Raising input impedance?

[HighFidelityGuy]

Hi,

Through trying out the passive output mod on my Caiman DAC I've come to realise that having a preamp with a high input impedance can be beneficial for a number of reasons, not least because it means you can buy lower value coupling caps for the DAC output which are cheaper.

I currently use a passive attenuator instead of a preamp and this has an input impedance of 10K ohms because it uses a 10K Alps Blue Velvet potentiometer. From what I've read it seems like the most common solution to this is to use a buffer before the attenuator to restore the proper high input, low output impedance balance.

I'm thinking about trying a buffer stage but before I do I wanted to understand why passive attenuators tend to have a lower value pot thus giving them a lower input impedance? For example what effect could I expect from swapping the 10K pot for a 100K version? Would it simply reduce the maximum output level of the attenuator?

Thanks. :)

[Reid Malenfant]

I'm thinking about trying a buffer stage but before I do I wanted to understand why passive attenuators tend to have a lower value pot thus giving them a lower input impedance? For example what effect could I expect from swapping the 10K pot for a 100K version? Would it simply reduce the maximum output level of the attenuator?
Frankly i'm not too sure why they tend to have lower values, though i can think of a couple of reasons why they might have.

1. If the passive attenuator uses a higher value resistor the way the volume control works will be a little odd. The reason being that the load that the attenuator sees will have a much greater influence on the volume setting & the attenuator will no longer follow a logarithmic law. This really isn't too important but it'd definately effect the way it works. Don't forget that whatever you are driving the attenuator with (in your case a beresford) will see the attenuator & the power amp as the load. If the attenuator has a lower resistance/impedance than the power amp input impedance it'll perform more like a conventional volume control.

2. Lower noise. All resistors generate there own noise when a current is present, this is known as Johnson noise & there is no getting away from it. The lower the resistance the less noise will be generated.

[HighFidelityGuy]

Thanks Mark.
Could you explain a bit more about how the source sees the impedance of the attenuator and the power amp? For example my power amps have an input impedance of 50K, so does that mean that my Caiman is actually driving a 60K load, or does it not work like that?

Thanks. :)

[Reid Malenfant]

No, it doesn't work out like that at all ;)

Your Caiman will see exactly the impedance of the attenuator when the volume is fully shut off, IE zero (10K ohm). The power amp is connected to the wiper of the attenuator so it'll be grounded when the volume is minimum & the Caiman sees 10K ohms. As you increase the volume the wiper of the attenuator is placed somewhere between the input & ground so this will load the Caiman a tad more. The higher the volume level the more it'll be loaded.

The maximum load the Caiman will see is when the volume is maximum & this would be 8k333 or 8333 ohms.

Effectively as you increase the volume your low frequency roll off will rise, but not by much ;)

[DSJR]

I wish I was good enough at maths to work these things out properly.

Just to remind folks that I'm having superb results from my upgraded MF X10-D line buffer, which offers nigh on 500K input impedance and an output impedance of, I think, 500 Ohms or so (without checking the spec sheet). With the uprated caps, 12VAC supply and recommended valves I can now say that this device is transparent to the signal as near as damn it and sounbds great into my headphone system...

[Reid Malenfant]

If you want the maths it's not too difficult, though showing it on here is a little more so :eyebrows:

To calculate two resistors of different values in parallel you need to do this:-

1/ 1/r1 + 1/r2

Or in other words in this case 1/r1 = 1/10000 = 0.0001. Then 1/r2 =1/50000 = 0.00002 & add them together to get 0.00012....

Then 1/0.00012 = 8333 :)

[HighFidelityGuy]

No, it doesn't work out like that at all ;)

Your Caiman will see exactly the impedance of the attenuator when the volume is fully shut off, IE zero (10K ohm). The power amp is connected to the wiper of the attenuator so it'll be grounded when the volume is minimum & the Caiman sees 10K ohms. As you increase the volume the wiper of the attenuator is placed somewhere between the input & ground so this will load the Caiman a tad more. The higher the volume level the more it'll be loaded.

The maximum load the Caiman will see is when the volume is maximum & this would be 8k333 or 8333 ohms.

Effectively as you increase the volume your low frequency roll off will rise, but not by much ;)

Balls, I knew it couldn't be that simple. :doh::lol:

So not only do I end up with a low input impedance it also changes depending on the position of the pot. So using a buffer stage before the attenuator will provide a constant load and therefore more consistency. I'm now starting to understand why a buffer stage is recommended, either that or an active preamp.

I've seen another preamp that I'd like to try but that's £1600, so I can get a buffer to use with my existing attenuator for a lot less than that, so I think I'll try that for now. THIS (http://www.autocostruire.com/catalog/product_info.php?products_id=304) is the buffer I fancy trying, I don't know if anyone here has tried it. I guess there's also the Pass B1 which I could probably make a clone of.

Thanks.

[Reid Malenfant]

Hi Dave, that looks very nice :) You could also use the Gator that is being developed for the Caiman ;) Take it from me that's all it is - a buffer, but it's a single ended class A buffer made with discrete components (apart from an op amp that'll likely be running very low gain & thus low feedback from the circuit overall) :)

The valve stage may look nice, but it'll use vast amounts more current (electricity) to do the same trick. The heaters are going to use some & that's just to make the tube function :doh:

It's up to you fella, but knowing the longevity of tubes i know where i'd put my money & it's cheaper :eyebrows:

I'm just mentioning this before you blow your money, take your time & think about it all ;) I can't go into the complexities of tube circuits as i don't know them, one thing i will bet though is that i doubt the tube stage has an active load (unless it's a constant current transistor) which i know the Gator will have. I have designed & built single ended class A power amps so i do happen to know a little about the subject ;)

[Jonboy]

I have just built this passive pre amp and stuck it between my 2A3 power amps and my Audionote pre because the power amps are so sensative you just need to breath on the volume control it blows you out of the room, the pot i have used is a 100k Alps black stepped pot bought from a certain forum owner some time ago, just set the volume control on the passive to about 3 oclock and the pre now has a lot more control over the volume with over driving the amps.
I have another passive fitted with a Alps blue but found it was a bit lean sounding and robbed some of the the bottom end, the black pot in this one is a lot better and has very little affect on the sound

It's all silver wired and built into a Maplins ali case


http://i419.photobucket.com/albums/pp272/jonboy_01/DSC_0241.jpg

[HighFidelityGuy]

Hi Dave, that looks very nice :) You could also use the Gator that is being developed for the Caiman ;) Take it from me that's all it is - a buffer, but it's a single ended class A buffer made with discrete components (apart from an op amp that'll likely be running very low gain & thus low feedback from the circuit overall) :)

The valve stage may look nice, but it'll use vast amounts more current (electricity) to do the same trick. The heaters are going to use some & that's just to make the tube function :doh:

It's up to you fella, but knowing the longevity of tubes i know where i'd put my money & it's cheaper :eyebrows:

I'm just mentioning this before you blow your money, take your time & think about it all ;) I can't go into the complexities of tube circuits as i don't know them, one thing i will bet though is that i doubt the tube stage has an active load (unless it's a constant current transistor) which i know the Gator will have. I have designed & built single ended class A power amps so i do happen to know a little about the subject ;)

I looked at the gator option but I don't need the headphone output or the extra few dB's of gain it provides, so I'd prefer to stick with the basic passive output.

I know what you mean about the valve issues. I'd prefer not to add the extra "issues" surrounding valves but at the same time I'm intrigued to see what difference adding a valve stage will make. I also liked the way that particular valve buffer allows you to adjust the bias to tune the sound.

On the solid state side of things there's the Pass B1 and the FETISHIZATOR, the latter supposedly having a valve sound. So I'll look into both of those as well.

[Reid Malenfant]

Hey Dave, i didn't know that the Gator was a gain stage, i thought it was just a buffer? I might be wrong, but i thought it was just a low feedback single ended class A buffer :confused:

[HighFidelityGuy]

Hey Dave, i didn't know that the Gator was a gain stage, i thought it was just a buffer? I might be wrong, but i thought it was just a low feedback single ended class A buffer :confused:

No, it definitely has gain. :) Here's a quote of how Stan described it:

"I wanted to find a solution for the low output of the passive mod and preserve the headphone output capability, but didn't want to go overboard with a complex discrete opamp type design. The gain required is only between 2X to 4X so a single ended class A design with just 10% negative feedback was all I felt that was necessary. No need to crack a nut with a sledge hammer."

I don't have any issues with low gain. I could actually do with a little less.

It looks like the FETishizator (http://www.lampizator.eu/LAMPIZATOR/FETISHIZATOR/JFET_CDout.pdf)needs to be tuned to the exact DAC chip being used and that requires more electronics knowledge than I have by the looks of it. I'd need someone more knowledgeable to figure out the component values. The Pass B1 (http://www.passdiy.com/pdf/B1%20Buffer%20Preamp.pdf) looks much simpler though. Saying that the B1 needs an 18V supply, so I'd have to use a separate one. The FETishizator only needs a 12V supply, so I could power that off the Caiman's internal 12V reg. Leo has tried the FETishizator and I've resurrected his thread on the subject to see if I can get some more info.

[DSJR]

Will you people never learn??? :D

Doesn't have valves in it so it can't possibly be any good, can it???

http://i132.photobucket.com/albums/q8/DSJR_photos/volume2_315x225.jpg

Around £220, so if the Gator doesn't appeal.......

[HighFidelityGuy]

I wish I was good enough at maths to work these things out properly.

Just to remind folks that I'm having superb results from my upgraded MF X10-D line buffer, which offers nigh on 500K input impedance and an output impedance of, I think, 500 Ohms or so (without checking the spec sheet). With the uprated caps, 12VAC supply and recommended valves I can now say that this device is transparent to the signal as near as damn it and sounbds great into my headphone system...

Thanks for the info.
Out of interest does your X10-D reduce the signal level at all between your source and amp? Cheers.

[Clive]

Thanks for the info.
Out of interest does your X10-D reduce the signal level at all between your source and amp? Cheers.
The recall the gain being 0.9, so yes it will reduce the signal level but only very marginally.

BTW, the Burson buffer is also very good.

[HighFidelityGuy]

The recall the gain being 0.9, so yes it will reduce the signal level but only very marginally.

BTW, the Burson buffer is also very good.

Thanks Clive, that's what I figured. So is a gain of 1 effectively no gain and is that what's refered to as unity gain? Cheers.

Thanks for the heads up un the Burson too, that's another to add to my list. :)

[Reid Malenfant]

So is a gain of 1 effectively no gain and is that what's refered to as unity gain?
Yes indeed :)

[HighFidelityGuy]

Yes indeed :)

Excellent, thanks. It's nice to know you're learning something. :cool:

[HighFidelityGuy]

I've just been reading some good info about why only low value pots like 10K are used in passive attenuators. I failed to realise up until now that because the pot is the only thing in the signal path that the amp sees the back end of the resistor as an output impedance. So using a higher value pot to get higher input impedance also gives you higher output impedance. Foe example a 100K pot has a maximum output impedance of 25K, which would require your amp to have an input impedance of 250K (following the 1:10 rule I also just read about). So when you add that issue to the issue of higher value resistors/pots causing higher levels of distortion (like Mark said) then then pretty well answers why low value pots must be used in passive attenuators.

So I'm now going to look into building a Pass B1 and maybe a Lightspeed attenuator to try out. :eyebrows:

Thanks to everyone that gave advice, I've learned a lot. :cool:

[Dave Cawley]

which would require your amp to have an input impedance of 250K (following the 1:10 rule I also just read about). So when you add that issue to the issue of higher value resistors/pots causing higher levels of distortion

Why would that be?

Dave

[kininigin]

hi dave you have pm