Hi all.

I was thinking of do a laboratory for electronic/electricity with the PC, mean that I will like to test diferent electric propierties.

I see that voltmeter and others are in some way easy, the one that should be a problem will be the oscilloscope.

I was thinking that I will need a thing or interface for connect it to my PC via parallel, serial or usb, this thing should send the data not in an analogic way but digital way.

Actually I know I will need a ADC and a PIC, and that the resolution of my oscilloscope will be if I have a ADC of 10Khz the max that I can output a graph is about 5Khz.

But I question, is posible make no use completely of this things?, ie.. do them by myself? and speed up them a little?


I will say you that I am not good at electronic I normally dont enter to that class, also was for that I decide do this thing, because I normally dont enter to the lab at school (and I cant specially use a oscilloscope at house) and is a way that I consider I can be interested in learn more about the subject... .

The questions that I have first are related to what subjects should I know?
- Digital signal processing.
- Clock of a circuit (speed up) .
- Compresion of data in circuits and conversion to a floating point format in circuits.
- Ports of the PC.
In the regard of electronic what more?
- Drivers for the app and the external interface.
- FFT or DFT or both.
- Sampling.
- Fast or direct access to memory (DMA in windows is posible?)
in the regard of software what more?


I have readed some post in this subforum, and I guess some people have that idea before, I will like to know if you ahve done it.


Actually I have finded some sources for oscilloscopes, but like I understand, they read a file in the disck and process it. I will like connect the coaxial and connectors to the interface and process the signal directly without save the signal on disck (but I gues I should use some like save some parts of data when not used...), more using only memory.


I talk with my teacher and he say that I need at less ADC804 with 10kHz and I can sample to 5Khz.

Then my questions are... if I will like to use some things that are around there ADCs PICs what I should must know?. And if I will like to do myself almost all without ADC and PIC help, what I must know?




I am a person that like to know the subject first before do nothing, then if you give my pointers I will first understand and then design.

Have a nice day or night.
Posted on 2004-09-23 08:24:04 by rea
Digital Oscilloscopes, built on a PC platform have been around for a while now... at least over 10 years, which makes em pretty obsolete ;-)

Two sources for them are National Instruments (good, expensive) or Measurement Computing (very good, bit cheaper). Dataq has some new productsthat are very inexpensive, while I've never used them they look first rate.

A 10KHz sample rate is very small, concidering an audio signal can go up to 20KHz (min 40KHz sample rate). CD's sample at about a 44KHz I believe.

Keep in mind if you're building your own A/D, you need to keep the input in range. Most ICs need to see 0-5V or maybe -5 to +5 at the input, so you'll have to scale up or down your signal. Also a nice high input impedance is needed so you do not load the thing being measured.

Any FFT or DFT processing is 'just software' done on the bit stream after it is inside your PC. That part of the project can wait till you get the hardware working.
Posted on 2004-09-23 15:22:13 by Ernie
As soon as I posted I noticed Microchip has a package for doing measurements over the USB. It uses the PICKit 1 as a base, plus a daughter board that does the conversion.

http://www.microchip.com/stellent/idcplg?IdcService=SS_GET_PAGE&nodeId=1406&dDocName=en020777&part=AC164120

You'll need:

PICtail Signal Analysis Daughter Board $25 at Digi-Key

PICKit 1 $36 at Digi-Key



I'm going to get me one of these soon as I like toys.
Posted on 2004-09-23 15:39:14 by Ernie
Uhm why don't you use the soundcard output/input for sampling? You can get SB Audigy for $43 (like I did) and you'll get 96kHz samplerate + low latency if you need it.
Posted on 2004-09-23 16:26:16 by Ultrano
Uhm why don't you use the soundcard output/input for sampling? You can get SB Audigy for $43 (like I did) and you'll get 96kHz samplerate + low latency if you need it.


WHA?!?!?! $43?!?! I paid more than that for my Sb Live! 5.1 last yr!!
Posted on 2004-09-23 18:31:27 by x86asm
Yeah, I was worried if there's some mistake, but now that I have this superb quality here I no longer doubt it :grin:
now my shitty TDA-based amplifier sounds better than the STK-based amp, playing with a Sony DVD player in the other room! Was a bit unexpected. And with its EAX support, games can sound MUCH more reallistic (there's a demo game included to show off the differences). Pity that the games I play don't use EAX ^^" .
Posted on 2004-09-23 18:39:59 by Ultrano
I guess that there where from time a go :). Altought I still whant to do it. I will like if you have the time answer the questions above remarked with blue ;).

Ernie, like I understand you say that I can do the ADC and still use a PIC? I have managed to find a www.DSPguide.com I need to read it for do the ADC? or this book will help me on the software part?, also checking the link, interesting, but I whant to pass litle time hurting my head ;), altought I will use that like a parameter or a objetive pheraphs.


Ultrano, I will see that option, altought the majority of motherboards have it and I can use them, but I was reading that normally will only serve for sound range, also if a "decent card" will introduce pheraphs some modification to the signal, but is a good choice if whant to do more easy in the hardware side.




I have a question, I was reading that for a normal use, ie projects at house a osciloscope would be 40Mhz, also after reading some things about other circuits I read that it would be about 250Mhz, 500Mhz or 1Ghz. I only whant question how posible is do a digital oscilloscope based on the PC of 40Mhz, 250Mhz, 500Mhz and 1Ghz?.
Posted on 2004-09-24 08:09:38 by rea
Here is a description of How Oscilloscopes Work.

It first describes analog scopes, and then it describes digital scopes.
Posted on 2004-09-24 23:54:17 by tenkey
btw, hgh, with Audigy you don't have distortion higher than 0.001% !!! It's got 100dB S/N ratio :D and it's 24-bit 96kHz sound, something you can never accomplish with USB or parallel port.
Posted on 2004-09-25 04:26:13 by Ultrano
Then work with some of this ports is not a option?

In the sound cards what about PCM, I dont know exactly what is, but I read somewhere that it afect the input? is that true? in what way...?

If not I was trying only for see if posible a PCI interface, I have finded this one http://www.electronicproducts.com/ShowPage.asp?SECTION=3130&PRIMID=&FileName=FEBHL13.FEB2001, also is some like what I whant. Only for know also the PCI you have a resource?.


I will start with the most basic some little experiments with parallel and serial and then let see how I can do with sound cards (I think I will not do for specific sound card...?), but get the capacity of the audio card.
Posted on 2004-09-25 08:04:55 by rea
hgb, I tell you - the soundcard is the best way. And I repeat, Audigy is an awesome soundcard, where you get 0.001% distortion. Otherwise you'll have at least 0.1% distortion with ANYTHING else you use! I have built many electronics connected to the parallel, and the speed there is drastically lower.

The MIDI keyboard is home-made, the PlayStation gamepad is also connected with the parallel port, also I have a DDC connection to my sister's PC via the parallel (I'd love to have 3 parallels instead of only one :grin: )
With the parallel one thing is sure - you will never get the data on perfectly same intervals of time - so sine waves will sometimes look like triangular if you create that project of yours!

Yes, you've read right about soundcards being distorting and stuff. But that's for every soundcard below SB Live 5.1. And the Audigy is 10 times better than SB Live :).

You can create PCI card, but it'll be too hard, and there is a risk of sacrificing your mainboard ^^" . And you'll have to put a lot of electronics to get the image right.
Posted on 2004-09-25 12:14:30 by Ultrano
I guess that there where from time a go :). Altought I still whant to do it. I will like if you have the time answer the questions above remarked with blue ;).

Ernie, like I understand you say that I can do the ADC and still use a PIC? I have managed to find a www.DSPguide.com I need to read it for do the ADC? or this book will help me on the software part?, also checking the link, interesting, but I whant to pass litle time hurting my head ;), altought I will use that like a parameter or a objetive pheraphs.


Ultrano, I will see that option, altought the majority of motherboards have it and I can use them, but I was reading that normally will only serve for sound range, also if a "decent card" will introduce pheraphs some modification to the signal, but is a good choice if whant to do more easy in the hardware side.




I have a question, I was reading that for a normal use, ie projects at house a osciloscope would be 40Mhz, also after reading some things about other circuits I read that it would be about 250Mhz, 500Mhz or 1Ghz. I only whant question how posible is do a digital oscilloscope based on the PC of 40Mhz, 250Mhz, 500Mhz and 1Ghz?.


Howdy HGB,
yeah, that is possible, but that kind of oscilloscopes are not getting all of the samples :), there's a little trick sampling might be 10MHz in an 1GHz oscilloscope, IF a sampling is made in every 100 s periods, which has to be made very preciese, it has to synchronize the ramp to the signal, then scan the period from the beginning to the end by getting a sample each time a little later. this is far good enough for most cases in High frequency applications, you can see the wavefrom's shape and frequency.
these are usually digital scopes.

as of i know now, there is no way to make a catode tube capable of directly displaying 5GHz signals, and might never be.

MATRIX
Posted on 2004-09-30 07:42:16 by >Matrix<
Currently,
with some help on the software side i could make a PCI CARD,
i can do cad, (develop PCB/circuits), make PCBs in small series, and i can assemble PCBs with SMD Parts, i know a little programming too

MATRIX
Posted on 2004-09-30 07:46:24 by >Matrix<
That would be nice :), altought I am now in the moment of adquisition of info about the project (reading the DSP book above, and reading some sources about signal processing), in the part of software I will do a research to, I can help but I first must know (mean that actually in that way I am beginer) how make drivers for the electric circuit.

For the moment based in the article "how oscilloscopes work" I see that the only thing that is not necesary here in the electronic part is the display..., altought I dont know how much process would be in the card and how much process will be passed to the software side and I think that would be nice also implement a compresion of data and get x-samples compressed, I remember I read that bitRAKE talk about a compresion algoritmh that is not necesary to uncompress totally for get the data (a direct access) .



>Matrix< I hope I can do it. You will try it yourself?, have something specific in mind?, pheraphs we can interchange information and pass fun time.

You think that you can say how many will be the cost of the PCI card?


Have a nice day or night.
Posted on 2004-10-01 08:11:47 by rea
May we continue PM?
send me email, then maeby AIM

MATRIX
Posted on 2004-10-01 10:41:59 by >Matrix<
That would be nice :), altought I am now in the moment of adquisition of info about the project (reading the DSP book above, and reading some sources about signal processing), in the part of software I will do a research to, I can help but I first must know (mean that actually in that way I am beginer) how make drivers for the electric circuit.

For the moment based in the article "how oscilloscopes work" I see that the only thing that is not necesary here in the electronic part is the display..., altought I dont know how much process would be in the card and how much process will be passed to the software side and I think that would be nice also implement a compresion of data and get x-samples compressed, I remember I read that bitRAKE talk about a compresion algoritmh that is not necesary to uncompress totally for get the data (a direct access) .



>Matrix< I hope I can do it. You will try it yourself?, have something specific in mind?, pheraphs we can interchange information and pass fun time.

You think that you can say how many will be the cost of the PCI card?


Have a nice day or night.


I thought i comment you here once more as long as you don1t write mail,
PCI has a Bandwidth of 32 bits, if runs at 66MHz it can easily do 80Mb / s
the hardware could be a single / double channell scope for example- an ADC. Hardware isn't that complicated, only a good PCI datasheet is needed. In software, you 'd have to write a device driver for it, first time it'd be nice to make it work for example on simple dos.
Oscilloscope: it is not a complicated thing however synchronization could be made with hardware or software, you know, in cathode tube scopes there is a triangle waveform which deflects the electron beam, which must be started at some trigger level or slope speed. ( edge / level trigger )
in some cases lissjaous curve is needed to determine some parameters, where you put 2 inputs one at y deflection amplifier and the other at x deflection amplifier. you'll see some curves going around.
Of course it is no problem to make any changes on the board, first one may be a single ADC for PCI communication testing ( 486 has PCI so testing is not a problem :) ) , then it may be upgraded and made professional.
prices are not high cause' there are many pci cards around here and ics can be used in some cases free. anyway it doesn't really matter how much it costs, it whould be surely below commercial product's price.
If everything is ok i will make the connectors gold plated so quality is not a problem.

ps.: currently i'm working on automatization ( like testing circouts with computer/ assembling boards / drilling holes ) cause' i will not have enough time to assemble my boards.

MATRIX
Posted on 2004-10-01 13:47:59 by >Matrix<
No, I work now I return from work(not programming or electronic related, altought I like work with my hands), and now only eat and go to school at day I have like 3 hours for me :S.

ps.: currently i'm working on automatization ( like testing circouts with computer/ assembling boards / drilling holes ) cause' i will not have enough time to assemble my boards.


Nice, I was thinking in some like that when you say:

make PCBs in small series, and i can assemble PCBs with SMD Parts


That is a nice project too :).

only a good PCI datasheet is needed
I wasnt able to find one in the past, altought not a deep search and pheraphs not the correct words "PCI interface".

ics can be used in some cases free
Are you talking about some like reciclate?



I send you a litle mail.
Posted on 2004-10-01 14:19:25 by rea
You don't need to use a PIC and you certainly shouldn't use an ADC0804 they are far too slow as your teacher suggests (although his estimate is terrible).

The ADC0804 has a 100uS conversion delay...
Take a look at the ZN448E basic specs:

8-bit successive approximation ADC
9us conversion time (110kHz max sampling rate)
+5 volt (4.5 to 5.5v) supply, 25mA current consumption
inbuilt 2.55 volt reference
Max error: +/- 0.5LSB
18-pin DIL package

You should be able to sample well over 100khz signals with this baby :)
Far more useful, yes? Still too slow for many things but for the price...
Posted on 2004-12-26 19:09:18 by Homer
I have done little in this way (very little), altought I think I understand more wha is there, I have watched a little project of my friends regarding to voice adquisition IIRC with a ADC 804 :), but I havent the knowledge for do what I whant unfortunately...
Posted on 2004-12-27 19:52:16 by rea