DC Voltage level sensing

Yes, one can try to install Windows-only stuff inside Wine, but that bring bloat and potential instability. For me ngspice works reasonably well. Bringing LTspice and Wine is rather high price to see circuits posted on the net...

I wrote "encoded": the information is encoded as ASCII and I found no explanation on the net what encoding is in use. It has _some_ superficial similarity to format of orignal Spice, but organization seem to be quite different.

For known encoding I cold easily write convertor to other format, but ATM LTspice files are useless to me.

-----------------------

** For several years I used "individual.net" which involved payment - but after moving up to Windows 7 found that no version of "Outlook Express" existed that was compatible.

A friend recommended "Thunderbird" which cannot access "individual.net" for some reason. So I went over to Google Groups.

Soon got used to it and it has a huge archive.

.... Phil

They charge my card once a year, or send me an email if the card parameters have changed.

Nice. What's their business model?

Yes, if nothing changes. I lost my card and was issued a new one with a different number, so I couldn't even find my old news service to give them the new number. My email also changed, so if they tried to contact me it bounced.

A year is a long time. Lots of things can happen.

It's a private project by Wolfgang M. Weyand in Bad Homburg, just outside Frankfurt. You can send donations via Paypal. It apparently has been running for many years.

It had a brief glitch several weeks ago but that was fixed in hours. Other than that, no problems.

The log on their home page shows entries going back to 2007.

T'bird is more picky than most about SSL certificate validation, and flatly refuses to report errors. However, you can add certificate exceptions and make it work, like I need to with thecubenet.com

Clifford Heath.

well I do.

NT

someone who uses google groups would want to

no.

This is true, I had to make some shims with hooks from a tin can to get my radio out last time, it took about half an hour, so yeah easily 10 times slower;

Not sure Phil requires any criticism to rant.

I know people like the TL431 for similar circuits, but is there really anyt hing wrong with using a common 12 volt regulator in the way I described? S eems pretty simple and effective to me. It also produces a regulated 12 vo lts which can be useful for other circuitry rather than driving it off the unregulated car power.

Vin E PNP C o--+-----------\___/-----+---o High == Vin > 13.4v | B| | | | | | \ \ | /10K /10K | \ \Add LED as indicator? | / / | | | | --- | | V | | L7812ABV --- V | +-----+ | +--| |----+--------------o 12V out +-----+ | 10K | +--\/\/--+ | | V V

Looks ok in Google groups if you select "Show original".

I've found LTspice to work fine under Wine in Ubuntu, and (display at least) the MacOS version can work with them too. If you just want to see the schematic, it's workable.

Nasty surprise, though, if you try a Mac trackpad laptop, and need the 'right mouse button'.

Details, please.

I find LTS runs totally fine in Linux under WINE.

Hi Win,

Your drawing was fine. Just for fun, I've added your R4 for hysteresis. Only viewable with club membership and our secret decoder rings :-)

Cheers, James Arthur

It took me a bit to figure out what this circuit is doing. This is the com parator circuit, fig 25 on page 22 of the TI data sheet with the FET and a resistor added for hysteresis.

I'm not sure why a FET is used in place of a bipolar device. The threshold of the FET becomes part of the accuracy/precision budget and they tend to not be so accurate or precise. A bipolar device has a much more defined tu rn on voltage and so will not impact the accuracy/precision nearly as much.

I would also suggest that hysteresis will also detract from this budget by definition. The desired threshold must be within a window of voltage set b y the state of charge of the battery, the temperature and potentially the l oads that are on at any given time. So the actual window for the threshold is not going to be very wide. Add in some lack of accuracy of the thresho ld setting and a window of precision as the parts range in temperature and you may find your circuit not working in some situations. Add hysteresis o n top of that and in one situation the circuit may turn on, but then fail t o turn off and vice versa in other situations.

So in lieu of having good data to know where your threshold needs to be, I would suggest you skip the hysteresis which actually accomplishes nothing ( either your threshold is good or it's not and noise in a car won't be filte rable by hysteresis) and makes it harder to fit the accuracy/precision of y our circuit to the threshold window requirement of your battery charging.

Actually, as I write this, I'm pretty sure you won't find a reliable approa ch by a circuit with a hard threshold. It is going to be hard to filter th e noise from this circuit and it is going to be hard make it work in all se asons. I'm thinking it would be better to use a small MCU to measure the v oltage and handle all the issues in software. Filtering can be done very e asily and a great deal of intelligence can be used to decide when the ignit ion is on or not. It can also add features like keeping the dashcam on for some time after leaving the car.

??? Figure 25 is a crowbar, with very different operation.

The gain for the TL431 is very high. Acting as a comparator, the change in the R2 R1 threshold voltage, dealing with the MOSFET's Vth variation, is very small. Using a BJT would require knowledge of the maximum load, which was not known.

Maybe, I'm not sure. My observations of battery voltage, once the engine is running, show the charging regulator in full control, setting the voltage above 14 volts. So whether a 13.5 or 13.7 or whatever voltage setting isn't critical. Some real-world testing will reveal the answer.

We must be looking at different data sheets. Fig 25 is a simple comparator which is what you are using it for. I just realized that means I was wron g about my comment that the threshold of the FET factors into the accuracy and precision of the circuit.

I think you will find that set point varies widely. While it does not requ ire the battery to be fully charged, there will be a load on the alternator when providing a heavy charge but more importantly there will be a large d ifference in voltage with temperature. This is the nature of batteries and regulators are built to sense the temperature and adjust the charging volt age.

The point is the expected voltage on the car power in all conditions won't be a single value, it will be a range depending on several factors and maki ng sure your circuit has accuracy and precision good enough to work properl y in all conditions will be a problem.

Maybe the OP lives some place like southern California where the temps don' t vary so much. But most places will have to deal with a lot. Both the ba ttery and the circuit will have variations and I'm not sure it is possible to make this work for all conditions without a lot more work or a simple di gital solution.

"Going digital" is against my religion and so can only be a very last resort. There *is* a considerable temp range to contend with here: zero celcius in winter to 140F in summer. However (I don't have Win's schematic to hand right at this moment so am guessing) there should be a way to introduce a simple 'summer/winter' switch into the design to cope with this. I don't mind doing this manually; tweaking a pot twice a year I can live with. :)