LME49810 PSpice Model available at TI

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The TINA ref design from the site works, once you have re-edited the shockingly badly drawn schematic to cure some nets that aren't properly connected. The ref design is very crude so I haven't yet checked how closely the model matches the real data sheets.
Are TI getting kids on 'work experience' schemes to do these ref designs?
 
BTW, the TINA SPICE model is also encrypted, so you can't back-engineer an LT-SPICE model from it.

If you get enough of the behavior you can back-engineer it.

I will bug TI about it - the same issue came up with respect to the model for the LM3886.

I pointed out an error on one of the TI datasheets -- an update for a product that's been in production for over 20 years -- they have yet to correct the error. Ergo, I am not optimistic that they are going to change the encryption of the models.
 
Having had a closer look at the TINA reference design, it's a load of rubbish! The compensation cap is set to 1uF! When I set it to 22pF, the FR doesn't change. also the output transistor quiescent current isn't changing when I alter the bias voltage. If I get some spare time tomorrow I will use the SPICE model to build up the data sheet ref circuit from scratch in TINA and see how it simulates. TI aren't doing too well with releasing valid SPICE models of the Nat-Semi audio chips so far!!
 
But of course there's nothing to stop you copying the LM49810 block from the ref design and pasting it into a new design.
Also with the free version of TINA, you can copy functional blocks that don't exist in that version from other TINA designs on the web onto a dummy schematic from which you can copy and paste them into your own design. The free version of TINA will simulate properly using them. It's just that those blocks aren't available on the menu bar in the free version.
 
Well, I tried using the model to build up a sim of what is basically figure1 in the Nat Semi data sheet for the 49810. Still no use at all. The SPICE sim amplifies, yes, but the 'biasP' and 'biasM' inputs sem to have no effect, neither does the value of the compensation cap. The only thing that seems to work is the operation of the MUTE pin!

Putting out SPICE models that are misleading, or in this case, of no use whatsoever is worse than not having the model in the first place.
 
www.hifisonix.com
Joined 2003
Paid Member
Hardly matters whether the simulator is LT Spice, TINA, or Pspice, if the model is wrong! Whether the SPICE engine is an industry-standard one (as in TINA or PSPICE) or a proprietory one (as in LT Spice), a wrong model will give answers that can't be trusted.

Yes you are right, but why put out a model, to run on ones own simulator that's clearly just a pile of nonsense. I just don't get it.

Separately, those TINA schematics . . . Oh no.
 
www.hifisonix.com
Joined 2003
Paid Member
MC10 rules. Period. ;)

Yes that too!

I remember using the version on a minidisc in about 91 or 92. You had to stick a special key disc in the drive to enable it to run, sort of like a dongle. I tried to model a sigma delta modulator. I eventually gave up, went with my gut, and got about 10 ppm on a 0-10 V input signal instrumentation, so DC to a few Hz only. Discrete of course.
 
It is reasonably popular but not really so much amongst technical discussions. Given that it is a high performing front end with likely many 'basic' problems (thermal compensations for e.g.) addressed would be a nice and quick tool to assess the merits of compensation and feedback methods as an example. But since not that much is known about what is inside and encrypted or not no detailed model is published, there is really not much to talk about.

I have built some discrete amps but maybe I don't have much DIY blood in me... the chip just seems to me to make a lot of sense to use ! I just want to get a feel for what I should expect from my configuration before I spend money on it!
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.