SpeakerTalk Forum Index SpeakerTalk
This forum has been set up to facilitate discussion of 1970s KEF speakers and drive units. The owner of the Forum has no connection with KEF Audio.
 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

KEF 105 vs 105/2

 
Post new topic   Reply to topic    SpeakerTalk Forum Index -> KEF Speakers from the 1970s
View previous topic :: View next topic  
Author Message
T.O. Chef
Senior Contributor 100+


Joined: 31 Aug 2007
Posts: 156
Location: Toronto Canada

PostPosted: Thu Sep 13, 2018 1:14 am    Post subject: KEF 105 vs 105/2 Reply with quote

I adore the 105. The original version. Twice I have had the good fortune to completely refurbish some. Both had the same filter layout, and both had the same rear head fixings. One set was owned by a professor in Ottawa CA, the second were my own rescue from e-Bay. (the professor is over the moon happy).

In a very weak moment, I have given my personal pair away (due to age advancement and available real estate for them). I regretted that sale immediately. A Ton of Love & Care went into that refurbishment and they were outstanding.

To get to the point ...
I have located and purchased another set, bit these are very different. And they sounded like utter BS when I got them.
The heads on these have a redesigned pair of covers for the rear. The crossover is mounted horizontal (not vertically as the others), but most importantly, the filter is strangely configured.

There's a CUT in the PCB trace at one point.
On the HF section There's a 30uF cap in parallel with a .028Mh coil on the input, then a wire jumper to the rest of the HF section. Next is a 3.3uF series, a 0.35uF to ground, and a 7uF series, and a 2.2uf / 47R to ground. Then the T-52 +.
There's an aB circuit using a 0.22 polyester cap. (Between the 3.3 and 7uF.)

On the B-110 layout, there's first a coil, then a 30uF cap, then the normal 10uF to ground, 2mH to ground, 0.25mH series, but then a Zobel with a 10R first and 7uF to ground. Then the T-52 +.

All of this seems to be completely original. All cap's were so close to each other (side to side) that I doubt there's been much if any drift over the years. One exception is the 2.2uF cap's that measured a ridiculous 3.93 and 6.08.

I have recapped using Falcon supplied sets. These replacements did not match original values of those removed, but they did match each other very very closely.
After install ... they still sound like crap.

So. I removed parallel coil with the 30uF input on the T-52. I removed the aB'cap. I paired the 7uF on that section to the 47R to ground, and replaced it with a 3.3uF as per the classic layout.

Things sound better, A lot better, But not ideal.

I have a million questions about these.
Speakerguru ... can you help here?
I'm exasperated.

I have a set of original 105.2 filters (version 1), and I have ordered a replacement cap set that I will very carefully mount in them. The topology seems more in line with the original 105 (version 1) crossover layout aside from the massive 480uF up front of all sections.
I am considering modifying the internal 105 cab's to make a secure mounting location for them, running cables up and out to the heads, and connecting the original 105 protection circuit board to the relay on the 105.2 board.
Can I assume that the protection circuit on the original 105 and the later 105.2 is basically the same??

Please.
I welcome any suggestions on this project. I've never been so frustrated with a refurbishment as I am with these treasures.

Thanks All.
_________________
Music is food for the soul.
Back to top
View user's profile Send private message Send e-mail
T.O. Chef
Senior Contributor 100+


Joined: 31 Aug 2007
Posts: 156
Location: Toronto Canada

PostPosted: Thu Sep 13, 2018 1:32 am    Post subject: Reply with quote

Further to the discussion.
The bass section of this later 105 pair is the classic + to 7.0mH series, then 80uF to ground, then 33R to ground, then 360uF series, then 3.5mH series then B300 +.

But ... I find it crazy interesting that my impedance scans after recapping both boards read 35 ohms at 10hZ, gently sloping to 10 ohms at 30hZ, A bouncy ride from there between 6-10 ohms, then a steep slope up from 500hZ to a whopping 25ohms at 3K. Then a steeper drop to 8 ohms at 4K followed by some normality.

I understand that the graph will never be near flat, but this roller-coaster is unlike anything I've seen with its huge peaks and valleys. And Never have I seem the bass section with a giant top curve at the front end.
What has anyone else experienced?
_________________
Music is food for the soul.
Back to top
View user's profile Send private message Send e-mail
audiolabtower
VIP Contributor 500+


Joined: 06 Jan 2009
Posts: 567

PostPosted: Thu Sep 13, 2018 2:20 pm    Post subject: Reply with quote

Your hf section looks like the aB version which some T52s required, presumably if the resonant freq was high Q and closer to crossover freq? Many T52s did not need it. I assume the 30uF is in C9 and the 0.22 in C11? and your typo is a 0.35mH coil to ground?

My hf section is the 0.028mH, a 3.5uF (measured), the 0.35mH to ground, 3.8uF (measured) to tweeter +ve and no zobel network.

My mid is 31.3uF, 1.3mH, 10.8uF to ground, 2mH to ground, THEN A SERIES 100UF (95.4 measured), then 0.25mH, then the driver +ve with the 7uF and 10 ohm zobel. If your 100uF is missing, since you don't mention it, this will definitely mess up your 400Hz crossover.

The protection on the Mk1 was special fuses, thermally matched to the driver coils, which you can't get any more, very different to the active S-Stop circuit on the Mk2. I have used mine without fuses for nearly 40 years with no problems, even when a power amp once went to 50V rail, the series caps saved them. If your search my threads you'll find my external crossovers near the amp terminals which have a lot of advantages, not least in vibration reduction and speaker cables.

Impedance - mine are around 22 ohm at 2.5kHz crossover, the 200Hz peak around 13 ohm and the 40 Hz peak around 17 ohm. This is normal and why you need a good damping factor in the amp, other wise if the amp has a high output impedance like valve amps or Naims that used to have the series output resistor, the potential divider of amp and speaker can mimic the curves of the impedance in the freq response, maybe only fractions of a dB, but maybe also a reason for subjective differences between amps. With the vast majority of amps this particular issue will not matter. as long as the dynamic impedance can be covered by current output, the voltage drive will not care about the ohm value, the higher the better as easier to drive, just at the expense of efficiency.
Back to top
View user's profile Send private message
speakerguru
VIP Contributor 750+


Joined: 18 Nov 2005
Posts: 990
Location: Green Hut, Tovil

PostPosted: Fri Sep 14, 2018 10:12 am    Post subject: Reply with quote

T.O. Chef wrote:
I understand that the graph will never be near flat, but this roller-coaster is unlike anything I've seen with its huge peaks and valleys. And Never have I seem the bass section with a giant top curve at the front end.
What has anyone else experienced?

High impedance is simply a by-product of deliberate mismatching the driver to the amplifier, by means of the xo, on the part of the designer, to reduce the output in that region.

This can be because of the driver being more efficient, directional or resonant in that region or it could be a function of the cabinet geometry. Some KEF models reduced this variation in impedance by conjugating the crossover with additional components. This made the load more resistive and therefore, having the voltage and current in-phase, easier for the amplifier to drive. It is more expensive, of course.
Back to top
View user's profile Send private message
T.O. Chef
Senior Contributor 100+


Joined: 31 Aug 2007
Posts: 156
Location: Toronto Canada

PostPosted: Fri Sep 14, 2018 12:49 pm    Post subject: Reply with quote

Thanks to you both for the insight.
Your answers have shed much light on the questions.

Yes. My 105's have the aB section on the T-52 using a .022uF.
The layout is: 31.7uF (measured) / 0.028mH parallel to 3.71uF (measured), 0.38mH to ground, 7.59uF (measured) , 2.2uF (measuring way off at 3.93 one side 6.08 the other) & 47R to ground, then fuse, then T-52 +.
I have not yet measured actual inductor values.

The mid section is first a1.3Mh series, then 32uF series, (measured), 10.9uF to ground,, 2mH to ground, 110uF (measured) series, .02mH series, 10R & 7.9uF to ground, to fuse, to B-110 +.

I have measured both the T-52 and B-110 and found all parameters to be very very close to published spec's.
In removing the aB section (and changing the 7uF series to 3.3uF and putting 7uF in series with 47R to ground) things sound much better but the 2800 Hz impedance peak jumped from 24 ohms to 32.
So, from what's been said here, I assume that in my case the purpose of the Zobel and or the aB circuit is to reduce that peak. Yes too, I can see on my graphs that the peak reduction with the original layout also smooths out the phase in that region.

At this point I'm beginning to think that I have not been careful enough to match my replacement cap values to the measured values of the old. (Excepting the obviously out of whack ones). Unfortunately I don't have enough stock to do this perfectly. The supplied set purchased came with measured values very close to what's printed on them. I'm now wishing I had specified standard 10% tolerances on my order!

I'll go back to the drawing board and try to fix this best I can. I'll report back on my results.

Oh. And one more question: Is it normal for this filter to have a steep upward impedance going south from 10 ohms at 30Hz to 35 ohms to 0 Hz? Is this intentional to reduce output in the sub sonic region?
_________________
Music is food for the soul.
Back to top
View user's profile Send private message Send e-mail
audiolabtower
VIP Contributor 500+


Joined: 06 Jan 2009
Posts: 567

PostPosted: Fri Sep 14, 2018 1:25 pm    Post subject: Reply with quote

Intentional, yes because a 12in unit in a 70 litre box can't give you below 20Hz at any significant power level compared to the design efficiency, so impedance should be high. I assume this impedance is also a by product of the large series cap which gave the special bass loading extension half way between closed box and ported.

If your tweeter crossover impedance has gone up with the change in values then the output in that region should have gone down, and if there was too much energy here it would explain a smoother subjective response. The tweeter zobel caps seem way off which would also throw out the balance.
Back to top
View user's profile Send private message
T.O. Chef
Senior Contributor 100+


Joined: 31 Aug 2007
Posts: 156
Location: Toronto Canada

PostPosted: Fri Sep 14, 2018 5:05 pm    Post subject: Reply with quote

Oh No Question the 2.2uF in Zobel on the HF were both wonky. But Can I be sure that 2.2uF was the actual value of what was used in these? Or, should I just be sure to match replacements at whatever values I have on hand that are as close as possible to each other for speakers A and B?
Actually. The question is mute since my only 2.2 100V LL cap's measure 2.36 and 2.41uF.

Another dumb question. With apologies for my ignorance.
I note that all schematics I've seen put the capacitor Ahead of the coil when in series. This is reversed in mine on the Mid input. Is this unusual? Does it make a difference?
Same situation with the notch filter on the B-110. The resistor is placed ahead of the cap. Of no consequence?
_________________
Music is food for the soul.
Back to top
View user's profile Send private message Send e-mail
audiolabtower
VIP Contributor 500+


Joined: 06 Jan 2009
Posts: 567

PostPosted: Fri Sep 14, 2018 8:47 pm    Post subject: Reply with quote

The zobel cap value will not be ultra sensitive and 5% will be fine. Also since the network is in parallel with the inductance of the drive unit it does not matter which way round the cap or resistor is at ground or +ve, they are just there across the drive unit to make the coil impedance look a little more flat above resonance. They aren't really a notch filter, you would need another inductor to compensate for resonance of the drive unit.
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic   Reply to topic    SpeakerTalk Forum Index -> KEF Speakers from the 1970s All times are GMT + 1 Hour
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2005 phpBB Group