|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
68.192.193.60
I'm putting together a two-way Altec 414c/802g/32a system. At the suggestion of Joe Roberts, I'm using a single 3uF oil cap on the 802 to bring it in at about 6-8kHz while letting the 414c run full range.
I need to attenuate the 802 to match the 414C. Rather than using a resistor-based L-pad, I wanted to get fancy with an autoformer made by Werner Jagusch custom made in 1dB steps (14 dB total range). Included is a connection diagram:
The autoformer has taps at 0-15, with the diagram indicating 0 (negative) and 15 (positive) to be connected at all times (they are bare wire). Tap 14 is in parallel with Tap 0 (negative terminal) and one of Taps 1-13 connected to the positive terminal.
Questions:
1. Can someone enlighten me as to how this works?
2. Is there a multi-position rotary switch (10W or so) that will be sonically transparent that will allow 12-position switching for the autoformer?
3. Has anyone had direct experience with Werner's autogformers?
Thanks,
Randy
Follow Ups:
So you're digging the full range 414 setup? Some do, some don't.
I started messing around with autoformers but haven't finished the experiments. Those German autoformers look very nice. Tempting. I've been using Peerless and old Triads to mess around with.
The way I see it, there are two directions to go, given that the autoformer will step up the impedance the cap looks into. This, of course, will change the crossover frequency for a given cap.
1) One could install a resistor across the primary of the autoformer to give the cap a more stable impedance to work into.
The transformed impedance of the driver will appear in parallel with this resistor Rl, so the impedance will not be perfectly stable with different taps but more so than without it. If/when you settle on a tap, you can calculate the resistor to give exactly the load resistance you need.
2) Let the autoformer transform the impedance to a higher value, measure it with DATS/Woofer Tester, and use a small, high grade cap at that high impedance.
Measuring this, I find maybe 75-125 ohms reflected at potentially useful settings of the autoformer, peaking like crazy with rising frequency. For that reason it might be useful to put a resistor on the secondary side of the autoformer to limit the rise.
The multiplicative effect of the impedance transformation enhances the rising tendency. It won't rise above the value of that added resistor though.
The phase diagram of an impedance sweep on this #2 setup looks quite strange. Not sure what to make of it.
Sounds good this way but I went with #1 because it was more straightforward.
In any case, the challenge is to control the reflected impedance that the cap sees when you change taps.
I need to get back to that project. Got sidelined. Among other things, I would like to measure the acoustic output of the various schemes.
Dave Slagle knows a lot about this, or at least he has been messing with it for a while, so maybe he'll pop in with his observations.
------------------------------
Free your mind and your ass will follow -- Parliament/Funkadelic
1) is generally the option Werner recommends, with a variable resistor in parallel with the primary of the autoformer. He indicates it is for "rolloff", presumably indicating that it is used for impedance compensation to control the rolloff frequency.I have another potential alternative: I have a Dennis Had SE KT-88 amp that can use anything from a 6V6 to a KT-150 to give 5-12wpc. I also have another amp on order from him, this one a parallel SE amp that will output 15wpc with KT-88's. Depending on tubes use and the amount of attenuation needed on the 802, I may be able to run the PSE amp on the 414 and the SE amp on the 802, with a 3-5 dB difference in power between the two.
Edits: 08/02/16
That pot scheme is cool if you want variable rolloff (granted that you can't vary the LF knee). I suppose it is also useful for trimming when slightly less than exact value or perfectly matched caps are used.
I saw it as more crucial to nail down the xover frequency against changing reflected impedances when taps are varied, but it all depends on what you want to achieve.
Let us know how these autoformers work out!
------------------------------
Free your mind and your ass will follow -- Parliament/Funkadelic
those are beautiful looking autoformers in pictures of custom crossovers. Maybe a screw-block strip and a spade lug on the horn driver wire would look professional and be practical.put the tap wires from the autoformer to one side of a screw terminal block, then move the wire with the spade lug which goes to your horn to the desired tap.
Lets say the striped red wires in the picture below are from your autoformer and the single aqua wire will go to your Altec 802/32a horn.
for a fixed crossover point, the capacitor value gets smaller as more attenuation is used as the reflected impedance to the "primary" is higher. you could also use an appropriate resistor in parallel with your 802 so the reflected load would be the same to the highpass cap and not have to change the capacitor.
Karlson Evangelist
Edits: 07/30/16 07/30/16 07/30/16
Thanks Freddy!
I had also thought of using a terminal block as illustrated - my crossover will be one autoformer, one capacitor (3 uF), and thanks to you, one l-pad with only two legs connected.
"you could also use an appropriate resistor in parallel with your 802 so the reflected load would be the same to the highpass cap and not have to change the capacitor."
Werner also sent me a diagram with the autoformers that showed an l-pad connected in parallel but with only two legs connected. I didn't know what it was for. His diagrams said "middle position normal". I assume this to be an 8-ohm l-pad attenuator, and since no signal goes through it the quality shouldn't matter that much.
Randy
I would think the L-pad would be in parallel with the driver and adjusted relative to the desired attenuation to keep the reflected load to the 3uF fairly constant. As more attenuation is applied, the resistance in parallel with the driver would be lowered. This parallel resistance would have an additional benefit of swamping the driver's resonant impedance peaks.
I guess the variable resistance could also be used at the autoformer's input. Which way is it shown in your diagram?
are you using KBG caps?
Karlson Evangelist
It is shown in parallel with the autoformer's primary winding (taps 0 and 15). The tweeter is connected to two secondary taps, so I suppose it is in parallel with the tweeter as well.It looks like the attached image (at least for the HF portion). Werner indicated on this (and my) diagram that the variable resistor was for rolloff. From what you are saying without it the value of my 3uF cap (6,500hz) may be seen as smaller (7,000-8,000hz). In my case, it is not critical because of the slow first order rolloff and the gradual natural rolloff of the Altec 414c, but being able to use the variable resistor to tweak the crossover frequency would be a nice feature. Any idea of the range it would vary?
Edits: 07/31/16 07/31/16 07/31/16
without the parallel resistance, your crossover frequency would drop. 1uF would get the crossover down to 3KHz at higher attenuation levels.if you have a multimeter or VOM, temporarily disconnect the L-pad and measure and record its values at the center, high, and low positions. That will give an idea of the impedance available seen by the capacitor for different taps.
looking at an old wirewound 8 ohm pad as as resistor, I can get around 43 ohms on the high range and below 1 ohm on the low end. Half way might be around 25 ohms.
If you didn't have the parallel input resistor, then 6dB attenuation would reflect 64 ohms impedance for a 16 ohm driver (assuming it was a pure resistance) and 0.33uF might give ~8KHz crossover with a 16 ohm load, or 0.68uF with an 8 ohm load.
those are very high-end looking autoformers - I've been happy with Bob Crite's model which allows 1dB steps.
Karlson Evangelist
Edits: 07/31/16 07/31/16
This is in 1 dB steps as well (custom order). I figured with 802's going for $500 a pair it was worth the extra cash.Is there a way to measure/calculate the impedance at each attenuation level? Does the autoformer as as a series resistor?
With the added impedance of the autoformer, I may be able to experiment with a smaller capacitor (i.e., a nicer capacitor since the value will be smaller). If instead of 8 ohm impedance,the total impedance is now 28 ohms, I can go from a 3uF to 0.82uF capacitor.
Am I completely off base?
EDIT: Dave Slagle in another thread stated "another interesting thing with the autoformers is you can do the crossover network at higher impedances making the cap values smaller."
Using an online l-pad calculator, if I need 6dB attenuation, R1 will be 4 ohms and R2 will be 8 ohms. Plugging this into a first order calculator gives me a 2uF capacitor (12 ohms) with an 8 ohm damping resistor.
Edits: 07/31/16
if you had a woofer tester or similar, you could directly measure the reflected impedance
here's Bob Crite's two way "first order" network for Eminence's Beta8cx with a Selenium D210ti compression driver. The required 12dB attenuation dropped the needed capacitor value to ~0.44uF for a nominal 3KHz crossover. That's nice as exotic caps would be more affordable. IIRC those autoformers used to sell for around $35US each and have 1dB steps - I think they have enough power handling for most "Klipsch style" situations.
Karlson Evangelist
That is interesting. Using the v-cap crossover calculator, a 3KHz first order with a 0.44uF capacitor woould require a reflected driver impedance of 120 ohms. Could that possibly be right?
Bob Crites shows a chart that gives you the turns ratio for different attenuation values. The reflected impedence is proportional to the square of the turns ratio. For 12dB the turns ratio is 4. Assuming an 8 ohm speaker, the reflected impedence is 8*4*4=128 ohms. Thats what the cap would see if in the circuit befor the attenuator. You can change the impedence value by paralleling a resistor across the attenuator inputs.
Or, you can move your cap between the autoformer and the driver and the crossover point will never change with different attenuation levels. The cap will only see the impedence of the driver.
Interesting - If I want to bring my 802 in at 8khz, with about 4 dB attenuation, I could use a 1uF Cap. That same cap would bring the crossover frequency down to 6khz at 5 dB attenuation, and 5khz at 6 dB.
With the 802G, this 1uF cap in series could easily work in a 3dB range.
you could also put a resistor directly in parallel across the driver - that would swamp some of the driver/horn impedance peaks, and not use the input shunt resistance. Sometimes if a horn/driver/speaker has a large impedance peak, it will crossover much lower than whats expected if the speaker was a pure resistance
oh - here's a Klipsch network with some reflected values noted
Karlson Evangelist
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: