Large Advent - Original Capacitors Compared to Axon Poly
Posted 13 April 2006 - 04:56 AM
It is interesting that it is fairly easy to produce white or pink noise digitally. A pseudo-random number (PRN) sequence is used and there are guys who spend a lot of time figuring out how to generate good PRNs. It is interesting that these number sequences usually repeat after some time interval because it is (was years ago) computationally expensive, when done in real (budget) hardware, to process the PRN. My point is that the pink noise on the Stereophile test CD, cycles at a few Hertz from memory, and I noticed a tonal change from the start of the sequence to the end. This made it very hard to hear small differences when switching since it depended on where it was in the PRN sequence. Anyway, the point is that this plays a factor in the audibility of the differences. The lesson, get a good noise source, I was too lazy to go to the lab and get a noise generator. FM interstation noise is also good if that's all you've got and there are test CDs with better noise.
I want to comment, that you sometimes hear people say after an A/B test experience that they thought they had it, then lost it. What I found with music, is that if I didn't switch during a segment with a lot of energy in the high end, I would not hear the difference. It turns out that with voice for example, the vowel sounds are what we call voiced, they're a fundamental with some harmonics, but not up into the 5 or 10 kHz range. Other sounds are unvoiced and are noise like with broad band energy, SSSes for example, we make a sound like steam escaping which is noise like and does have very high frequency content. In music there are short periods that we often miss during a switch where there is enough HF energy to hear the difference. This is in contrast to a noise test source where we always have high frequency energy, and this is why noise makes it so much easier to hear.
Someone mentioned/asked if I broke in the poly caps - nope, right out of the box. I'm not surprised, in all my years of analog and digital design, we never broke in caps. We burned in equipment to catch infant mortalities and to verify that the design worked over voltage and temperature, especially for high reliability and military designs. Once design errors are corrected failures are usually mechanical within components due to thermal expansion and contraction.
It has crossed my mind that perhaps the Advents are actually good for comparing caps because their forward voicing and low crossover point mean that much of what we're hearing comes through the cap under test.
The forward voicing tends to minimize masking that might otherwise occur due to more woofer output. I don't like the forward voicing but it may help us hear very small capacitor differences.
Well, I got some fresh new bi-polar electrolytics to try. There was no value near 15 or 16 uF in the Bennics so I got 15 uF M.D.L. caps 100V 5% DF, and also 1 uF. Don't know about this brand.
Posted 13 April 2006 - 12:55 AM
I have read much of Jerry's stuff in the A.A. I am a bit jealous of you guys because I don't really have quite the patience to dig into my Advents and change things around several times etc.
Also, I'm one of those guys who likes to leave original alone and you guys like to tweek and improve (not that there's anything wrong with that :^)
One thing I will say about sealing is that when I bought my 1976 utility pairs, both were still sealed with the (apparently) original Mortite and it was all hard and falling apart. Not sealing well at all. In fact, when I got the first pair I thought that they sounded like crap, Not like I remembered Advents sounding. No deep bass and "honky". This is when I removed the woofers to see what was going on and discovered the bad Mortite. I ran right out and bought some new and installed. Like night and day!
I'll let you guys know how the new Mortite is doing in 30 years :^)
Posted 12 April 2006 - 01:09 PM
Posted 12 April 2006 - 01:07 PM
I didn't mean for this to be an ultimate rebuild or anything, I consider this more or less stock. Perhaps we should start a thread for thoughts on ultimate rebuilds, since there seem to be serveral people with ideas and things that they've tried. I do have some thoughts and I'll probably put them in a new thread since this is mainly about capacitors.
Yes, I never found the decrease position of any use, but have left it in.
I never bought much into cabinet bracing as a major factor, however I must say that I've found cabinet mid bass buzz issues with many of these older speakers. I'd replace the thin crossover board with .75" thick wood, probably make it run the full width as a brace, with a perpendicular piece as a half or partial shelf. Front to rear dowels are probably a good idea also, or I might just add a few more partial shelf braces.
I do also often use low cost gold posts, to minimize corrosion problems and to provide for larger wire.
I believe that the 16/8 Rev2 crossover is probably best, however when I tried the .45 mH inductor with this early Rev1 Advent it sounded even brighter in the lower midrange, probably due to higher Q or moving the drivers further in phase causing more peaking due to more in-phase vector addition of the drivers. I'm going to measure the frequency response at some point and will have more to say.
I notice that you used .55 mH, did you measure the original? I don't actually have a sample of that one to measure and the file in the library here shows .43 mH as I recall.
I do also prefer air core coils as I've said in other threads, but have not tried them on the Advents, yet.
Thanks for the input on the 16/16/4R, never heard of that one, I suppose we could call this Rev0. Do you know if the inductor was the .8 mH ~2.3 ohm on that one?
>I've taken mine a little further than you did. I
>mirror-imaged both pair, and added a front to back brace of
>1.25" diameter hardwood dowel between the front and back
>panels to prevent breathing. Soft foam strips around the mesh
>grill on the tweeter to "soften" the edges of the
>tweeter hole in the grill panel. Soldered all internal
>connections (very important). Redid the crossover to the
>final 16/8 version. Rewired the tweeter switch so the two
>poles are in parallel to cut contact resistance, and
>eliminated the Decrese setting, replacing that with a second
>Extended with 5.6 uF. New black stretch grill cloth without a
>second layer (looks great on a walnut Advent). Replaced the
>8-32 studs and nuts with gold plated 5-way binding posts. Air
>core woofer inductors. The tweeter parallel inductor is now a
>0.55 mH air core 20 gauge inductor to get the same R as the
>original coil used in the third generation crossover. Sound
>is very open and imaging is improved. I use these in the
>Extended setting with the 8 uF resistor bypass cap. Caps are
>a 15 uF Dayton (Bennic)non-polar elctrolytic paralleled with a
>1 uF Daytom polypropylene for the 16 uF cap, and the 8 uF
>bypass is a Dayton polypro.
>So far, I have found three different crossovers for the
>original Advents, not two as usually reported. The first was
>16/16 uF with a 4 ohm series R for the tweeter used on very
>early models before 1971; second was 16/16 uF with a 3 ohm
>used until mid 1975; and the third was 16/8 with 3 ohms used
>from mid 1975 to the end of production in 1977. The
>resistance of the parallel inductor on the tweeter also
>changed betwen version 2 and 3, dropping from around 2.2 ohms
>to around 0.5. The higher resistance is not needed for
>damping on the third version with 8 uF and the corner on the
>tweeter at rolloff is sharper with the lower DCR value.
Posted 12 April 2006 - 03:09 AM
Posted 12 April 2006 - 12:39 AM
So far, I have found three different crossovers for the original Advents, not two as usually reported. The first was 16/16 uF with a 4 ohm series R for the tweeter used on very early models before 1971; second was 16/16 uF with a 3 ohm used until mid 1975; and the third was 16/8 with 3 ohms used from mid 1975 to the end of production in 1977. The resistance of the parallel inductor on the tweeter also changed betwen version 2 and 3, dropping from around 2.2 ohms to around 0.5. The higher resistance is not needed for damping on the third version with 8 uF and the corner on the tweeter at rolloff is sharper with the lower DCR value.
Posted 10 April 2006 - 04:56 AM
Posted 10 April 2006 - 03:34 AM
>After trying many sealers over the years, I've settled on
>Mortite "Weatherstrip and Caulking Cord". I've
>found this to become slightly less soft over 10 years or so,
>not bad even after 30 years. I don't think hardening has ever
>caused a leak as far as I noticed. This is good enough for me.
> I did once use something that looked the same and it was
>clearly more of a paintable putty designed to harden, this was
>I've had trouble finding the black stuff, and it is better in
>that it blends in with a black baffle board. I did locate
>some at an automotive store but it was sold in a fairly large
>The Mortite seems good enough as I see it.
>I tried silicone sealer as one member here mentions, the
>problem is try and remove a driver after using it. I only did
>this once, and the silicone did not give on removal, the wood
>was pulled apart. Never again.
>>Pete...thanx for printing all that and the time to test
>>My only real issue is the Mortite. The stuff is a window
>>glazing compound and as such, it's manufacturer does not
>>if it dries up some over time. Every old original Advent
>>been into shows signs of the Mortite drying up. I'm open
>>suggestions, but I found a 3M product (black) that comes
>>rope caulk down to 1/4 inch-some sort of poly. They use
>>product in automotive glass applications. I think this is
>>better product and much more stabile over the long term.
>>think the speaker builders use a product now that is more
>>foam and less like caulk, but I have no experience with
>>The real test on the speaker 'seal' is to gently push in
>>the woofer cone and it should come back very slowly if
>>everything is sealed properly.
Has anyone tried the speaker sealer products offered by Parts Express? I've had much success with both the black, rubber sealing caulk (p/n 269-300)and the speaker gasketing tape (p/n260-540)
Remember, it's all about the music
Carl's Custom Loudspeakers
Carl's Custom Loudspeakers
Posted 09 April 2006 - 10:57 PM
After trying many sealers over the years, I've settled on Mortite "Weatherstrip and Caulking Cord". I've found this to become slightly less soft over 10 years or so, not bad even after 30 years. I don't think hardening has ever caused a leak as far as I noticed. This is good enough for me. I did once use something that looked the same and it was clearly more of a paintable putty designed to harden, this was no good.
I've had trouble finding the black stuff, and it is better in that it blends in with a black baffle board. I did locate some at an automotive store but it was sold in a fairly large quantity.
The Mortite seems good enough as I see it.
I tried silicone sealer as one member here mentions, the problem is try and remove a driver after using it. I only did this once, and the silicone did not give on removal, the wood was pulled apart. Never again.
>Pete...thanx for printing all that and the time to test etc.
>My only real issue is the Mortite. The stuff is a window
>glazing compound and as such, it's manufacturer does not care
>if it dries up some over time. Every old original Advent I've
>been into shows signs of the Mortite drying up. I'm open to
>suggestions, but I found a 3M product (black) that comes in
>rope caulk down to 1/4 inch-some sort of poly. They use this
>product in automotive glass applications. I think this is a
>better product and much more stabile over the long term. I
>think the speaker builders use a product now that is more like
>foam and less like caulk, but I have no experience with this
>The real test on the speaker 'seal' is to gently push in on
>the woofer cone and it should come back very slowly if
>everything is sealed properly.
Posted 07 April 2006 - 03:46 PM
The real test on the speaker 'seal' is to gently push in on the woofer cone and it should come back very slowly if everything is sealed properly.
Posted 07 April 2006 - 02:19 AM
Great write-up as always, thank you for all your efforts.
Well worth reading and printing out for referrence.
Posted 07 April 2006 - 12:49 AM
Here's what I've done:
1. Refoamed the good woofer and reported the T&S parameters in another thread. Note that RSSOUND foam gave very low and correct Fs and in system Fc.
2. Remounted loose woofer XO inductor, used hot glue around the base for extra stick and to minimize vibration. Keep this inductor as far away from the tweeter inductors as possible to minimize coupling.
3. Replaced internal wiring with 16 ga, Carol wire, eliminated wire nuts, and soldered directly to speaker terminals.
4. Checked all connections, switch, and resistor with an ohmmeter, all was fine.
5. Checked that the inductors passed a DC resistance check -fine.
6. Removed both original Advent electrolytic caps.
Wired up the tweeter normally, putting the original inductor across the tweeter, caps are still out of circuit. I used the original inductor since it is an odd high resistance unit and I wanted these to be completely original.
Ran a bead of carpenters glue around all the internal box joints, and around the rear input plate.
Replaced the fiber glass and closed them up using fresh mortite on the woofer. Tweeter screws were fine, woofer screw holes needed some repair.
Outboarded the tweeter crossover, with an Eagle 3 ohm resistor and a 15 uF new Axon poly (two 7.5 uF in parallel) to one terminal post, and the original 16 uF electrolytic to another post also from the resistor. The other end of the resistor was wired to the front positive terminal on the tweeter.
Power amp (100W high current) out went straight to the Large Advent to drive the woofer, and also to an A/B speaker switch. The A output went to one cap, and the B output to the other cap.
The only replaced component is the Eagle resistor and it is shared, it measured 3.0 ohms.
Note that the Axon was 15 uF and the original 16 uF, I did this simply because that was the available value, if this test showed a difference I would have paralleled a 1 uF.
Playing music and switching, it was often as if the switch was not connected, almost nothing.
During clapping (similar to broadband noise) in live music, and certain forward sounding recordings one sounded just a bit more forward. Very difficult to tell.
Next, I tried pink noise from Track 15 Stereophile test CD #2. Obvious that one speaker was very slightly more forward.
I expected it to be the higher 16 uF electrolytic, but it was the Axon probably due to lower ESR.
I added resistance to the Axon and adjusted by ear, 1 ohm high, .5 ohm - high, .33 ohm - low, .4 ohm made them so close I'm certain it would be impossible to tell the difference.
I'm not trying to prove anything about 15 uF vs. 16, I just chose the closest value and prefer to round down most of the time.
I'd say that the difference without the .4 ohm resistor would be inaudible without the A/B switch and probably well within production variations for the original design. It was most often inaudible even with the A/B switch.
I should probably point out that I tested a few of these original Advent electrolytic caps, after some use that might have reformed them, and they were within tolerance and had reasonable ESR. We also do not know if the ESR has gone up over time, it usually does and I might try some fresh Bennic electrolytics as another comparison just out of curiousity. Still, we are splitting hairs here, they might have had .2 ohms ESR when new in 1972 compared to about .4 now, not much of a difference and this is of course speculation.
I'll probably include the .4 ohm resistor because I plan to use these speakers as a reference pair for the original stock Large Advent sound. I probably would not bother for speakers that end up for use in a workshop or background music system, the difference is so small.
I did most of these tests without BSC:
and without the "Increase" bypass cap thus in the "Normal" position. I did have both increase caps available, fresh Axon (15 uF) and the original electrolytic. I listened again with BSC and with both bypass caps coming to the same conclusion about the .4 ohm resistor. I didn't notice as much of a difference for the original electrolytic vs. Axon "increase" cap, but did not use A/B testing. Again, the system sounds much better with BSC.
I would normally do a capacitor test with much better speakers, however, I did this quickly because there seems to be so much interest in, and controversy toward the Advents. I would normally do more measurements but these are not my primary interest.
The results here are my observations and opinions, some will not take this as a conclusive test, it was good enough for my purposes.
Constructive comments are welcome.
1 user(s) are reading this topic
0 members, 0 guests, 0 anonymous users