A nice(?), simple chip amplifier

[PJPro]

OK. I'll take that as a no then. I'd have thought that lots of people on the forum would be using TINA-TI and would have encountered (and overcome) this problem. Seems I am wrong.

 

[cobaltmute]

BTW, the LM317 is a TI part as well.

AFAIK, Tina-TI is a spice simulator - you should get be able to add the spice model to it.

 

[PJPro]

Yes, thanks cobaltmute. However, TI don't seem to have a model for it and TINA-TI seems to come with a very limited set of models.

I've tried plugging a LM317 spice macro behind a symbol for the TLE2425 (which is in TINA-TI). It will compile and I've linked up the pins with model script but I'm getting some crazy results which suggests all is not well.

TBH, I'm getting a little hacked off with TINA-TI. And LTSpice might be unlimited but the front end is completely counter intuitive (for my simple mind).

 

[PJPro]

Hmmm. I'm just gonna have to learn LTSpice.

 

[diditmyself]

Sorry I can't help you with TINA-TI. I think you'll get more support from LTSpice users, and I think you're going to need it. Like you said there's nothing intuitive about it, a real technocrat product.

 

[PJPro]

Quote:

Originally Posted by diditmyself /img/forum/go_quote.gif ......you'll get more support from LTSpice users, and I think you're going to need it.......

Hmmm. That sounds worrying.

However, I have managed my first limited success with LT Spice. I've placed a 3rd party model into the library of the LM317, created a circuit using the LM317 and ran a DC simulation to arrive at expected results i.e. the correct voltage coming out of the LM317 given the adjustment applied.

Just need to do the same with the OPA637, TLE rail splitter and BUF634

 

[diditmyself]

Like I said, you seem like a very fast learner. You have more patience with this stuff than I do. The more here that knows LTSpcice the better.

 

[PJPro]

No as fast as I'd like. I've got the same problem with LTSpice that I hit with TINA-TI. For some reason neither will accurately model the function of the rail splitter......or, perhaps more likely, I don't understand how the TLE works!

 

[00940]

With rail splitters, you have to be very careful about where you place the ground in the simulation.

A very good thread about LTspice: Things you should know about LTSpice - diyAudio

Btw, I've learned first with LTspice. I then tried Tina-TI but really prefers LTspice.

Edit: here's an amp with virtual ground in LTspice. The true ground is linked to the voltage source and all "grounds" to a trace named VG.

 

[jcx]

a problem with macromodels - Boyle op amp macromodel in particular is the use of Spice node 0 internally for voltage gain reference

sometimes this simplification is OK, other times it misses important performance characterisics

also no Spice allows floating nodes, everything has to have a dc path to node 0 - although it can be GigaOhms

 

[PJPro]

Thanks for your comments guys but it's a little more fundamental than that. The rail splitter isn't splitting the rails or arriving at 0V on the output.

I'm feeding it a single supply eg 0, 20VDC. I'm expecting +/-10VDC on the rails and 0VDC on the output but I'm getting 20VDC on the positive, 0VDC on the negative and 10VDC on the ouput. This is the same behaviour I witnessed on TINA-TI....which sort of suggests I'm doing something wrong, although I am using the same (TI) macro in both.

 

[cobaltmute]

I can't say for LTSpice or Tina-TI, but in Multisim, it depends on where you attach the "ground".

If you attach it to the negative end of the power supply, you set that as zero potential and the middle of the rail splitter is 10V

If you attach it to the middle of the rail splitter, you get a +- 10V

I might guess and could be wrong that the other SPICE programs work the same. Your rail-splitter is doing its job - it is just what point are you measuring 0 from.

 

[MisterX]

Quote:

I'm getting 20VDC on the positive, 0VDC on the negative and 10VDC on the ouput.

That is what you should be "getting".

Quote:

Originally Posted by PJPro /img/forum/go_quote.gif This is the same behaviour I witnessed on TINA-TI....which sort of suggests I'm doing something wrong, although I am using the same (TI) macro in both.

This is what you are doing wrong---->

Quote:

I'm expecting +/-10VDC on the rails and 0VDC on the output

 

[wap32]

The TLE is a "rail splitter" in the sense that it splits the rail-to-rail voltage in half, in your case 20V.

So you would have the split voltage, 10V, as the output, not 0.
By using that output as your virtual ground you will effectively have the equivalent of a +/-10V supply (20V-10V-0V), with the exception that the virtual ground won't be able to source/sink much current using the TLE alone.

Someone please correct me if I'm wrong.

 

[PJPro]

Quote:

Originally Posted by wap32 /img/forum/go_quote.gif The TLE is a "rail splitter" in the sense that it splits the rail-to-rail voltage in half, in your case 20V. So you would have the split voltage, 10V, as the output, not 0. By using that output as your virtual ground you will effectively have the equivalent of a +/-10V supply (20V-10V-0V), with the exception that the virtual ground won't be able to source/sink much current using the TLE alone. Someone please correct me if I'm wrong.

Well, as far as I can tell, this is the approach used by the Pimeta, PPA and M3. In fact I've just noticed that AMB states, with regard to the M3, that a dual supply is unecessary as the supply is split into two rails with the virtual ground being at one half the supplied voltage (in my case 10VDC).