz to 8.82 @ 10 hz. At around the 5 kHz tweeter crossover level where the 6.0 uF cap should be close to 6 uF it tests over 7, about 18% high. This is the reason vintage caps should be replaced.
Now note the relatively level blue phase line is on the ERSE 6.0 uF cap. ESR is significantly lower - in the mili-ohm range. This cap uF value tested about 5% low but varied only 0.2 uF over the entire frequency range. Whereas, the original Ind. Cond. Corp. cap varied 2.0 uF over that same range.

I've added a second 3a 6.0uF cap test done 7 months earlier. Note how similar the phase line 'smile' is to the May 2011 test. This second DUT had an even worse ESR performance; starting out a over 0.3 ohms @ 20 kHz. and a similar 2 uF range of uF data.
My apologies for the scientific notation in the data colums of this test.

I'd put the 3a 6.0 uF cap in my 'BAD' catagory. Uglies yet to follow.....

post-100237-0-44937300-1325391546_thumb.

post-100237-0-06443000-1325391551_thumb.

post-100237-0-19661100-1325423835_thumb.

Link to comment
Share on other sites

It's evident from the many restos I've done of vintage KLH speakers that the designers made, in restrospect, some bad decisions regarding crossover caps. I think most will agree, if one acquires a KLH-5 or 6 or 12, etc, the first thing they do is recap them as most contain those black caps with the red ends along with some brown ones with 3 wires exiting them. Those are the double duty caps.

Well, those black jobs with the red ends fall into my catagory 'the uglies'. I believe all were made by the same OEM supplier who labeled them with different names like "TEMPLE", "MEXICO", "CALLINS" and others. I'm sure you recognize some of these names if you've ever done a recap and replaced those 'uglies'. A pic of some Temple caps is at lower right.

Below are screen shots of tests I've done on them and it's clear they do have high variance in phase and uF values over the freq. range tested. Also, ESR is high - similar to NPE's. Look at the 3.0 Temple cap test. ESR starts out at 1.2 ohms and sky rockets up from there to almost 600 ohms at 10 hz. :o Note the uF values in the 25 uF cap. They start out at almost 27 uF at 20 kHz and rise to over 32 uF at 10 hz. Certainly, not very consistent performance.

With data like this, you should now have a better unterstanding of why vintage caps (and especially these 'uglies') should be replaced by modern ones. In the next post, I'll show tests I've done on some 'good' caps used to replace the 'uglies'.

post-100237-0-18922700-1325510897_thumb.

post-100237-0-68382700-1325510901_thumb.

post-100237-0-75358800-1325510905_thumb.

post-100237-0-75358400-1325510910_thumb.

post-100237-0-15387100-1325510915_thumb.

post-100237-0-12885700-1340191722_thumb.

Link to comment
Share on other sites

And now for some 'GOOD' caps. The tests below are of Dayton film caps used to replace the originals tested in the previous post. Study the consistency of the data in the phase, ESR and uF columns. Pls. disregard the minus signs in the ESR columnar data. It's evident the new, film caps are much improved over that of the previous post.

This type of data is typical for film caps I install during xover upgrades and restorations.

post-100237-0-57189200-1325523072_thumb.

post-100237-0-40045700-1325523077_thumb.

post-100237-0-92723300-1325523081_thumb.

post-100237-0-40797200-1325523086_thumb.

Link to comment
Share on other sites

  • 3 weeks later...

Here is another test of a Fisher XP-9C 12 uF 25V cap branded as "WHALE". It's very bad with ESR ranging from almost 2 ohms @ 20 kHz to 80 ohm @ 10 hz and uF ranging from below 8 to 11.9. The phase range was one of the worst I've seen.

It's replacement, a Carli 12 uF 250V film cap's ESR ranged from 36 miliohms and was still less than 1 ohm @22 hz. and uF ranging from 11.92 going over 12 slightly in the middle freq. and ending at 11.97 @ 10 hz.

post-100237-0-21239000-1326900016_thumb.

Link to comment
Share on other sites

Vintage but GOOD caps. Yes, there are some. Over some years of doing restorations and subsequent cap testing, I have found the AR Compulytic caps made by Sprague have stood the test of time. So now, when I retore an AR with these caps, I simply leave 'em in there. Below are tests done on the 150 uF 50 V and 50 uF 50V Compulytics found in some later AR3a's accompanied by tests run with the same calibration on ERSE 100V NPE caps of equivalent uF values.

The 150 uF Compulytic cap that's wired in parallel with the 11 inch 3a woofer creates a 2nd order electrical rolloff. I estimate it operating passband to be in the 10 to 800 hz range. Here we see the range of uF to be 156 to 142, well within the10% tolerance printed on the cap. ESR is 2.1 ohms to 58 miliohms. The ERSE 150 uF cap's test showed similar results. However, the ERSE cap had a brader range of phase angle over the entire test range.

The 50 uF Compulytic cap is an important cap for the midrange because it's wired in series with the midrange driver that operates from about 300 hz to over 5000 hz. It's uF range over that hz range was only 52.3 to 50.4 uF; quite impressive. ESR was .40 ohms to 19 miliohms. The ESR 50 uF cap performed similarly.

The bottome line: Compulytics don't need to be changed out. OTOH, the Chicago Industrial caps, as was shown in a previous post should be replaced.

post-100237-0-61799600-1327094682_thumb.

post-100237-0-18402200-1327094686_thumb.

post-100237-0-49351600-1327094689_thumb.

post-100237-0-97531800-1327094692_thumb.

Link to comment
Share on other sites

When I restore KLH 6s and 17s, I tend to replace the 8mfd caps with $4 or $5 5% Daytons or Solens. But I leave the two (what are they, 2 mfd?) caps on the board alone, due mostly to laziness. The 8s are easy to access (they generally float among the fiberglass) and it takes only a few minutes to replace them. Would there be sonic improvements if I replaced the other caps as well? I generally use the "reduce" setting on the three-position tweeter control, and if I'm understanding the crossover schematic right, it looks like the signal doesn't go through the other caps in this setting. Any thoughts?

Link to comment
Share on other sites

When the switch is in the decr. position, only the 8 uF cap is in the circuit. The other two 2 uF caps don't come into play unless the switch is in normal or incr. position.

The model 17 also has two 2uf and one 8 uF cap in the xover. However, with that one, at least one of the two 2 uF caps is active with the 8 uF cap in either of the two switch positions (according to the schematic drawin by Zilch).

Link to comment
Share on other sites

  • 1 year later...

I had the opportunity to test my first AR4x 20 uF Sprague Compulytic cap. As expected, it tested very well varying only 1 uF over 500 to 20,000 hz range. It stayed in the 20 to 21 uF range. ESR also was steady in the 0.1 to 0.3 ohm range. Elect. phase angle was somewhat dissapointing though. It fell from -90 deg (target for capacitors) to -126 deg over the 500 to 20 kHz range. This parameter is not as critical as the ESR and target uF and thus I'd have to conclude once again that Sprague Compulytic caps don't need to be replaced because they don't deteroriate over time like most NPE caps or, the old paper-wax Chicago Industrial caps used by AR.

Link to comment
Share on other sites

Even though I am unable to follow all of the tech talk here (phase angle?), I appreciate your having conducted these tests to confirm what many have suspected - - that some of these old Sprague caps somehow retain their intended performance values over these many years. Thanks for the good work, Carl, it is another bit of helpful information for those who do, and for those who don't understand all the tech talk but wish to keep alive or restore these old classic speakers.

Link to comment
Share on other sites

When I restore KLH 6s and 17s, I tend to replace the 8mfd caps with $4 or $5 5% Daytons or Solens. But I leave the two (what are they, 2 mfd?) caps on the board alone, due mostly to laziness. The 8s are easy to access (they generally float among the fiberglass) and it takes only a few minutes to replace them. Would there be sonic improvements if I replaced the other caps as well? I generally use the "reduce" setting on the three-position tweeter control, and if I'm understanding the crossover schematic right, it looks like the signal doesn't go through the other caps in this setting. Any thoughts?

IMHO that's a big mistake. The 2uF caps are cheap (Madisound has some for $0.35 each) and not difficult to replace. The original black/red PVC caps are awful. Here is a photo (I've posted this before) that shows a KLH 17 xo. The "exploded" cap is a 2uF.

Just my 2 cents....

Kent

post-101828-0-88421000-1370107761_thumb.

post-101828-0-83775400-1370107770_thumb.

Link to comment
Share on other sites

  • 2 weeks later...
  • 7 months later...

Here's a reason to seriously consider recapping those OLA's with the Unicon 16 uF NPE capacitors. I recently recapped a pair of OLA's with Ser. #'s 817XX. For those of you visiting these WT2 caps for the 1st time, go back and review post #4 where I explain the test results.

The first test on the left is a test of the Unicon cap. Note the large, undesirable but typical swing in electrical phase. MFD ranged from 13.4 to 17.9 uF. ESR was in the normal to elevated range.

The ERSE PET (MET) film cap has a much more consistent electrical phase near the -90 degrees. The one particular sample I tested had a uF range from 14.7 to 15.8, tighter than the original but below target a smidge. Note ESR was mostly in the miliohm range and didn't elevate into the Unicon range until about 100 hz and below.

post-100237-0-33314200-1389473683_thumb.

post-100237-0-55249600-1389473694_thumb.

Link to comment
Share on other sites

  • 2 months later...

Here is one test on an AR4x 20 uF paper wax capacitor. Those are the ones most often found in 4x's. It's clear from this particular sample it was not in good shape and certainly a candidate for replacement. Just look at the capacitor uF values in the far right column and how they range broadly over the freq. range. Also note the extreme waviness to the blue phase angle line in the chart.

The other test posted here is of a Dayton 1% tol. 20 uF PP film cap. It is very good with the uF column staying very consistent over the entire 10-20,000 hz. Also note the significantly lower ESR values (vs the paper wax test) in the column with the heading "Real....".

So, the bottom line here is, the one tests definitely shows the potential for needing to replace the original paper/wax 20 uF cap with a modern film type. Because the caps sits atop the woofer coil, I suggest you cut the wires and leave the original cap in there and attach the wires to the new, film cap.

post-100237-0-96305200-1394990769_thumb.

post-100237-0-74754700-1394990780_thumb.

Link to comment
Share on other sites

  • 3 months later...

Nice work Carl and just what I needed to read (amongst a few other threads). It will certainly help me with my choice of caps for my AR91 project. I also appreciated the info on the Sprauge Compulytic and how their ESR is low; much like a film cap.

It would be great to see some readings on the large Callins can caps (vs the Sprauges).

Link to comment
Share on other sites

IMHO that's a big mistake. The 2uF caps are cheap (Madisound has some for $0.35 each) and not difficult to replace. The original black/red PVC caps are awful. Here is a photo (I've posted this before) that shows a KLH 17 xo. The "exploded" cap is a 2uF.

Just my 2 cents....

Kent

Kent,

Did the explosion of the cap cause any collateral damage? Like to the amp or any other components (drivers) in the speaker?

Link to comment
Share on other sites

One issue that should be considered when old electrolytics are replaced is ESR of capacitor. Film and foil caps have generally lower ESR`s than non polar electrolytics. When ESR of cap or DSR of choke do change, this will also change slope of crossover and this will change phase relationship near crossover frequency, especially if higher slope crossovers are used. So... upgrade can actually be downgrade, if the original design was made properly.

This is not too important in designs like AR2ax where slopes are first order ones, When restoring designs like 5 or 3a this might have some effect. Lower cap ESR or inductor DCR can usually be compensated with suitable series resistors. Personally I do think that it is not too wise to buy better and more expensive part and downgrade it by adding series resistor. Crossover slopes can also be restored by recalculating crossover. I have done this with excel for one 2nd order low pass filter, when I replaced woofers series inductor with lower DCR unit. I needed to unwind 3mH series inductor to 2,85 mH and replace 100uF + 1,5ohm resistor combination with 91uF + 1,8ohms to restore original slope. Much to do for proper 3 way upgrade.

You should also consider that amplifier output impedance does have similar effect to crossover slopes, as it is in series to signal and at least to some crossover parts. So there are no free lunches or easy answers. Upgrades especially to vintage gear can also be problematic, as dry and hot climate might age drive units quite differently than damp and cold climate.

But anyway, I love this and would not buy new speaker or amplifier.

Best Regards

Kimmo

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...