HELP! Infinity Crossover Information Needed

[bob p]

Just thought I'd ask in case somebody has a collection of old HiFi service data:

I need information on the crossover for a vintage HiFi speaker system. The model name is The Infinity Reference Standard 9 Kappa, circa 1985-1989.

Specifically, I need a schematic and/or a service manual for the speaker. The speaker is a 5-way, and it uses crossover frequencies of 80, 800, 4500, and 10k. I have an owners manual but no service data and Harman International isn't being helpful.

My problem is that the supertweeter (10KHz-45KHz) appears dead, but the driver works when jumpers are attached from the other speaker's supertweeter. Hooking a 'scope up to the "problem" supertweeter terminals with a music signal shows no signal coming out of the crossover. Signal injection tests confirm appropriate frequency signals at every speaker terminal except the one that goes to the 10k-45k supertweeter driver. It looks like one seciton of the crossover just died.

Just as you might expect, a self-proclaimed professional was inside of the speaker doing a re-foam on the woofers, and he appears to have yanked on the wires and damaged the crossover's circuit board. The board is bent upward and some of the L-pad connections are bent. There's no signal going to the L-pad for the supertweeter.

I've tried tracing the board, but there are some problems: although the tweeter caps seem to be 2uF film/foils, the inductors are all custom wound and unlabeled. Making matters worse, everyting is covered in hotmelt glue and its next to impossible to trace the circuit or to identify components or even frequency specific sections of the board.

I've tried reverse engineering the 10K crossover frequency to get component data. As best I can tell, the Z for the tweeter is maybe 4 or 6 ohms (based upon the rated nominal impedance for the speaker). This would suggest that the highpass filter to the supertweeter might use a 0.19mH inductor shunted across the tweeter, with what looks like a 2 uF series cap.

I've even tried comparing signal injection traces on the good board vs. the bad board. The problem is that there's so much hotmelt glue all over everything that its hard to tell what's what.

If anyone has data for the crossover system, or any helpful ideas, I'd appreciate hearing from you. FYI Harman-Kardon has no technical data that they're willing to share, and they don't stock replacement crossover boards.

[bob p]

Well, I'm happy to say that I've been able to procure a copy of the crossover schematic from the folks at Harmon parts, which was a very pleasant surprise.

I have some questions about it, and I'm hoping that some of you might be able to help. The circuit is attached as a .png file. (Let me know if you need a different file format)

My Questions:

1. Is this a 3rd Order Butterworth filter? It looks like one based on the topology, but when I use web-based crossover calculators to determine values for a 3rd-Order Butterworth filter, I always get a higher value for C3 than shown in the schematic -- 5uF instead of 3 uF.

2. Why are the 2 uF and 3 uF PP caps bypassed by 0.01 uF PP caps?

3. Why does the attenuation circuit use two 1N5338B Zener diodes between the signal and ground, oriented cathode to cathode?

Any insights would be greatly appreciated!

[bob p]

Sorry -- the graphic mysteriously disappeared. Here's the .JPG version:

[bob p]

Well, I disassembled the crossover assembly. The thing is HUGE, heavy, and so laden with wires and coated in hotmelt glue that its not possible to do any diagnostic work without taking it apart. I had to de-solder a bunch of inductors and switches, and remove several globs of hotmelt glue so that I could flip the board over and examine it. (I hate double sided boards!) I have to admit it was a bit of a PITA but then there wasn't any way to do definitive testing without doing some major disassembly.



Here are my signal tracing results:

- A 15 KHz sine wave applied to the speaker's main input shows up at the beginning of the "high tweeter" crossover section. That is to say, the signal passes through the fusible link and arrives at the input to caps C1 and C2. There's no signal at the distal end of the caps, or at any point farther downstream.

- Injecting the test signal at the distal end of the caps produces an appropriate signal at the output (a resistor that I have wired in place of the tweeter). This suggests that the remainder of the crossover system is working.

- The inductor has continuity to ground; its Z is too low to measure accurately.

- The L-pad measures an appropriate change in resistance across its operating range.

- The diodes appear to test OK, according to the diode test function on an el-cheapo DVM.

I'm a little confused by this failure mode, as it seems to suggest that BOTH C1 and C2 have failed with each cap failing open-circuit. (having BOTH caps fail that way seems unlikely, unless one cap's failure caused the other cap to fail, but who am I to argue with data?)



The good news is that I had a hunch that the 2 uF cap was the culprit, based on some rudimentary testing that I had done before I obtained the schematic. Luckily, I ordered a replacement 2 uF 5% PP cap yesterday, and I have a suitable 0.01 uF 5% PP cap on hand, so I'll probably just replace both caps and re-test the circuit.

If anyone has insights that would explain the two cap failure mode, or insights about questions 1-3 listed above, I'd love to hear from you!

Edit: I guess only the 2uF cap is likely to have failed; if the 0.01uF cap is still functional, it should be providing a crossover frequency in the range of 2-4 MHz. Although I think a single cap failure is more likely, I'm still at a loss to explain why the HiFi types insist on bypassing everything with a 0.01 uF cap, as that cap would create a filter that has an ultrasonic crossover frequency in the MHz range. :-/