Sony SS-NX1 Speaker disassembly - Late 90's audiophile mini system speakers

[Ken Tajalli]

I should note however, that this cap change was accompanied by changing the burned out tweeter. Replaced the burned out 6OHm with an 8Ohm, and then had to balance with a couple of wire wound cement ceramic resistors so the rest of the cross over saw the same X-over point - so I think the cap issues probably pale in comparison to the havoc I've wreaked there
I should note, the caps and tweeter were on the 'other' speaker (not the one I took the measurements from - outside of the full setup (third graph) which was both plus sub.
That's an interesting point - would it be possible from the circuit diagrams and the info I have to take an educated stab at the X-Over frequency as designed?
I didn't notice any swelling in the larger caps - can these degrade without showing external symptoms? Of course if i'm going to start buying better components to swap out - the bigger question would be is it worth it - or should i be investing in a better set of modern speakers.

The xover frequency is approximate, and the actual number is not really important here.
If you have a dip, and you can not correct it by tilting the speaker a few degrees up or down, then change the 12uf cap on the bass filter to something like 10uf or a little less.
Tune by ear.
Just a little trick if the threads on the chipboard cabinet should come loose with repeated use.
Thin common wood glue a drop of water, use a straw to pickup a few drops and inject it into the holes, give it an hour or, put the screws in but do not tighten fully! leave it overnight, then tighten gently.

 

[mhardy6647]

Bucking = shielded
for our British members!

.... for our British members still using CRT's

Cheers,
JaapD.

Not surprisingly, we didn't retire our CRT TV until COVID days, replacing it with a small LCD flat-panel television our son found on the side of the road near his home in Virginia.
We're not much into television.

 

[Ken Tajalli]

Does that include television, sir?

 

[mhardy6647]

Just for completeness

 

[Ken Tajalli]

Just for completeness

Does the calendar go beyond 2020?

 

[mhardy6647]

Does the calendar go beyond 2020?

Yup! It's still working as of 12 Aug 2024. EDIT: That screen is the VCR's clock setting screen (not the TV's), by the way.
I don't know how much longer it's got... I haven't pushed the ^ YEAR button on the remote to find out where it ends, in case it's a Mayan calendar kind of thing, you know?

 

[Ken Tajalli]

Yup! It's still working as of 12 Aug 2024. EDIT: That screen is the VCR's clock setting screen (not the TV's), by the way.
I don't know how much longer it's got... I haven't pushed the ^ YEAR button on the remote to find out where it ends, in case it's a Mayan calendar kind of thing, you know?

You have a working VCR too?
Are you on the other side of a time tunnel?
Must admit, I do have a brand new, unused super-VHS machine in packaging in the loft. I am keeping it, if one day I feel like transferring the tons of home videos we have.
That day hasn't come yet!

 

[Xmall]

Thanks all for the tips on the caps - and the screws! I had noted the screws on the X-Over had a little glue on them - and they are now getting loose since i cracked that. Probably need to sort that out to sort vibration.

I'll see if i can get hold of some better caps - swap them out and post the results just for fun

In that case, moving to an Audyssey XT32 or Dirac enabled setup will likely give you years of enjoyment!

Thank GXAlan - I will throw this in the mix. I imagine i'll also get nice things like wifi - and music streaming which would be nice!

If I understand correctly you measured the unrepaired one.
Could you provide measurements of the repaired item?

Yeah that is correct. The results unfortunately were not over pretty.

Bear in mind - I think Aydessey corrected this to some degree and brought my hacks somewhat under control

Effectively the modified circuit on my franken-speaker looks like this:

And measurements

Blue Trace: Original Speaker
Yellow Trace: Franken (Repaired) - Speaker

Y Axis (SPL): 40-95
1/12 smoothing, no adjustment or EQ.

Based on the above, I think i can probably conclude that the crossover is somewhere around the 2-3k mark?

I was expecting there to be a bit more influence in the lower frequencies to be honest as I designed this more as a DC fix.

 

[Xmall]

If you have a dip, and you can not correct it by tilting the speaker a few degrees up or down, then change the 12uf cap on the bass filter to something like 10uf or a little less.
Tune by ear

Thanks Ken, this is interesting.

Can you explain (in layman's terms) what the second capacitor actually does? I thought it 'blocked' the low frequencies - so would I be correct in assuming that dropping the cap to a 10uF would effectively push the crossover point slightly higher - allowing the woofer to be producing more sound at that point rather than the tweeter?

 

[Ken Tajalli]

Thanks Ken, this is interesting.

Can you explain (in layman's terms) what the second capacitor actually does? I thought it 'blocked' the low frequencies - so would I be correct in assuming that dropping the cap to a 10uF would effectively push the crossover point slightly higher - allowing the woofer to be producing more sound at that point rather than the tweeter?

The second cap sets the xover frequency.
The cap filters higher frequencies. Smaller the cap, higher the frequency.. In essence, allows more xover frequency to come through the bass driver.
If you have such a big dip, try removing the 12uF , just to test the waters.
We are talking about the only cap in bass filter.

 

[JaapD]

W.r.t. your query, the second capacitor, in combination with the inductance and the other capacitor, is part of the third-order cross-over filter; 18db/octave. All three parts together define its cross-over point.

Just a thought, since you're questioning the cross-over frequency while able to measure the frequency curve, why not block-off the tweeter with some foam or heavy damping material (or disconnect it)? Then you'll get the woofer plot, and dentify its12 dB falloff point? Or vice versa, disconnect the woofer then measure the frequency curve and identify its 18db falloff point.

Cheers,
JaapD.

 

[Ken Tajalli]

Not that I am arguing with JaapD, but there are two 12uF capacitors in that circuit.
One is on the tweeter circuit, and another on the Bass filter!
I was talking about the bass filter, NOT the one JaapD is on about.
In case there was a confusion.

 

[Ken Tajalli]

Blue Trace: Original Speaker
Yellow Trace: Franken (Repaired) - Speaker

Y Axis (SPL): 40-95
1/12 smoothing, no adjustment or EQ.

View attachment 386184

Take out the resistors on the Franken repaired version, it is not doing any good. All it attempts to do is impedance match. But in this speaker, it would be irrelevant.
It may even dull the sound a little, unless you prefer it that way.
The bass output is drooping off too early, hence the dip. We can either bring the tweeter response closer to it, which is not that wise, or lift the bass output at xover region, hence my suggestion of lowering the sole capacitor on the bass xover filter.
If it sounds right, leave it alone!

 

[JaapD]

Not that I am arguing with JaapD, but there are two 12uF capacitors in that circuit.
One is on the tweeter circuit, and another on the Bass filter!
I was talking about the bass filter, NOT the one JaapD is on about.
In case there was a confusion.

Thanks for your clarification. Indeed a mixup between the two different filters, from my side caused by the fact that you were mentioning the 'second' cap.

You mention "The second cap sets the xover frequency". Can we conclude that the xover frequency of the bass section is determined by both the inductor as well as the capacitor through the formula 1/(2 Pi sqrt (LC)) ?

Cheers,
JaapD.

 

[JaapD]

Take out the resistors on the Franken repaired version, it is not doing any good. All it attempts to do is impedance match. But in this speaker, it would be irrelevant.
It may even dull the sound a little, unless you prefer it that way.
The bass output is drooping off too early, hence the dip. We can either bring the tweeter response closer to it, which is not that wise, or lift the bass output at xover region, hence my suggestion of lowering the sole capacitor on the bass xover filter.

I see on the franken speaker an impedance match as well as a sensitivity match (between woofer and tweeter, attenuation through R2 and the tweeter impedance).

Is this a desired or unwanted attenuation?


Cheers,
JaapD.

 

[Ken Tajalli]

You mention "The second cap sets the xover frequency". Can we conclude that the xover frequency of the bass section is determined by both the inductor as well as the capacitor through the formula 1/(2 Pi sqrt (LC)) ?

Ofcourse, but he is stuck with the unknown inductor, so I explained in laymen's terms as he requested.
This is a cheap, fun restoration. I tried to keep it simple.
Just change one cheap capacitor to fill the dip.
But again, if it sounds right, don't do nothing but enjoy it.

 

[Salt]

Thanks for the update.
The yellow line presents some interesting information: there is a typical LP rolloff of the woofer with approx. 18dB/octave starting sub 2 kHz ( 2 1/2 kHz is typical for crossover frequency and thus the woofer's rolloff would fit), given by no addition of signal around 2 to 3kHz to the signal from the tweeter.
So the crossover frequency of tweeter HP is too high or/and too steep. Additional maybe some phase or polarity issues occur.
Astonishingly both tweeters, new and used, rolloff pretty steep up from 7 kHz what I thougt was due to the age of the old tweeter and should not happen with the new one.

Did you measure on level of the tweeter, 0°, on axis?

 

[Xmall]

I see on the franken speaker

Glad to see my nickname stuck!

Thanks for your clarification. Indeed a mixup between the two different filters, from my side caused by the fact that you were mentioning the 'second' cap.

Thanks @Ken Tajalli and @JaapD - the capacitor mixup came from me - but you have both answered important questions for me.

Take out the resistors on the Franken repaired version

I felt these were necessary in this case - tested without to start and the additional volume out of the tweeter was painful (i didnt know which frequency range however). I just chalked it down to different speaker design, so the resistors were to match impedance, and volume.

Did you measure on level of the tweeter, 0°, on axis?

No I didn't on this one - its one question i never thought to ask. Where should you setup the mic? I always assumed mid speaker - but given the base is more omni-directional (as pointed out to me in the other thread) - I might be screwing my measurements.

That said, I see the same dip in my over-all listening position test -so I'm guessing it is something X-Over related.

I'm away for work all next week, so I wont be able to sort out any new caps for awhile.


Out of interest, i did remember that JayCar had a frequency response for the new tweeter on the website - could this help in deciding the best approach with the X-Cover?

 

[Paweł L]

To build proper X-over you would have to properly measure each driver and use CAD soft like Vituix. This way you can try different topologies. Just forget all the text book formulas because they're very much useless in real word xover design. At this point is hard to assess the crossover without measurements for midwoofer and tweeter section (X-O point, slope, phase rotation) to make educated guess what components would have to changed to achieve better freq. response. It doesn't look terrible. Sure there's dip around 3k, followed by peak at 4.5k, but some of it might be due to diffraction. One thing I would try is to reverse polarity of the tweeter just to make sure phase is not reversed as it will create large dip at crossover point.
Increasing the 3.3uF will set the high pass lower. Increasing the second cap, 12uF will lower the Q factor of the network (shallower slope). Increasing the value of the inductor will also move crossover point lower in frequency. All of the of course has effect on the phase rotation. There are many combinations of values/topologies available to achieve desired results, however some minimum proper set of measurements and CAD is a must.

Please reverse the tweeter and give us some measurements, this way we have more info what is really going on.

 

[Ken Tajalli]

Please reverse the tweeter and give us some measurements, this way we have more info what is really going on.

+1
by reverse, he means positive wire to negative terminal on the tweeter and so on.