4th time constant for LCR RIAA

the road not taken.
Post Reply
docali
Posts: 58
Joined: Thu Oct 16, 2008 7:16 am

4th time constant for LCR RIAA

Post by docali »

Hello,

does anyone know the correct mathematics how to calculate or incorporate the 4th time constant to an LCR? I did some simulations by trial and error (see attachement).

But there must be a way to calculate these values including the real series resistors of the inductors.

Best regards,
docali
Attachments
LCR600_4thTC.asc
LCR-600
(2.86 KiB) Downloaded 657 times
ClausNew
Posts: 4
Joined: Wed Mar 13, 2013 10:00 pm

Post by ClausNew »

I'm working on a white paper regarding mathematics background and an Excel tool for exact calculation of each component with free choice of the impedance (e.g., 600 Ohm), but this will take a couple of weeks more.
Please find attached a LTSpice simulation with imported values from my Excel tool. The tool finds the best values for goog accuracy and under consideration of the E-Series (e.g., E96 for R, E12 for C).
Maybe my example helps you so far or please wait for my white paper.
Attachments
RIAA_T6_Generic.asc
LTSpice LCR generic
(7.04 KiB) Downloaded 454 times
docali
Posts: 58
Joined: Thu Oct 16, 2008 7:16 am

Post by docali »

Thank you very much, Claus!

Where will you publish your white paper?

Best regards,
docali
ClausNew
Posts: 4
Joined: Wed Mar 13, 2013 10:00 pm

Passive RIAA filter design including tool

Post by ClausNew »

Sorry for answering so late. I published my white paper in another forum. In the attachment you can find a copy. I hope to get some response for discussion.
Attachments
RIAA_WP_1.0.zip
RIAA white paper, tool and simulation examples
(3.03 MiB) Downloaded 1882 times
docali
Posts: 58
Joined: Thu Oct 16, 2008 7:16 am

Post by docali »

Thank you very much, Claus, for sharing this with us!

It must have been an huge amount of work!

What I do not fully understand actually are the resitors RL1A and RL2A. I suppose that these have to soldered in series to the inductors as additional external resistor?

Is there any chance to change the calculation without these addditional resistors giving up some accuracy?

Best regards,
Bernd
jazbo8
Posts: 3
Joined: Fri Feb 01, 2013 1:01 am

Post by jazbo8 »

+1 on the great work done by Claus.

RL1 and RL2 are the estimated or measured series resistances of the inductors, so you do not need to add them.

Jaz
docali
Posts: 58
Joined: Thu Oct 16, 2008 7:16 am

Post by docali »

Hello Jazbo8,

i think you are not right. RL2 is the sum of RL2A.1||RL2A.2 and RL2.Cu.

The copper resistance of L2 is RL2.Cu.

Best regards,
docali
jazbo8
Posts: 3
Joined: Fri Feb 01, 2013 1:01 am

Post by jazbo8 »

Yes, you are right, I got it backwards :oops: , R.cu is the part that needs to be measured, and the R.a part added to form RL.

Jaz
ClausNew
Posts: 4
Joined: Wed Mar 13, 2013 10:00 pm

4th time constant for LCR RIAA

Post by ClausNew »

Hello docali,
I confirm that RL2 is the sum of RL2A.1||RL2A.2 and RL2.Cu. and RL1 is the sum of RL1A.1||RL1A.2 and RL1.Cu.
Under consideration of fixed time constants and fixed impedance (e.g., 600Ohm) there is a clear dependency between the inductances and their series resistances; means if you change one of them the other has to be changed. All the dependencies you can see in chapter 6 of my document where I summarized the equations. Therefore I cannot modify my formulas; they should be exact.
Nevertheless, with the help of my tool you can find a compromise if you modify the values of RLxA.y. First you input all necessary values and press “Calculate Components�. If you get all valid results, you can modify all yellow fields. In this case set RL1A.1 and RL2A.1 to “0� as well as RL1A.2 and RL2A.2 “blank�. You will see directly result in the failure plot. Of course, the failure will be higher and the impedance of the filter is not constant over frequency as possible.
On the other hand I see no great disadvantages using the series resistances compared to advantages. The filter can be tuned very precise, also in the inductances are higher as specified. Maybe you loose a little bit of gain, you get no additional access noise because no DC flows via the network. Of course, you have to invest in these resistors but if you are not using Z-foil resistors, the investment will be very low compared with the inductors.
Hopefully, I could help you
Claus
docali
Posts: 58
Joined: Thu Oct 16, 2008 7:16 am

Post by docali »

Hello Claus,

thanks for your explanation!

From my testings I found that you can get a similar error by leaving RL2A out and reducing C2 slightly and adjusting the other values around it.

But, and this seems to be the important point, which was not clear to before: you loose the contstant impedance of the network? If this is relevant in the real world I cannot assess.

Best regards,
docali
ClausNew
Posts: 4
Joined: Wed Mar 13, 2013 10:00 pm

Post by ClausNew »

docali wrote:Hello Claus,

thanks for your explanation!

From my testings I found that you can get a similar error by leaving RL2A out and reducing C2 slightly and adjusting the other values around it.

But, and this seems to be the important point, which was not clear to before: you loose the contstant impedance of the network? If this is relevant in the real world I cannot assess.

Best regards,
docali
I was ask some questions regarding the filter impedance of the network. I updated my tool. Now you are able to switch the plots: Filter gain (linearity) or Impedance (seen from the source into the filter with correct Ro). Please find the new version attached. I have not yet updated the article.

Claus
Attachments
CalcRIAAComp1.1.zip
(464.15 KiB) Downloaded 588 times
Post Reply