Simulating absolute value circuit in TINA-TI

I am looking to build an absolute value circuit using a pair of op-amps, hopefully using a single supply and would like to simulate my design using TINA-TI.

The ultimate aim is to measure the voltage difference between an input and a reference voltage that is not necessarily ground. The modified LT1001 circuit in the answer to this question would be okay, but it is dual supply.

TI provide a useful application bulletin with various designs for absolute value circuits. Figure 5 in the application bulletin offers a nice single supply variant. I have reproduced it here in case of a stale link later.

I have entered the design into TINA-TI - I think without errors, but the simulation is not working with the output swinging immediately to mid-supply.

Why aren't I seeing the waveform as shown in the application bulletin?

Also at the bottom of the clip (below Figure 5.1) it contains the sentence "Notice that the input range of the circuit is 4V below the power supply rail.". What does this mean - the diagram shows the input to be -4V to +4V and the output 0 to 4V for a 5V supply?

• Well, for one thing, your input voltage is substantially less than that used in the example design, where the input was an 8Vₚ₋ₚ sine wave. By 4V below the supply rail it means you can safely have an input of -4V; 4V below the negative supply rail (which is ground in this case). – Hearth Feb 2 '19 at 16:32
• I'm also inclined to ask, where did you get these specific values for the resistors from? I haven't bothered doing any math, but it's possible your problem lies there. – Hearth Feb 2 '19 at 16:35
• Thank you, that makes sense now (0V being a power supply rail in this case). The values come from a table at the back of the application bulletin. There are various options, I used the "Lowest Cost" one. Originally, I had an OPA2337 in the circuit but changed it as per the note under Figure 5.1 (both gave the same [wrong] output waveform). – iwbnwif Feb 2 '19 at 16:39
• I have updated the output image to show the 8Vp-p input. The output remains the same. – iwbnwif Feb 2 '19 at 16:41