Oscilloscope and vertical offset

[cyberfish]

The vertical offset adjustment knob on oscilloscopes - does amplification happen before or after vertical adjustment?

I don't have access to an oscilloscope right now, so can't check myself.

From my memory it's amplify then shift, but wouldn't that make shifting kind of useless?

I thought the point of vertical shift is to be able to see, say, a small AC signal with big DC bias. But I guess AC coupling would do, too.

By the way, I'm trying to design an oscilloscope, that's why I'm asking.

If it's amplify then shift I wouldn't need any circuitry for shifting right?

Thanks

[TrickyNekro]

The vertical offset adjustment knob on oscilloscopes - does amplification happen before or after vertical adjustment?

I don't have access to an oscilloscope right now, so can't check myself.

From my memory it's amplify then shift, but wouldn't that make shifting kind of useless?

I thought the point of vertical shift is to be able to see, say, a small AC signal with big DC bias. But I guess AC coupling would do, too.

By the way, I'm trying to design an oscilloscope, that's why I'm asking.

If it's amplify then shift I wouldn't need any circuitry for shifting right?

Thanks

To see small AC signals you need AC coupling as you said... vertical offset won't do.

You need vertical offset for mainly 2 reasons.... (at least for analog scopes)

First of all to get better resolution.... instead of 2V and 4 divs, you can get 1V to 8divs..
That's nice :-)

The other reason is to display multiple channels on the same screen (of course you only have one screen)...

That's mainly it... ;-)

[Soeren]

Hi,

Vertical shift won't give you any better resolution (that's handled in the vertical gain), but shift is handy in so many situations, like when you need to measure a waveform, you can bring the interesting part of the waveform to the measuring strips (X and Y) in the middle of the screen. When you need to just study a part of the waveform, you can move anything else out of the screen, to make it easier to read - both the X- and Y-shift is a feature that you don't want to be without. I cannot tell you all the situations where I use it, since it's something that you don't think about, when you have lived most of your life with 'scopes on your bench and you use shift together with gain and timebase, but I wouldn't even consider using a 'scope without that facility.

Then again, I wouldn't consider building a 'scope, as it will cost more than getting a brand new one, even if I set the value of my own work during the incredible amount of hours it would take to $0.00.

[TrickyNekro]

Yes... that's what I meant in the first post... Ehhh.... why must I get beat down when I'm bored to write down some more... words...
Vertical swift "alone" won't give better resolution...
But when using vertical swift along with gain "not amplification  " you can get better resolution...
If fact gain is much more appropriate cause it's logarithmic in nature.
It can never be zero, it just tends to zero "too much to see" 

Last time I checked when using 10V/div signal is not amplified if the signal is in fact below 10V.... (or multiplied with a greater than 1 factor)... 
Isn't that... nice humor 

Anyways, just joking here...
Soeren is pretty much right about the scope too, unless you plan to make something easy just for fun....
Like the Axon scope.... no hard feelings here guys

Best Regards, Lefteris
Greece

[cyberfish]

Yeah I know it's important, just not whether I should implement it in hardware (op amp summer) or just software.

I think I will do hardware, though, otherwise the amplification range with offset will be very limited. (in hardware, I can add then amplify)

The purpose of the project is mostly for learning, but if all goes to plan, I think I will be able to make something cheaper than anything comparable on the market.

It's going to be a 100Msps/~30MHz 2 channels USB scope (about 66666 times faster than axon scope ).

The cheapest comparable scope I can find on eBay goes for $200.
http://cgi.ebay.ca/100MS-s-PC-Based-Digital-Oscilloscope-DSO-2090-DSO2090_W0QQitemZ370384313233QQcmdZViewItemQQptZBI_Oscilloscopes?hash=item563ca01b91#ht_5112wt_1049

and has no Linux support.

I just don't see why they have to be so expensive.

Here's what I'm planning on using -
$9 Spartan 3AN FPGA (smallest TQFP-144, highest speed grade, can probably be clocked to 200MHz)
$20 THS7002 dual channel programmable gain amplifier (70MHz -3db BW)
$9.5*2 ADC9283 100Msps ADC
$8 AD7399 dual channel DAC (for setting offsets)
$4.5 FT245 USB to FIFO chip
$30 for PCB (BatchPCB, $10 setup fee + ~$20 for 9 sq. in. 2 layer)
$15 for 2 cheap BNC probes (BW >= 40MHz, just about anything)

The rest should all be negligible (passive components, $1 chips, etc).

Microcontroller and RAM for buffering can both be synthesized on the FPGA (it has ~7KBytes block RAM). Amazing what a $10 FPGA can do.

Even cheaper for me because I can get most expensive chips for free as samples (most companies have no problem giving students free samples).

I don't see it going anywhere near $200, especially if PCB is mass produced (not that I currently have plan for that =P).

Of course, this is all assuming my time is free, which is a pretty close estimate .

Software part should be easy, since that's my specialty. Will be cross-platform and open sourced along with the rest of the design if I ever finish it. I'm surprised no one has done anything like this yet. Or I didn't look hard enough.