autoformer and polarity

[John Warren]

I was just looking to understand this more and see if anyone had test results on tap selection to share.

For you...

http://www.northreadingeng.com/Forums/index.php?topic=4.0

[babadono]

Thanks John for taking the time to do these tests.

Hey OP did you see these results?

I know I can hear the level difference at the different taps, but the phase?? I suppose an elaborate test could be set up that would adjust the amplitude independent of the tap selected so the amplitude to the driver is constant. No, wait how could that be done? A variable gain amp after the crossover?

[down_2_earth]

Find two railroad tracks, lie across them, and wait.

Two trains, one on each track, come along. Both are right side up and both hit you at exactly

the same time. The trains are in polarity and in phase. - no...you just got hit by two trains

The same thing happens again and both trains hit you at exactly the same time. However,

this time one train is upside down. That is a polarity reversal. - no...not at all...one is an inverted signal...the other non inverted

The third time both trains are right side up but one hits you first and the other hits you shortly

after the first. That is a phase shift. - nope again.....its a time delay

The last time the second train is upside down and hits you later than the first. That is both a

polarity reversal and a phase shift. - absolutly not...clearly this is an inverted time delay

[mikebse2a3]

I was just looking to understand this more and see if anyone had test results on tap selection to share.

For you...

http://www.northreadingeng.com/Forums/index.php?topic=4.0

Interesting results....Thanks John..!!!

miketn

[mboxler]

John...

As a newbie, I'm curious if the phase changes are related to the autoformer, or to the increasing load on the capacitor created by the tap changes. At -12dB, the load would be 160 ohms, vs the 20 ohm load using the -3dB tap. The -12dB tap would alter the frequency rolloff to around 75hz (?).

I only ask this because to OP referred to Al's universal network, which would include a swamping resistor to create a consistent load regardless of the tap setting.

Thanks, Mike

[John Warren]

I have more measurements to do including phase measurements for a -3dB L-pad equivalent with the re-sized series cap to account for the impedance change.

I'll look at the so-called "swamping" resistor too. We'll see what it does to the crossover frequency, output, phase, etc.

I had to travel this week so couldn't do much in the way of this project. I also owe Dean some measurements.

Will post results as they're completed on the forum I put up in the earlier post.

Cheers!.

.

Edited May 2, 2014 by John Warren

[ClaudeJ1]

I was just looking to understand this more and see if anyone had test results on tap selection to share.

For you...

http://www.northreadingeng.com/Forums/index.php?topic=4.0

Interesting results....Thanks John..!!!

miketn

This is the reason why Roy uses active Xovers primarily or designs his passives with resistors. He never like the phase shifts of Autoformers.

[Deang]

What he said was that people don't take the phase issues into account when they design with them. He likes LR24 with correction (LCRs) - and when you're using something like LEAP, I'm sure it's a lot easier to just not use them. The simple filters with the autotransformers sound great in spite of whatever phase related issues are going on. I find this stuff interesting, but I'm sure not going to lose any sleep over it. It's great seeing people embrace the complexity of steeper slopes and tackling time domain issues with active filtering, but you don't need those things to get great sound, and it's my opinion that the majority of listeners are going to experience the best results with the least amount of intervention that they can get away with. I still think the Type A sounds the best, and all of the technical bullshit reasons of why it can't doesn't change the fact that it does.

I guess this is a good thread to do this: there is a new autotransformer on the way. I was asking Mark Cooper down at UT about some of these things and the next thing I know he starts unloading all of these cool stories. He begins sharing memories of PWK coming down and visiting his Dad to talk about what he wanted, and is so often the case, what you want just costs too damn much, so PK paid for the most important stuff. As it turns out, the "most important stuff" ended up being getting the DCR as low as possible and wax impregnation for the windings. Cooper said PK was adamant about this, as he didn't want any air gaps in-between the windings. If you're paying attention, you'll notice wax impregnation mentioned on many of Klipsch' schematics. I asked the obvious question, "then why do you use varnish?". He said that no one ever asked for it, and that varnish is really good enough. He said the varnish protects the steel better than the wax, and that even though varnish doesn't fill in the air gaps, it does seal the windings. I asked him if he was an audio guy doing this for himself, what would he use?

"Wax."

"Is there anything else you would do?"

"Yeah, I can think of a couple of things".

"How about you just wind this thing as if you were building it for you, and I'll buy a bunch, does that work?"

Paraffin has an interesting quality: it's difficult to contain. They had to have a special tank built to deal with it. That was an interesting story in and of itself. This one is worth using Google for. There's some good YouTube videos on it as well.

I should have the prototype within the next week or two. The unit will have a heavier gauge wire (so the frame will be larger), and since the DCR will be lower, I'm curious to see what the impact might be as it relates to attenuation (lower DCR means less loss). UT is going send me a spec sheet with the unit, but I've also called Bob, who has agreed to curve it against the 3636. I could just use independent listening tests, but when it comes to new product development, I think we can all agree that we want as many numbers as possible.

I was told that it will be big and ugly.

I was told the waxing process will only add a bit more to the cost, but like an idiot I never asked what the total cost of the unit would be. I'll release more details as I get them.

Apologies in advance for adding to the phasing dilemma!

[muel]

Here ya go, Dean. http://jeffsplace.me/wordpress/?p=5811

I'm sure Frederik would be glad to customize to the specifications of your choosing.

I think I'd have to go the cheap route with the copper version of these autotransformers.

Perfect for your next flagship crossovers! You might want to get some of these in stock... but you would need to take out a 2nd mortgage. The price is steep! If you are a person who already spent $20k+ on your speakers this might not seem like a lot of cash.

[John Warren]

Ok, it's traffic time.

New set of measurements added

10 Ohm non-inductive resistor across the autoformer input taps (0, common)

13uF series cap on autoformer input tap changed to 60uF, additional capacitance necessary to maintain similar filter behavior with the 10 Ohm shunt resistor .

Load on taps was 10 Ohm non-inductive resistor.

http://www.northreadingeng.com/Forums/index.php?topic=4.msg8#msg8

Edited May 3, 2014 by John Warren

[Deang]

Cameron, audiophiles don't have to be morons unless they choose to be. I can think of the perfect speaker for those though.

Thanks John.

[muel]

So being a moron is a choice? How encouraging!

[Deang]

Lol, quit twisting my phase, uh, polarity, whatever.

[John Warren]

The oscilloscope photos provide some idea as to how the phase of the signal is changed without and with the "swamping" resistor. The phase angle does not change much when the swamping resistor is used so phasing errors caused by changing the attenuation level of the mid are minimal.

The downside for having the ability to modify the mid attenuation is the cost of the larger capacitor and resistor. An Lpad works quite well also:

http://www.northreadingeng.com/Forums/index.php?topic=4.msg11#msg11

Edited May 4, 2014 by John Warren

[Deang]

This is what I woke up to this morning:

'"This will not change your mind regarding the use of autoformer attenuation because you have convinced yourself otherwise. That said, they really are "much ado about nothing".'"

Lol, where's my DFH on this.

Seriously though, why were PK's numbers so different?

Autotransformers don't waste power. Even with the swamping resistor, less power will be wasted than using an L-pad. I might get whacked for this one, but I'm pretty sure that's true.

[Deang]

"The Problem with Attenuators"

Bob, Gil, Chris -- please post this for me if you can, I won't have access to my PC until late tonight.

Edited May 4, 2014 by DeanG

[John Warren]

You do not have to post that article, I have the original.

Remember the "sock puppet"?

[Deang]

Yes. Are you attempting to terrify me, lol.

[John Warren]

See the two posts (#7 and #8) in this thread, link below. I think this was PWKs "issue" with L-pads.

http://www.northreadingeng.com/Forums/index.php?topic=4.0

Edited May 4, 2014 by John Warren

[ClaudeJ1]

I still think the Type A sounds the best, and all of the technical bullshit reasons of why it can't doesn't change the fact that it does.

I agree with you on the type A. The AA was created because as amplifier power went up, tweeter were fried. Since Klipch and Associates had a lifetime warranty on their speakers (unwritten), they got tired of replacing tweeters for free.

My own setup uses only one cap value at the K402 rolloff point. 18 db/octave acoussitcally AND 6 db per ocatave electrically. This is essentially what PWK did with the A network WITH the autoformer. I'm sure if he had found a 104 db squawker driver, her would have skipped the autoformer altogether. But it was a simple and elegant solution that wasn't "overkill" like so many things in "high end" hi fi (shall we discuss $30,000 cables?)

In my case I'm using active bandpass on the woofer section only with a bi-amp situation. Instead of attenuating the midrange, I'm boosting the horn bass votage drive to compensate for a 5-7 db difference in efficiency. Based on today's measurements, I need to add about +4 db to my tweeter so Audyssey doesn't have to. I will probably end up with an extra 6 db because I will line up the voice coils and maintain my 2/3 Passive-1/3 Active Hybrid network by moving the tweeter to line up with the K-1133 driver. I will have to go tri-amp to do it. No big deal, I just got B&K Power amp with 6 stereo amps in it, each with it's own gain control for a total of 12 channels. How much fun will I have with that! I think they are about 60 Watts per channel, but since I normally use about 50 woofer Milliwatts for 85 db at my sweet spot, I still have 30 DBwatts of headroom before clipping (that is 1,000:1 power ratio). Can you say DYNAMIC, boy and girls?

Still mostly First Order, still mostly passive here. Basically I'm NOT against anything that works well (I use Autoformes in all my other speakers), regardless of brand, type, or technology. Peace.

Edited May 5, 2014 by ClaudeJ1