Learning electronics: Release the Magic Blue Smoke

The Something Awful Forums > Discussion > Hobbies, Crafts, & Houses > Learning electronics: Release the Magic Blue Smoke

«‹›781 »

PDP-1: Oct 12, 2004; It's a beautiful day in the neighborhood.

Since it's o-scope chat day, can anyone recommend a scope with ~1GS/s, 8-bit or better for one channel for measuring a 0-5V analog pulse? I'd need to be able to interface it with a PC somehow, just enough to collect the trace data after collecting one shot. Price matters too, this particular company hates spending money on silly engineering tools.

I would absolutely love it if I could find a PCIe card scope with these specs but nothing I've seen in my (admittedly brief) search even comes close. Next best option would be a standalone used unit with a GPIB/USB interface.

# ? Mar 22, 2012 14:52

Adbot: ADBOT LOVES YOU

# ? May 12, 2024 17:51

taqueso: Mar 8, 2004

Jonny 290 posted:

Hey cool, jonny290 at gmail dot com, if you don't mind.

The bench meter is just an aesthetic thing and for ease of use, really. Bench space is absolutely not a concern; I have many square feet of shelves to fill up and want to get a nice looking test bench to go next to the radio desk. I have an Amprobe handheld DMM that works very well, for sure, I would like a second DMM though for flexibility. Maybe just get another handheld?

I think they look great on a bench, too. Some of the older Keithley 5.5 to 6.5 digit multimeters go for $100-200 on ebay. I think the best one is the model 196, which is more like $200-300.

# ? Mar 22, 2012 16:16

sixide: Oct 25, 2004

PDP-1 posted:

Since it's o-scope chat day, can anyone recommend a scope with ~1GS/s, 8-bit or better for one channel for measuring a 0-5V analog pulse? I'd need to be able to interface it with a PC somehow, just enough to collect the trace data after collecting one shot. Price matters too, this particular company hates spending money on silly engineering tools.

I would absolutely love it if I could find a PCIe card scope with these specs but nothing I've seen in my (admittedly brief) search even comes close. Next best option would be a standalone used unit with a GPIB/USB interface.

If price is your primary concern, get one of the budget Rigol scopes. They meet all your requirements. I'm assuming you don't need a huge bandwidth if you're only looking for 1 GS/s.

# ? Mar 22, 2012 18:16

CapnBry: Jul 15, 2002; I got this goin'; Grimey Drawer

Slanderer posted:

Well, for one, a bench DMM is big and probably a waste of space. Get a Fluke or maybe just a BK Precision (they're not that bad, I swear!)

I have a BK Precision 2709B I got online for $90 I think. The problems I have with it are the loving 90dB beeps this thing puts out every time you turn the selector switch and you'd better move the probes from current to volts before you switch back to volts. (BWEEEEEEEE) "I goddamn know, I'm moving them!" That and when it is set to measure mA, it might read 0.9mA and when you switch it to uA it reads closer to 500uA.

It does seem to autorange pretty quickly, and the continuity tester is latched so it doesn't beep halfassedly.

# ? Mar 22, 2012 20:35

Krenzo: Nov 10, 2004

ANIME AKBAR posted:

Because sophisticated instruments like oscilloscopes cost a great deal of money to develop, and demand is low, so they don't enjoy economies of scale. The cost of the actual components don't make up most of the cost, not by a longshot.

Things are probably going to get cheaper from here on out due to the emergence of relatively low cost high performance off-the-shelf ADCs and FPGAs, allowing smaller companies to feasibly create high performance hardware (up until a few years ago most of the high speed stuff in a digital oscilloscope was ASIC technology). But the lack of overall demand still makes it hard for new competitors to actually gain a foothold.

Bummer, well I was thinking instead of spending $1k on a scope, I could make my own and do it open source style since individual components are not that expensive.

# ? Mar 22, 2012 23:39

sixide: Oct 25, 2004

I'd love to make test equipment--as a career. The amount of engineering required to do a halfway decent job is a bit much for a hobby, I think.

# ? Mar 22, 2012 23:57

horse_ebookmarklet: Oct 6, 2003; can I play too?

Last page someone said the extech 330, and its a great BUT

CapnBry posted:

It does seem to autorange pretty quickly, and the continuity tester is latched so it doesn't beep halfassedly.

It doesn't latch. The beeper is so gritty

# ? Mar 23, 2012 08:54

taqueso: Mar 8, 2004

NotHet posted:

Last page someone said the extech 330, and its a great BUT

It doesn't latch. The beeper is so gritty

It doesn't bother me at all. The slowish autoranging is a bigger annoyance to me. For a really gritty beeper, try a BK 2703B. The 330 at least has some tiny amount of minimum on time.

# ? Mar 23, 2012 16:19

the wizards beard: Apr 15, 2007; Reppin

4 LIFE 4 REAL

ANIME AKBAR posted:

I also got a pair of those smart LCR tweezers (the $300 ones, not the lovely chinese clones), and it's turned out to be a good purchase. It looses accuracy when you go below ~25pF/100nH, but it's still enough that I can easily identify unmarked SMD capacitors and inductors, which is the reason I got it. Keep in mind that's when I know that the components can only be a certain number of values (so if it says 980pF then I know it's a 1nF cap). I don't use it for actual quantitative measurements. For that I think you really do need a decent benchtop LCR.

I bought the low-end Agilent Capacitance meter for work a while ago and I was able to get it with a set of SMD tweezers that can be used as meter probes. This might be a better option than smart tweezers if you already have a nice meter.

http://uk.farnell.com/agilent-technologies/u1782a/smd-tweezers/dp/1710933

# ? Mar 23, 2012 19:20

PDP-1: Oct 12, 2004; It's a beautiful day in the neighborhood.

sixide posted:

If price is your primary concern, get one of the budget Rigol scopes. They meet all your requirements. I'm assuming you don't need a huge bandwidth if you're only looking for 1 GS/s.

Thanks, I'll check them out. I haven't used Rigol stuff before, does anyone have experience with them who can comment on the build quality, software interface, etc?

sixide posted:

I'd love to make test equipment--as a career. The amount of engineering required to do a halfway decent job is a bit much for a hobby, I think.

I once thought about trying to come up with a line of hobby-level stuff myself. My plan was to make a generic powered backplane with USB connection, then have modules that would clip into it to provide functions like voltmeters, mini-oscilloscopes, low-wattage power supplies, etc. Mix-n-match the modules to build the lab kit you wanted, expand it later as needed. Control and readout would all be done on a PC connected to the USB port. Physically it'd be somewhat like the National Instruments RIO system, but without the controller unit or calibration traceability to keep things cheap.

:effort:

# ? Mar 23, 2012 19:47

Krenzo: Nov 10, 2004

PDP-1 posted:

Thanks, I'll check them out. I haven't used Rigol stuff before, does anyone have experience with them who can comment on the build quality, software interface, etc?

http://www.eevblog.com/2009/04/05/eevblog-1-rigol-ds1052e-oscilloscope-reviwed/

# ? Mar 23, 2012 20:48

sixide: Oct 25, 2004

I have an 1102E on my bench. Build quality is far better than I would expect for the price. The interface is on par with or better than the entry level Tek DSOs these days. I wish there was an auto-hide feature for the menu, but otherwise good.

They won't put fear into LeCroy anytime soon, but otherwise no real complaints.

Buy third-party probes, though.

# ? Mar 23, 2012 23:46

horse_ebookmarklet: Oct 6, 2003; can I play too?

Thanks for the ideas on the lcr meters. I guess I'll just keep a lookout for a ~$200 one popping up on ebay.

My local surplus had these for $3 each. I bought 5, they had 30. How could I say no? :science:

No real ideas for applications now... Unless I want to buy a few more and kill myself with a 1kv+ charge pump.

# ? Mar 24, 2012 03:28

ANIME AKBAR: Jan 25, 2007; afu~

You definitely should have bought all 30. You never know when you might need a large capacitor bank. And you could always resell them on ebay too.

# ? Mar 24, 2012 05:02

insta: Jan 28, 2009

NotHet posted:

Thanks for the ideas on the lcr meters. I guess I'll just keep a lookout for a ~$200 one popping up on ebay.

My local surplus had these for $3 each. I bought 5, they had 30. How could I say no?

No real ideas for applications now... Unless I want to buy a few more and kill myself with a 1kv+ charge pump.

You get your whore rear end back to that shop and buy the remaining 25 of them RIGHT GODDAMN NOW. They're going for 10x that on eBay, and are just cool to have around. Rectify mains into DC and feed into all 30 of those caps in parallel, then short it out with a puck SCR into a three-coil loop of copper to smush cans in half.

# ? Mar 24, 2012 16:10

Corla Plankun: May 8, 2007; improve the lives of everyone

Lets Play Arson posted:

Thought i'd say thanks again. Got it wired up and its working beautifully now. I just need to make a shroud for it so I can duct it outside.

I've got another question though, and probably a stupid one that gets asked all the time. But are there any good textbooks for learning electronics from the perspective of microcontrollers? I've had a look at the online guides that get linked in stickies everywhere and I felt they were all lacking. Alot of things treat microcontrollers as a hobby in and of itself but i'm just planning on using them as a means to an ends. So perhaps they're out of my scope and I should stick to stuff I can understand like cam timers...

Making Things Talk is the closest thing I have ever seen to a textbook that just does microcontrollers. It was required reading for an art course at SMU so at least one professor likes it.

If you want to focus on a specific suite of controllers, I am sure there are tons of books that focus on introductory knowledge; PICAXE's manuals are free and there is basically an infinite amount of words about arduinos out there, too.

# ? Mar 24, 2012 16:34

Krenzo: Nov 10, 2004

PDP-1 posted:

...Rigol...

After looking around the eevblog forums, it looks like there are some very good, cheap real-time scopes that have surpassed the Rigol DS1052: the Owon SDS7102, the Hantek DSO5102, and a few others.

eevblog forum posted:

The device [Hantek DSO5102] itself, compared to Rigol DS1102E, is in my opinion a "bit" better :
- 7" 800 x 480 pix display
- "real" 100Mhz ADCs (yes, exact the same AD9288 but 100MHz rated)
- 2500 wfrms/s (Rigol E 800 wfrms/s)
- Screen refresh 30/40/50 frames/sec (Rigol ?)
- DPO-like functionality (waveform persistent for 0.2sec-> 8.0sec)
- dedicated good known application SOC and Linux OS

Here's a thread dedicated to people hacking its software. It has me interested in picking up a Hantek scope since I was looking for something that had at least 200 MHz of BW.

Also, Rigol is going to release a new set of oscilloscopes in June: the DS2000 series. Unfortunately, they're said to be aimed for the $1-2k price range.

# ? Mar 25, 2012 07:44

ANIME AKBAR: Jan 25, 2007; afu~

It always bothered me that people tend to talk about DSOs in terms of just things like sampling rate/memory depth/refresh rates. You almost never hear about things like noise floors, dynamic range, or ENOB.

Personally I'm waiting for a mid range scope with 12 bit ADCs, and with analog frontends to take advantage of that. For the work I do, I need a dynamic range of >50dB, and 8 bit ADCs simply can't do it.

# ? Mar 25, 2012 19:14

Delta-Wye: Sep 29, 2005

I've been helping some students with an Atmel-based project and I am totally stumped by why all their examples and header files are set up to use the following style of bitmasking:

DDRC |=(1<< PORTC5); //output
PORTC |= (1<<PORTC5); //high
PORTC &= ~(1<<PORTC5); //low

Why not just define PORTC5 as 0x20 and just do DDRC |= PORTC5? PIC and MSP430 platforms both seem geared this way. I'm fairly confident that the compiler turns those into the proper constants anyways so it's not like it adds processing time, it just looks like rear end. What is the deal?

# ? Mar 27, 2012 16:24

longview: Dec 25, 2006; heh.

I'm doing a project for school now and part of it involves sensing an analog temperature sensor (AD590), and the most promising ADC chip we have available is the AD7575 but the datasheet was somewhat confusing:

It only speaks of the Vref as being 1.23V but the datasheet also says it can be as high as Vdd, does anyone if it will work properly with the Vref at say 2-4V instead?
Same thing with the AGND, it's allowed to be as high as VDD according to the datasheet, and it'd be very practical to just set that to about 2V to get the most out of the digital range.
So do I need to ensure my voltage range is 0-2.46V as the datasheet says or can I just adjust the references to higher values?

And yes, I know I could just use a built-in ADC on the microcontroller, but it's not an option.

# ? Mar 27, 2012 17:04

taqueso: Mar 8, 2004

Delta-Wye posted:

I've been helping some students with an Atmel-based project and I am totally stumped by why all their examples and header files are set up to use the following style of bitmasking:

DDRC |=(1<< PORTC5); //output
PORTC |= (1<<PORTC5); //high
PORTC &= ~(1<<PORTC5); //low

Why not just define PORTC5 as 0x20 and just do DDRC |= PORTC5? PIC and MSP430 platforms both seem geared this way. I'm fairly confident that the compiler turns those into the proper constants anyways so it's not like it adds processing time, it just looks like rear end. What is the deal?

I can think of a couple reasons, but I don't think they are actually that useful. It lets your define the bits as 0, 1, 2, 3... so you don't have to think about the numeric representation (0x01, 0x02, 0x04...). It also guarantees that the constants only include one bit, so you couldn't accidentally have PORTC5 be 0x21 or something.

But, if someone considers that important, why not make the PORTC5 constant (1<<5) or (1<<BIT5)?

# ? Mar 27, 2012 17:46

Slanderer: May 6, 2007

taqueso posted:

I can think of a couple reasons, but I don't think they are actually that useful. It lets your define the bits as 0, 1, 2, 3... so you don't have to think about the numeric representation (0x01, 0x02, 0x04...). It also guarantees that the constants only include one bit, so you couldn't accidentally have PORTC5 be 0x21 or something.

But, if someone considers that important, why not make the PORTC5 constant (1<<5) or (1<<BIT5)?

I'm still kind of confused what the actual headers Delta-Wye is using say...

Regardless, after figuring it out, I became somewhat fond of the headers used by the XMEGA, both for having specific macros for bit-positions, bit masks, group masks, and all kinds of stuff. Some people don't like the use of structs for all the peripherals, but whatever.

http://www.atmel.com/Images/doc8075.pdf

Notably, listed at the very end under " Alternative Ways of Writing Code":

code:

PORTB_OUT = (1 << PORTB_OUT0_bp); // Set PORTB_OUT, bit0

quote:

The bit position definitions are included for compatibility reasons. They are also
needed when programming in assembly for instructions that use a bit number

# ? Mar 27, 2012 19:35

Delta-Wye: Sep 29, 2005

Slanderer posted:

I'm still kind of confused what the actual headers Delta-Wye is using say...

Regardless, after figuring it out, I became somewhat fond of the headers used by the XMEGA, both for having specific macros for bit-positions, bit masks, group masks, and all kinds of stuff. Some people don't like the use of structs for all the peripherals, but whatever.

http://www.atmel.com/Images/doc8075.pdf

Notably, listed at the very end under " Alternative Ways of Writing Code":
code:
PORTB_OUT = (1 << PORTB_OUT0_bp); // Set PORTB_OUT, bit0

It looks something like this:

code:

#define PORTD _SFR_IO8(0x0B)
#define PORTD0 0
#define PORTD1 1
#define PORTD2 2
#define PORTD3 3
#define PORTD4 4
#define PORTD5 5
#define PORTD6 6
#define PORTD7 7

My understanding is you go:
PORTD |= (1<<PORTD5);
to set portd 5 high. It doesn't work as expected though so I don't know. It certainly looks strange to me!

# ? Mar 27, 2012 22:08

Bad Munki: Nov 4, 2008; We're all mad here.

I just found an unused $50 gift card for Radioshack on my dresser. If only their parts were reasonably priced.

# ? Mar 28, 2012 00:23

Tobasko: May 15, 2006

Hey! I've just started learning electronics, I'm in a DC circuits class at my community college. Eventually I want to build little interactive robots and fun artsy-toys with arduinos and stuff, but circuits are such a foreign thing to me I had to start with a class. This thread is awesome, reading through the first 2 pages have answered a ton of questions left open by my super crappy class (We're 3 months in and have only just started parallel resistors).

For my semester project I've chosen to put together a Drawdio kit. Today I was going through the pieces and schematic trying to figure out what each piece does and I've got some questions I was hoping someone could help me out with.

What's the little person looking component? Is it literally a person? I know that the body and graphite are used to complete the circuit in a drawdio... Where do the open Vcc arrows go? If the little person dude isn't the person do the open arrows go to the pencil/person? Q1 is the transistor, right? Why is it just floating out there?

# ? Mar 28, 2012 00:34

SolidElectronics: Jul 9, 2005

Tobasko posted:

Hey! I've just started learning electronics, I'm in a DC circuits class at my community college. Eventually I want to build little interactive robots and fun artsy-toys with arduinos and stuff, but circuits are such a foreign thing to me I had to start with a class. This thread is awesome, reading through the first 2 pages have answered a ton of questions left open by my super crappy class (We're 3 months in and have only just started parallel resistors).

For my semester project I've chosen to put together a Drawdio kit. Today I was going through the pieces and schematic trying to figure out what each piece does and I've got some questions I was hoping someone could help me out with.

What's the little person looking component? Is it literally a person? I know that the body and graphite are used to complete the circuit in a drawdio... Where do the open Vcc arrows go? If the little person dude isn't the person do the open arrows go to the pencil/person? Q1 is the transistor, right? Why is it just floating out there?

I think the person symbol is actually a person.
As for the VCC arrows, they're all connected together (same with all the GND lines). It's just written that way to make the diagram cleaner. So in this circuit all the Vcc arrows go to battery + and all the gnd lines go to battery -.

# ? Mar 28, 2012 01:26

Slanderer: May 6, 2007

SolidElectronics posted:

I think the person symbol is actually a person.
As for the VCC arrows, they're all connected together (same with all the GND lines). It's just written that way to make the diagram cleaner. So in this circuit all the Vcc arrows go to battery + and all the gnd lines go to battery -.

Except they somehow managed to reverse the battery symbol...

# ? Mar 28, 2012 01:44

sixide: Oct 25, 2004

Is there some reason there's a PNP for the buffer transistor? I guess someone must want the output inverted, though that really only makes sense if the 555 output spends a lot of time near Vcc.

# ? Mar 28, 2012 02:05

ante: Apr 9, 2005; SUNSHINE AND RAINBOWS

Considering it's a speaker output, it doesn't really matter whether or not the buffer inverts or not. Maybe they haven't been taught what NPN transistors are yet.

# ? Mar 28, 2012 02:45

Bad Munki: Nov 4, 2008; We're all mad here.

So I'm in the process of buying a house which I'll be moving into at the end of May. I'm excited to do some random poo poo around the place, like a switch on the sliding glass patio door which would trigger some soft-on (i.e. fade-on) landscape lighting around the patio, temperature sensors around the place that would report online, basically useless-but-fun stuff like that. A doorbell that tweets, I dunno.

What silly poo poo would you do to a house if you had very few restrictions and time/money weren't really a problem?

# ? Mar 28, 2012 03:24

taqueso: Mar 8, 2004

Automatic (time based or light sensing or both) blinds. I've wanted to do that for a long time, but it never quite seems worth it.

# ? Mar 28, 2012 04:13

ante: Apr 9, 2005; SUNSHINE AND RAINBOWS

Alarm clock espresso machine

# ? Mar 28, 2012 04:38

Tobasko: May 15, 2006

Awesome, thanks SolidElectronics!

ante posted:

Considering it's a speaker output, it doesn't really matter whether or not the buffer inverts or not. Maybe they haven't been taught what NPN transistors are yet.

We haven't been taught about transistors yet but I pulled that schematic right off the drawdio website, so hopefully whoever designed it knows about transistors :x

# ? Mar 28, 2012 05:37

SnoPuppy: Jun 15, 2005

sixide posted:

Is there some reason there's a PNP for the buffer transistor? I guess someone must want the output inverted, though that really only makes sense if the 555 output spends a lot of time near Vcc.

If the default state is the person disconnected, then the output will be high (JK flop in the 555 is held in the set condition because the threshold pin is less than 1/3 Vcc). When a person grabs it, the C1 is allowed to charge/discharge, and the output will toggle. Given that the collector resistor is only 10 ohms, you really don't want to have that transistor default to "on."

I'm more concerned with the lack of a base resistor on the transistor. It's just a diode to the rail, so when the 555 is driving 0 significant current might flow.

# ? Mar 28, 2012 08:39

Nerobro: Nov 4, 2005; Rider now with 100% more titanium!

I'll start off with my real questioon first.

How would you drive a laptop LCD as a computer monitor? I'm looking for a less than $75 solution, given the cost of LCD monitors. The target panel is one out of an eee pc. With the potential for some others, depending on how many dead laptops I still have hanging around.

It seems i'm not the only one tryign to drive lcds.

Here's someoen driving a conventional LCD display from a beagleboard.
http://axio.ms/projects/beaglelcd/

And here's some more information discussing LVDS.
http://forums.afterdawn.com/thread_view.cfm/98801

It turns out the motherboard in my webserver has a LVDS port. But I'm yet to see a laptop display with a complete, normal lvds connector. ARe there adapters?

This seems to come to the same conclusion I did. IN that I need some sort of controller to go from hdmi/vga/composite to the LVDS that a laptop display is looking for.
http://ask.slashdot.org/story/00/06/29/0345250/can-you-reuse-a-laptops-lcd

The price seems a bit steep on this, but these guys sell a controller card for LCDs. Also the fixed resolutions may be a problem.
http://store.earthlcd.com/LCD-Products/LCD-Controller-Cards

I think life would be a lot easier if i could find the datasheet to one of these LCD displays.

Bad Munki posted:

So I'm in the process of buying a house which I'll be moving into at the end of May. *sni8p*
What silly poo poo would you do to a house if you had very few restrictions and time/money weren't really a problem?

I am two weeks from closing on a house.

Thermostats are "dumb". I like the idea of saving money on heating and cooling. It wouldn't be hard to coordinate the heating, cooling, and windows to provide the cheapest temperature control. A few thermistors and motors are going to be on the menu. A threaded rod, and a dc motor, with end stop switches will make a dandy device for opening windows.

The house also has a crawl space exhaust fan, it will be interesting to see what that can do to house temps.

We'll see what I come up with. I plan on having compressed air available throughout the house, so it's also not out of the question to use cheap/small/tiny solenoids with pneumatic cylinders to operate things. Hmm.... that's actually a good idea. That could allow windows to "fail closed" instead of "fail stationary."

Mmm home automation.

# ? Mar 28, 2012 09:28

insta: Jan 28, 2009

I think a likely time for a DIY automated window system to fail is when the rest of the house is burning down. Might want to reconsider how they fail

# ? Mar 28, 2012 14:10

movax: Aug 30, 2008

Nerobro posted:

How would you drive a laptop LCD as a computer monitor? I'm looking for a less than $75 solution, given the cost of LCD monitors. The target panel is one out of an eee pc. With the potential for some others, depending on how many dead laptops I still have hanging around.

This gets difficult mostly due to the lack of documentation and standardization. The connectors and cables differ from model to model, but LVDS itself is fairly simple (with appropriate level-shifters/etc when needed). I think most modern connectors will be difficult to breadboard.

quote:

It turns out the motherboard in my webserver has a LVDS port. But I'm yet to see a laptop display with a complete, normal lvds connector. ARe there adapters?

This seems to come to the same conclusion I did. IN that I need some sort of controller to go from hdmi/vga/composite to the LVDS that a laptop display is looking for.
http://ask.slashdot.org/story/00/06/29/0345250/can-you-reuse-a-laptops-lcd

Software gets a bit more difficult, especially on the x86 side. On something like the Intel 5 Series, the integrated GPU (Ironlake) generally gets configured through Video BIOS calls (INT15h/INT10h). I setup target displays with information like color depth, backlight polarity, spread spectrum, ALS support, etc through those calls. I also can setup many different panel types in the Video BIOS when I compile it, and in each of these I can setup SSC, monitor timings (dot clocks, etc), power-up sequences, etc.

There's no way for the end-user to really get this information (for their LCD), and the datasheets are usually buried deep in a mound of useless Chinese search results. I think we had to pull teeth to get the information we needed out of the random Chinese place this screen came from.

quote:

I am two weeks from closing on a house.

Star Trek doors everywhere

# ? Mar 28, 2012 15:36

movax: Aug 30, 2008

longview posted:

I'm doing a project for school now and part of it involves sensing an analog temperature sensor (AD590), and the most promising ADC chip we have available is the AD7575 but the datasheet was somewhat confusing:

It only speaks of the Vref as being 1.23V but the datasheet also says it can be as high as Vdd, does anyone if it will work properly with the Vref at say 2-4V instead?
Same thing with the AGND, it's allowed to be as high as VDD according to the datasheet, and it'd be very practical to just set that to about 2V to get the most out of the digital range.
So do I need to ensure my voltage range is 0-2.46V as the datasheet says or can I just adjust the references to higher values?

And yes, I know I could just use a built-in ADC on the microcontroller, but it's not an option.

Hey your post got lost I think, just reposting it here so people can see it again.

Your temp sensor (datasheet) looks like it outputs a current proportional to the temp. If you can change the sensors, something like the DS18B20 (1-wire bus) or the Analog Devices TMP3x series would likely be easier to interface. By using the Dallas part, you could bypass the ADC entirely I think.

To answer your VRef question though, it'll need to be at least as high as the input voltage you're measuring. In the 5V range, a 12-bit DAC would be overkill (something ~1mV of resolution I think).

# ? Mar 28, 2012 15:46

longview: Dec 25, 2006; heh.

movax posted:

Hey your post got lost I think, just reposting it here so people can see it again.

Your temp sensor (datasheet) looks like it outputs a current proportional to the temp. If you can change the sensors, something like the DS18B20 (1-wire bus) or the Analog Devices TMP3x series would likely be easier to interface. By using the Dallas part, you could bypass the ADC entirely I think.

To answer your VRef question though, it'll need to be at least as high as the input voltage you're measuring. In the 5V range, a 12-bit DAC would be overkill (something ~1mV of resolution I think).

Thanks for the repost. It's not really practical to change any of the components since we don't have the option of ordering anything in, the lab has that sensor and that (plus a few other) ADCs in stock and the assignment specifically says we have to use a dedicated ADC chip. I'm going to use my own Arduino dev. kit for the micro since it's much easier than PIC18 assembly for prototyping.

Other options include the AD570 ADC which requires a -15V supply and I want to keep this on a 5V single supply. We can also use thermistors to measure the temperature, but I suspect linearizing them will be more effort than using the AD590.

By putting a 10k resistor in series with the sensor it should give a roughly 2-4V measurement range so I wanted to put the AGND at 2V and VREF at ~4V (though the datasheet says max input is 2*VREF, presumably that's 2*(VREF-AGND) if it's possible).
Otherwise we can use some opamps to subtract a few volts (subtracting 2V or so should work), but I'm concerned about the accuracy then so I want to avoid it if possible.

# ? Mar 28, 2012 16:16

Adbot: ADBOT LOVES YOU

# ? May 12, 2024 17:51

Slanderer: May 6, 2007

longview posted:

I'm doing a project for school now and part of it involves sensing an analog temperature sensor (AD590), and the most promising ADC chip we have available is the AD7575 but the datasheet was somewhat confusing:

It only speaks of the Vref as being 1.23V but the datasheet also says it can be as high as Vdd, does anyone if it will work properly with the Vref at say 2-4V instead?
Same thing with the AGND, it's allowed to be as high as VDD according to the datasheet, and it'd be very practical to just set that to about 2V to get the most out of the digital range.
So do I need to ensure my voltage range is 0-2.46V as the datasheet says or can I just adjust the references to higher values?

And yes, I know I could just use a built-in ADC on the microcontroller, but it's not an option.

Let me just say that that ADC is an antique piece of crap, and you would have been much better off getting free samples of a slightly-more modern DIP ADC, or spending $5 on a modern ADC on a breakout board.

That said, you can't do what you want, at least not really. The absolute max rating on the AGND pin says it can be high, but that doesn't mean poo poo will work. AGND and DGND are not isolated, not really--they still need to be connected somewhere. However, the chip breaks out both so you can connect AGND to a small AGND net/plane (it mentions single point, so, I dunno, maybe a net routed as a star?) However, this still needs to be connected to your digital ground--but only at a single point. Through magic that few can learn without going truly mad, this helps isolate digital and analog noise...but will cause lovely ground loops and stuff if done wrong.

If you really, really want to change your negative reference voltage, you'll need to build a differential DC amplifier to shift the signal down (and amplify/attenuate if desired) so you end up with a single-ended ground referenced signal.

Or you could go the easy route and use the loving thermistor. The equations are easy as hell, and if you don't want to compute them on the fly, use approximations or a lookup table for the range of interest.

EDIT: I'm also surprised that you don't have any simple TI voltage-output 3 lead temperature sensors. Despite their age, they are hilariously cheap, and have example circuits for everything from remote sensing to conversion between temperature units with just resistors and elbow grease (or converting to a scaled voltage that you can read from a standalone voltmeter display)

Slanderer fucked around with this message at 20:00 on Mar 28, 2012

# ? Mar 28, 2012 19:56

The Something Awful Forums > Discussion > Hobbies, Crafts, & Houses > Learning electronics: Release the Magic Blue Smoke

«‹›781 »