I'm experimenting with a DBX 234 crossover to set the xover frequencies before to assemble a fixed frequency crossover. I want to do the experiment with my usual listening chain. My current tube line stage doesn't like low impedance, and based on experience, I think it likes at least 50KΩ.
There was something strange when I first tried the DBX 234 almost bypassing the entire signal. The DBX manual says that "Input impedance is >25k ohm"... so I said let's put an active buffer stage between preamp and DBX. I borrowed a buffer from a friend and everything changed for the better. Now it sounds really good.
There is any simpler way to increase the input impedance "passively". What would you suggest?
There was something strange when I first tried the DBX 234 almost bypassing the entire signal. The DBX manual says that "Input impedance is >25k ohm"... so I said let's put an active buffer stage between preamp and DBX. I borrowed a buffer from a friend and everything changed for the better. Now it sounds really good.
There is any simpler way to increase the input impedance "passively". What would you suggest?
Only by attenuation. 100k:10k divider would do it at the cost of 21dB loss in level... Active buffer is the right approach.
Well, probably I should have put my question differently, I correct my wrong sentence
"There is any simpler way to increase the input impedance "passively"(*). What would you suggest?"
with
"There is any simpler way to modify the input buffer of these pro crossover to increase input impedance? What would you suggest?"
(*) my mistake was here, I wrote "passively" in quote marks to mean not-really "passively", sorry for that.
"There is any simpler way to increase the input impedance "passively"(*). What would you suggest?"
with
"There is any simpler way to modify the input buffer of these pro crossover to increase input impedance? What would you suggest?"
(*) my mistake was here, I wrote "passively" in quote marks to mean not-really "passively", sorry for that.
I think there is some confusion here. If a unit has 25k input impedace spec, it does NOT mean it needs to be fed with something that has a 25k output impedance.
If a unit has a low output impedance like 1k or 100R, it does NOT mean that it should feed somnething that has an equally low input R.
The usual routine is that the input R of a unit should be at least (preferably more) than 10x the output R of the source unit.
Jan
If a unit has a low output impedance like 1k or 100R, it does NOT mean that it should feed somnething that has an equally low input R.
The usual routine is that the input R of a unit should be at least (preferably more) than 10x the output R of the source unit.
Jan
Thanks Jan, I know it. In fact in my original post I wrote:
The line stage has an output impedance of about 3K (a), and according to the well known rule that you mention, 30K would be ok. What I noticed over the years is that 30K is ok but it really shines with loads >=50K say. The problem with these pro-type units is that they all work with Z around 20K (may be a standard) and I wonder how I can do some small modification to overcome that issue.
- - -
Footnote (a). I know this 3K may sound strange but that preamp was originally designed to feed tube amps with Z_in as high as 100K... one may say, why not buffer the preamp? I see this should be the right thing to do in a commercial product where one cannot predict how the preamp will be used. But I don't want to have an output buffer when I know that the next device (only less than 25cm of cable and connectors in the middle) is buffered.
The line stage has an output impedance of about 3K (a), and according to the well known rule that you mention, 30K would be ok. What I noticed over the years is that 30K is ok but it really shines with loads >=50K say. The problem with these pro-type units is that they all work with Z around 20K (may be a standard) and I wonder how I can do some small modification to overcome that issue.
- - -
Footnote (a). I know this 3K may sound strange but that preamp was originally designed to feed tube amps with Z_in as high as 100K... one may say, why not buffer the preamp? I see this should be the right thing to do in a commercial product where one cannot predict how the preamp will be used. But I don't want to have an output buffer when I know that the next device (only less than 25cm of cable and connectors in the middle) is buffered.
Put a series resistor in input of your preamp/DBX234. It will attenuate the signal but will increase the input impedance seen by tube preamplification.
There is no other passive method. Except if you modify the DBX 234. What you want to do, without lowering the level, requires amplification and cannot be done passively.
There is no other passive method. Except if you modify the DBX 234. What you want to do, without lowering the level, requires amplification and cannot be done passively.
OK, I see. Then a solution is to add 20k to the output of the preamp as noted, but you will lose almost half of the signal level.
If you can compensate by turning up the volume control, you're home free.
If you have the schematic diagram of the amp, there may be a simple change to increase it's input impedance. 30k sounds like a simple resistor that sets it, and replacing it with 50k may be the ticket.
Jan
If you can compensate by turning up the volume control, you're home free.
If you have the schematic diagram of the amp, there may be a simple change to increase it's input impedance. 30k sounds like a simple resistor that sets it, and replacing it with 50k may be the ticket.
Jan
The preamp may produce excess distortion with a high impedance source, series resistor doesn't lower the impedance the preamp sees, which is the aim here. Most stuff expects to be driven by 100 ohms or so, and most loads are 10k to 47k, so that the level loss is tiny and you don't have to worry about non-linear input current leading to distortion.
Even very high performance amps can be let down with a high source impedance.
Also adding series resistance will add Johnson noise anyway.
Even very high performance amps can be let down with a high source impedance.
Also adding series resistance will add Johnson noise anyway.
Yes, agreed. I personally would look at the amp input circuitry to see what sets Rin, see if it can be modified.
I also am sceptical that the sound of a 3k output preamp deteriorates when loaded by 30k, but I know little of typical tube circuits.
It doesn't sound like a competent design though.
Jan
I also am sceptical that the sound of a 3k output preamp deteriorates when loaded by 30k, but I know little of typical tube circuits.
It doesn't sound like a competent design though.
Jan
I should get the circuit of the DBX 234... which I can't find on the web. I contacted this guy here... he sells the schematic saying it is an official DBX document, but I have no guarantee that this is true... not sure it is worth it.
I suggest to switch to a proper preamp that doesn't have a goofy high output resistance. 🙂
That said, Jan is correct and a 3k output resistance shouldn't be bothered by a 30k input resistance. Assuming you don't have some super-long high-capacitance interconnect cables.
This sounds a meaningless statement to me:
"My current tube line stage doesn't like low impedance, and based on experience, I think it likes at least 50KΩ."
I have no clue what that means. I assume this is based on subjective evaluation.
Dave.
That said, Jan is correct and a 3k output resistance shouldn't be bothered by a 30k input resistance. Assuming you don't have some super-long high-capacitance interconnect cables.
This sounds a meaningless statement to me:
"My current tube line stage doesn't like low impedance, and based on experience, I think it likes at least 50KΩ."
I have no clue what that means. I assume this is based on subjective evaluation.
Dave.
Sorry, maybe I was not clear, let's try to do it better:
Of course I understand the technical considerations. Is this an "improper design", it may be, but it sounds fantastic to me provided that it is paired with the right amp. Is it a non transparent unit? Is it a colored sweetner? It may be, but I don't mind if the music sounds right to me.
- The crossover's Z is about 25K that's the problem. My current problem is to interface the preamp with the xover.
- I mentioned 30K as the input Z of an amp I used in the past, saying that 30K is ok for my preamp but it does much better with loads >=50K
Of course I understand the technical considerations. Is this an "improper design", it may be, but it sounds fantastic to me provided that it is paired with the right amp. Is it a non transparent unit? Is it a colored sweetner? It may be, but I don't mind if the music sounds right to me.
Confirmed by the designer?
Or did you actually modify it and listened to the various versions?
Jan
Yes, my preamp was designed without an output buffer intentionally. I think that even in the manual it is written that the preamp should be paired with amplifiers with sufficiently high input Z.
So intentionally designed wrongly or, to be more gentle, designed intentionally for the audiophile suckers looking for the right extra expensive golden cables.
@piecor You are talking to guys here who just look at the numbers and circuit design. We have no way to equate your subjective analysis to the situation.
I suggest to just add a 25k resistor in series. You will lose 6db but you'll have the 50k input resistance you desire.
How that will sound?.........I have no clue. 🙂
Dave.
I suggest to just add a 25k resistor in series. You will lose 6db but you'll have the 50k input resistance you desire.
How that will sound?.........I have no clue. 🙂
Dave.
I use simple cheap simple cables and I love the sound of my system. The most expensive cable that I own is a digital BJC that costed less than 40$.
I consider a transformer to be "passive"; others may disagree. If you install a transformer with a turns ratio of 2:1 at the input of your analog crossover, the input impedance will increase by 4 times. (the square of the turns ratio). The signal will also be attenuated by a factor of 2.
30k is high with respect to 3k. You normally look for at least a factor 10.
I gather that the 50k is just 'an idea' but has no basis in reality?
Jan