Who is using passive preamps and why?

I don't think anyone has mentioned cable length and cable capacitance yet in this thread. Those certainly being factors that need to be kept to a minimum on the output side of a passive preamp. High output impedance into high cable capacitance (which is proportional to length, as well as varying widely among different cables) will result in upper treble rolloff, and long cables will increase the likelihood of other adverse cable effects as well.

Regards,
-- Al

Grant -- I'm not particularly familiar with the VRE-1, but I recall that it includes a jfet (active) buffer stage. Assuming that is in the signal path and is followed by a coupling transformer at the output, I'd expect the resulting output impedance to be much lower than in the case of a purely passive preamp. Perhaps that is why the long cable is not a problem for you?

Best regards,
-- Al

I can understand frequency abberations due to impedance matches, but why would a passive ever be any less dynamic?

One explanation which I suspect underlies those perceptions in many cases would be upper treble rolloff caused by the high output impedance of a passive preamp interacting with cable capacitance that is excessive in relation to that output impedance.

Upper treble rolloff would result in dull and sluggish transients, which very conceivably may create the subjective perception of reduced dynamics.

Regards,
-- Al

Pubul57: Does the 50kohm of my attenuator address that issue, or does it cause problems on the other end into my 100kohm amp load?

Herman's comments accurately addressed the impedance seen by the cdp looking into the resistive volume control, which as he explained is dependent on the rvc, it's setting, and the amplifier input impedance.

A separate issue, which I think is what the quoted question is addressing, is the relation between amplifier input impedance and the output impedance seen looking back from the amplifier into the output of the rvc. That will be dependent on the rvc, it's setting, and the source component's output impedance (if it is high enough to be significant).

With the volume control turned all the way down, the impedance looking into the output of the rvc (I'll call it Zo) will be essentially 0. With the volume control all the way up, it will be the parallel combination of the volume control's end-to-end resistance (in this case 50K) and the output impedance of the source component. If the source component has a low output impedance, such as 50 ohms, then the impedance of the combination will be essentially the same as the source component's output impedance.

Zo will be at a maximum when the volume control is set to the mid-point of its resistance range (which is NOT the 12 o'clock position; it will be just a few steps down from the maximum volume position). At that setting Zo in this case, where the rvc has an end-to-end resistance of 50K, equals the parallel combination of 25K with (25K + 50 ohms), or about 12.5K.

The ratio of your amp's input impedance (100K) to the 12.5K worst case output impedance that is driving it is 8:1. That is slightly short of the so-called 10x rule that you are no doubt familiar with, but considering how close it comes to satisfying that rule, and that the ratio will be better at any other volume control setting, that all seems comfortable.

The other criterion that Zo should satisfy is that it should be considerably lower than the capacitive reactance (which is an impedance, measured in ohms) of the cable that connects the rvc to the amp, at the worst case frequency within the audio band, which is 20kHz.

That number can be calculated by taking the capacitance per foot of the cable, multiplying by the number of feet, and plugging into the formula Xc (capacitive reactance) = 1/(2 x pi x f x C), where f is 20,000 and C is capacitance in farads. The resulting Xc will be in units of ohms.

For example, a 5 foot cable having a low capacitance of 20 pf (picofarads) per foot results in Xc = 1/(2 x 3.14 x 20000 x 5 x 20exp-12) = 79.6K, which would be a satisfactory result in relation to Zo = 12.5K.

Best regards,
-- Al

so 2 meters of Cardad Golden Reference 12pf/ft should be no problem for the passive approach, with with a 2 meter IC?

Yes, that is excellent, due to the very low cable capacitance per unit length. Xc calculates to 101K at 20 kHz for the 2 meter length, which I calculate will result in a roll-off of 0.066 db at 20kHz, in combination with the worst case source impedance of 12.5K. In other words, completely negligible.

Best regards,
-- Al

Yes, that strikes me as a good way to look at it, Herman.

I notice that at this page of the Goldpoint site the following statement is made, which I believe is incorrect and misleading:

When choosing the stepped attenuator value for an in-line level control or "passive preamp" (such as the Goldpoint Level Control Boxes), the attenuator value is chosen to match the input impedance of whatever it will be controlling. example: If the amplified monitor speakers or power amplifier you will connect the output of your your passive preamp to has an input impedance of 20K, then order a 20K stepped attenuator for that application.

Although they then qualify that with this statement:

You can usually use a level control value which is LESS than the rated input impedance of the gear it will be controlling.... 25K is usually a good choice for both vacuum tube and solid-state equipment

Best regards,
-- Al

Also, and I think that I have read this above, just want to make sure I understand, with a 10k pot the output impedance rises as the knob is rotated clockwise until it reaches 10k. Did I get this right?

No, Herman's relevant statement was:

In my example they were both 10K so it is 1/2 of the 10K. With the volume close to zero it rises to about 10K. and falls as the volume goes up.

In your case, with a 10K pot and a 100K amp Zin, at minimum volume control setting the input impedance (not output impedance) looking into the input side of the pot would be 10K. With the control turned all the way up, the impedance would be 10K in parallel with 100K, which is 9.1K.

The source component would see a load impedance equal to that value (dependent on the volume control setting) factored by square of the ratio of transformer primary turns to secondary turns (assuming the transformer is "ahead" of the volume control).

As I indicated a couple of posts ago, the OUTPUT impedance of the passive, which must be kept low in relation to the amp input impedance, will be at its worst case maximum when the volume control is set to the mid-point of its resistance range (assuming the source component's output impedance is small). Apart from the presence of the transformer, with a 10K pot and a 50 ohm source impedance, that output impedance would be 2.5K. The step-up transformer will raise that value a little, by stepping up the (very small) contribution of the 50 ohms. But for any reasonable turns ratio the resulting output impedance is likely to still be well under 5K, and therefore fine in relation to the amp's 100K Zin.

BTW, calculating the resistance of a parallel combination of two resistances is easy. It's just the product (multiplication) of the two numbers, divided by their sum.
If more than two resistances are in parallel, it's a little more difficult, the result being equal to the reciprocal of the sum of the reciprocals of the individual resistance numbers. (Reciprocal = the number divided into 1).

Best regards,
-- Al

Almarg: In your case, with a 10K pot and a 100K amp Zin, at minimum volume control setting the input impedance (not output impedance) looking into the input side of the pot would be 10K. With the control turned all the way up, the impedance would be 10K in parallel with 100K, which is 9.1K.

The source component would see a load impedance equal to that value (dependent on the volume control setting) factored by square of the ratio of transformer primary turns to secondary turns (assuming the transformer is "ahead" of the volume control).

To make sure it's clear, I should add that if you have a transformer ahead of the volume control which steps UP the incoming voltage from the source component, the source component will see a load impedance that is REDUCED from the 9.1K or 10K number, in proportion to the square of the turns ratio.

For example, if the transformer provides a 4x (12db) voltage step-up, the source component would see a load impedance of only 568 ohms to 625 ohms, depending on the volume control setting. Obviously that gets into worrisome territory.

Best regards,
-- Al

Is it safe to assume that the Goldpoint is shunt-type?

No, they are series types. From the Goldpoint website:

Goldpoint Level Controls used to offer Ladder and Shunt type stepped attenuators too. The fact is that with the very low noise, 0.1%, thin film resistors we are now using, we believe that there is no real advantage to Ladders and Shunts - and so have ceased offering them.

And a schematic of their Mini-V series attenuator:

http://www.goldpt.com/schm_ser.html

Best regards,
-- Al

Hi Clio,

I took a look at it. I see that several different transformer ratios are offered. 1:1 and 1:2 will certainly be no problem for low impedance sources. The other two ratios, 1:8 and 1:13, will, as I speculated earlier, result in an overall input impedance that is extremely low. 1:8 will divide the 10K attenuator impedance down to 156 ohms. 1:13 will divide it down to 59 ohms. Addition of a 100K amplifier load may reduce those numbers slightly further, depending on the volume control setting.

Whether or not that heavy a load will result in good sound is dependent, of course, on how well the source component can handle having to supply relatively high currents, and on how flat its output impedance vs. frequency curve is. But in general I don't think that those two ratios can be counted on to perform well.

Also, I don't understand the statement about output impedance remaining constant as a function of attenuation setting. If the attenuation is set for minimum volume, the output impedance looking back from the amplifier into the pva will be zero or very close to it. If the attenuator is set fully clockwise (max volume), the output impedance will be 10K in parallel with the source component's output impedance times the square of the transformer's step-up ratio. As I indicated in one of my earlier posts, if the attenuator is set to the middle of its resistance range, the output impedance will be around 2.5K (the worst case), assuming the source component's output impedance and the transformer's step-up ratio are small.

BTW, moving the transformer to the output side of the pot, as you may realize, would most likely not be helpful, because it would probably raise the pvc's output impedance to levels that would be too high relative to the amp's input impedance. Also, it would increase the range of amplitudes over which the transformer would have to operate, which MIGHT compromise its performance to some small degree (I'm not knowledgeable enough in that area to be able to say).

The bottom line, imo: Assuming (as I do) that parts quality is good, it looks like an excellent product at 1:1 or 1:2 transformer ratios, but performance at 1:8 or 1:13 will be highly dependent on the characteristics of the source component, and good performance at those ratios cannot be counted on.

Best regards,
-- Al

Herman -- As I understand it, it's not a TVC. It's a step-up transformer followed by a resistive attenuator.

Best regards,
-- Al

Cdc: Why does the I/C from the source to passive pre have to be as short as possible while from the passive pre to power amp can be longer?

Actually, the reverse is true. The high output impedance of a passive preamp will form an RC low pass filter in conjunction with the capacitance of the cable that is connected to its output. The longer that cable is, and the higher the capacitance per unit length of the cable, the more likely it is that significant upper treble rolloff will occur.

The relatively low input impedance of a passive preamp will produce a minor increase in the amount of current flowing through the cable between source and preamp, thereby perhaps increasing some subtle cable effects if the cable is long, but those effects will be minor in comparison to the effects of excessive capacitance in the cable connecting passive to power amp.

Clio09:Question: Jack tells me the transformers are wound for 100 ohm impedance, meaning the source must have an output impedance of 100 ohms or less. Any of this make sense?

Transformer impedance ratings, as distinguished from the reflected impedances corresponding to the connected source and load impedances and the turns ratio, have always been something of a mystery to me. Perhaps someone else will comment more knowledgeably. But I believe that the rated impedance of the transformer itself corresponds to how lossy the tranformer is in terms of flux leakage and dc resistance, and represents the amount of current limiting that would occur in the primary with the secondary short circuited (and a voltage source with near-zero output impedance connected to the primary). The higher the output impedance of the source component, the greater the losses that would result from that transformer impedance.

Best regards,
-- Al