Modified AR9 crossover

[kkc]

I have now measured all passives in the XO from one of my AR9s.

Some observations - see attached diagram:

1. The actual circuit is an original AR XO but is slightly different then the published spec: there is one cap at a different then expected spec (based on printed value on cap) and 2 extra inductors
   
2. I was able to measure all but the largest two caps. All measured caps were way out of spec. All of them were considerably HIGHER in value then the spec printed on the cap itself. I would have expected the capacitance to decrease as the dielectric breaks down. But presumably, the ESR values would have increased too in this case since the capacitance has increased. I don't know... out of my expertise here...!
   
3. Inductors were fine
   
4. The woofer and UMR circuits were different from the published schema. The woofer has an additional inductor and reduced inductances in the other two inductors
   
5. The UMR circuit has an additional inductor and DIFFERENT capacitor (8uF instead of 6uF). The cap reading was completely unreadable....!
   
6. Note the serial number of the AR9 at the top of the drawing.
   

For me its very clear, the caps need replacing...! I will continue other comments on the outboarding the XO under the older thread.

Rgds, KKC.

[johnieo]

I have now measured all passives in the XO from one of my AR9s.

Some observations - see attached diagram:

The actual circuit is an original AR XO but is slightly different then the published spec: there is one cap at a different then expected spec (based on printed value on cap) and 2 extra inductors

You do not have extra woofer inductors! The 10-mH inductors are each made from two (count them 2) 3.84 mH coils stacked close together so that their series-aiding mutual indutance (M = 1.07 mH) produces 10 mH. You MUST re-install these with the windings in the same relative direction as they were removed or you will ruin the crossover. The total inducance of the pair is L1 + L2 + 2M = 3.84 + 3.84 +2.14 = 10 mH. If you stack them series opposing, you will get L1 + L2 - 2M = 3.84 + 3.84 - 2.14 = 5.54 mH. Mind your directions! Check the pair with an L meter to ensure they are stacked correctly.

The cap in the UMR drawing in archives reads 8-uF, not six as you indicate.

I was able to measure all but the largest two caps. All measured caps were way out of spec. All of them were considerably HIGHER in value then the spec printed on the cap itself. I would have expected the capacitance to decrease as the dielectric breaks down. But presumably, the ESR values would have increased too in this case since the capacitance has increased. I don't know... out of my expertise here...!

Electrolytic caps can first increase in value as oxide becomes thinner due to etching. Running the capacitor will restore some of the change but after this many years, trash them. It is not so easy to "reform" a bi-polar cap as a polarized cap.

When the electrolyte is sufficiently dry, C will decrease- eventually just two metal electrodes separated by air!

Inductors were fine

as they should be -- except you must re-install the pairs properly.

The woofer and UMR circuits were different from the published schema. The woofer has an additional inductor and reduced inductances in the other two inductors

see first comment about woofer inductor.

The UMR circuit has an additional inductor and DIFFERENT capacitor (8uF instead of 6uF). The cap reading was completely unreadable....!

can't help about added coil, but cap is 8 uF on drawing and, if unreadable, is completey shot.

Cheers,

[kkc]

You do not have extra woofer inductors! The 10-mH inductors are each made from two (count them 2) 3.84 mH coils stacked close together so that their series-aiding mutual indutance (M = 1.07 mH) produces 10 mH. You MUST re-install these with the windings in the same relative direction as they were removed or you will ruin the crossover. The total inducance of the pair is L1 + L2 + 2M = 3.84 + 3.84 +2.14 = 10 mH. If you stack them series opposing, you will get L1 + L2 - 2M = 3.84 + 3.84 - 2.14 = 5.54 mH. Mind your directions! Check the pair with an L meter to ensure they are stacked correctly.

Hi John

I was as surprised as you. I just double checked. In fact both La and Lb are DOUBLE STACKED coils (see pic), as you rightly stated, but their inductance values are DEFINITELY as I have stated, 5.25mH each. I also measured half the stack and that came to 2.2-2.3mH. The new coil Lc = 2.9mH is a single stack. Even in the published AR9 schematic there is an inductor shown where I have placed Lc, but no value is given.

See attached pic. Am I still missing something? I do take your point about re-connecting the right way. Thanks for letting me know that.

Rgds, KKC.

[johnieo]

Hi John

I was as surprised as you. I just double checked. In fact both La and Lb are DOUBLE STACKED coils (see pic), as you rightly stated, but their inductance values are DEFINITELY as I have stated, 5.25mH each. I also measured half the stack and that came to 2.2-2.3mH. The new coil Lc = 2.9mH is a single stack. Even in the published AR9 schematic there is an inductor shown where I have placed Lc, but no value is given.

See attached pic. Am I still missing something? I do take your point about re-connecting the right way. Thanks for letting me know that.

Rgds, KKC.

both La and Lb better be double stacked.

I believe you measured their values incorrectly! Look at the circuit section attached here.

I think you did not remove one lead from each coil so as to prevent the meter from measuring their parallel combination (two ten mH in parallel = 5 mH) the very low Z of the 2500 uF cap will only slightly affect its measurement.

An L-C meter measures with an a.c. frequency sometimes 800 Hz, sometimes 8000 Hz or other depending on brand. It simply applies an a.c. voltage and measures an a.c. current. then it divides the two values and gets an impedance. The number shown on the display depends on which you have dialed. L, or C At the frequency of measurement, the 2500 uF cap is essentially a dead short, so you are measuring two Ls in parallel. Did you notice that you measured the identical value for each coil? That is why. you were making the same measurement!!

You likely had the same problem with the 1.37 mH coil reading 1 mH in the UMR.

Notice that the 2.88 mH woofer coil measurement was correct, that is because you disconnected the woofers and one end of the L was open as it should be to measure.

I am sure the coils are fine, just replace the caps.

[kkc]

John

Your explanation makes complete sense. I haven't the possibility of checking the XO right now. But I measured these values four times and in the active presence a knowledgeable colleague. We both checked that points A and B were not connected - so that we would not be measuring the inductors in parallel.

What I will do is check the values once the components have been completely disconnected from each other and then report back.

Thank you very much for your time and effort.

Rgds, kkc.

[johnieo]

John

We both checked that points A and B were not connected - so that we would not be measuring the inductors in parallel. Rgds, kkc.

HOW did you both "check" that points A and B were not connected? Did you:

• CUT the wire on one side of each stacked pair with a pair of side cutters? or
  
• UNSOLDER a connection on one side of each stacked pair?
  

If you did not do one of the above, then your measurements are wrong. I think you also made this same mistake in measuring the 1.37 mH and obtaining 1.0 mH. the 1.37 has L and C in parallel with it--this lead must be cut as well. The capacitors in the two circuits do not mean the incductors are "not connected!"

If you cut the wires to one side of each stacked pair or unsolder and measure 5.25 mH then they were stacked series opposing, but I seriously doubt AR would have a mistake like that.

Cheers,

[ar_pro]

You are absolutely correct...the woofer section would not function properly if the coils were stacked opposing, and the problem would be audible.

[johnieo]

Hi John

I also measured half the stack and that came to 2.2-2.3mH. The new coil Lc = 2.9mH is a single stack.

Rgds, KKC.

This tells me that you for sure did not disconnect or move the coils!! You cannot measure the inductance of one half of a stacked pair without moving it at least a few inches away and unsoldering one lead!

Each coil that comprises half of the stack is 3.84 mH and if next to its mate, it will have also 1.07 mH of induced mutual inductance, but because you put the meter across only half of the stack, the meter current in the two halves of the stack is flowing in opposite directions, so it is L-M or 2.77 mH for one of the stacks. The other half of the stack (2.77 mH, for same reason) is in series with the other 10 mH (total 12.77 mH) and that series combo is in parallel with the 2.77 of the half stack. if you parallel 2.77 and 12.77 mH you get 2.27 mH, which is precisely what you measured!

You don't need new coils; you need a new friend.

Cheers,

[kkc]

OK; guys. Can we stop beating the dead frog....! Once I receive the new caps, all components will be cleaned and separated. At that time, I will measure in complete isolation and post results here...!!!

In the mean time, as I stated in my earlier response, lets assume you are correct and I made silly mistakes. But please stop banging me on the head about it...! I can't hear the music...!

Thank you. kkc.

[Diamonds&Rust]

In the mean time, as I stated in my earlier response, lets assume you are correct and I made silly mistakes. But please stop banging me on the head about it...! I can't hear the music...!

John would never, ever bang anyone over the head about anything. He isn't so much correcting you as acting like the retired Professor Emeritus he is. He's not saying "You made a mistake, bad you, look how smart I am!"

He's saying, "You made a simple mistake, here is the natural consequence of that, here is the math to show how you got what you got, DO YOU SEE?"

The "DO YOU SEE?" part separates the jerks from the teachers.

Misunderstanding his intentions would not only hurt his feelings, but would also make for a much poorer experience.

There aren't many like John. In fact, they may have broken the mold before he was cast.

Bret

[kkc]

My apologies if I offended anyone. My point was -that there is no more need to discuss this until after I fully test the components completely isolated from each other. I fully understand John's points and respect and appreciate his help a lot. As well as yours Bret. Rgds, kkc.

[kkc]

John and everyone, I apologize for much confusion. I finally managed to take a couple of hours and completely dismantle the XO from one of the AR9. This gave me a chance to measure the values of all components.

I HAD INDEED MEASURED THE INDUCTANCE WRONGLY.... CIRCUITS WERE STILL CONNECTED AT ONE POINT.

I also measured the big caps by paralleling with some lesser values.

The actual measures are shown in the attached diag.

Regards and buckets of thanks to all for pointing out my errors.

kkc.