ODD Spice! (LT1213 model error?)

[Lostgallifreyan]

Please can someone tell me if they can replicate this (in a default install
of LTspice)?

It's a differential amp set to gain of 0.25 (though unity or other shows same
effect). It drives the output hard to the supply rail and stays there! If you
replace the LT1213 model with the LT1215 (faster slew rate, twice supply
current but otherwise very similar IC), you get the expected standard
differential amplifier behaviour.

WHY? I've looked to see if there's anything in the data sheet to indicate
why, and there isn't, they really look like they should both behave very
similarly here.

I've looked at LT's site to see if there's a later (or fixed) model but it's
the same, and LT1214 is also the same as 1213 with same error.


Copy below, save as LT1213 Test.asc...

Version 4
SHEET 1 880 680
WIRE 192 128 176 128
WIRE 240 160 224 160
WIRE 96 176 -48 176
WIRE 176 176 176 128
WIRE 192 176 176 176
WIRE -112 192 -128 192
WIRE 272 192 272 128
WIRE 272 192 256 192
WIRE -128 208 -128 192
WIRE -48 208 -48 176
WIRE 96 208 32 208
WIRE 192 208 176 208
WIRE 176 224 176 208
WIRE 224 272 224 224
WIRE 256 272 224 272
WIRE 256 288 256 272
WIRE -128 304 -128 288
WIRE -48 304 -48 288
WIRE -48 304 -128 304
WIRE 32 304 32 288
WIRE 32 304 -48 304
WIRE 176 304 32 304
WIRE 224 304 224 272
WIRE 224 304 176 304
FLAG -112 192 12V
FLAG 240 160 12V
FLAG 256 288 0
SYMBOL voltage 32 192 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V1
SYMATTR Value 5V
SYMBOL voltage -48 192 R0
WINDOW 3 -2 -39 Left 0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V2
SYMATTR Value PULSE(0 5V 0 0 0 0.49ľ 1ľ)
SYMBOL res 176 144 R270
WINDOW 0 12 82 VTop 0
WINDOW 3 6 56 VBottom 0
SYMATTR InstName R1
SYMATTR Value 7.5K
SYMBOL res 80 192 R270
WINDOW 0 12 83 VTop 0
WINDOW 3 8 56 VBottom 0
SYMATTR InstName R2
SYMATTR Value 30K
SYMBOL res 80 224 R270
WINDOW 0 12 83 VTop 0
WINDOW 3 9 56 VBottom 0
SYMATTR InstName R3
SYMATTR Value 30K
SYMBOL res 160 208 R0
WINDOW 0 17 26 Left 0
WINDOW 3 19 79 Left 0
SYMATTR InstName R4
SYMATTR Value 7.5K
SYMBOL voltage -128 192 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V3
SYMATTR Value 12V
SYMBOL Opamps\\LT1213 224 128 R0
WINDOW 0 -19 64 Left 0
SYMATTR InstName U1
TEXT -162 130 Left 0 !.tran 1m startup

 

[Lostgallifreyan]

Ok, I found where the problem happens but I don't understand why, yet.

If you remove the 'startup' from the spice directive it works, so it looks
like it latches up if it gets a signal on it before a power supply.

Anyone know why though? I'm not sure if anything in the data sheets make it
clear why the LT1215 doesn't have this problem while the LT1213 does. Also,
does this difference indicate a weakness in the models, or a strength? It
looks like fine detail, but I don't know if it's correct. For one thing,
probing the non-inverting input in the unedited version I posted shows a very
unlikely 40 volts which doesn't seem right, model or no model!

I'll stick with the LT1215 anyway though, I want that fast slew rate.

 

[Helmut Sennewald]

"Lostgallifreyan" <no-one@nowhere.net> schrieb im Newsbeitrag
news:Xns9C5B34CEBC91Czoodlewurdle@216.196.109.145...

Ok, I found where the problem happens but I don't understand why, yet.

If you remove the 'startup' from the spice directive it works, so it looks
like it latches up if it gets a signal on it before a power supply.

Anyone know why though? I'm not sure if anything in the data sheets make
it
clear why the LT1215 doesn't have this problem while the LT1213 does.
Also,
does this difference indicate a weakness in the models, or a strength? It
looks like fine detail, but I don't know if it's correct. For one thing,
probing the non-inverting input in the unedited version I posted shows a
very
unlikely 40 volts which doesn't seem right, model or no model!

I'll stick with the LT1215 anyway though, I want that fast slew rate.

Hello,

This behavior isn't with the real device. It's only a convergence
issue if the "startup" option is used. You can either don't use
the "startup" with this circuit or you could defeat gmin-stepping.

..options gminsteps=0

This option will give you the correct result even when using "startup"
in the .TRAN-analysis.

Best regards,
Helmut

http://tech.groups.yahoo.com/group/LTspice/


The test circuit "test.asc" is below.
If the last line is splitted, please correct it with a text editor.

Version 4
SHEET 1 880 680
WIRE 48 96 -96 96
WIRE 80 96 48 96
WIRE 208 96 160 96
WIRE 272 96 208 96
WIRE 416 96 352 96
WIRE 352 160 320 160
WIRE -208 176 -224 176
WIRE 320 192 320 160
WIRE -224 208 -224 176
WIRE -96 208 -96 96
WIRE 208 208 208 96
WIRE 288 208 208 208
WIRE 416 224 416 96
WIRE 416 224 352 224
WIRE 464 224 416 224
WIRE 80 240 16 240
WIRE 176 240 160 240
WIRE 208 240 176 240
WIRE 288 240 208 240
WIRE 16 256 16 240
WIRE 208 256 208 240
WIRE -224 320 -224 288
WIRE -96 352 -96 288
WIRE 16 352 16 336
WIRE 16 352 -96 352
WIRE 208 352 208 336
WIRE 208 352 16 352
WIRE 320 352 320 256
WIRE 320 352 208 352
WIRE 320 384 320 352
FLAG -208 176 12V
FLAG 352 160 12V
FLAG 320 384 0
FLAG -224 320 0
FLAG 464 224 out
FLAG 48 96 in1
FLAG 176 240 in2
SYMBOL voltage 16 240 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
WINDOW 0 38 35 Left 0
WINDOW 3 34 86 Left 0
SYMATTR InstName V1
SYMATTR Value 5V
SYMBOL voltage -96 192 R0
WINDOW 3 -14 183 Left 0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR Value PULSE(0 5V 0 0 0 0.49ľ 1ľ)
SYMATTR InstName V2
SYMBOL res 256 112 R270
WINDOW 0 40 56 VTop 0
WINDOW 3 -3 53 VBottom 0
SYMATTR InstName R1
SYMATTR Value 7.5K
SYMBOL res 64 112 R270
WINDOW 0 38 53 VTop 0
WINDOW 3 -2 56 VBottom 0
SYMATTR InstName R2
SYMATTR Value 30K
SYMBOL res 64 256 R270
WINDOW 0 39 52 VTop 0
WINDOW 3 -6 52 VBottom 0
SYMATTR InstName R3
SYMATTR Value 30k
SYMBOL res 192 240 R0
WINDOW 0 46 30 Left 0
WINDOW 3 38 70 Left 0
SYMATTR InstName R4
SYMATTR Value 7.5K
SYMBOL voltage -224 192 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
WINDOW 0 34 38 Left 0
SYMATTR InstName V3
SYMATTR Value 12V
SYMBOL Opamps\\LT1213 320 160 R0
WINDOW 0 -19 64 Left 0
SYMATTR InstName U1
TEXT -224 48 Left 0 !.tran 1m startup
TEXT -224 16 Left 0 !.options gminsteps=0
TEXT -224 -16 Left 0 ;"gminsteps=0" is only required with the LT1213 if
"startup" is used

 

[Lostgallifreyan]

"Helmut Sennewald" <helmutsennewald@t-online.de> wrote in
news:h539e6$qck$00$1@news.t-online.com:

This behavior isn't with the real device. It's only a convergence
issue if the "startup" option is used. You can either don't use
the "startup" with this circuit or you could defeat gmin-stepping.

.options gminsteps=0

This option will give you the correct result even when using "startup"
in the .TRAN-analysis.

Best regards,
Helmut

Great answer, works perfectly. I noticed that extra spice directive on the
'test fixture' for the LT1213 but I didn't realise it was for something
beyond the filtering circuit behaviour for that circuit so I didn't yet try
it..

I found some cheap ones so I'll buy them anyway now I know they won't latch
up. Or if they do, it won't hurt much finding out.

http://tech.groups.yahoo.com/group/LTspice/

I just bit the bullet of having to sign up to Yahoo, just so I could join
that one...