[jimt]

I was reviewing the C20 schematic, RIAA feedback filter. With the selecter
set for RIAA, the ckt switches in 180K and 430pf for the high frequency -
approx 75us (77.4us). However, the low frequency comp for RIAA switches in
a 3.9M and 1800pf. This computes to 7020us. I thought the RIAA at this
point should be 3180us.

Here is the link to a pdf of the schematic.
http://www.pmillett.com/file_downloa...sh/C20_sch.pdf

Am I misunderstanding the circuit and or the RIAA compensation filter
specification? Thanks for any advice.

[John Byrns]

In article ,
"jimt" wrote:

I was reviewing the C20 schematic, RIAA feedback filter. With the selecter
set for RIAA, the ckt switches in 180K and 430pf for the high frequency -
approx 75us (77.4us). However, the low frequency comp for RIAA switches in
a 3.9M and 1800pf. This computes to 7020us. I thought the RIAA at this
point should be 3180us.

Here is the link to a pdf of the schematic.
http://www.pmillett.com/file_downloa...sh/C20_sch.pdf

Am I misunderstanding the circuit and or the RIAA compensation filter
specification? Thanks for any advice.

I believe that the longest time constant in the RIAA curve is heavily
dependent on the actual gain of the amplifier without feedback, as well
as on the values of the feedback network components. I believe that
with the right amplifier gain it is actually possible to achieve a time
constant of 3180 usec. using only a capacitor with no resistor. So my
guess as to why the resistor in the C20 is larger than you expected is
that it is simply being used to trim the final time constant to the
required 3180 usec. value, and that the amplifier gain is such that it
has an equal or slightly greater effect on the final time constant in
the circuit of the C20.


Regards,

John Byrns

--
Surf my web pages at, http://fmamradios.com/

[Chris Hornbeck]

On Fri, 15 Dec 2006 09:50:33 -0600, John Byrns
wrote:

I believe that the longest time constant in the RIAA curve is heavily
dependent on the actual gain of the amplifier without feedback, as well
as on the values of the feedback network components. I believe that
with the right amplifier gain it is actually possible to achieve a time
constant of 3180 usec. using only a capacitor with no resistor. So my
guess as to why the resistor in the C20 is larger than you expected is
that it is simply being used to trim the final time constant to the
required 3180 usec. value, and that the amplifier gain is such that it
has an equal or slightly greater effect on the final time constant in
the circuit of the C20.

This is all exactly right. And another way to think about it is
to just ignore the op-amp/ time constant model and to consider the
open loop amplifier.

Practical sound-quality oriented valve phono stages have very little
more gain than is required closed loop below the 50Hz pole. In the
extreme case of *no* excess gain available for feedback, the
resistor would be infinitely large.

Much thanks, as always,

Chris Hornbeck
"Too soon oldt; too late schmardt."

[tubegarden]

Chris Hornbeck wrote:
On Fri, 15 Dec 2006 09:50:33 -0600, John Byrns
wrote:

I believe that the longest time constant in the RIAA curve is heavily
dependent on the actual gain of the amplifier without feedback, as well
as on the values of the feedback network components. I believe that
with the right amplifier gain it is actually possible to achieve a time
constant of 3180 usec. using only a capacitor with no resistor. So my
guess as to why the resistor in the C20 is larger than you expected is
that it is simply being used to trim the final time constant to the
required 3180 usec. value, and that the amplifier gain is such that it
has an equal or slightly greater effect on the final time constant in
the circuit of the C20.

This is all exactly right. And another way to think about it is
to just ignore the op-amp/ time constant model and to consider the
open loop amplifier.

Practical sound-quality oriented valve phono stages have very little
more gain than is required closed loop below the 50Hz pole. In the
extreme case of *no* excess gain available for feedback, the
resistor would be infinitely large.

Much thanks, as always,

Chris Hornbeck
"Too soon oldt; too late schmardt."

Hi RATs!

This is an interesting thread. I do not pretend I understand it, but,
it is a pleasure to see you all sharing information

I got some old Dr. Bruce Edgar ehorns recently. I have much headroom,
even if I lack storage

Happy Ears!
Al

[jimt]

Thanks for all of the information and insight. I guess the best way to see
if it all sums up is to build a reverse RIAA ckt or purchase one of those
test disks that have a compensated sweep signal, input it to the C20, and
scope the output.

jimt

"jimt" wrote in message
...
I was reviewing the C20 schematic, RIAA feedback filter. With the selecter
set for RIAA, the ckt switches in 180K and 430pf for the high frequency -
approx 75us (77.4us). However, the low frequency comp for RIAA switches in
a 3.9M and 1800pf. This computes to 7020us. I thought the RIAA at this
point should be 3180us.

Here is the link to a pdf of the schematic.
http://www.pmillett.com/file_downloa...sh/C20_sch.pdf

Am I misunderstanding the circuit and or the RIAA compensation filter
specification? Thanks for any advice.

[Arny Krueger]

"jimt" wrote in message

Thanks for all of the information and insight. I guess
the best way to see if it all sums up is to build a
reverse RIAA ckt or purchase one of those test disks that
have a compensated sweep signal, input it to the C20, and
scope the output.

No. The best way is to use a good PC audio interface and appropriate
software. The software will produce a good high-resolution graph of the
frequency response of the UUT, which you should compare to one of the
readily-available curves of the standard RIAA curve.

[Arny Krueger]

"Bret Ludwig" wrote in message
ups.com
Arny Krueger wrote:
"jimt" wrote in message

Thanks for all of the information and insight. I guess
the best way to see if it all sums up is to build a
reverse RIAA ckt or purchase one of those test disks
that have a compensated sweep signal, input it to the
C20, and scope the output.

No. The best way is to use a good PC audio interface

Such as an AP System One or Two.

They are stand-alone devices, not primarily PC audio interfaces. Wonderful
boxes if you have the megabucks to obtain their use. But, overkill for
people who work with audio contstruction projects for the fun of it.

Not a sound card.

Thanks for demonstrating your ignorance and irrational fear and hatred of
computers for us yet again, Bret.

[Arny Krueger]

"Bret Ludwig" wrote in message
oups.com
Arny Krueger wrote:
"Bret Ludwig" wrote in message
ups.com
Arny Krueger wrote:
"jimt" wrote in message

Thanks for all of the information and insight. I
guess the best way to see if it all sums up is to
build a reverse RIAA ckt or purchase one of those
test disks that have a compensated sweep signal,
input it to the C20, and scope the output.

No. The best way is to use a good PC audio interface

Such as an AP System One or Two.

They are stand-alone devices, not primarily PC audio
interfaces. Wonderful boxes if you have the megabucks to
obtain their use. But, overkill for people who work with
audio contstruction projects for the fun of it.

Not a sound card.

Thanks for demonstrating your ignorance and irrational
fear and hatred of computers for us yet again, Bret.

I have no hatred of computers. i own more of them and of
more different kinds than you do, probably.

Hey Bret, do please brag about whatever it takes to make you feel adequate.

But my
original suggestion was the cheapest and simplest-a HP200
and a DMM accurate to above the audio band, of which
there are several, or a EVM-a VTVM or FET-VOM likewise.

Flawed on any number of counts.

To measure a RIAA curve adequately requires precision on the order of 0.1 dB
20-20 KHz.

If you can do it with an audio interface that you already have, and a copy
of RMAA 5.5, then your out-of-pocket cost is zero. OK, so you will need to
do a loopback measurement of the audio interface to get a handle on what to
subtract out of whatever you measure from the preamp.

Rotsa ruck if you think you're going to get that sort of precision out of a
HP200 and a VTVM with anything like a reasonable amount of effort.

[Chris Hornbeck]

On Mon, 18 Dec 2006 08:37:55 -0500, "Arny Krueger"
wrote:

"jimt" wrote in message

Thanks for all of the information and insight. I guess
the best way to see if it all sums up is to build a
reverse RIAA ckt or purchase one of those test disks that
have a compensated sweep signal, input it to the C20, and
scope the output.

No. The best way is to use a good PC audio interface and appropriate
software. The software will produce a good high-resolution graph of the
frequency response of the UUT, which you should compare to one of the
readily-available curves of the standard RIAA curve.

You'll mostly be measuring the hum in your test lash-up.
Signal levels are low and gain is high at low frequencies,
so PC interfaces are problematical in the real world.

Combining the PC interface with a reverse RIAA network
is a good start, but ground loop issues will still make
for some hair pulling.

All good fortune,

Chris Hornbeck
"Too soon oldt; too late schmardt."

[Arny Krueger]

"Chris Hornbeck" wrote in
message
On Mon, 18 Dec 2006 08:37:55 -0500, "Arny Krueger"
wrote:

"jimt" wrote in message


Thanks for all of the information and insight. I guess
the best way to see if it all sums up is to build a
reverse RIAA ckt or purchase one of those test disks
that have a compensated sweep signal, input it to the
C20, and scope the output.

No. The best way is to use a good PC audio interface and
appropriate software. The software will produce a good
high-resolution graph of the frequency response of the
UUT, which you should compare to one of the
readily-available curves of the standard RIAA curve.

You'll mostly be measuring the hum in your test lash-up.
Signal levels are low and gain is high at low frequencies,
so PC interfaces are problematical in the real world.

Contrary to your apparent beliefs, I've actually done what I recommend. It's
not all that problematical.

If you do these measurements with a FFT, were it necessary you could obtain
accurate measurements in the presence of considerable hum. Of course you'd
get responses at hum-related frequencies, but that is pretty easy to manage.

Combining the PC interface with a reverse RIAA network
is a good start, but ground loop issues will still make
for some hair pulling.

One significant problem with inverse RIAA networks is that they are only
accurate when driven with a given impedance.

[Chris Hornbeck]

On Tue, 19 Dec 2006 20:27:23 -0500, "Arny Krueger"
wrote:

You'll mostly be measuring the hum in your test lash-up.
Signal levels are low and gain is high at low frequencies,
so PC interfaces are problematical in the real world.

Contrary to your apparent beliefs, I've actually done what I recommend. It's
not all that problematical.

You lead a more charmed life than me. I find measuring
phono stages difficult even with a floating reverse RIAA
network and differential testing inputs.

Obviously, whatever works works, but IME it ain't a
walk in the park getting believable numbers. Interesting
stuff happens 60 and 80 dB down from a coupla milliJolts
signal. Coupling this to a computer's noisy grounding,
etc. is *not* trivial, for folks like me.

I can make "noiseless" transfers into and out of the
computer at line level, and am glad to get that. At
microvolt levels I run into trouble.


One significant problem with inverse RIAA networks is that they are only
accurate when driven with a given impedance.

Here we flip-flop on what we consider problematical. I
can easily make a reverse RIAA network largely indifferent
to source and (minor) variations in load impedance. The
40dB range and the light 47K ohm expected loading
conspire to ease the burden.

All good fortune,

Chris Hornbeck
"Too soon oldt; too late schmardt."