Advent Crossover - Please help

[stan461]

I have a set of crossovers for Large Advents. I believe they don't work properly but I'm not sure. It's the 3 way (Normal-Decreased-Extended)

type.

How can I test this crossover in the 3 positions without having to install it or hook it up? I have a multimeter.

[Carlspeak]

I have done a few Large Advent xover upgrades recently and, in the process, actually tested a 3 position switch for continuity with a VOM so I could better understand which components came into play for the purpose of drawing a schematic.

Attached is a scan of a table I made which includes a sketch of the bottom and side view of the 3 position switch. I did the testing with no components attached.

For example, at position "A" of the switch, a "0" ohm reading indicates there is continuity (closed circuit) between tabs marked 1&2. However, there is NO continuity (infinite ohms) between tabs 1 and 4. Connector posts (i.e. tabs)1, 2 and 4 are the key ones that carry the incoming signal (via the 16 microfarad cap) thru to the tweeter so they are the ones of interest in checking continuity. Going to position "B" there is no apparent continuity in the bare switch with tab 1 and tab 2 or 4. However, there is a small inductor coil that eventually connects tabs 1 and 2 when the xover is assembled....and then so on test pos "C" which shows that both 8 microfarad cap and 3 ohm resistor come into play.

I saw something in your photo that looked like the 3 ohm resistor wire going thru tab 2 to tab 3 was not soldered to tab 2. It must be if it isn't. The wire connecting tab 1 to tab 5 must not touch tab 3 or the wire coming from tab 2 to tab 3.

Also attached is a photo of the brand new, upgraded xover I assemble for a customer. I includes all air coil inductors, Metallized PP Solen caps and a non-inductive 3 ohm resistor.

Carl

Carl's Custom Loudspeakers

http://www.classicloudspeakerservices.com

[stan461]

Thanks for your help. I appreciate it.

What is the price on your Advent Xovers (please specify pair or each)?

Would you need the old Xovers sent to you?

I would install them. Shipping the speakers to install Xovers is not an option for me.

Thanks again.

[Pete B]

I'd like to see a picture from the other side of the switch, if it's not too much trouble, I can't see the decrease inductor in that picture.

What we have here is a picture of the 16 and 8 uF crossover from Stan, and of the earlier 16/16 uF crossover from Carl.

The red and yellow .5" dot things are the inductors in Carl's picture with wire so thin that it is hard to see.

The inductor has heavier guage wire in Stan's picture as I suggested in the other thread:

http://www.classicspeakerpages.net/dc/dcbo...g_id=&page=#687

The inductors all look to be iron core.

Pete B.

[stan461]

I'm assuming you meant me. I hope this is OK.

[dynaco_dan]

Hi Stan;

Just a comment on the excellent quality photo you took ofthe crossover.

[Pete B]

>I'm assuming you meant me. I hope this is OK.

That's perfect, thanks!

I've seen the 16/16 and the later crossover first hand but not the 16/8 so this helps.

Did you resolve your problem with Carl?

You might have noticed that I'm suggesting to always change electrolytic caps that are old like these.

Pete B.

[Carlspeak]

Attached is a photo of a completely rebuilt Large Advent xover. This particular one utilized good quality 16 and 8 microfarad metallized PP caps, a new switch and all new air coil inductors. The components are mounted to a piece of cardboard and shipped to the customer who had no problem removing the cardboard and installing it.

Carl

Carl's Custom Loudspeakers

e-mail .... carlspeak@aol.com

[stan461]

I am considering Carl's crossovers.

The Advents are really wonderful speakers. I can't believe how good they are. I was using the AR3a's for a while and the Advents are a nice change of pace. Better or worse, I can't say yet. And if I did, that would only apply to me and my amp, my preamp, my room and of course - my ears.

My negatives on the Advents are a harhness in the highs when pushed and a little hollowness in the midrange. The crossovers may help with these issues. The AR3a's have a huge midrange and a very smooth high end. Almost never harsh.

I'll report back on what the new/upgraded crossovers did for my Advents.

This is a work in progress. I don't expect miracles and I realize the crossovers may not be the end of my tweaking of the Advents.

And once again, thank you.

[Pete B]

The crossover information in the other thread seems to be getting off topic and this is a better place so I'm posting it here:

There seem to be 3 revisions of the Large Advent crossover with no model change to identify them. I've used the capacitor values as an easy way to identify them:

Rev1 - Two 16 uF caps

The first original revision with the second 16 uF being switched in for the increase position. The cap is too large causing the increase to extend down well into the midrange and it raises the Q of the crossover. This crossover also has small "iron dot" inductors with very thin wires. The red shunt inductor does not handle much power and could burn out shorting or opening. There is even a service bulletin in the library stating that this inductor should be checked for a DCR of about 2 ohms, I measured 2.35 ohms.

The caps are 16uF/50V 10% non-polarized electrolytics.

The woofer inductor is 1.5 mH/.28 ohm powdered iron core.

Rev2 - 16 and 8 uF caps

This revision lowers the increase cap to 8 uF and replaces the red dot inductor with a much heavier guage (estimated at 20) inductor. I'm not sure if the .43 mH value is correct in the library but it is certainly in the ball park.

Caps are the same type as Rev1.

Woofer inductor is the same as Rev1.

Rev3 - 13 uF

I taken a look at one of these but not in a lot of detail. This was a significant change with the ferro fluid tweeter, same "donut" look but flush mount. The woofer inductor became air core and several have commented that the crossover frequency was increased. The system is stuffed with foam rubber not a fiberous material, and this is a major mistake probably to raise rather than lower Fc for more 50-60 Hz mass market thump.

I A/B compared a Rev1 system to the Rev3 system with BSC compensation and while they were similar I preferred the Rev1 system.

Pete B.

[Pete B]

Recap?

I've put a lot of thought into this and considering that it is established that electrolytic caps do not age well, and after measuring several that did not meet their voltage specs and in one case had highly increased capacitance, I now suggest always replacing electrolytic caps that are more than 10 or 20 years old. The exception to be considered on a case by case basis is the large computer grade caps found in some systems, and perhaps some of the more exotic types. But always the electrolytic caps to the mid and tweeters which are more likely to fail due to a leaky cap:

http://www.classicspeakerpages.net/dc/dcbo..._id=7195&page=4

http://www.classicspeakerpages.net/dc/dcbo..._id=&page=#7710

http://www.classicspeakerpages.net/dc/dcbo..._id=&page=#7728

http://www.classicspeakerpages.net/dc/dcbo..._id=&page=#7713

I measured the electrical transfer function of the large Advent Rev1 tweeter crossover and the Q is on the high side, somewhat over 1 if I remember correctly. I was surprised to find this and it does make the circuit more sensitive to ESR, however the shunt inductor has a high resistance, around 2 ohms from memory, and this should lessen the sensitivity to ESR.

Mods for Rev1 and Rev2 systems:

I suggest replacing the caps with reasonably priced film caps, Mylar or poly are fine. I'd go with 15 uF which is within the 10% tolerance of the original cap but will lower the Q of the circuit just a bit. It is best to lower the cap across the increase switch to 8 uF since this was a good refinement to the design. I plan to look at this a bit more but I don't know how soon that will be.

It's also suggested that you measure the DC resistance of the shunt inductor (Rev1 only) as it was mentioned earlier in the thread that it cannot handle much power and turns shorted in one poster's system. It is best to replace this inductor with a low cost air core type, then add a resistor in series to provide a total DCR equal to the original. This will lower the distortion at high levels in the tweeter circuit. Note that the additional resistor is not needed if the increase cap is reduced to 8 uF since the 8uF value lowers the Q of the network and it is better without the resistor. This is my recommendation rebuild Rev1 crossovers to Rev2 type but use an air core shunt coil as the replacement.

Speaking of this distortion, I always noticed distortion in the mids/highs when turning up the Large Advents, harshess or a sort of crying quality, it adds to our perception of loudness and tells us to turn it down. I tend to wonder if this was deliberate by Henry Kloss to save drivers from overload. I'm not sure if it's due to the very small iron core inductor, or the tweeter having zero Xmax. It is worst at fairly high levels and not much of an issue for normal listening but does take away from the transparency of the system at moderate-high levels.

There seem to be more people interested in mods for the Advents and we should also consider the woofer inductor. The powdered iron inductor probably has saturation distortion issues when used with 50W or greater amplifiers. Measurements of Advent systems showed much more bass distortion than AR systems which used air core inductors.

We must consider the DCR of this inductor since it effects low bass damping or the Qtc of the system. The Advents have ripe low bass, and if you like this then the woofer inductor should be replaced with one of the same RDC, and it turns out the 14 gauge Solen is the same at .28 ohms. I would NOT go with a higher gauge inductor, this is a place where a heavy inductor is called for.

If you want to tighten up the bass slightly a lower DCR coil can be used. The 14 gauge Goertz foil inductor has a DCR of .25 ohms and is only a buck or so more than the Solen. The 12 gauge Goertz foil has a DCR of .16 ohms and is about 5 bucks more than the Solen. The increase in damping will be very small even with the 12 gauge but if you want the most damping the lower DCR the better however we are near the point of diminishing return. North Creek also has 18 through 8 gauge coils but they're much more expensive even for the same gauge. There is really no engineering sense to using these but if you have money to burn, go ahead.

These are all more or less "stock" upgrades as far as not significantly changing the voicing of the system, just cleaning up the distortion. I might work on a mod to incorporate the baffle step compensation into the internal crossover and then new components will probably be needed, however it could be years before I get to this.

Also, test the resistor and make sure it is still 3 ohms, reflow solder joints to eliminate cold joints. Check the switch for low resistance in each position.

Pete B.

[Guest Ken Perkins]

Hi Pete,

I've tried a variety of coils on the Advent woofer ranging from Erse and Sledgehammer laminates to Solen and Erse 14 gauge air cores, all ranging in DCR from .15 to about .4 ohms (Okay, I was bored). The best sounding one I tried was the Jantzen 15 gauge aircore from Parts Express. It's spec'd at .38 ohms but measures about .33 ohms.

http://www.partsexpress.com/pe/pshowdetl.c...r=255-426&DID=7

I say "best" from the standpoint of overall balance with the tweeter and the actual character of the bass Q of the system. Measurements with this coil indicate a Q of .83 with an Fb of @41 hz. Both speakers match exactly.

Also interestingly, I've got a collection of Advent crossovers (five in all including an odd "parts" board) that ALL have the small bar coil of @.45 mH with a DCR of .25 ohms across the tweeter. These are all the second version of the crossover with the 8uf extention cap bypassing the padding resistor. There must truely have been a run of bad coils to have this many that far off from the "spec'd" 2 ohms according to the service bulletin.

One more thing...last summer I did a quick and dirty crossover redo to apply some baffle step by using a 2.2mH coil in the woofer, leaving the highpass function alone and increasing the series resistor to 4 ohms as well as a "true" lpad of 3 ohms series/20 ohms parallel. It showed promise in that the missing balance was restored somewhat when the speakers are pulled away from the wall for better imaging/soundstage but I ended up putting them back stock to work on something else I had going at the time. I'll probably revisit this later but in the meantime, I'm workin up (in my head, anyway) a modern day Advent "update" using the Peerless 850146 and one of the Tangband mid/tweeters with a Fc of @600 hz...stay tuned.

[Carlspeak]

Thanks Pete for the cap history on the large advents. I thought I was going crazy when I got two different caps (e.g. the 16 and 16 m-farad version and the 16, 8 m-farad version) from 2 different customers. The 16 & 16 m-farad speaker had a ser. #A79279. I will try and get the serial no. of the 16 & 8 m-farad speaker.

Carl

Carl's Custom Loudspeakers

[Pete B]

Hi Ken,

Good to hear from you again!

Just wondering if you measured those bar coils at .45 mH?

It would be great to confirm the schematic in the library.

I was also a bit confused about this because I didn't know that they also changed the inductor in the 16/8 crossover having never seen one.

If you look at Stan's and Carl's pictures one shows the bar inductor used in the 16/8 crossover, the other shows the "red dot" inductor with very fine wire. It makes sense that the DCR of the bar coil would be lower, this is normal due to the heavier wire and it suggests that no series resistor is needed with a replacement coil because the lower 8 uF cap lowers the electrical Q. The lower DC resistance provides better rejection of low frequency material. I'm positive that the service bulletin only applies to the fine wire "red dot" inductor.

Sounds like your using the woofer inductor to "tune" the overall balance which will depend on amplifier source impedance. I tend to run solid state and I think my suggestions are good for people looking for a more damped response, especially if they use the BSC.

I've also been thinking along the same lines that it would be nice to be able to duplicate the Large Advent with modern drivers. Thought that the Peerless 1727 that I measured in the AR-2ax box would work well as a replacement woofer or better the CSC version and I've got a few tweeters in mind - none that seems to be a perfect fit. I'm beginning to think that there might not be a tweeter that will do the job and was looking at the Fostex FE107E as a full range "tweeter". Not my kind of solution to the problem but it seems like it would do the job. Similar concept to Kloss using the Altec full range in the early AR speaker. Just some rambling thoughts.

Pete B.

[Pete B]

I've done reliability analysis for advanced electronic design and the first thing that comes to mind as most important are contacts, because they're subject to corrosion. There are two types of connections in a broad sense, what we call gas tight and non-gas tight connections. Properly done solder joints, welds, and wire wrap are examples of gas tight connections. Gas tight is where the metals bond on a molecular level and air or any corrosive gas does not readily enter the contact surfaces. One may wonder how wire wrap can be gas tight, the wire wrap post is square with sharp edges, the force of winding the wire around the post actually welds the two metals together at the small contact area point where the force per square inch is extrememly high.

Something else to consider that I've never seen done in mass-market commercial products is to chemically wash the rosin off of soldered joints because the rosin is actually corrosive over time. It is required in high reliability design. Obviously, it is also important to follow good construction and soldering techniques when building and repairing electronic equipment.

There are many examples of non-gas tight contacts, any removable connector, screw connections, riveted connections, and so on.

I bring this up because there is an excellent example in Stan's picture of his Advent crossover. The solder lugs on the switch are connected with rivets through the case to the internal contacts. There are probably several non-gas tight connections here and corrosion can raise the resistance or even produce a non-linear diode effect. The red arrow in this picture points to a highly corroded rivet:

http://members.aol.com/basconsultants/SWT_COR.jpg

The early Genesis/EPI tweeter has a brass rivet mounting the input terminal in a way that is very similar to the above switch, and it is not soldered. I have one that is also highly corroded and I measured .2 ohms across the contact. I cleaned and soldered the connection which eliminated this resistance. I used non-acid paste rosin to help get the best solder connection.

It is possible that the rivet used in that switch has a very tight fit into the solder post so that it is a much better contact than the surface connection might indicate, this is pure speculation on my part, but the better companies do their best to improve reliability.

One could solder that rivet to the solder lug but I'm not sure if the case or internals will tolerate the heat, so, do so at your own risk.

Pete B.

[Guest Ken Perkins]

Hi Pete,

The crossovers I'm talking about all have the small gauge button-like inductor in series with the tweeter in the "decrease" position of the switch. Otherwise, it is not in the circuit in any other switch position.

The larger gauge bar inductor is parallel to the tweeter's + terminal and ground. It measures between .43 and .45 mH and, which are close enough for government work. All of mine measure @.25 ohms. I used the 20 gauge, .45mH Jantzen here and I added a 2 ohm resistor to restore the impedance to a total of about 2.5 ohms as per the service bulletin.

My crossovers are complete replacements, totally bypassing the original and hardwired for the "normal" position of the switch. In other words, I'm not using the "decrease" button coil OR the 8 uf resistor bypass used in "extended". The woofer coil is a Jantzen 15 gauge, the tweeter's cap is made up of a 15 uf + 1 uf Sonicap pair, the resistor is a 3 ohm Mills and the shunt coil in the highpass is the 20 gauge Jantzen plus a 2 ohms Mills.

My original crossovers all look like the one posted above by Stan.

http://www.classicspeakerpages.net/dc/user_files/830.jpg

[Pete B]

Hi Ken,

Thanks for confirming the bar type inductor value of ~ .45 mH .25 ohms. I measured the red dot Rev1 type at .8 mH 2.35 ohms but I've only measured one and want to confirm it with a few more.

I agree Ken that we are not talking about the yellow dot decrease inductor. However, in the 16/16 crossover the shunt inductor is a red dot type, whereas in the 16/8 I've learned from seeing these pictures that it is a much higher gauge bar type.

When we first talked about keeping the series resistance the same when replacing the shunt inductor I had not seen a 16/8 and you, I think, had not seen a 16/16 and this caused some confusion. Your reading of resistance was correct and much lower than mine because the wire is much thicker in the bar type.

The inductors are different and should not have the same DCR, it was part of the Rev2 design change and therefore no additional series resistor is need when the bar type inductor is being replaced.

Rev1 crossovers that are upgraded to Rev2 by changing the 16 uF to 8 uF must also have a new low DCR .45 mH inductor installed without an additional series resistor. An inexpensive 19 gauge air core type should be fine.

If one is simply repairing a Rev1 16/16 keeping it stock with 16/16 caps and a new inductor is needed, the old one should be rewound, or a newer one used with a series resistor to bring the total DCR up to about 2.35 ohms.

Here's a picture of a completely stock 16/16 early Rev1 crossover, note the yellow dot decrease inductor and the red dot tweeter shunt inductor, the woofer inductor is mounted to the back of the box. Note that the rivets holding the switch terminals in place are also badly corroded in this picture:

http://members.aol.com/basconsultants/ADV1_STK.jpg

[Guest Ken Perkins]

Hey Pete.

Now it makes sense and thanks for that pic of the Rev1 xo!

I've never seen one and had no idea that version had the red button coil. Makes sense that it would have a dcr so high. Since I have the Rev2 versions, I will remove that unnecessary resistor now that the confusion has cleared up. I simply couldn't understand why that bar coil could be bad on all these crossovers!

I think we've just made a very important discovery that should be more well documented and is must-have info for anyone upgrading/modding these two versions of crossover. We need to get a schematic with proper parts values posted for BOTH version of the crossover.

[Pete B]

Yes, it's hard to know without seeing all the different versions. Good that we got it all sorted out and thanks very much for all your input. I agree it would be nice to update the library.

Pete B.

[mrdb]

On the KLH Five crossovers that I changed the caps on a couple of years ago, I noticed a similar problem: the components were soldered between terminal strips that were riveted to a metal plate, with the plate supplying the ground connection between all terminal strips. One rivet was loose so I ran a hardwired connection between all strips and the input terminal ground. I have a later Five crossover that is implemented on a PC board. Evidently with the quantities KLH was making by that time, it was cost-effective and certainly more reliable.

I work in the implantable medical electronics field where workmanship is critical, and after seeing how the crossovers were implemented in the AR, KLH and Advent speakers, I am surprised that they have survived as long as they have. I agree with you about nonlinear, diode-type effects. Unless severe, they could cause distortion that might be attributed to other causes. I repaired an old Scott receiver years ago that had no FM. The problem turned out to be corrosion on the IF coil wires where they were soldered to the terminal pins. Plenty of flux that had not been cleaned and moisture over the years eventually led to the failure.

Bob

[Pete B]

Hi Bob,

Yes it's good to always try to eliminate non-gas tight connections, for me, I like doing it because I know I won't have to open it up again because of contact corrosion or have degraded sound quality.

I mentioned the Genesis tweeter above where the rivet is part of the conduction path. In contrast the Dynaudio D21AF tweeter uses a hollow rivet and brings the tinsel lead wire through the hole with it soldered to the terminal post which provides a gas tight path from the post to the voice coil.

Pete B.