autoformer and polarity

[ClaudeJ1]

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

Nice work John. Thanks for doing this and sharing. Maybe this helps to explain why McIntosh has always had HUGE BiFilar Autoformers on their Power Amplifiers! Reactance Isolation?

[John Warren]

I was interested in the comparison between the L-pad and autotransformer responses which, based on the acoustic measurement, showed no significant difference. PWK claimed about a 3dB difference in the response peaks in the article "The Trouble with Attenuators". Perhaps on a horn with much larger response peaks a difference would be apparent(?). On the K400 with the PD5VH not so, no 3dB differences..

If the autoformer provided any advantage sonically I do not think Delgado (or Klipsch) would have eliminated it.

[beeker]

whaooooo you fellas may as well be typing chinese...im outta here

[babadono]

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.

All I can say is both Dean's and ALK's workmanship make this guy's stuff look like a monkey did it. Yes the components are awesome I'm sure (better be for $5k a crack) but are you going to leave THAT out in plain sight? Not me.

[ClaudeJ1]

I was interested in the comparison between the L-pad and autotransformer responses which, based on the acoustic measurement, showed no significant difference. PWK claimed about a 3dB difference in the response peaks in the article "The Trouble with Attenuators". Perhaps on a horn with much larger response peaks a difference would be apparent(?). On the K400 with the PD5VH not so, no 3dB differences..

If the autoformer provided any advantage sonically I do not think Delgado (or Klipsch) would have eliminated it.

Well, it could be an efficiency thing leftover from the old 10 Watt amplifier days. L-pads turn the energy to heat, but increasing the impedance and using a smaller capacitor uses less energy.

[Deang]

Autotransformers are expensive, and if you're obsessed with phase like Roy is, maybe it makes sense to just move to resistors. I will say this, autotranformers are transparent and predictable - at least with the simple filters. I did build a set of Type As once, replacing the 3619s with some L-pads, and was surprised by how grainy it sounded. I didn't really give them a fair shake I suppose, just using some cheap wirewounds.

I was recently told that at the time that PWK was working on his crossover designs, the autotransformer cost less than the capacitors he would have needed if he would have designed without them.

Edited May 13, 2014 by DeanG

[ClaudeJ1]

I was recently told that at the time that PWK was working on his crossover designs, the autotransformer cost less than the capacitors he would have needed if he would have designed without them.

Makes perfect sense to me. This is way before the price of copper went up to the point where people steal plumbing pipes from vacationers or kill themselves climbing up a telephone pole to look for copper.

[Deang]

If the autoformer provided any advantage sonically I do not think Delgado (or Klipsch) would have eliminated it.

There is no way in hell you're that naive. : ) You get rid of it because a resistor costs a dime and an autotransformer costs more than a 100 times that.

So, how do you plug an autotransformer into LEAP?

http://www.audioheritage.org/vbulletin/showthread.php?846-autotransformer

[John Warren]

"L-pads turn the energy to heat, but increasing the impedance and using a smaller capacitor uses less energy."

The plot below

GREEN is the 13uF cap on autotransformer, -3dB tap

PURPLE is 60uF series cap on autotransformer with 10 Ohm MILLS NI resistor shunt, -3dB tap

RED is 60uF series cap, -3dB Lpad, no autofromer.

Each plot, the amplifier level was not changed, only the connections to accommodate the desired filter.

For all intents and purposes the three responses are same sensitivity.

Edited May 17, 2014 by John Warren

[John Warren]

If the autoformer provided any advantage sonically I do not think Delgado (or Klipsch) would have eliminated it.

There is no way in hell you're that naive. : ) You get rid of it because a resistor costs a dime and an autotransformer costs more than a 100 times that.

You speak as though testing this theory requires an act of Congress. It's as simple as having a 60uF cap on hand, a couple of resistors, a few patch cords and, of course, a pair of Klipschorns.

Measurements and auditions of nets with and without autotransformers has convinced me that using a resisitve Lpad is a perfectly acceptable option for this particular horn.

The fact that it's a lower cost option and doesn't introduce phasing errors, however small, is only icing on the cake! A win-win if you will.

Sorry.

[John Warren]

So, how do you plug an autotransformer into LEAP?

http://www.audioheritage.org/vbulletin/showthread.php?846-autotransformer

If you'd like ot know how engineers model an autotranformer in either MATLAB or SPICE I'd be happy to show you. Predicited voltages, reactances, driver volume velocities, power dissipation real and imaginary) can be accurately determined.

How do you think these voltage responses (shown below) were predicted? These are three simulations of the measurement conditions shown above. The autoformer and it's phase response predictions are provided in simulation.

(Note the simulation below are voltage simulations at the voice coil, not acoustic simulations).

What's LEAP? (only joking, I know what it is).

Edited May 17, 2014 by John Warren

[Deang]

"L-pads turn the energy to heat, but increasing the impedance and using a smaller capacitor uses less energy."

Should have said:

L-pads attenuate by turning the unused energy into heat, but an autoftransformer is an impedance multiplier, so attenuation is achieved by simply raising the impedance the amplifier sees, thereby reducing current delivered to the driver. Attenuation is achieved without wasting power in the form of heat. The higher impedance also allows the use of smaller value capacitors, which means I can use really cool capacitors. Who in the hell wants a POS 50mfd Jantzen, Solen, etc. in their network working as the primary cap for the circuit.

From your perspective, that post may be as bad as the first - which by the looks of it might have been one of my four in the morning endeavors. Sorry.

Roy is kind of upset with me right now, and deservedly so I suppose. I owe him an email and if I can get back on his good side, maybe I can get him to tell me why they stopped using them - I'm pretty sure it was due to the injection of additional phase error. In all seriousness though, in a network that only uses FOUR parts, why not go all out? Your plot actually shows that the autotransformer is better, not by much, but still, it's better. Besides, I'm not convinced that you know what in the hell you're doing over there. : )

I'm sure you've seen the thread with the new unit, and I really appreciate you not pissing in my Cheerios. One unit is going to Bob for some tests and to evaluate it's suitability in our work. I was sent two units - would you like to repeat your test using the new unit?

[BEC]

"One unit is going to Bob for some tests and to evaluate it's suitability in our work."

Hurry up and get that thing down here to me. I am actually interested in seeing how that works.

[ClaudeJ1]

Thanks, John. You have restored my faith in resistors and L-pads, even though I choose active networks with simple Voltage level controls (potentiometers at the power amp front ends)with the consistent slopes on each driver with time delay (digital active), acoustic feedback from the room itself(Audyssey MultiXT/32), and minimal parasitics from imperfect passives in the signal path (just for proctection from DC offsets). High resolution digital trumps 1929 technology any day. How's that for crapping in the corn flakes.

Edited May 18, 2014 by ClaudeJ1

[Deang]

Get a grip Claude. Someday we will have a showdown: my 1920s approach to your fingernails across the chalkboard approach.

Bob, old beater 1920-something awesome sounding autotransformer coming your way on Monday.

[ClaudeJ1]

Get a grip Claude. Someday we will have a showdown: my 1920s approach to your fingernails across the chalkboard approach.

Bob, old beater 1920-something awesome sounding autotransformer coming your way on Monday.

I'm glad you get the fact that I was specifically just "yankin' yer chain" Dean. ROFLMAO on the chalkboard thing!

After listnening to a presentation on Blue Ray recordings by Mark Waldrup of AIX at Axpona in Chicago, I'm convinced that DSP's will continue to evolve in all consumer products. Cheap or expensive.

Having owned one of the very first CD players in Michigan (back when it was a Sony or nothing over 30 years ago) to get the full dynamics from my Khorns (no more ticks and pops from bad vinyl and dust), I still recognize that it's A to D and D to A conversions that make recorded music possible. There are SOUND arguments either way for passive vs. active speaker driver control (pun intended). Sine waves still rule.

But when either one is done WELL, even self proclaimed expert listeners like most of us, cannot honestly say anything sounds BAD.

Edited May 18, 2014 by ClaudeJ1

[Deang]

I felt bad after I posted that, but then I remembered that hyperbole aside, I was being truthful. I do agree that when any implementation is done well, it's going to sound good - but I haven't experienced that yet. So far, it's always been noisy, grainy, and somewhat flat.

My medium is all digital, it sounds good. I really do believe the problem lies with the corner cutting employed in the "professional" units combined with very unforgiving solid state.

No more Klipsch Gatherings, so we will have to organize our own. I do very much feel it's time to settle some arguments (that's what we used to us them for) -

You're a good sport to put up with me though, I know I can be pretty pig headed. Sorry.

Edited May 18, 2014 by DeanG

[Schu]

in...

Edited May 18, 2014 by Schu

[John Warren]

Besides, I'm not convinced that you know what in the hell you're doing over there. : )

That's beside the point.

So I ran more curves, fresh off the press. Again, the 13uF + Autoformer @ -3dB (the AA configuration?) and compared (again) to the L-pad equivalent. Microphone was on-axis about 1,4m from the debris screen of the PD5VH. This time I tweaked the capacitor to get a closer response (57uF rather than 60uF done earlier). Amplifier is Crown (solid state). Earlier measurements were MC2300. The hardware was on for about 2 hours before taking the measurements.

Red is Autoformer attenuation and light blue is L-pad. Each major division is 1dB.

Edited May 18, 2014 by John Warren

[John Warren]

Here's the above plot using the stock-brokers scale.

Edited May 18, 2014 by John Warren