Digital oscilloscope Project

[fobyellow]

digital oscilloscope +fpga

any one have done a sample one?
i've made a acquisition board with 2*AD9215, and now writing code for UF32 USB-chip, next week the PCB might be OK

 

[angus]

spartan3 based dso

I really like all of the ideas presented here in this forum. I have been thinking recently about tackling a similar project and it's nice to know others think it is possible.

Instead of using a PC display interface, though, I have been thinking about having a built-in LCD display. I'm pretty new at all of this, so how would you do this? Thanks.

 

[Gonzakpo]

protel 99 projects

Why everybody use fpga ????

I'm also "trying" to design my personal DSO, but I found almost imposible to use a fpga because of 2 reasons:

1. Expensive (I've never used them so I don't have even the programmer) And also I'm from Argentina and I can't afford almost 500 u$s for a board...
2. I don't know how to use or program them :S......BUT it would be nice to learn about.

As far as I've researched, I can connect the adc directly to the usb ic and send samples all the time at a speed of 30Ms/s (16-bit) or 60 Ms/s (8-bit)....
And leave the trigger issue for the software....

That was my initial plan, but I wasn't convinced with the sampling rate.....

And now I don't know what to do, I'm kind of lost. I want to achive at least a 100 Ms/s rate!! (12-bit)...

I've seen lots of projects that uses memories to memorize all the samples and then send them but I really don't see the point of doing that because at 100 Ms/s you need a huge memory to store a 1Hz signal (for example). Ok, lets supose now that when the memory is full you transmit the stored data to the pc and when the memory is emty again you start receiving samples again. But what happends with all the samples you lost while you were sending data to the computer??...

Maybe I'm misunderstanding something.......Explain it to me if I'm wrong.

DOES ANYONE KNOW A BETTER WAY OF ACHIEVING SAMPLES RATES OF 100 Ms/s (or higher)??? ...(remember, without using a fpga because I can't afford them)..

 

[CMOS]

how to build digital oscilloscope

Why everybody use fpga ????

Because they are the fastest available devices in the market today.

1. Expensive (I've never used them so I don't have even the programmer) And also I'm from Argentina and I can't afford almost 500 u$s for a board...

Check out http://www.digilentinc.com/Products/Detail.cfm?Prod=S3BOARD&Nav1=Products&Nav2=Programmable the sell FPGA boards at $99 !!

2. I don't know how to use or program them :S......BUT it would be nice to learn about.

www.fpga4fun.com will be a good starting point.

I've seen lots of projects that uses memories to memorize all the samples and then send them but I really don't see the point of doing that because at 100 Ms/s you need a huge memory to store a 1Hz signal (for example).

They use variable sampling rate. So you dont' require large memory for low frequency signals.

But what happends with all the samples you lost while you were sending data to the computer??...

Thats call blindness problem and it is present in almost all PC based DSO's unless they are using large FIFO buffers, capturing and sending data simultaneosuly.
I read somewhere, PC base scopes remain blind for 99% of the time because they are busy sending data to PC!!!

DOES ANYONE KNOW A BETTER WAY OF ACHIEVING SAMPLES RATES OF 100 Ms/s

Not without FPGA/CPLDs and huge memory...preferably a FIFO.

 

[Gonzakpo]

portable oscilloscope avr

thanks for the answer.....

I have a doubt about the variable sampling rate.....
How can I know the frecuency of the input signal ??....The only solution I can imagine is to change the sampling rate depending on the time base you've selected.
But I just doesn't make sense because I wouldn't be able to measure high frecuency signals using a large time base....(because of the low sampling rate)...

There must be another solution, like autodetection, but what would happen if the input signal isn't a "repetitive" signal.

The variable sampling rate doesn't convince me :S...maybe you know a better way to aply variable sampling rates???......

Regarding the FPGA's .....which model should I chose to start? (and keep using it for the rest of my life ....I just don't like the idea of buying several boards JUST to see which is better) ......and...does the starter board comes with a manual???...I need something to learn from!

Last question, does anyone have a nice input circuit schematic for the a/d ?? (with "over tension" protections)

Thanks.

 

[CMOS]

circuit schematic oscilloscope pc

The only solution I imagine is to change the sampling rate depending on the time base you've selected.

Thats right!

But I just doesn't make sense because I wouldn't be able to measure high frecuency signals using a large time base....(because of the low sampling rate)...

Actually variable sampling rate should be used along with variable low pass filter to avoid aliasing problems.

There must be another solution, like autodetection, but what would happen if the input signal isn't a "repetitive" signal.

Well in that case intelligent hardware trigger design and good software calculations will be the only solution

The variable sampling rate doesn't convince me :S...maybe you know a better way to aply variable sampling rates???......

Go for large memory (GBs) then

Regarding the FPGA's .....which model should I chose to start? (and keep using it for the rest of my life ....I just don't like the idea of buying several boards JUST to see which is better) ......and...does the starter board comes with a manual???...I need something to learn from!

Spartan-3 from Xilinx are good FPGAs to start with and cheapest too!

Last question, does anyone have a nice input circuit schematic for the a/d ?? (with "over tension" protections)

Check out download section of www.bitscope.com , they have old bit-scope schematic available.

 

[Gonzakpo]

avr scope project

WOW! that was fast! je

Thank you soooo much for your answer...

I'll consider learning the fpga stuff, It's a nice idea. But, when you desing something HOW CAN YOU SOLDER THAT ???? (I know that the word "solder"
doesn't exist, but I wasn't sure how to say that.......just to be sure that you've already noticed, I speak spanish )


Two more questions:

As I've said before, I can design the dso to sample all the time and send the samples to the pc at the same time. Without fpga and memories. (using a sample rate of only 30 Ms/s) ...but the question is: I've read somewhere that I CAN'T do that because the OS doen't work in real time, so I can't receive samples all the time....... Is that true? ...(I HOPE NOT :s)

How should I chose the correct trigger level?......


Well, I think that's all for now..... Thanks!

EDIT: I've been exploring the bit scope web and I found out that the sell 8bit oscilloscopes!!, that's a really poor resolution!....I was planing to use at least 12-bit, or maybe 14. ....

 

[CMOS]

digital scope circuit

I'll consider learning the fpga stuff, It's a nice idea. But, when you desing something HOW CAN YOU SOLDER THAT ???? (I know that the word "solder" doesn't exist, but I wasn't sure how to say that.......just to be sure that you've already noticed, I speak spanish )

You solder them using hot-air soldering station.

As I've said before, I can design the dso to sample all the time and send the samples to the pc at the same time. Without fpga and memories. (using a sample rate of only 30 Ms/s) ...but the question is: I've read somewhere that I CAN'T do that because the OS doen't work in real time, so I can't receive samples all the time....... Is that true? ...(I HOPE NOT :s)

No windows is not a Real-Time Operating System. If you want to make a scope without memory, you should use Real Time ports of Linux like RTAI or RTLinux.

How should I chose the correct trigger level?......

That depends on your input signal level. If you want auto-trigger, the trigger level should be 0V or half the ADC resolution.

EDIT: I've been exploring the bit scope web and I found out that the sell 8bit oscilloscopes!!, that's a really poor resolution!....I was planing to use at least 12-bit, or maybe 14. ....

Thats not poor....you'll hardly need resolution better than 8-bits.

 

[Gonzakpo]

avr oscilloscopes

why you say that 8.bit resolution is not poor???

For example, if I want to mesuare signals of 60vpp with only 8 bits i would be having a resolution of aprox. 0.2 V !! ....
I think that's a lot.

I'm disappointed with the matter about real time transmition :S :S
That way everything was soooooooo simple..... only an adc and a usb peripheral controller...
Is really imposible to do that??

Thanks!!

 

[CMOS]

cheapest fpga

For example, if I want to mesuare signals of 60vpp with only 8 bits i would be having a resolution of aprox. 0.2 V !! ....
I think that's a lot.

Would you ever require 0.2V accuracy at 60V? Think over it. If it of that much impostance you should go for 10 or 12-bit ADCs then.

I'm disappointed with the matter about real time transmition :S :S
That way everything was soooooooo simple..... only an adc and a usb peripheral controller...
Is really imposible to do that??

I didn't say impossible. It is possible on RTAI/RTLinux platforms but not on windows.

 

[Gonzakpo]

digital oscilloscope using fpga

Hello

Does anybody know an operacional capable of allowing higher tensions in the input than the power suply ones?

My idea is the next one: I want to develope an eficient protection for my adc. The idea is to power the operational between the adc conversion limits so if a higher tension input is received, this tension will be cut between the suply limits in the output of the operacional protecting the adc.
And also it's an useful idea because I can visualize instantly from the oscilloscope screen that something is wrong in the input signal (it is cut).

If anybody knows a similar way for protecting the adc please tell me!....
(I want some protection that doesn't stop the oscilloscope operation)

Thanks!

 

[Nimer]

20mhz digital oscilloscope schematics

hi

didnt understandyour idea well...elaborate more

regards

 

[Gonzakpo]

bitscope dso slow

It's simple,

The idea is to suply an operational amplifier (OP) between the adc input maximun voltage ranges....
For example between +0.5V and -0.5V.... (maximun input voltage range of my adc is 1Vpp)
Also, configure the GAIN of the OP to 1. (A=R2/21)

And finally, (and this is the feature I'm looking for).... This OP MUST allow higher voltages (in the input) than the suply onces (-0.5V / +0.5V).....
For example 30V!!!

The purpose of this idea is to "cut" the part of the signal wich exceeds the input maximun voltage ranges.
I'll will upload aa example drawing for you. The problem is that I don't know how this process is called in english.

Remember, the OP I'm looking for MUST allow higher voltages (in the input) than the suply onces!.

Thanks!.

 

[Ave_Rapina]

pc oscilloscope circuit schematics

Hello
why not use an simple zenner with the tensio you want to protect
for exemplo maximum is 5V you can put an 5.5v zener with an lower resistence a limit the input tension
if yo need for + and - tension just use 2 in serial in reverse one of the other.
This increase a little bit the limit tension but if you use an 5V zener the limit tensio is 5V +0.7V from the other zenner

The reverse is equal

Regards

 

[Gonzakpo]

usb digital oscilloscope schematics

Yes, it's a posible solution, but I don't know why, I don't trust in zeners

It's a very expensive design to trust in a simple zener,

Anyway, I'll keep that solution in mind. Probably I'll end up doing that.

Thanks.

Any more ideas?

EDIT: I've just figure out something. My Adc limits are +0.5V to -0.5V so a zener or a diode aren't useful in this case because the adc limits are lower than any zener or diode "triger voltage"... :S

 

[jackson1985]

firewire oscilloscope cable

hi guys can ny1 help me in guiding me a good project (final year) related to my subject i.e electronics n instrumentation...

 

[melc]

www.jyetech

Why everybody use fpga ????

Because they are the fastest available devices in the market today.

Except maybe the Ghz range DSPs... which aren't so affordable but available on the market:
[I]**broken link removed**[/I] or: [url]http://focus.ti.com/lit/ml/sprt415/sprt415.pdf

or much better if you need to be in the top of the high speed DSPs:

**broken link removed**
[/url]

 

[Giuss]

epm3064 programming

I'm working to an homemade DSO too and my prototipe is working but needs improvents
This is a short summary of my project :

ADC TLC5540
CPLD Xilinx XC9572
RAM Cypress CY7C199
CPU PIC16F877
30 Mhz clock

Now I'm working to improve it :

Going to 40 Mhz clock

Add a digital trigger

Add an USB connection

More to come...

 

[funghome]

open source digital oscilloscopes

i am interesting to this project.
http://www.cmccord.co.uk/FYP/index.htm
but while file is final hex file in
http://www.cmccord.co.uk/FYP/DownLoad/PicCode.zip

 

[CMOS]

pc oscilloscope scheme

i am interesting to this project.
h**p://www.cmccord.co.uk/FYP/index.htm
but while file is final hex file in
h**p://www.cmccord.co.uk/FYP/DownLoad/PicCode.zip

It is a very low frequency scope (5KHz) and practically of no use!!