captainbeefheart

Gain - 53/53 Distortion

I have 6 smooth plates left. A quick reminder how I test preamp tubes, normally one will either have a cheap emissions tester or better a transconductance tester, the latter being better but I do not think this gives anyone really any indication at life or performance in circuit. Why? First of all transconductance or "gm" is not constant, it changes and along with the plate impedance (rp) make up the mu. Now "mu" is the only constant and; gm*rp=mu So gm is only half the story when calculating gain in circuit, you need to know plate impedance which also is not a constant and changes like gm depending on specimen and operating point. You can have two sections, one with a gm of 1300 and the other 1500 but in the same circuit you will get the same gain. Having matched gain is most important in circuit. I test for microphonics and noise, gas, leakage, and distortion. Distortion will give you an idea how linear the specimen is, this will be more prevalent the more output swing you need so having if you want your channels to be matched and perform similar you want matched gain and similar distortion profile. Test circuit: B+ = 250v load = 100k bias = 1.5v Distortion measurements are taken at the same output voltage for all sections.

Capacitors and inductors have their voltage and current's 90° out of phase from each other. With a capacitor the current leads voltage and the opposite for an inductor. Now the amount of phase shift will vary depending on reactance and frequency. Take for example the 13uF capacitor on an AA network, the lower in frequency you go the more phase shift you will have. The higher in frequency you go phase shift will reach a point where it will be 0°. Reactance is also known as impedance, so the lower the impedance of the reactive element you get less signal is across the two terminals, where you get to the point the impedance is so low and for modeling AC terms behaves just like a wire you get the same potential across it between it's two leads, this is where 0° phase shift occurs. The best way to view this would be at each driver output from the network and plot phase shift vs frequency. You will see the woofer will have greater phase shift as frequency increases and the tweeter network will have the most phase shift as frequency decreases. Mid-horn will have phase shift at both high and low frequencies and the least at middle frequencies.

Very true! I should have said the more turns (or higher impedance) the harder it is to wind a high quality transformer. James is really good. I have a set of 3k Transcendar Gerry wound for me I am liking right now. Honestly our tired old ears won't notice -1.5db at 20kHz anyway. For my age I still do ok on hearing tests because I have always stayed away from high volume anything and wear ear protection. I have always been extremely sensitive to poor quality high frequency sounds, ringing, dissonant distortions, etc.. it's not just fatigue it actually hurts my ears, it even makes my skin and teeth feel very weird I hate. It's more of a bragging rights thing to have extreme bandwidth amplifiers, well extreme on the top end anyway bass is good. I don't see the need for a stereo amp to need to put out full power at 100kHz. I mean it is a good indication the amplifier is of good design but I don't really think there is much information past 40kHz that we will perceive. I say 40kHz because of harmonics and them being part of the bigger picture.

Probably another reason many 300b amps use 3k load transformers, less turns will give less cross winding capacitance and better high frequency response compared to 5k.

Wire inside speakers are very small length. With the price of raw materials including copper increasing going with copper clad aluminum is a way to keep prices of products competitive in the tough market. Engineers sit around and try and beat the cost curve so when you can cut the price of wire in half it's kind of a no brainer when performance is the same.

Yes constant DC with + voltage type systems the conductor becomes an anode and so with the constant DC current applied anodization occurs much faster. I would think this would be not an issue with AC speaker system with no DC, of course many SS amps still have a very small amount of DC offset at the output which may still may have enough current to anodize the conductor. 20mV of DC across an 8 ohm voice coil will produce 2.5mA of current, probably enough to cause problems. Old cars that ran -6vdc systems do not suffer the anodization issue like the cars today, of course the body would then become the anode and oxidize faster which is why they switched to + type power, easier to replace wire than the frame and body.

What's wrong with copper clad aluminum? Your overhead power distribution lines are aluminum? The fact aluminum wire is half the price of copper makes it very appealing. For speaker wire we are going to oversize the wire for best performance, I can see why we strayed away from aluminum inside our walls of our home but for cheap speaker wire for a DJ gig it won't make any difference. For power distribution if you have an unacceptable loss across the wire aluminum is inferior as it will be less stable compared to copper when temp increases but this isn't an issue for this application when sized correctly. Since it is half the price you could even jump up to 12awg and it will be cheaper than 14awg copper and have less impedance. Both copper and aluminum oxidize, aluminum oxidation is harder non-conductive crusty stuff so just make sure you make good connections but since it's usually copper clad aluminum this point is moot.

The laws of physics do not care what is written on the label for marketing. The majority of the public do not know anything about conductors/wires/cables so it's smart to write "speaker wire" on the outside and push the price up. Speaker signals are large signals, they do not need anything special, you are most concerned with losses and damping factor so the only variable it comes down to is impedance. The larger the gauge wire the lower the impedance per foot, for 30' 14awg is a good choice, you could even get away with 16awg but I would go with 14 for slightly lower losses and slightly better damping. There is absolutely no need to purchase anything specifically marketed as "speaker wire". You could even use Romex but solid wire is tough to bend and coil up etc.. As you noticed I chose a 14/2 conductor that has an extra outer layer of insulation around both wires, this is a wise choice if you are planning on DJ'ing, this is what professional sound engineers would choose for this application. Signed electrical engineer

Nice Job!!! Load will depend on B+ and what kind of power vs distortion you want. For voltages around 200v-250v you can use 2k-2.5k load impedance. When you start to get up to 400v you will have a higher load like 5k or even 6k. There is no rule, you can use what you want. For example: 400v 3k load gives you almost 12 watts of power but distortion will be higher of course, if you choose 5k you only get 8 watts but distortion is -6db lower. Granted if you measured at the same output power, say both the 12 watt amp and the 8 watt amp measuring at 8 watts output they will be close for distortion so most designers tend to just go ahead and use the lower load impedance to get maximum power say capable of 12 watts but still only rate the amp at 8 watts output because that's where it is at 5% THD which is considered to be about the maximum you want with a SET amplifier. They get away with that much because it is second harmonic dominant and fairly benign. I notice you are -1.5db down at 20kHz, have you found where this is happening? Before or after the output transformer?

For 30' I would use 14awg wire, at .0025 ohms per foot that works out to ~.075 ohms. Will work good. You don't need anything exotic you will not hear any difference. Just go into any Home Depot or Lowes and they will have something like this that you can purchase by the foot so go in and have the guy cut 65' of the stuff. Terminate how you like or just strip, twist and insert into speaker and amp binding posts. Easy peasy https://www.homedepot.com/p/Southwire-By-the-Foot-14-2-300-Volt-CU-Black-Flexible-Portable-Power-SJOOW-Cord-55812199/204725134?MERCH=REC-_-searchViewed-_-NA-_-204725134-_-N&

More stuff to think about if we haven't enough to think about already. With our Klipsch speakers that are inductive voice coils, the DCR will take over at low frequencies but the series inductance at high frequencies will cause increase of impedance which you would think is easier for the amplifier right? Well I would say less than 5% of the people are running fully open loop amplifiers, probably less than 1% really. With negative feedback (like amps the 99% of us use) as loading becomes easier (higher impedance) open loop gain increases, when open loop gain increases so doesn't the amount of negative feedback which in turn reduces the amplifiers phase margin to the point of instability. I hate blanket statements but the flatter the load on the amplifier (pure resistance) the better the results.

Of course we use inductive voice coils and the big difference being there is a straight resistance due to the DCR of the voice coil so it isn't just a purely reactive load like an ES panel. But in regard to the discussion an ES panel for the most part is a purely reactive load and capacitive by nature which is not how the majority of amplifiers are designed to power. Besides the voice coil inductance and DCR, the passive filters in the crossovers are still in the circuit and make the load even more difficult for the amplifier. I know it's a tough habit to break but with AC and reactive elements you cannot strictly think in terms of only voltage,current, and resistance like you would with DC. Adding another vector makes things interesting 😜

Oh another great example that relates to audio is if you have ever heard of amplifiers that do not like or have even let the magic smoke out when paired with electrostatic speakers? It's because an ES speaker is for the most part a purely capacitive load and very difficult for many amplifiers to drive. When using ES speakers choose your amplifier wisely, many are not up to the task.

Both capacitors and inductors are reactive loads, current and voltage are out of phase (I believe this is what you mean by the waves don't line up) unlike with a resistive load where voltage and current are in phase and all the power from the source goes into the load. Theoretical perfect caps or inductors voltage and current are 90° out of phase but in the real world there is a loss angle (tangent) which complicates things further, they won't be exactly 90°. Example with capacitors are: loss angle = esr/reactance Many have experience with this with electrical motors which are an inductive loads, this is probably what you remember because it is the most widely used example and you are correct you would use capacitance to correct it. The opposite for a capacitive load where you would use an inductor to correct it. There are pros and cons to placing them in parallel or in series but for the most part you will almost always see a parallel approach, so placing a capacitor in parallel with an electric motor. A series example is with an amplifier and a capacitive load you will add an inductor in series with the output which you see many times in SS amplifiers. For example my well pump is 3/4 HP and would pull 8.4 amps after it came up to speed, after adding a capacitor in parallel it now pulls 5 amps. The extra 3 amps was being wasted and didn't go into the load. I made the system more efficient and easier on the source by increasing the power correction. The idea is you want your source (amplifier) to not waste power and also have an easier load to drive. I won't even get into reactive loads changing phase margin of the amplifier. That's a whole different issue and topic.

I am not trying to say ALK networks are the way to go, I only mentioned his method of power correction for reactive loading. You cannot think in terms of solely current and voltage, there is a time vector and loss angles with reactive loads. I am not even mentioning stability and phase margin of amplifiers, that is completely different problem but ideally you want to get your power factor as close to perfect as possible and you want to make the load as linear and as close to a pure resistance as possible. In case you need a brief refresher on the subject; https://electrical-engineering-portal.com/reactive-power-power-factor-correction

The way you describe your issue to me sounds more of a problem of the amplifier just not "happy" with the tough load these speakers can present. There are ways to fix this, one is adding a bunch of parts to the crossover to power correct the reactive loading, this is the approach ALK uses with his crossover design. The amplifier gets a much easier load to drive and you get less losses in the passive networks. The other solution of course is changing amplifiers until you find one that can drive difficult loads. I know I say this quite a lot, "many people don't know" but I repeat myself again many people don't know that when companies design your amplifier and test it they do so on a very easy to drive purely resistive load. Yes some well educated designers will make sure their amps can drive difficult loads, some claim they can but they are vague and typically are talking about one variable which is can deliver extra current with dips in the impedance plot but that's only one issue. To rule this out of course I would try at least 3 different amplifiers with the speakers and if the problem persists than it is within the speakers or room acoustics and a room sweep with a calibrated mic and software would be the next step to see what's going on.

Good catch! Thank you!! It was a typo, I meant to write 1/2 watt not 1/5 watt. My book 400mW with a 1/2 watt resistor is yes within limit but not good engineering practice. I will go back and edit the original post.

I too have put much thought into this. If you look into the equations of fluid dynamics you will notice they are nonlinear, these equations are used for sound waves through liquid or gases. Because of this nonlinearity, sound waves are being distorted as they travel. For audio acoustics it's a gas form in the air and localized pressure changes, basically the waveform will travel faster during the high pressure portion of the waveform and slower during the low pressure part of the waveform, this is the nonlinearity. For a typical sine wave the peaks will be the high pressure portion and the troughs the low pressure portion. Because of this nonlinear distortion other frequencies are introduced shown by a Fourier analysis which are even harmonics, mainly second order harmonic distortion. This is how our hearing has evolved, we obviously naturally hear in nature a single ended, nonlinear distorted wave form and this has always been natural to us. I 100% believe this is why people tend to prefer single ended amplifiers that produce nonlinear distortions and the subsequent even harmonics added. This is because it is natural to us, we are use to processing this type of information because that is how it is propagated in nature. Push pull amplifiers are more efficient and look great on paper BUT cancel even harmonics and so a Fourier analysis shows odd harmonic dominance which isn't present in nature and so can sound out of place, mostly fatiguing I presume because since it isn't natural our subconscious is analyzing and being worked harder to process the information. It's like doing a particular physical workout for millions of years and then changing to a different workout program, the next day those new muscles will be sore that haven't been worked out.

I can understand the need for more power with many speaker offerings being in the mid 80's for sensitivity ratings but to put it into perspective the 135 watt amplifier will put the Epic CF4 at 123db which is uncomfortable and for many painful volume. I just can't imagine needing more than that for in home use. There is never any free lunch in engineering and not to be a broken record in repeating that for a 300 watt amp you will not be in the sweet spot with a 102db sensitive speaker. Very few amplifiers have a linear diminishing distortion content the further you reduce the output power, this is the hallmark of what you want in an amplifier. The way they get a 300 watt amplifier to put out that power with such low distortion actually causes increased upper harmonic distortion at lower power output levels, it isn't linearly diminished distortion vs power, or the reciprocal a linear increase in distortion vs increase in power out. For an 86db speaker and 135 watt amp you will more likely run the amp in it's sweet spot at moderate levels. When anyone tells me that their high power amplifier isn't cutting it with very efficient speakers I half the time think they may need to bring the amplifier to a tech to make sure it is functioning correctly. Take your hearing seriously as when you get older and lose it from abuse you will wish you hadn't.

Exactly why his amplifiers have become so popular with music enthusiasts, they sound great in the power band we commonly listen to our music at especially with efficient speakers.

FWIW a little tidbit of information to think about. The higher powered the amplifier is the worse it sounds at lower output powers because it will actually have higher order harmonics at a higher level present at the critical lower power where most do their listening with Klipsch speakers. This is why so many have done away with the arc welder power amps especially with Klipsch speakers because they are so efficient you just do not need the power. Take a look at the product literature and you will never see an FFT graph of these high powered amps for 1 and 10 watts output, they only show distortion figures for rated maximum power which tells us nothing how the amp will sound in the real world. It's just the nature of feedback with these amps, they need it to get the low figures at high power but at low power it actually makes for bad sound. So unless you are a commercial movie theater I would stay away from these types of amps. Of course if all you do is watch movies disregard what I have said, I mainly am talking about music reproduction quality.

The CF4 is an 8 ohm speaker, this means your amplifier is producing 135 watts not the rated 200 watts into a 4 ohm load, well unless you plan to parallel the 4 speakers into a 2 channel setup but your home theater amp tells me this isn't the case and you are using 4 speakers for surround sound. I have a similar powered amp for sale, not sure why you need more power but it has more bells and whistles for integration into a home theater setup, it is a Pioneer Elite VSX-90TXV and it is dead quiet on very efficient klipsch speakers.Here are the details; https://www.pioneerelectronics.com/pio/pe/images/portal/cit_3424/476124747VSX-90TXV.pdf

It was so long ago I can't remember fully either. I do remember some guys putting larger tweeters on top and time aligning a little better, this was the first I heard about increasing the 2uF cap to let lower frequencies into the tweeter since the K-77 is pushed up pretty high at 6kHz. From memory with just a broader bandwidth on the bigger tweeters the sound was an improvement. Kinda like why we like the less steep, more "blended" simpler A and AA crossovers. They say music lives in the mid-range and now the tweeter and mid-horn are presenting the upper mid-range frequencies in a more blended fashion between two drivers. I think Bob was the first to add the inductor and drop the mid-horn down after doing measurments of it and seeing higher distortion up there from it and since the tweeter is passing down near 4kHz now the speaker can be cleaned up by getting rid of the distortions from the mid-horn in these upper mid-frequencies. But some of us like non-linear distortion as it sums to even harmonics. I wish Bob was around to make a call and ask him. I don't know his son that well but Bob would chew my ear off back in the day.

You would know better than me that's for sure. I had an old drawing of an A network with the 2uF crossed out and 4uF written in for a value. I was always under the impression this was kinda the original mod when upgrading the tweeter with the eminence driver found in CT125 and now the CT120, which will handle lower frequencies better so the larger cap drops the filter to lower cutoff. As a bonus the next step was to clean up the mid-horn by lowering the filter down to 4500 since the tweeter is now covering this range and does a better job at it so the inductor was added. It's been so long and the internet now makes everything so much easier to keep track of. I know now just about everyone adds the inductor and of course Crites sells them with the inductor. Have you heard the type A with just a larger value tweeter cap and no mid-horn inductor?