Reply by Tom Swift July 7, 20152015-07-07
Phil Hobbs <pcdhobbs@gmail.com> wrote:

>>Here are some examples where simulation can be very valuable. > > <snip> > > Sure thing, I didn't intend to imply otherwise. I do the math first, > though, whenever feasible. > > The things that are hard to do by hand are often also things that > SPICE models are bad at, e.g. overload recovery and CM slew > distortion. > > There are things where simulation is really the only cost-effective > method, e.g. checking the strictly-linear response of an active filter > made with non-ideal components, or (what I'm often interested in) > looking at the sensitivity of a high-frequency and/or high-Z circuit > to strays. > > I don't trust behavioural IC models at all. There are too many > noiseless op amps with zero input capacitance and perfectly-clean slew > limiting out there in spiceland. > > Transistor models are much better--I built a BFP640/pHEMT cascode TIA > a couple of months ago. Simulation correctly predicted the collector > current where it started oscillating, within about 10%, and got the > frequency more or ledd right too (about 6 GHz). You do have to put in > realistic strays, though. > > Cheers > > Phil Hobbs
Ok, thanks for the response. I just dropped in to offer some ideas on error-proofing LTspice. Your comments have been interesting, but I gotta get back to work. Lots of circuits to build and put into service. Thanks
Reply by John Larkin July 7, 20152015-07-07
On Tue, 07 Jul 2015 12:35:56 -0400, krw <krw@nowhere.com> wrote:

>On Tue, 07 Jul 2015 08:29:18 -0700, John Larkin ><jlarkin@highlandtechnology.com> wrote: > >>On Mon, 6 Jul 2015 22:29:40 -0700 (PDT), dakupoto@gmail.com wrote: >> >>>On Sunday, July 5, 2015 at 7:58:07 AM UTC-4, Bill Sloman wrote: >>>> On Saturday, July 4, 2015 at 12:23:27 PM UTC+2, Bill Sloman wrote: >>>> > I've been using the Analog Devices Spice model in LTSpice to model the AD734 running with a current output - see Figure 25 on page 13 of the Rev E AD734 data sheet >>>> > >>>> > http://www.analog.com/media/en/technical-documentation/data-sheets/AD734.pdf >>>> > >>>> > When running a roughly 15kHz sine wave through the device, the positive current output limits at something between +200uA and +280uA. >>>> > >>>> > The voltage at the W and Z1 outputs of the AD734 is well below the rail. >>>> > >>>> > I've tried 2.2k, 6.8k and 15k current setting resistors. Only with 15k did the voltage at the W and Z1 outputs get high enough to be interesting. >>>> > >>>> > The current clipped at +200uA with 2.2k, +282.77uA at 6.8k and +267.94uA at 15k. >>>> > >>>> > The negative-going excursions looked perfectly sinusoidal, and went down to -350uA. >>>> > >>>> > Working in another region of operation, with more head-room, the currents clamped a lot higher, at about +800uA, when the negative currents were getting down to -2.4mA. >>>> > >>>> > It looks very much as if there's some kind of silly mistake in the AD734 Spice model (which would interest Jim Thompson, who wants to sell Analog Devices better Spice models). >>>> > >>>> > If the actual device acted like the model, the data sheet wouldn't talk about +/-10mA output current limits (as it does on page 13). >>>> > >>>> > It's easy enough to hand-edit .cir files, if you kno0w what you are doing. Any advice will be gratefully received. I probably should have raised this with Analog Devices directly, but the price they charge for the AD734 means that they can't be selling many of them, which doesn't suggest that I'd get a prompt response. >>>> >>>> Oops. It looks as if the defect wasn't in the Analog Devices model, but in my circuit diagram - a connection to +15V seems to have been edited out at some point and the circuit was getting it's positive power supply from its inputs, making the model inconveniently realistic - I've had that happen on real circuits, and it can take a while to work out what's going wrong. >>>> >>>> My apologies to one and all. >>>> >>>> -- >>>> Bill Sloman, Sydney >>> >>>I have always felt that text input based SPICE versions(HSPICE, NGSPICE) are so much better >>>than GUI input based SPICE versions(LTSPICE) >>>The tiniest error in the text netlist input causes the simulation engine to crash, forcing >>>the oser to ensure that the initial input netlist >>>is correct. I believe that LTSPICE also allows >>>text file input. I have used the BSIM 4.6.5 >>>sub-micron device models in netlists that are >>>approximately 750 - 900 lines long, with both >>>HSPICE and NGSPICE, with only a few whimpers >>>of protest initially, but absolutely no issues >>>afterwards. >> >>Text entry of netlists is barbaric. Imagine a 1000-line netlist to sim >>some part of a product. Imagine needing to revisit it after a year or >>two of inactivity. What are you going to do to figure it out? Read it >>and try to draw the schematic! > >Pull out the D-sized schematic from your archives. Don't worry, it's >wrapped around the matching card deck.
All our schematics are scaled for B size, so our digital copier can print or scan them. I can open a 20 year old PADS schematic or layout, from the company server, instantly, with the latest PADS software. I don't think we've ever lost a schematic or layout or BOM in about 30 years of designing and selling stuff.
> >>Sure, Spice will catch you if you mistype "V10", but LT Spice won't >>misspell V10 when it saves a netlist. >> >>I can enter a presentation-quality LT Spice schematic in a fraction of >>the time I could type a netlist. I can reopen it a year later and see >>what's there in seconds. And it has a nice parts library, and HELP, >>right there in plain sight. > >I thought you were still into velum schematics.
I draw my original designs on vellum, decorated with design notes, equations, rev notes, whatever. I like to draw. My CAD folks enter them into PADS and we iterate an review from there. I tend to lose little scraps of paper, but I never lose a D-size vellum. Funny, but my newest engineering hire likes to draw, too. My layout people prefer to enter the schematics from hand-drawn input, so they can follow their own standards. It all works. -- John Larkin Highland Technology, Inc picosecond timing laser drivers and controllers jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Reply by krw July 7, 20152015-07-07
On Tue, 07 Jul 2015 08:29:18 -0700, John Larkin
<jlarkin@highlandtechnology.com> wrote:

>On Mon, 6 Jul 2015 22:29:40 -0700 (PDT), dakupoto@gmail.com wrote: > >>On Sunday, July 5, 2015 at 7:58:07 AM UTC-4, Bill Sloman wrote: >>> On Saturday, July 4, 2015 at 12:23:27 PM UTC+2, Bill Sloman wrote: >>> > I've been using the Analog Devices Spice model in LTSpice to model the AD734 running with a current output - see Figure 25 on page 13 of the Rev E AD734 data sheet >>> > >>> > http://www.analog.com/media/en/technical-documentation/data-sheets/AD734.pdf >>> > >>> > When running a roughly 15kHz sine wave through the device, the positive current output limits at something between +200uA and +280uA. >>> > >>> > The voltage at the W and Z1 outputs of the AD734 is well below the rail. >>> > >>> > I've tried 2.2k, 6.8k and 15k current setting resistors. Only with 15k did the voltage at the W and Z1 outputs get high enough to be interesting. >>> > >>> > The current clipped at +200uA with 2.2k, +282.77uA at 6.8k and +267.94uA at 15k. >>> > >>> > The negative-going excursions looked perfectly sinusoidal, and went down to -350uA. >>> > >>> > Working in another region of operation, with more head-room, the currents clamped a lot higher, at about +800uA, when the negative currents were getting down to -2.4mA. >>> > >>> > It looks very much as if there's some kind of silly mistake in the AD734 Spice model (which would interest Jim Thompson, who wants to sell Analog Devices better Spice models). >>> > >>> > If the actual device acted like the model, the data sheet wouldn't talk about +/-10mA output current limits (as it does on page 13). >>> > >>> > It's easy enough to hand-edit .cir files, if you kno0w what you are doing. Any advice will be gratefully received. I probably should have raised this with Analog Devices directly, but the price they charge for the AD734 means that they can't be selling many of them, which doesn't suggest that I'd get a prompt response. >>> >>> Oops. It looks as if the defect wasn't in the Analog Devices model, but in my circuit diagram - a connection to +15V seems to have been edited out at some point and the circuit was getting it's positive power supply from its inputs, making the model inconveniently realistic - I've had that happen on real circuits, and it can take a while to work out what's going wrong. >>> >>> My apologies to one and all. >>> >>> -- >>> Bill Sloman, Sydney >> >>I have always felt that text input based SPICE versions(HSPICE, NGSPICE) are so much better >>than GUI input based SPICE versions(LTSPICE) >>The tiniest error in the text netlist input causes the simulation engine to crash, forcing >>the oser to ensure that the initial input netlist >>is correct. I believe that LTSPICE also allows >>text file input. I have used the BSIM 4.6.5 >>sub-micron device models in netlists that are >>approximately 750 - 900 lines long, with both >>HSPICE and NGSPICE, with only a few whimpers >>of protest initially, but absolutely no issues >>afterwards. > >Text entry of netlists is barbaric. Imagine a 1000-line netlist to sim >some part of a product. Imagine needing to revisit it after a year or >two of inactivity. What are you going to do to figure it out? Read it >and try to draw the schematic!
Pull out the D-sized schematic from your archives. Don't worry, it's wrapped around the matching card deck.
>Sure, Spice will catch you if you mistype "V10", but LT Spice won't >misspell V10 when it saves a netlist. > >I can enter a presentation-quality LT Spice schematic in a fraction of >the time I could type a netlist. I can reopen it a year later and see >what's there in seconds. And it has a nice parts library, and HELP, >right there in plain sight.
I thought you were still into velum schematics.
Reply by John Larkin July 7, 20152015-07-07
On Mon, 6 Jul 2015 22:29:40 -0700 (PDT), dakupoto@gmail.com wrote:

>On Sunday, July 5, 2015 at 7:58:07 AM UTC-4, Bill Sloman wrote: >> On Saturday, July 4, 2015 at 12:23:27 PM UTC+2, Bill Sloman wrote: >> > I've been using the Analog Devices Spice model in LTSpice to model the AD734 running with a current output - see Figure 25 on page 13 of the Rev E AD734 data sheet >> > >> > http://www.analog.com/media/en/technical-documentation/data-sheets/AD734.pdf >> > >> > When running a roughly 15kHz sine wave through the device, the positive current output limits at something between +200uA and +280uA. >> > >> > The voltage at the W and Z1 outputs of the AD734 is well below the rail. >> > >> > I've tried 2.2k, 6.8k and 15k current setting resistors. Only with 15k did the voltage at the W and Z1 outputs get high enough to be interesting. >> > >> > The current clipped at +200uA with 2.2k, +282.77uA at 6.8k and +267.94uA at 15k. >> > >> > The negative-going excursions looked perfectly sinusoidal, and went down to -350uA. >> > >> > Working in another region of operation, with more head-room, the currents clamped a lot higher, at about +800uA, when the negative currents were getting down to -2.4mA. >> > >> > It looks very much as if there's some kind of silly mistake in the AD734 Spice model (which would interest Jim Thompson, who wants to sell Analog Devices better Spice models). >> > >> > If the actual device acted like the model, the data sheet wouldn't talk about +/-10mA output current limits (as it does on page 13). >> > >> > It's easy enough to hand-edit .cir files, if you kno0w what you are doing. Any advice will be gratefully received. I probably should have raised this with Analog Devices directly, but the price they charge for the AD734 means that they can't be selling many of them, which doesn't suggest that I'd get a prompt response. >> >> Oops. It looks as if the defect wasn't in the Analog Devices model, but in my circuit diagram - a connection to +15V seems to have been edited out at some point and the circuit was getting it's positive power supply from its inputs, making the model inconveniently realistic - I've had that happen on real circuits, and it can take a while to work out what's going wrong. >> >> My apologies to one and all. >> >> -- >> Bill Sloman, Sydney > >I have always felt that text input based SPICE versions(HSPICE, NGSPICE) are so much better >than GUI input based SPICE versions(LTSPICE) >The tiniest error in the text netlist input causes the simulation engine to crash, forcing >the oser to ensure that the initial input netlist >is correct. I believe that LTSPICE also allows >text file input. I have used the BSIM 4.6.5 >sub-micron device models in netlists that are >approximately 750 - 900 lines long, with both >HSPICE and NGSPICE, with only a few whimpers >of protest initially, but absolutely no issues >afterwards.
Text entry of netlists is barbaric. Imagine a 1000-line netlist to sim some part of a product. Imagine needing to revisit it after a year or two of inactivity. What are you going to do to figure it out? Read it and try to draw the schematic! Sure, Spice will catch you if you mistype "V10", but LT Spice won't misspell V10 when it saves a netlist. I can enter a presentation-quality LT Spice schematic in a fraction of the time I could type a netlist. I can reopen it a year later and see what's there in seconds. And it has a nice parts library, and HELP, right there in plain sight. -- John Larkin Highland Technology, Inc picosecond timing laser drivers and controllers jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Reply by Phil Hobbs July 7, 20152015-07-07
>Here are some examples where simulation can be very valuable.
<snip> Sure thing, I didn't intend to imply otherwise. I do the math first, though, whenever feasible. The things that are hard to do by hand are often also things that SPICE models are bad at, e.g. overload recovery and CM slew distortion. There are things where simulation is really the only cost-effective method, e.g. checking the strictly-linear response of an active filter made with non-ideal components, or (what I'm often interested in) looking at the sensitivity of a high-frequency and/or high-Z circuit to strays. I don't trust behavioural IC models at all. There are too many noiseless op amps with zero input capacitance and perfectly-clean slew limiting out there in spiceland. Transistor models are much better--I built a BFP640/pHEMT cascode TIA a couple of months ago. Simulation correctly predicted the collector current where it started oscillating, within about 10%, and got the frequency more or ledd right too (about 6 GHz). You do have to put in realistic strays, though. Cheers Phil Hobbs
Reply by July 7, 20152015-07-07
On Sunday, July 5, 2015 at 7:58:07 AM UTC-4, Bill Sloman wrote:
> On Saturday, July 4, 2015 at 12:23:27 PM UTC+2, Bill Sloman wrote: > > I've been using the Analog Devices Spice model in LTSpice to model the AD734 running with a current output - see Figure 25 on page 13 of the Rev E AD734 data sheet > > > > http://www.analog.com/media/en/technical-documentation/data-sheets/AD734.pdf > > > > When running a roughly 15kHz sine wave through the device, the positive current output limits at something between +200uA and +280uA. > > > > The voltage at the W and Z1 outputs of the AD734 is well below the rail. > > > > I've tried 2.2k, 6.8k and 15k current setting resistors. Only with 15k did the voltage at the W and Z1 outputs get high enough to be interesting. > > > > The current clipped at +200uA with 2.2k, +282.77uA at 6.8k and +267.94uA at 15k. > > > > The negative-going excursions looked perfectly sinusoidal, and went down to -350uA. > > > > Working in another region of operation, with more head-room, the currents clamped a lot higher, at about +800uA, when the negative currents were getting down to -2.4mA. > > > > It looks very much as if there's some kind of silly mistake in the AD734 Spice model (which would interest Jim Thompson, who wants to sell Analog Devices better Spice models). > > > > If the actual device acted like the model, the data sheet wouldn't talk about +/-10mA output current limits (as it does on page 13). > > > > It's easy enough to hand-edit .cir files, if you kno0w what you are doing. Any advice will be gratefully received. I probably should have raised this with Analog Devices directly, but the price they charge for the AD734 means that they can't be selling many of them, which doesn't suggest that I'd get a prompt response. > > Oops. It looks as if the defect wasn't in the Analog Devices model, but in my circuit diagram - a connection to +15V seems to have been edited out at some point and the circuit was getting it's positive power supply from its inputs, making the model inconveniently realistic - I've had that happen on real circuits, and it can take a while to work out what's going wrong. > > My apologies to one and all. > > -- > Bill Sloman, Sydney
I have always felt that text input based SPICE versions(HSPICE, NGSPICE) are so much better than GUI input based SPICE versions(LTSPICE) The tiniest error in the text netlist input causes the simulation engine to crash, forcing the oser to ensure that the initial input netlist is correct. I believe that LTSPICE also allows text file input. I have used the BSIM 4.6.5 sub-micron device models in netlists that are approximately 750 - 900 lines long, with both HSPICE and NGSPICE, with only a few whimpers of protest initially, but absolutely no issues afterwards.
Reply by Tom Swift July 7, 20152015-07-07
Tom Swift <spam@me.com> wrote:
 
> Here are some examples where simulation can be very valuable.
> 1. Use universalopamp2 and set the voltage gain, GBW and slew rate to > see the effect on open and closed loop response, gain and phase > margin.
This is a bit garbled. It should read 1. Use universalopamp2 and set the voltage gain, GBW and slew rate to see the effect on open loop gain and phase margin and the effect on closed loop response.
Reply by Tom Swift July 7, 20152015-07-07
Phil Hobbs <pcdhobbs@gmail.com> wrote:

>>I try to name all nets in a schematic. &#4294967295;It makes probing the layout >>much easier.
> Hmm. Normally I don't spend that much time on a given > simulation--there are usually only a couple of things I need to find > out that SPICE can tell me. > > I might spend more effort on an all-discrete design, but IME most IC > models give only an impressionistic view of actual chip behaviour, so > it's hardly worth the effort.
> Cheers
> Phil Hobbs
Here are some examples where simulation can be very valuable. 1. Use universalopamp2 and set the voltage gain, GBW and slew rate to see the effect on open and closed loop response, gain and phase margin. This can help you decide on how much performance is needed in the op amp and how much money to spend on one. I'm sure you could do this in your head, but most people need help and LTspice does a good job. It gets more complicated when there are multiple op amps involved. 2. Check pll loop damping and step response with variations in op amp gan and bandwidth, as above. 3. Develop new circuit configurations and check the response before going to breadboard. Pencil and paper is a good place to start, but LTspice gives a much better picture of what is going on and where the weak points are. You used this technique yourself when you developed a calibrator for one of your wideband circuits. You found that a tiny amount of series inductance was enough to destroy the performance you were seeking, and ended up putting a bunch of parts in parallel to reduce the inductance. 4. Document circuit response for troubleshooting in manufacturing. Techs can easily see how the circuit is supposed to work and check waveforms at different points. Then compare with the results on a defective pcb to see what needs to be replaced to solve the problem. Finally, check the repaired circuit to verify it performs as expected. I think it would be professional to include node names at every node to help the techs find their way around the circuit. There are many more examples, but I think you will find most engineers feel that LTspice simulation is a valuable tool. Until you get screwed over by sloppy or nonexistent node names.
Reply by krw July 6, 20152015-07-06
On Mon, 6 Jul 2015 19:14:30 -0700 (PDT), Phil Hobbs
<pcdhobbs@gmail.com> wrote:

>>> pretty simple circuit! > >>I try to name all nets in a schematic. &#4294967295;It makes probing the layout >>much easier. > >Hmm. Normally I don't spend that much time on a given simulation--there are usually only a couple of things I need to find out that SPICE can tell me. > >I might spend more effort on an all-discrete design, but IME most IC models give only an impressionistic view of actual chip behaviour, so it's hardly worth the effort. >
I probably use SPICE for even less than you because I don't give the models that much credit. Even the discrete models suck. It's too easy to trust the simulation. OTOH, when I did chip design I relied on SPICE (well, ASTAP at the time). IBMs models were superb. I was referring to the board schematic, though. Naming all the nets makes reviewing the layout much easier. Though not all net names get printed on the schematic, it's easier to deal with them than a thousand similar strings.
Reply by Phil Hobbs July 6, 20152015-07-06
>> pretty simple circuit!
>I try to name all nets in a schematic. &#4294967295;It makes probing the layout >much easier.
Hmm. Normally I don't spend that much time on a given simulation--there are usually only a couple of things I need to find out that SPICE can tell me. I might spend more effort on an all-discrete design, but IME most IC models give only an impressionistic view of actual chip behaviour, so it's hardly worth the effort. Cheers Phil Hobbs