From: John Larkin
Subject: Re: High speed ADC evaluation
Date: Wed, 06 Nov 2002 20:45:40 -0800
Organization: Posted via Supernews, http://www.supernews.com
X-Newsreader: Forte Agent 1.91/32.564
On Thu, 07 Nov 2002 03:56:33 GMT, "Thomas Magma"
>We are doing some ADC development and have selected the Texas Instruments
>TLV5535, 8 Bit, 35 MSPS ADC Single Channel to suit our needs. We have
>purchased the eval kit TLV5535EVM to hopefully evaluate some sub-sampling
>techniques and also evaluate the added noise of the device.
>It appears that this eval kit is virtually useless because it uses the
>TLV5535 to convert a signal to digital and then has a DAC to convert it back
>to analog which you can then hook up to a spectrum analyzer and view the
>spectrum. So you don't know which device is adding noise to the signal. The
>DAC also has an unfiltered output.
>I haven't spent any time playing with this eval kit yet but it doesn't seem
>like I would be able to trust any of the results I get to properly evaluate
>the performance of the TLV5535. And I don't know how this going to effect
>our ability to evaluate sub-sampling techniques.
>Anybody have any suggestions or comments?
Unless you want to make a major project out of it, it's easier to just
read the datasheet and believe it, design your product, and test the
prototype. That is probably the fastest route to a working product.
Your alternate is to build your own 'demo' board (or maybe kluge
theirs), stash the digital data somehow (logic analyzer? custom fifo
or something) and import the burst data for numerical analysis.
With a whole lot of work, you could duplicate the datasheet results.
I've found most such ADCs to behave at least as well as advertised
(with an exception or two!) so tend to just design my product and test
it. The actual product prototype is the best test bench.
Bit if you're going to do serious undersampling, the DAC will probably
be a small contributor to net error, so the TI eval board isn't a bad