Leo's strange and wonderful resistor tweak

[Deang]

About a half a year ago Leo and myself were passing emails back and forth about the impedance characteristics of our RF-7's. There were several threads on this as well.

Are they 102db/w @ 8 ohm speakers, or 99db/w @ 4 ohm speakers? Something in between maybe? The question at hand was - what taps to put these things on with a tubey?

I'm sure Klipsch won't mind me saying that the RF-7 is not the most tube friendly speaker around. Oh, they sound good with tubes alright, but because of the impedence curve -- the response is not quite flat. In spite of sounding great, they are prone to sounding a little peaky at times. Leo and myself take advantage of the parallel notch filter Klipsch placed in the crossover to modulate the tweeter's broad peak, and have made a slight adjustment that works great at smoothing out the treble for 2-channel music.

The question remained however regarding the best way to accomplish the smoothest, flatest response from top to bottom. I had thought that by maybe driving down the DCR on the woofer inductor, the amp would have better control of the drivers, and the slight increase in bass SPL would work well in accomplishing this goal. This didn't work out very well. In spite of the bass being ever present, it sounds somewhat plodding and thick. My attempt at using a lower AWG inductor to do this failed, and I will be buying some Madisound Sidewinder air cores with the correct DCR to replace the spools of copper sitting on the boards.

So, feeling somewhat dejected I started cleaning up the room, which had pretty much become trashed from the latest endeavor. While cleaning, I ran across a small bag with two 8 ohm resistors, and at first I couldn't remember what they were for. Then I remembered the words of Obi-Wan,

"Another thing you should really try, especially with all that pp power, is: Get 2 8-ohm 5-watt resistors (closest value to 8 between 6.8 and 10 Ohms is fine). Connect your RF-7s to the 4-Ohm amp out. Connect the resistors in parallel. This is very easy engineering logic .. not a trick. Just do it."

The best time to try something new is when you're feeling sorry for yourself after just mucking something up.

I broke out the Klein crimper (my new favorite toy) and terminated the resistor ends with some brazen copper Paduit ring tongues. These fit very nicely into the screws of the terminal strips at the amp's outputs. I stood back and admired the strange looking concoction. My how the mighty have fallen -- zip cord and 8 ohm resistors hanging off of my amplifier outputs! I turned on the amps and let them warm up. I went for a smoke. Dejected people do that.

I came back in for a listen. I have new amps and new speakers. It's the damnest thing I've ever experienced. The bass still doesn't sound "right", but it actually sounded better than it did, and Craig -- you're going to love this -- but it sounds "flat". I don't know how to explain it except to say I like it.

Now Leo, you are a wild one. What the hell is going on with my stuff? How can it sound so...different? The sound is more "relaxed" -- does that make any sense?

I know you ordered the 8 ohm transformers for the Moondogs based on some information that changed your mind about the 4 ohm thing, and you never said what that was. This thing that I'm doing, can I do it because of the power that I have? What will happen if I drive it a little (besides the resistor getting really hot)? Would you mind explaining what the amps are seeing -- they sure seem to like it.

[leok]

Dean,

I'm glad you finally tried it and the results are so good. What happens is, the amp sees a more resistive, easier to drive load, and the speaker sees a more controlled, better damped drive. All that should give you a better controlled sound. The amps, speakers, and Dean are all happier.

I didn't do this with the 2A3 SET amps because they don't have the power to spare. Wish I could.

leok

[Chris Robinson]

I can't wait for his reply ...

[Deang]

Is that the reason you went with the 8 ohm Cobalt transformers -- because the RF-7's ride along more at 6 ohms and above then 4 ohms and below? The power issue? 4 ohm transformers would have taken you down to 1.5 watts -- is that correct?

As far as my situation goes, and I just can't remember -- did I just cut my power in half? The amps run at 60 watts whether I'm on the 8 ohm taps or the 4 ohm taps -- but does adding this resistor to each amp change that?

[leok]

Dean,

Right. The output voltage would have been lower and less overall power available. I reviewed it with the transformer designer who recommended I stay with the 8 Ohm version.

leok

[fini]

The Good Old Sunbeam T-20

fini

[Deang]

LOL, that's great.

About the time I've come to the point of totally losing my sense of humor -- you guys come in a bail me out. Very nice. Thanks!

I didn't use 5 watt resistors, I used 12 watt -- and I think I've learned something interesting. I can push my volume control dial to just past 10 o'clock before the resistors start to heat up. If I dial back just a hair -- they cool down to lukewarm. If I go to 10:30, which is where I listen when doing intense therapy -- the resistors heat up to the point where you really just don't want to leave your finger on them more than 10 seconds. I played the risk game and did a whole song this way -- certain the resistors would give out. I think what I learned is that even in my loudest listening session (about 95db) -- I'm probably using less than 12 watts continuous. Have I surmised correctly?

[leok]

Therapy level sounds ok indefinately. Power resistors can run hotter than you can touch at all, so of you can hold on for 10 seconds, they're not too hot.

I would guess that means over 5 watts continuous. I'm impressed

leok

[fini]

I think it's a little unusual to have English muffins past 10 o'clock, unless you're a late riser.

fini

[leok]

That's 5 Watts for each speaker and 5 Watts for each resistor. Amps producing 10W / channel. Maybe everything is a little higher .. probably not 12W in each resistor or they would be hotter. Anyway, whatever you're burning in the resistors you're using in each speaker.

leok

[Ray Garrison]

Hummmmm...... this makes for an interesting thought experiment.

If I understand the logic here, ... wait a minute, I couldn't figure out the proper sequence of events here, so I'm just going to list a few things that occur to me in no particular order, then take a look at what we've got.

1. Amplifiers typically have some output power level where the combination of total harmonic distortion plus noise reaches some minimum value. This can be determined by running an analysis of THD while ramping power up over time. Here's a graph of a decent little integrated amp, the Creek 5350, showing this effect.

2. At lower power levels, many amps (particularly solid state) have fluctuating levels of THD and noise that do not track linearly with power levels. Here's another graph of that Creek that shows this somewhat better:

3. Most speakers, and certainly this is especially true of Klipsch, draw low amounts of power from the amp most of the time under most circumstances.

4. As the amp operates at this low level, the distortion values from loud to soft passages might change by a factor of, uh (looking squinty eyed at latter graph) 13 (?) dB, or a factor of *20* fold.

5. At the same time, as this noise is a larger portion of the signal relative to the output level at this low power level, it will be much more noticeable that it would be at a higher output level.

6. By placing a power resistor in parallel with the speaker, we force the amp to pump out more power, moving it into an area of more linear performance.

7. The fluctuation of the relatively (compared to the resistors) low amount of power going into the speaker respresents a much smaller variance in overall power level, thus resulting in lower fluctuations in harmonic envelope and noise levels.

8. In addition to lowering the levels, the reduction in variance allows us to ignore the distortion more easily - the human ear/brain is very good at filtering out and ignoring signal contamination, providing that the unwanted signal is somewhat stable. If it's varying all over the place, you can't filter it out as effectively, so it becomes much more noticeable.

That's really cool. It seems like you ought to be able to calculate the particular resistance value that would provide the optimum load for any given amp/speaker/volume level, and optimise the system to sound best at that level.

[leok]

Ray,

What you have posted there is what I consider the fundamental ss amp problem: distortion as a % increases as power drops. I don't think the parallel resistor approach is all that effective in the ss case because I think the ss distortion is somewhat a fixed voltage problem.

In Dean's case, the goal is to make the speaker load look more like a constant resistor rather than the varying complex reactive load that it really is. Also, the lower impedance seen by the speaker helps.

I've experimented with a voltage divider to push ss amp operation up into a lower distortion region (operating it at 40 Watts or while feeding the speaker less than a Watt using a 10:1 resistor divider). This was very effective, and made the ss amp sound much better, but not nearly as nice as a PWM or tube amp.

leok

[Erik Mandaville]

This is also sometimes referred to as a 'swamping resistor.' It just helps the amplifiers 'see' a more even impedance curve over a broader range of frequencies. Lowther owners have been using them for awhile, and I have used 20 ohms (My drivers have an impedance of 15 ohms)in parallel with my PM2As and Moondogs. I wish I could report the sonic improvements that Dean experienced! It can make for a better match between amps and speakers, and I felt good about that, despite the fact that there wasn't an obvious difference in the sound.

And Dean! I hope you get the chance to listen to a really good SET amp. You seem to be curious, and I hope you have the opportunity in the near future. But Leo could be right in the sense that if you play music at really loud levels, single-ended distortion figures at those levels may not be your cup-o-tea! Only you can decide that. With the right speakers, they can be wonderfully musical. I'm working on what will probably be my last amp for awhile -- that is until I can upgrade to the cobalt OPT. What might be an interesting experiment is to try a low-power SET on your high frequency horns, and something with a little more drive (you already have those amps!)on the low end.

[leok]

Mark,

I'm refering to the amp shown above, and many other ss designs I've seen reviewed in Audiophile mag. and the NAD I gave up on.

The Tripath TA1101B exhibits 0.03% distortion in a flat line from 6 Watts down to below 0.5W where the spec. stops. At that point the % distortion is still a flat line. I'm sure it eventually takes a turn up, but a power much lower than most linear ss amps including the one shown above.

Distortion for a Tube triode circuit without feedback approaches 0% as the power approaches 0 Watts. Their distortion curves are not "U" shaped (until powers so low that they are completely insignificant).

I believe the fundamentally different low power distortion characteristics of the Tripath and tube triode designs are evidence of why they sound so good driving efficient horn speakers.

leok

[djk]

Some random thoughts about the topic under discussion:

Enamel coated wirewound resistors usually have a 250*C rise above ambient at their rated power.

That would be 527*F at 25*C room ambient.

The reason a resistor in parallel improves the sound is that without the resistor there is inadequate inductance in the primary of the output transformer to drive the high impedance peaks of the loudspeaker near resonance.

The lower the Qts of the woofers, the worse the problem is.

IE: the RF7 woofers have much lower Qts than the Cornwall woofer does.

The simple solution for the RF7 is a 15 ohm 25W resistor wired in parallel with the loudspeaker and use the 4 ohm tap on the amplifier.

Ported speakers have twin peaks, one between 10hz~20hz that we can usually ignore due to lack of program material in that region, the second is around 60hz ± 10hz or so. This is the killer. An Altec 515 in a vented box hits over 200 ohms around 60hz. 15 ohms in parallel with 200 ohms is 14 ohms, that would be the new maximum impedance at resonance. Say the minimum impedance is 5 ohms, a typical value for a Klipsch speaker. 15 ohms in parallel with 5 ohms is 3.75 ohms, a good match for the 4 ohm tap on most tube amplifiers.

I can give you a design for a trap that will flatten the resonant peak without adding any extra load on the amplifier, but someone will have to offer to receive snail mail, scan, and post it for me.

[Deang]

"The simple solution for the RF7 is a 15 ohm 25W resistor wired in parallel with the loudspeaker and use the 4 ohm tap on the amplifier."

This presupposes that the nominal impedance of the RF-7 is 6 ohms, is this correct? I suspected as much. When Leo and myself discussed this a while back I had mentioned this, and had also wanted to try a 12 ohm resistor -- which puts it at 4 ohms dead on. You're suggested 15 ohm resistor brings it in at 4.3 ohms.

"I can give you a design for a trap that will flatten the resonant peak without adding any extra load on the amplifier, but someone will have to offer to receive snail mail, scan, and post it for me."

I still do snail mail, and I'm sure Leo is interested in this as well. What do you need from us to do this? If I post the information I have on the RF-7 at this forum, Bob Gassel will hang me from a tree.

On the other thing, I was wondering what you thought about the following? Is this an oversimplification for my situation? I'm ordering new inductors, but did run across this and have been thinking, which typically gets me into more trouble than it gets me out of.

"If you read the fine print in driver specification sheets you will find that the suggested alignments assume a series resistance of between 0.5 ohms and 0.7 ohms ( ie. DCR or sometimes referred to as Rg). Speaker box design computer programs also assume this resistance and none that I've seen allow you to change it which is probably fine for ported boxes but for sealed boxes it will cause major errors in the box volume..."

"...As you can see the ported box volume hardly changes, although it does change enough to change the tuning of the box/port combination. As DCR falls you have to increase (slightly) the box volume to compensate and if we'd assumed an DCR of 0.6 ohms and built a 22 litre box and then decided to throw the passive crossover away and do it properly then the box would be 2 litres smaller than it should be. This will cause f3 (the frequency at which the response is 3db down) to shift slightly higher and the roll off would be slightly shallower, which probably isn't a bad thing. Add in the extra damping and you've got a much better controlled driver/box combination."

[djk]

"I still do snail mail, and I'm sure Leo is interested in this as well. What do you need from us to do this? If I post the information I have on the RF-7 at this forum, Bob Gassel will hang me from a tree."

They're too busy fooling with garage doors to worry about small fry like us.

You could e-mail your snail mail address to me if you want to.

spamdkleitsch@spamyahoo.comspam

If you remove the spamspamspam I'll get the

message.

"If you read the fine print in driver specification sheets you will find that the suggested alignments assume a series resistance of between 0.5 ohms and 0.7 ohms ( ie. DCR or sometimes referred to as Rg). Speaker box design computer programs also assume this resistance and none that I've seen allow you to change it which is probably fine for ported boxes but for sealed boxes it will cause major errors in the box volume..."

This guy is on another planet. Madisound works up box designs with three different Rg values(regular inductor, big wire inductor, bi-amp or no inductor). But they have to do this. Programs don't assume anything like this, they have to be told to do it. And he has it backwards anyway, ported speakers are much more sensitive to this than sealed boxes.

Back to the resistor thing. I was going to suggest a series LCR tuned to 60hz to be placed in parallel with the speaker. This way the load only shows up where you need it, not broad band. The inductor is a large value, but we can get away with the smallest wire we can find as we are going to put a big resistor in series with it anyway. The cap is a bit of a problem as it is a big value too, but needs to be of high quality. Back-to-back polarized electrolytics look like the best here, with battery bias. With a low-leakage cap the life of the battery should be close to the shelf-life of the battery(no actual current flowing).

I think MIT has a note on this at their capacitor site.

If you need help figuring it out you can drop me a line and I will draw it up for you.

[Bill H.]

fini,

Love your CD Burner..............

[Deang]

Interesting thread for sure.

Since last night, I'm having some trouble with this myself. I borrowed my friends decade old, first generation Adcom 555 to give the DQ-10's their baptism by fire -- and it sounded very good. Curiousity got the best of me, and I hooked this monster up to the RF-7's -- not to blister my ears, but to actually try it out at low volume levels. The first thing I noticed was how the bass had tightened up dramatically, almost alleviating my criticism of the bass (damping factor?). The second thing I noticed was enjoying the music. Certainly not the sparkle and sweetness of the Quicksilver midrange and treble, but not the edge and harshness I expected either -- In fact, it sounded very clean. Of course, I'm running all of this through a $2400 Cary preamp, and now I can't help but wonder where all of this solid state harshness is really coming from. Since I don't hear anything above 15Khz, it may just be that I'm not hearing it.