[Paul[_13_]]

Since the TR5, TR6, and TR8 are all pretty
similar, we can just look at the TR8 for this
discussion:


https://www.gearslutz.com/board/atta..._schematic.pdf

Ok first off, the TR5 is similar to the TR8 on the front end, with
the TL072 Opamp configured as an inverting, unity-gain buffer amp.

So it's normal to NOT see any signal on the input differential pins 2
and 3, right? I remember reading about the virtual ground on the input
pins of this topology.

Secondly, on all these models, the so-called "power" switch only
turns on the LED, and turns off the mute on the two power amps,
which means the rail voltages are ALWAYS ON the Opamp and the final
amp, no matter the state of the switch, as long as the 3 prong AC
cord is plugged into the unit. This means the leakage currents in
the ICs are always there, which would explain why the back plates
of these units stayed warm, even when I turned them "Off."

Is this what they typically do with these smaller reference speakers
to maybe save money for the bean counters? Why not put a switch
with more poles, so you can turn the rail voltages off too?

[Phil Allison[_4_]]

Paul wrote:



https://www.gearslutz.com/board/atta..._schematic.pdf

Ok first off, the TR5 is similar to the TR8 on the front end, with
the TL072 Opamp configured as an inverting, unity-gain buffer amp.


** Normally called a one op-amp, differential input stage.


So it's normal to NOT see any signal on the input differential pins 2
and 3, right?

** Depending how the inputs are driven ( one, the other or both) you may or may not see any signal on op-amp pins 2 and 3.


Secondly, on all these models, the so-called "power" switch only
turns on the LED, and turns off the mute on the two power amps,
which means the rail voltages are ALWAYS ON the Opamp and the final
amp, no matter the state of the switch, as long as the 3 prong AC
cord is plugged into the unit. This means the leakage currents in
the ICs are always there, which would explain why the back plates
of these units stayed warm, even when I turned them "Off."

Is this what they typically do with these smaller reference speakers
to maybe save money for the bean counters?


** It is commonly done like this for several reasons.


Why not put a switch
with more poles, so you can turn the rail voltages off too?


** That would require a much larger switch connected in the AC wiring and make a loud thump when operated - while the "mute" function on the LM3886 operates quietly.

Some designs even have a SMPS running continuously, generating lots more heat and often failing in a few years of use.

The Event TR series use conventional power supplies and simple circuits - so get my approval.



..... Phil

[Paul[_13_]]

On 10/14/2018 9:58 PM, Phil Allison wrote:
Paul wrote:



https://www.gearslutz.com/board/atta..._schematic.pdf

Ok first off, the TR5 is similar to the TR8 on the front end, with
the TL072 Opamp configured as an inverting, unity-gain buffer amp.


** Normally called a one op-amp, differential input stage.


So it's normal to NOT see any signal on the input differential pins 2
and 3, right?

** Depending how the inputs are driven ( one, the other or both) you may or may not see any signal on op-amp pins 2 and 3.

I was driving only pin 2, from the unbalanced RCA input. But
there didn't appear to be any signal on either pin 2 or 3.

My sig gen signal only appeared before R20, and on the output pin 1.

I recall learning about virtual ground in this topology.



Secondly, on all these models, the so-called "power" switch only
turns on the LED, and turns off the mute on the two power amps,
which means the rail voltages are ALWAYS ON the Opamp and the final
amp, no matter the state of the switch, as long as the 3 prong AC
cord is plugged into the unit. This means the leakage currents in
the ICs are always there, which would explain why the back plates
of these units stayed warm, even when I turned them "Off."

Is this what they typically do with these smaller reference speakers
to maybe save money for the bean counters?


** It is commonly done like this for several reasons.


Why not put a switch
with more poles, so you can turn the rail voltages off too?


** That would require a much larger switch connected in the AC wiring and make a loud thump when operated - while the "mute" function on the LM3886 operates quietly.

Some designs even have a SMPS running continuously, generating lots more heat and often failing in a few years of use.

The Event TR series use conventional power supplies and simple circuits - so get my approval.


Ok, that seems to be a decent reason, and I assume the leakage
current is low enough, that it doesn't matter too much for the
electric bill. But I still might plug both AC cords into a switchable
extension cord, so I can turn on both at the same time, and avoid
the wasted energy.

I know, I'm anal about efficiency! Probably due to working
in the cell phone industry!

[Phil Allison[_4_]]

Paul wrote:



So it's normal to NOT see any signal on the input differential pins 2
and 3, right?

** Depending how the inputs are driven ( one, the other or both) you may or may not see any signal on op-amp pins 2 and 3.

I was driving only pin 2, from the unbalanced RCA input. But
there didn't appear to be any signal on either pin 2 or 3.


** Pins 2 and 3 stay near identical in normal operation.

If you had driven XLR pin 3, op-amp pins 2 &3 would both have half the signal voltage on them.


My sig gen signal only appeared before R20, and on the output pin 1.

I recall learning about virtual ground in this topology.


** Yes, that is what you had.



..... Phil

[Scott Dorsey]

In article , Paul wrote:
https://www.gearslutz.com/board/atta..._schematic.pdf

Ok first off, the TR5 is similar to the TR8 on the front end, with
the TL072 Opamp configured as an inverting, unity-gain buffer amp.

So it's normal to NOT see any signal on the input differential pins 2
and 3, right? I remember reading about the virtual ground on the input
pins of this topology.

If it's being driven by a transformer, you will see signal between them but
not necessarily signal between one of them and ground. Because the signal is
floating and has no ground reference.

If it's being driven by a modern differential output, you will see signal
on them because both sides are being driven with respect to ground.

If it's being driven by an unbalanced output or an impedance balanced output,
you'll see signal on only one of them.

Secondly, on all these models, the so-called "power" switch only
turns on the LED, and turns off the mute on the two power amps,
which means the rail voltages are ALWAYS ON the Opamp and the final
amp, no matter the state of the switch, as long as the 3 prong AC
cord is plugged into the unit. This means the leakage currents in
the ICs are always there, which would explain why the back plates
of these units stayed warm, even when I turned them "Off."

Is this what they typically do with these smaller reference speakers
to maybe save money for the bean counters? Why not put a switch
with more poles, so you can turn the rail voltages off too?

It saves money, and it also means "instant on" operation so that you just
touch the switch and there is the signal. This allows people to be able to
use the power switch as a mute button too.

In a perfect world you'd have a mute button like that AND a big switch on
the primary of the power transformer. You might also have some filtering
and input protection on the primary side of the transformer too. But you
will notice that there is a price difference between the Event and the Tannoy.
It has to come from somewhere, and I'd rather them cheap out on stuff like
this than on stuff that will affect the sound.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Scott Dorsey]

My sig gen signal only appeared before R20, and on the output pin 1.

I recall learning about virtual ground in this topology.

** Yes, that is what you had.

Think about it this way. R20 and R24 are identical, and both are coming
from sources that have effectively infinite current available.

Pin 1 is reverse polarity from the input, so pin 2 is part of a summing
junction with equal parts of the input and inverted input summed together.
So in normal operation there's nothing there.

On the other hand, pin 3 has the inverted input coming in from the source,
so you should see signal there. R21 and R25 just form a divider to halve
the voltage.

It's not really a very good differential input topology, because you wind
up having to use fairly high value resistors in order to get the same load
impedance on both sides (which you want for noise rejection). But it works
well enough for the job and it only takes one op-amp. Line level is high
enough that noise isn't a concern.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Paul[_13_]]

On 10/15/2018 7:03 AM, Scott Dorsey wrote:
My sig gen signal only appeared before R20, and on the output pin 1.

I recall learning about virtual ground in this topology.

** Yes, that is what you had.

Think about it this way. R20 and R24 are identical, and both are coming
from sources that have effectively infinite current available.

Pin 1 is reverse polarity from the input, so pin 2 is part of a summing
junction with equal parts of the input and inverted input summed together.
So in normal operation there's nothing there.


Correct. There is a "virtual ground" on pin 2.

On the other hand, pin 3 has the inverted input coming in from the source,
so you should see signal there. R21 and R25 just form a divider to halve
the voltage.

Pin 3 had no signal either, but that could have been because
I was only driving it unbalanced, from pin 2.

It's not really a very good differential input topology, because you wind
up having to use fairly high value resistors in order to get the same load
impedance on both sides (which you want for noise rejection). But it works
well enough for the job and it only takes one op-amp. Line level is high
enough that noise isn't a concern.
--scott


Here's a good video on "virtual ground":

https://www.youtube.com/watch?v=0Maaa6gcXpw

The subsonic filter looks to be a unity-gain, non-inverting
topology. The series cap C6 is obviously a high-pass filter, but
how are the values for R10 and R11 chosen? Where is the "RC time
constant" that determines when the low frequencies start rolling off?

I assume it's usually -3dB at around maybe 20Hz? Or something
like that?

[Phil Allison[_4_]]

Paul wrote:


The subsonic filter looks to be a unity-gain, non-inverting
topology. The series cap C6 is obviously a high-pass filter, but
how are the values for R10 and R11 chosen? Where is the "RC time
constant" that determines when the low frequencies start rolling off?



** The active filters used are all "Sallen-Key" types.

http://sim.okawa-denshi.jp/en/OPstool.php

Similar S-K filter stages are connected around the two LM3886 power amps - after scaling each output (eg R7 & R8) to give unity gain. This doubles the roll off slopes to 24db/octave.


..... Phil

[[email protected]]

On Friday, October 19, 2018 at 11:09:25 PM UTC-4, Paul wrote:
On 10/15/2018 7:03 AM, Scott Dorsey wrote:
My sig gen signal only appeared before R20, and on the output pin 1.

I recall learning about virtual ground in this topology.

** Yes, that is what you had.

Think about it this way. R20 and R24 are identical, and both are coming
from sources that have effectively infinite current available.

Pin 1 is reverse polarity from the input, so pin 2 is part of a summing
junction with equal parts of the input and inverted input summed together.
So in normal operation there's nothing there.


Correct. There is a "virtual ground" on pin 2.



Maybe confusion re pins 2/3 on the XLR vs pins 2/3 on the IC U2A

The differential input signal appears across Pin 2/3 of the XLR.

On Pins 2/3 of IC2UA there will be two "virtually" identical signals ie. virtually no voltage across pins 2 and 3 of the IC. Each will be 1/2 of the input measured to ground.

The differential input signal will be on the ICU2A pin 1 output measured relative to ground.

m

[Paul[_13_]]

On 10/21/2018 12:17 AM, Phil Allison wrote:
Paul wrote:


The subsonic filter looks to be a unity-gain, non-inverting
topology. The series cap C6 is obviously a high-pass filter, but
how are the values for R10 and R11 chosen? Where is the "RC time
constant" that determines when the low frequencies start rolling off?



** The active filters used are all "Sallen-Key" types.

http://sim.okawa-denshi.jp/en/OPstool.php

Similar S-K filter stages are connected around the two LM3886 power amps - after scaling each output (eg R7 & R8) to give unity gain. This doubles the roll off slopes to 24db/octave.


Oh, that's perfectly it, thank you!

I knew there was a reason I ask you folks!

The sub-sonic is a high-pass filter, and after
I plugged in the TR5 values (they are different from the TR8 values,
because the TR5 omits the active crossover U3A and U3B),
the cut-off was somewhere around 55 Hz.

http://sim.okawa-denshi.jp/en/OPseikiHikeisan.htm

And here is a good review of how to derive the transfer
functions of these filters (I was only able to find a video
for the low pass topology):


https://www.coursera.org/lecture/ele...function-rX3Ys

If the roll-off slope for the TR8 is -24dB/octave, and there is
-6dB/octave per element, then it looks like this can be considered a 4th
order filter, right? Where did you get the -24dB/octave?

[Phil Allison[_4_]]

Paul wrote:

Phil Allison wrote:


Similar S-K filter stages are connected around the two LM3886 power amps - after scaling each output (eg R7 & R8) to give unity gain. This doubles the roll off slopes to 24db/octave.



If the roll-off slope for the TR8 is -24dB/octave, and there is
-6dB/octave per element, then it looks like this can be considered a 4th
order filter, right? Where did you get the -24dB/octave?



** Look at the TR8 schem carefully - the LM3886s have additional feedback paths around them aside from the usual gain setting networks.

U1 has a filter path with the same component values used for HF crossover U3A while U4 has one the same as LF crossover U3B.

As the filters are in series, the ultimate roll of slope is double that of each - or 24dB / oct.


..... Phil

[Paul[_13_]]

On 10/29/2018 12:56 AM, Phil Allison wrote:
Paul wrote:

Phil Allison wrote:


Similar S-K filter stages are connected around the two LM3886 power amps - after scaling each output (eg R7 & R8) to give unity gain. This doubles the roll off slopes to 24db/octave.



If the roll-off slope for the TR8 is -24dB/octave, and there is
-6dB/octave per element, then it looks like this can be considered a 4th
order filter, right? Where did you get the -24dB/octave?



** Look at the TR8 schem carefully - the LM3886s have additional feedback paths around them aside from the usual gain setting networks.

U1 has a filter path with the same component values used for HF crossover U3A while U4 has one the same as LF crossover U3B.

As the filters are in series, the ultimate roll of slope is double that of each - or 24dB / oct.


I see. I checked the roll-off of the HF crossover U3A,
and indeed it is about -12dB/oct, or -40dB/decade. The cutoff
frequency is about 2,729 Hz, which matches roughly what I heard
on the bench.

So crossover U3A is a 2nd order filter.

Can you safely assume the order of this type of active filter will
be the same as the number of reactive elements (capacitors, or
inductors), which in this case is two caps?

Just like you can simply count reactive elements in a totally
passive filter, to get the order of the filter?

[Phil Allison[_4_]]

Paul wrote:

Phil Allison wrote:


** Look at the TR8 schem carefully - the LM3886s have additional feedback paths around them aside from the usual gain setting networks.

U1 has a filter path with the same component values used for HF crossover U3A while U4 has one the same as LF crossover U3B.

As the filters are in series, the ultimate roll of slope is double that of each - or 24dB / oct.


I see. I checked the roll-off of the HF crossover U3A,
and indeed it is about -12dB/oct, or -40dB/decade. The cutoff
frequency is about 2,729 Hz, which matches roughly what I heard
on the bench.

So crossover U3A is a 2nd order filter.

Can you safely assume the order of this type of active filter will
be the same as the number of reactive elements (capacitors, or
inductors), which in this case is two caps?

Just like you can simply count reactive elements in a totally
passive filter, to get the order of the filter?


** Normally yes.

If the changing impedance of a cap with frequency is reflected in the stage's output, then it is likely adding 6dB / oct to the final slope.

18db/oct active stages normally have 3 caps.



..... Phil

[Paul[_13_]]

On 10/29/2018 11:19 PM, Phil Allison wrote:
Paul wrote:

Phil Allison wrote:


** Look at the TR8 schem carefully - the LM3886s have additional feedback paths around them aside from the usual gain setting networks.

U1 has a filter path with the same component values used for HF crossover U3A while U4 has one the same as LF crossover U3B.

As the filters are in series, the ultimate roll of slope is double that of each - or 24dB / oct.


I see. I checked the roll-off of the HF crossover U3A,
and indeed it is about -12dB/oct, or -40dB/decade. The cutoff
frequency is about 2,729 Hz, which matches roughly what I heard
on the bench.

So crossover U3A is a 2nd order filter.

Can you safely assume the order of this type of active filter will
be the same as the number of reactive elements (capacitors, or
inductors), which in this case is two caps?

Just like you can simply count reactive elements in a totally
passive filter, to get the order of the filter?


** Normally yes.

If the changing impedance of a cap with frequency is reflected in the stage's output, then it is likely adding 6dB / oct to the final slope.

18db/oct active stages normally have 3 caps.


Ok, excellent. There is order in the universe!

Thanks for sharing your knowledge, Phil, it's been a great help!

[TheIronChicken]

Hi Everyone, sorry to wake this thread up, but I recently acquired a set of TR5's and it seems that one of them has probably a bad input sensitivity pot - very low volume and it feels like the wiper isn't making contact whatsoever which seems to be pretty common. On the working one, I can feel the drag of the wiper, but this one feels open. I'd like to replace them both, but having a hell of a time sourcing a drop-in replacement. They're 2k it looks like, but small, rectangular and it looks like the long shaft to clear the back panel is integrated. I can't find anything similar anywhere and don't really want to mod anything at all if I don't have to, so just wondering if anyone knows off the top of their heads where to get them, or at least a specific PN (it looks like it has "DI YA B2K" and 0406 on it.) I've seen where people have replaced them out there, but of course no mention of what they replaced them with (Grrr.)

Thanks, and again sorry to jump on this, but you guys sound like you may be my last and only hope... =]

Terry
(Some guy in Iowa)

[TheIronChicken]

Originally Posted by TheIronChicken

Hi Everyone, sorry to wake this thread up, but I recently acquired a set of TR5's and it seems that one of them has probably a bad input sensitivity pot - very low volume and it feels like the wiper isn't making contact whatsoever which seems to be pretty common. On the working one, I can feel the drag of the wiper, but this one feels open. I'd like to replace them both, but having a hell of a time sourcing a drop-in replacement. They're 2k it looks like, but small, rectangular and it looks like the long shaft to clear the back panel is integrated. I can't find anything similar anywhere and don't really want to mod anything at all if I don't have to, so just wondering if anyone knows off the top of their heads where to get them, or at least a specific PN (it looks like it has "DI YA B2K" and 0406 on it.) I've seen where people have replaced them out there, but of course no mention of what they replaced them with (Grrr.)

Thanks, and again sorry to jump on this, but you guys sound like you may be my last and only hope... =]

Terry
(Some guy in Iowa)

Never mind I think - I'm going to try these: https://www.ebay.com/itm/1PC-2090F-2...53.m2749.l2648
and cut the old plastic shaft off and JB-Weld it to the new one or something.