GPS NMEA, Failover Circuit upgrade attempt.

[HellasTechn]

So am i. There must be something we are missing here. Why would it work fine and then burn when i remove and re attach the rs232 inputs ?

 

[HellasTechn]

he he, well take a look here :
http://electronics.stackexchange.co...3232-overheating-burnt-after-connecting-to-pc

check the last comment.

 

[CDRIVE]

he he, well take a look here :
http://electronics.stackexchange.co...3232-overheating-burnt-after-connecting-to-pc

check the last comment.

Great find! That certainly is a mind blower! I'm putting those links in the same folder that I keep the MAX232.

Chris

 

[HellasTechn]

Yes. If you scroll down that page you can also find this link :
http://e2e.ti.com/support/interface/etc_interface/f/392/t/233847


Bottom line is that the MAX232 chip is not a reliable solution. I will stick with the resistors solution which so far have never failed, not even the slightest problem with them.

 

[CDRIVE]

Personally, I never understood why you thought you needed it. It would make sense if you were transmitting data from the Picaxe (0 to +V logic) to an RS232 port (±V logic) but you're not. That said the MAX232 has been around a very long time. It has an excellent track record. Used by possibly millions and featured in most, if not all, serial port "How To" publications. This includes Jan Axelson, the undisputed "Queen of Ports"!
http://janaxelson.com/
This is the first time I'm reading of these failures.

Chris

 

[chopnhack]

Yes. If you scroll down that page you can also find this link :
http://e2e.ti.com/support/interface/etc_interface/f/392/t/233847

That is very interesting. Curious for a large company to have a critical batch failure like that though... I wonder if it would not be more likely a counterfeit batch that got out?

Constantine, did you repeat the tests the last post on the stack exchange mentioned to check for resistance on the pins?

 

[HellasTechn]

Personally, I never understood why you thought you needed it.

Only because i wanted to make our project as reliable as it can be, so i thought "signal conditioning" would be a good idea. I couldnt imagine this chip can fail that easily.

It would make sense if you were transmitting data from the Picaxe (0 to +V logic) to an RS232 port (±V logic) but you're not.

I has actually able to transmitdata serial data from the picaxe to the PC without a MAX232. I only had to remove the 180 Ohm resistor that picaxe uses on serout pins.

I wonder if it would not be more likely a counterfeit batch that got out?

Most probably but would the chips actually bare the MAXIM logo ?

 

[CDRIVE]

I has actually able to transmitdata serial data from the picaxe to the PC without a MAX232. I only had to remove the 180 Ohm resistor that picaxe uses on serout pins.

I had to think long and hard before replying to this and it's possible that what I say may be all wet but I don't think so. Do you remember this post?
https://www.electronicspoint.com/th...t-upgrade-attempt.278866/page-12#post-1707306

I believe modern RS232 ports, especially USB/RS232 converters don't require logic 0 to be a negative level as they once did. In other words they now conform to standards more closely resembling TTL logic levels. My guess is that a Picaxe would not be able to directly drive one of those old serial ports, as the port would not recognize 0V as logic 0. I don't think they would recognize +5V as logic 1 either.

There are still many industrial control systems today using RS232. Some of which are very long runs within very noisy (electrically) environments. My bet is that in those situations they conform to the older (±30V) standards. Somehow my "Serial Port Complete" by Jan has vanished but I do believe she covered these issues quite thoroughly.

As I stated "I may be all wet", so I'm fully prepared to be barraged with arguments to the contrary. It wouldn't be the first time.

Chris

 

[HellasTechn]

That kind of make sense to me...

By scoping the PC's serial port i can see that in idle condition Voltage is -5 volts and when i transmit it pulses to +5
I will confirm that with new scoping and i will post the picture here to see.
The PC i use is a Panasonic toughbook cf-48 year 2004 model i think.

The GPS output though if memmory serves well, was -13 Volt and pulsing to +13. Still i am not 100% sure about that. I dont have the GPS receiver at the moment to scope it but i will as soon as i get it.

Constantine, did you repeat the tests the last post on the stack exchange mentioned to check for resistance on the pins?

No but i will later tonight.

 

[hevans1944]

Most probably but would the chips actually bare the MAXIM logo ?

Yep, they would. That's why it is called counterfeiting. Big problem with semiconductors sourced from Asia. Not just Rolex watches and Gucci handbags anymore: everything is counterfeited in Asia. Quality varies, but always caveat emptor. Even authorized distributors have been scammed. Only way to tell if its the real deal is to decap and compare with a genuine article, sometimes using very sophisticated techniques such as scanning electron microscopes. Only the guv'ment and the military can afford to do this. The Chinese in particular are rapidly advancing their semiconductor fabrication technology, so some of the counterfeits are quite excellent. They just don't have to bother with paying for original design, quality control, or royalties to the originator of the device that is counterfeited. Has a pretty nice effect on their profit margin.

 

[HellasTechn]

That i didn't know. So in other words companies know that their products are being copied and sold as original but can do nothing about it, right ?

 

[HellasTechn]

Constantine, did you repeat the tests the last post on the stack exchange mentioned to check for resistance on the pins?

I read 5.2K from pin 8 to 15 and the same value from pin 13 to 15. That is correct but also i get a 38K from pin 8 to pin 16 and an other 38K from pin 13 to pin 16 which according to the chip's block diagram should not be there.
From pins 10 and 11 to pin 16 i read nothing (open circuit) while i should be getting a 400K reading.
All measurements taken from unused chips.

After measuring the used chips i have got pretty much the same readings.

 

[hevans1944]

That i didn't know. So in other words companies know that their products are being copied and sold as original but can do nothing about it, right ?

They are trying to do something about it, both technically and diplomatically. It is a serious problem best discussed in a new thread rather than this one.

Same applies to my proposed effort to build a DIY survey grade GPS with sub-centimeter accuracy thereby saving tens (if not hundreds) of thousand of dollars versus the cost of a comparable commercial system.

I will start such a thread when I know more about the problem, but to answer your previous question, more data averaged over longer periods of time does not appear to be the answer, although post-processing is routinely used. The "best" solution seems to be to use CORS (Continuously Operating Reference Stations) to augment and correct your local GPS. A CORS station has a precisely known (surveyed) location for its GPS receiver, so any pseudo-distance errors between the GPS satellite and the ground station, caused mainly my ionospheric and tropospheric variations affecting the time-of-flight delay, can be measured by CORS and corrections broadcast.

In my South Venice Florida location there are at least three CORS, each transmitting corrections on at least an hourly basis, some more often. Two of these (McDill AFB and Ft. Myers) are very nearly on a light-of-sight that comes close to the property, so perhaps that will be of some benefit. I just don't know enough about it yet. But I will be following your work on this thread!

Hop

 

[chopnhack]

I read 5.2K from pin 8 to 15 and the same value from pin 13 to 15. That is correct but also i get a 38K from pin 8 to pin 16 and an other 38K from pin 13 to pin 16 which according to the chip's block diagram should not be there.
From pins 10 and 11 to pin 16 i read nothing (open circuit) while i should be getting a 400K reading.
All measurements taken from unused chips.

After measuring the used chips i have got pretty much the same readings.

8 to 15 - Check
13 to 15 - Check

How are you getting 38k ohm reading between your receive in pins? Are they not isolated from Vcc?? Assuming that your incoming RS232 feeds are not part of this circuit how are they coupled? And for both channels to have this is highly suspect... Your next readings have again two failed channels, for unused chips I would think you have a bad batch or a counterfeit batch.

Where did you buy these from?

 

[hevans1944]

one last thing ..... a serious omission ... you show lines crossing and others that are "T" joints

This is a reply to a much earlier post on this thread, but I doubt it will be "one last thing"...

There is a schematic drawing convention, probably invented because of the "disappearing dots" problem, that crossing lines NEVER represent connections. ALL connections MUST be "T" connections. If this is rigorously applied there can be no misunderstanding of what connects to where. Of course it takes a pretty good proof-reader's eye to vet a complicated schematic for adherence to those two rules.

Who's on First? may offer additional ambiguity in certain situations, mainly confined to baseball.

 

[HellasTechn]

I will start such a thread when I know more about the problem, but to answer your previous question, more data averaged over longer periods of time does not appear to be the answer, although post-processing is routinely used. The "best" solution seems to be to use CORS (Continuously Operating Reference Stations) to augment and correct your local GPS. A CORS station has a precisely known (surveyed) location for its GPS receiver, so any pseudo-distance errors between the GPS satellite and the ground station, caused mainly my ionospheric and tropospheric variations affecting the time-of-flight delay, can be measured by CORS and corrections broadcast.

In my South Venice Florida location there are at least three CORS, each transmitting corrections on at least an hourly basis, some more often. Two of these (McDill AFB and Ft. Myers) are very nearly on a light-of-sight that comes close to the property, so perhaps that will be of some benefit. I just don't know enough about it yet. But I will be following your work on this thread!

What you say is correct but we do not have such ground stations here in Greece so i can not use that. Though i think that GPS receiver would take into account any broadcast from such station (actually theese GPS recievers are DGPS capable) and calculate DOP accordingly.

In my South Venice Florida location there are at least three CORS, each transmitting corrections on at least an hourly basis, some more often. Two of these (McDill AFB and Ft. Myers) are very nearly on a light-of-sight that comes close to the property, so perhaps that will be of some benefit. I just don't know enough about it yet. But I will be following your work on this thread!

The fact that the stations are close to the location is a good thing. You only need to make sure the GPS receiver or receivers you will use can utilise the CORS signals.
Please read through the posts and let me know of any questions you may have.

 

[HellasTechn]

How are you getting 38k ohm reading between your receive in pins? Are they not isolated from Vcc??

They should be according to the block diagram.

Assuming that your incoming RS232 feeds are not part of this circuit how are they coupled?

I have no idea.

Your next readings have again two failed channels, for unused chips I would think you have a bad batch or a counterfeit batch.

Where did you buy these from?

E-bay chineese seller !!!

P.S.
Hopefuly the MAX3232CPE i have ordered now (again from e-bay but different chineese seller) will at least be of good quality.

 

[HellasTechn]

This is a reply to a much earlier post on this thread, but I doubt it will be "one last thing"...

There is a schematic drawing convention, probably invented because of the "disappearing dots" problem, that crossing lines NEVER represent connections. ALL connections MUST be "T" connections. If this is rigorously applied there can be no misunderstanding of what connects to where. Of course it takes a pretty good proof-reader's eye to vet a complicated schematic for adherence to those two rules.

Who's on First? may offer additional ambiguity in certain situations, mainly confined to baseball.

I take it as common sence that T joints are "connected" while crossing lines without a dot are "not connected". maybe it is just me thinking about it like that. I do agree though that the program should be useing a little bigger dot to where lines are connected.

 

[hevans1944]

just sayin'... if the two "rules" were always followed, dots would not be necessary. However, if the two rules are followed and dots are added at the "T" junctions, then there is no possible ambiguity, and the schematic is more readable. So, do all three: make connections at "T" junctions ONLY, never make connections where wires cross, and "dot" all the "T" connections. I wonder if this is where some wag came up with the expression of "crossing all the eyes and dotting all the tees," instead of the usual, other way around, of "dotting all the eyes and crossing the tees" as an expression for completeness?

Oh, well... with the demise of the MAX232, I think this thread is ready to die too.

 

[chopnhack]

E-bay chineese seller !!!

Oy!
Constantine, I use ebay too, but this is a critical piece of your hardware, please consider getting from a reputable source so that you are not chasing your tail. Digikey has free shipping to Greece on $100 order or more, I am sure you can do better with European outfits, have a look around Germany, England, etc. for better rates.