I hope that the new 997 has very low output impedance and high input impedance. The 990 had a 1K ohm output impedance and caused it to sound veiled compared to other pieces I own. I would like to see a 100 ohm output impedance and 50k ohm input impedance. This pre/pro needs to be able to control other equipment and that means it has to have a low output impedance. Any ideas on these specifications Outlaws?

Tim

HUH? Do you have 100 foot interconnects?

Yes, I tested it against other preamps in stereo mode. I used the 7125 or an NAD C350 as the amps and switched preamps ONLY. All of audio is impedance matching and having a 1K ohm output impedance feeding into a 50K ohm input impedance in the 7125 or the NAD sounded veiled compared with a pre-amp with a 100 ohm output impedance. I was not the only one who heard this. I am not knocking the 990, it is a very good pre/pro, but it is not as good in stereo as many other pre-amps. Do a comparo yourself!

In general you want to keep the output impedance of a preamp at 100 ohms or below. The input impedance should be around 50K ohms or better.

Tim

Something else must be in play here. A 1000 ohm output impedance in itself is meaningless without knowing what the input impedance of the associated gear and the capacitance of the cable is. Neither of these are an issue with typical gear you would connect the 990 to, and certainly would not be an issue with an Outlaw power amp.

Have you performed any objective measurements such as frequency response? If the response were effected by a component mismatch, that would explain the change in sound quality.

There is no difference in anything in set-up. My speakers are dead flat from 100hz to 14KHz +/-1 db. The matching is in impedance on everything in audio. Capacitance and inductance play some role, but that is more of an issue with speaker cable than IC's.

If you guys want to experiment try it yourself, you will see. Impedance is the same reason a tube amp sounds the way it does. A tube amp will put a bump at the xover points on your speakers due to it's impedance interaction with the speakers. On a typical two way you are crossing over around 2KHz to 3KHz, so that area gets emphasized. the bass also gets emphasized due the high impedance at the Fs of the driver. The output impedance is usually expressed as damping factor (8/output impedance). What happens is when the output impedance is not low enough to "drive" the component it is hooked up to, you get some frequency irregularities at the high impedance peaks. Sometimes this is because of back EMF. Remeber as your system plays impedance changes.

Trust me you want a very low output impedance, and high input impedance...physics do not lie! One component is driving the other, you want the low impedance component driving the high impedance component.

Tim

Sorry, but your assumptions are not completely accurate. Modern components operate in what is known as "unloaded" conditions, and this is quite unlike the older days of tubes and transformers where impedance matching was vital for maximum power transfer (power transfer was important as it is intimately related to voltage transfer).

Simply stated, in an unloaded condition you want the input impedance of the receiving component to have at least an order of magnitude higher impedance than the sending device; in other words a 1:10 ratio or more. In the case of a component with a 1000 ohm output impedance, there will be no interaction (frequency response error etc) as long as the receiving device has an input impedance of at least 10k ohm. All currently available consumer components easily meet this requirement. I know for certain that the Outlaw amps meet this requirement by a large margin.

Interactions can occur with the capacitance of the cable, but with the 1000 ohm output impedance you refer to, it would require a cable capacitance of .008 uF to have appreciable drop in response at 20kHz. For reference, a typical coaxial cable has a capacitance of roughly 30 picofarads per foot - 30000 times less, and that's far too little to have any effect.

Modern preamps almost universally use operational amplifer ICs in their gain and buffer stages, and these have exceptionally low output impedance. The effective output impedance of these preamps, except for the most exotic, is determined by the resistance of buildout resistors prior to the output jack. These are included to provide short circuit protection and to provide electrical isolation, which helps avoid ground loops.

I'm afraid that that 1000 ohm output impedance of the 990 is inconsequential as far a having any ability to influence the sound you are hearing. As long as you are terminating the output of the 990 with at least 10,000 ohms, which is the case in almost all common consumer gear, and your interconnects are not hundreds of feet long (unlikely) there will be no audible "veiling" of the sound you are hearing.

Tube power amplifiers sound the way they do because of their distortion spectra, not their output impedance or lack thereof. Frequency response deviations can occur because of simple voltage divider interaction between the output impedance of the amplifier with the speaker impedance, but these are not what gives a tube amplifier their unique sound. Not all speakers offer a load which is complex enough to have interaction with the output impedance of a typical tube amplifier to a significant degree. In the case of systems which use active crossovers and consequently do not have passive crossovers within the speakers at all, the load is quite simple and response variations because of output impedance interaction are inconsquential. In any event, if somebody has a speaker which presents a complex, low impedance load, I would not recommend that a tube amplifier be used at all.

Tube amplifiers produce predominately low order distortion components which are dominated by the even harmonic products. These usually are greatly diminished by the 5th harmonic. A typical solid state amplifier, particularly a "balanced" amplifier (more accurately called a "differential" amplifier) will have an almost total lack of even order harmonic distortion components, with the total harmonic spectra having significant contributions well beyond the 10th harmonic. The even order harmonic components are important because they tend to mask the presence of all other harmonic distortion products. This is the reason that a tube amplifier can be producing, say, 1% total harmonic distortion, yet the ear cannot tell that it is producing distortion at all.

I am not going to argue with you, it is a waste of time. By the way the 10:1 rule came from the era of tube amps. The impedance interactions are real and documented. In fact, Stereophile has been doing measuements into dummy speaker loads for years. You can go back and look at their data. Here is a link to an article printed by them in 1994 which sums up most of what I am talking about as far as speakers/amps. http://www.stereophile.com/reference/810/

If you want to hear what a tube amp sounds like put a 1/2 ohm resistor in series with your speakers. Or to get really close calculate out what you need to get the DF to tube amp levels, something belwo 20, in the case of some a DF of 4 such as Cary. By the way, modern speakers are using more and more drivers with wild impedance swings. You can go to Seas or Tymphany's website and see the impedance plots of the drivers. While some manufactuers use zoebel's to flatten the impedance, 98% do not. I do agree though that active xover's generally do not have wild impedance swings, as they generally use an impedance flattening circuit.

My ic's are 1 meter long and not garden variety, they are not esoteric either. If I recall corectly their capicatance ranges from 28 pico farads to around 100 pico farads depending on which ones. The ones I use most commonly are the 28 pico farad to the CDP(this does not change, as my CDp is sensitive to capacitance on cables) and the amp to preamp use a pair of 50 pico farad cables.

We can go on disagreeing forever, I have been through this too many times. I have done the testing myself and have had it proven over and over to others through testing and results. I will just agree to disagree with you.

Tim

I said in my last post that I don't disagree about impedance interactions. However they are not what gives the character to a tube amp. If it were only amplitude changes, then you could simply use an equalizer to make any old solid state amplifier sound just like a tube amplifier. Not likely.

Of course unloaded circuit networks have existed forever, but it has become the de facto interface model with the advent of solid state electronics which can achieve low output impedances with ease, this especially being the case with IC Op Amps.

Again, have you tested your equipment with lab equipment? I think that would sort out the problems you are having more than any amount of conjecture.

In any event, if you are concerned about output impedance of your preamp, why don't you simply replace the buildout resistor with a different value, all the way down to zero ohms if that's what you like? Similarly, the input impedance of gear is generally established by a resistor at the input of the input buffer stage - why not just replace it with a value you deem appropriate?

By the way, active crossover systems don't have to use "impedance flattening circuits" or any other compensation mechanism. In this type of setup, the diver is connected directly to the output of the power amplifier, so impedance swings are limited solely to that of the driver, and these swings tend to be very mild compared to a speaker system with complex passive crossovers. The crossover is at line level, before the power amplifier, so there is no possibility of interaction with the speaker's impedance magnitude (my system has been actively tri-amped for decades).

Yes, speakers measured with calibrated microphone. Interconnects measured with DMM. I agree that impedance is not the only thing which make a tube amp sound different, but it is what contributes to a majority of the sound. Tube amps also sound "warm", because of the higher distortion. Distortion actually makes things sound "warmer", whereas very low distortion sounds more "clinical". Hence the terms some use when describing SS and tube equipment.

It might not be just the output impedance, but in my experience, 1000 ohms such as in the 990 is very high! 100 ohms is much better! In fact 1000 ohms o/p impedance is the highest I have seen on SS gear. The testing was done in both stereo modes, and the "veiling" was there in both modes. It could have something to do with something else in the circuit, but I do not a have a schematic to make any conjectures.

Tim

Then, like I mentioned before, why don't you simply replace the resistors which establish the output and input impedances with something you deem more appropriate? You shouldn't need a schematic to be able to determine the location of these parts.

That would change the voltage drop across the resistor and depending on the circuit, the output voltage/current. In many pcs of equipment there is also an RC filter network that is part of the output. Some have an emitter follower circuit, and others use this resistor as part of a negative feedback network, changing the resistor opens up a whole can of worms. Designing it with low output impedance in the first place is a much better idea. If by just changing the resistor Sherwood(the manufacturer of the 990) could have lowered the o/p impedance, don't you think they would have done this to keep it low, we both agree lower is better? It is not as simplistic as you make it out to be. You just can't change one component and have no worries about other effects.

Tim

Depending on the 990/997 'differential' output design, one might adapt to find a very low impedance output available, even if some might stand aghast at such a suggestion.

I've been involved in the design of audio equipment for over 30 years, so, yes, I am very familiar with what the circuit path of home theater preamps looks like, and know what the ramifications of these circuit changes would be.

I think you would find the 990 circuitry as I described, and any competent engineer could change the value of these resistors without potential problems. Any RC type filtering like you describe would be performed within the feedback loop of the op amp, and not passively after the output pin. I haven't seen an emitter follower used in any modern piece of equipment except for very high end designs which use discrete components.

So, if you don't like the output impedance of the 990, change it, or find somebody who knows how to do it. It's not as big a deal as you seem to think it is.

Even in a differential output stage, there will still very likely be an isolation resistor in each leg of the signal. This is just good design practice.