Lundahl LL1545A LTSpice Simulation Issues

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Transformer LTSpice simulation issue.

I tried to LTSpice simulate a Lundahl LL1545A transformer to the AK4490 differential dac output, but I have FR simulation issue.
I have done an internet search about how is the right transformer LTSpice setup but the information are less.

For example, see for the Lundahl LL1545A the right LTSpice setup will be like 2 primaries coils and 4 secondaries coils with K L1 L2 L3 L4 L5 L6 0.999 spice directive.

I have done the following schematics.

shematics.jpg

If you done a real FR measurement with 20K load on the above setup by Spectrum Analyser or oscilloscope or rms multimeter, you will see that FR output is flat btw 20-15KHz, with a small increase voltage btw 17K-50K and then there is significant decrease until cut-off.

The best FR output that I have with LTSpice is the following, that is wrong.

FR.jpg

The LTSpice have a dialogue window for Coil setup like this.

LTSpice Coil Setup.jpg

But for the example, if you measure each primary coil by DER EE DE-5000 you have the following table.

LL1545A primaries.jpg

How is the right setup on LTSpice, can someone help?
I have included and the LTSpice .asc file
 

Attachments

  • AK lundahl Super VI Simulation.asc
    3.9 KB · Views: 68
A couple of issues you have - firstly the series resistance of the windings needs to come from Lundahl's own data (305R for primaries, 122R and 182R for secondaries). Sorting this out fixes up the frequency response very nicely (I zeroed your shunt resistance and capacitance though to get clear what's happening initially). With the correct series R I get -1dB at almost 400kHz.

Second the inductances of a transformer depend on the frequency (and to some extent the amplitude used to measure with). The turns ratio comes out of the ratio of inductances so its best to tweak the secondary inductance to be in line with what the turns ratio is.
 
@ abraxalito = From Lundahl's datasheet the 305R for primaries, 122R and 182R for secondaries, are described like static resistance and not series resistance.
Probably it is the same, but I confused with the Rs (Series Resistance) and Rp (Parallele Resistance) of De-ee measurings.

@ pieter t = like 5H on primaries inductance and 20H at the secondaries (in series)?

Abraxalito, can you attach me the file that you described like -1dB at 400K?
Here a new .asc file with the Lundhal's datasheet values.

Thanks for the help guys!
 

Attachments

  • AK lundahl Super VI Simulation, data from datasheet.asc
    3.8 KB · Views: 64
Oops, it is clear to me now!!
pieter t you had right, my sim is wrong, I have the R//C at the output.
Now, it is very close to the real measurement.

I have used the Series Inductance that I have measured with De-ee at 1KHz, 8.3H in primaries and 2H at primaries.
With the inductance ratio 1:4 (Prim. to Sec. ratio 1:2) which is the right?
 

Attachments

  • FR.jpg
    FR.jpg
    206.8 KB · Views: 249
Please confirm that you plan to use the transformer with 1:1 ratio; that's what I make out of your sim.
When you measure 8.3H in each primary, the total inductance of the combined primary (the 2+2) is 8.3H x 4 = 33.2H.
Your combined secondary (1+1+1+1) also has a total inductance of 33.2H as the number of windings of combined primary and secondary is the same (4 coils in series with 2H inductance each make 32H total inductance; the difference with 33.2H is some measurement inaccuracy).
In your sim I guess it is better to draw only one primary and secondary coil, with the same DCR and the same inductance.
 
Last edited:
Please confirm that you plan to use the transformer with 1:1 ratio; that's what I make out of your sim.
When you measure 8.3H in each primary, the total inductance of the combined primary (the 2+2) is 8.3H x 4 = 33.2H.
Your combined secondary (1+1+1+1) also has a total inductance of 33.2H as the number of windings of combined primary and secondary is the same (4 coils in series with 2H inductance each make 32H total inductance; the difference with 33.2H is some measurement inaccuracy).
In your sim I guess it is better to draw only one primary and secondary coil, with the same DCR and the same inductance.

Peter t, thnx for all, it is more clear to me now the right setup of this.

Here you go though looks like you've got it all sorted out now.

Ooops, I noticed I reduced the loading shunt cap at the output to 47pF while I was playing.

Thnx, I had make a similar sim yesterday with almost identical results.
The major problem has dissolve, it was a sim mistake with R//C at the secondary but our messages are very informative to me about how transformers work.
Thanks guys!
 
There's one nagging question I had about your sim - how did you come up with the coupling coefficient (0.999) ? Have you made a measurement of the leakage inductance? If not, its not a hard one to do for a reality check - the HF bandwidth will be determined by the leakage inductance.
 
No, I haven't done any measurement of the leakage inductance.
I have read from the internet that the 0.999 is the good practice for transformers like lundahl.
Can you explain it with more details, how the measurement can be done?

For me the Lundahl on AK4490 was a project. I would like to known how this "super v/i" implementation vs normal opamp stage or discrete.
From what I saw, the acoustic result is very good, the only problem is the output impedance, it is almost 1.2K and a buffer stage is must.

Additionally, I compared this "super v/i" stage with the simple direct method, like each AK differential output goes to each Lundahl winding at the primaries. From what I saw the super v/i implementation gives more low frequencies linearity than the direct method and at the LTSpice Simulation there is a little better phase linearity.
 
Last edited:
Can you explain it with more details, how the measurement can be done?

It isn't too difficult - the leakage inductance is the inductance of the primary when all secondaries are short-circuited. The test frequency should be mid-band (say 1kHz) and the amplitude kept as low as possible. The ratio of the leakage inductance to the primary inductance (measured with the same stimulus) gives a small fraction, call it T. Then your LTspice K value is set to (1-T).

Additionally, I compared this "super v/i" stage with the simple direct method, like each AK differential output goes to each Lundahl winding at the primaries.

In my experience with DACs which have opamp output stages the load impedance in the bass has a significant impact on the bass dynamics. Since trafos present a low impedance at LF I'd put a buffer between DAC and transformer.
 
Thanks for the info.
I make some measurements with the all primaries in series like the schema that I followed. The inductance measurement was from primaries in series e.g 10-16 with 9+15 short.
From what I saw it is very difficult to make a precisely measurement.
To short all secondaries right must be soldered, I had some deviation on my measurements with no secondary soldered. The best that I had was Primary Ind. = 30.18 and Leak.Ind=29.03 with my Der-ee at 1KHz and Ls inductance measurement.
It is 0.9619 ratio, with that the FR on LTSpice has a decrease from 1KHz, that it isn't true.

Only with K=0.999 the FR is almost like real FR measurement.
 
Why do you take the secondaries (9-10 and 15-16) as primaries, and the primaries as secondaries?
It makes a difference, and Lundahl is clear about this!
When you use the transformer 1:1 (but for some reason you have not been able to confirm that until now) the only thing to do is to short the two wires of your combined secondary for measuring leakage inductance, no need to "short all secondaries" as you should have created just one secondary....
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.