74HC4046 Model Development

I have been notified by Helmut Sennewald that the LTspice encryption process hides the node sequence. So please revisit...

for a corrected version.

This is just the VCO portion for now, and just the variable part; suitable only for LTspice.

Please try it out and let me know your comments, etc.

Since I'm new to working with LTspice, please send me any error messages, so I can refine my skills >:-} ...Jim Thompson

Thanks for taking on my suggestion from a while back, Jim. Looking forward to trying it out!

My aim is to make the model as close to the silicon as possible.

For amusement, since it had been mentioned, I'm also doing the uA741 at the same time. It may be done tomorrow or Monday, since it's rather easy.

The 4046 will take awhile to get all the pieces working together. ...Jim Thompson

Am 21.10.2012 02:02, schrieb Jim Thompson:

...Jim Thompson

why is it encrypted??

I'm not in the teaching business, I'm in the money-making business, so I don't want to divulge my techniques. ...Jim Thompson

-- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

So you want us to review it for free so that you can go make money off of it?

Hmm.

If you will remember a prior post, I asked which devices had poor Spice models. The 74HC4046 was mentioned multiple times. I am supplying a model that I presume will behave more accurately. Since the 74HC4046 is complex, I am asking, as I go, "Does this portion fit?"

If no one really wants a good 74HC4046 model I'll go do something else. ...Jim Thompson

Please do that.

Otherwise, if you want to help and want help yourself, then work with these people.

I'll think (more than) twice before I use the 4046 in a design again. As Phil and you pointed out its pretty crappy.

What upsets me most is that it's a copy of my 1965 design, the MC4024 (bipolar devices, oscillator core is PECL, TTL output). And they bungled the copy :-( Unlike my original version, the CMOS version dumps the capacitor is an uncontrolled manner, and the thresholds and current mirrors are abominations of "engineering". ...Jim Thompson

Yes ;-) ...Jim Thompson

Are bungled abominations more difficult to model?

RL

I really do hope someone will try out the model and let me know how it simulates... rather than blathering that it's encrypted. ...Jim Thompson

Am 22.10.2012 01:28, schrieb Jim Thompson:

No. if you don´t want to give you won´t receive.

You want to earn money with it, then spend some money for testers.

I dunno. Making the model will probably help some folks, if they don't expect its horrible nonlinearity to be horrible in quite the same way as the production parts. (For a variety of reasons, I'm also in favour of people being able to make a living.)

Cheers

Phil Hobbs

Hi Guys, OK I just finished test9ng the difference between the TI and NXP versions of the VCO in the HC4046. The good news (for me) is there's hardly any difference between the two. (This is with a VCO 'bias' resistor of 4.02k ohm.) (And four different caps, 2.7nF,

Here's the data for the TI version. First column is the VCO voltage (in volts) and the next four are the VCO frequencies in kHz. (Frequencies only measured on a digital 'scope so don't believe all those decimal places :^)

George H.

I GAVE you a model. I didn't charge for it. I encrypted it to conceal my "tricks of the trade".

I know my model works under LTspice...

my submission was simply to make sure it also works after encryption.

Quite frankly I don't give a rat's ass whether you use the model or not. I find the 74HC4046 a piece-a-crap, a rather poor CMOS copy of my original design, the quite linear MC4024 of nearly 50 years ago. ...Jim Thompson

I have found that the variation of parameters the 74HC4046 from one manufacturer to another is such that I cannot agree with such a sweeping statement. I found the Philips/NXP variant a very well behaved device, whereas others had/have poor VCO control linearity etc.

I'm just going from reports here, primarily Phil Hobbs' bad experiences. ALL of the data sheets look wonderful ;-)

I can certainly think of ways to improve it, particularly the current mirrors, though a complete architecture change could easily eliminate lots of the non-linearity issues.

Whoever rolled the CMOS copy totally missed an important point, don't dump the capacitor into a substrate diode. ...Jim Thompson