Inside the GDS-2204 Digital Oscilloscope
The GW Instek GDS-2000 Series is intended as a direct competitor to the Tektronix TDS2000 Series Digital Storage Oscilloscope (DSO). The purpose of this page is to document what is known about the Instek GDS-2204.
Background
The Tektronix TDS2000 Series was introduced in April, 2002, and the TDS2000B with USB interface was introduced in September, 2006. Here is a recent TDS2000B product specification (433KB pdf).
The Instek GDS-2000 Series was announced in June, 2006. The GDS-2000 Product Comparison (244KB pdf) shows how Instek planned to dominate the low end of the DSO market with their new product. They expected it to replace 80% of sales of their own GDS-800 Series oscilloscopes (See page 6.)
However, the GDS-800 Series (500KB pdf) is still available on their web site two years later in June 2008, and it is also marketed as the Tenma 72-7235 and 72-7240. Here is another Tenma ad showing both versions (767KB pdf). (scroll down near the bottom.)
The GDS-2204 User Manual
The GDS-2204 User Manual (3MB pdf) on the Instek site has the pages rotated sideways. Here is an older version (3.8MB pdf) that is a bit easier to read. There are some differences between the two versions. For example, a spec on page 120 in the old version states the single shot bandwidth is 100MHz. This limit does not appear in the current version, and the risetime of a single shot signal appears to be the same as the normal 1GS/sec signal.
GDS-2204 Construction
The GDS-2204 construction is simple but quite robust. Here is what it looks like inside. Of special concern is the use of eight interleaved ADC's that are pushed 25% past their maximum spec in order to reach 1Gs/s. Please see Fig. 3 for more information.
Digitizing Scopes
Despite the additional memory, the GDS-2204 suffers from the same problems as all digitizing scopes. They suffer from wideband noise, poor resolution, noticeable quantization steps on the waveforms, and various artifacts due to aliasing. They silently reduce the sample rate as you change horizontal sweep time, which makes ringing, crosstalk, and glitches from fast logic switching transients disappear from the display, and they are generally unsuitable for work on precision circuits and systems.
You can improve the apparent resolution and SNR by averaging the waveform, but it often takes many averages to get a clear picture. This makes the waveform acquisition too slow to be useful, and it still doesn't solve the problem of glitches disappearing due to sparse sampling.
To illustrate this, a short 13 second mpg movie was made to show the effect of averaging. The first 6 seconds shows the normal noise without averaging. Watch the menu on the right side to see when the averaging starts and when it changes. The movie is available here
Can you see any glitches or signals buried in the waveform? Yes, a switching transient is in the waveform. It occurs just after the peak of the sawtooth, and is due to crosstalk from an adjacent circuit. The amplitude is more than enough to terminate the oscillator cycle at the wrong time, which is how it was found. But you can't see it, since it is lost in the sparse sampling. And anyone who uses a digitizing scope probably won't see it either.
After trying to use this scope for a week or so and finding that digitizing scopes are totally unsuitable for advanced development work or debugging precision electronics, it was returned and two Tek 2467's were purchased on eBay for about the same price. They are in excellent condition and should last much longer than the five years expected with a new GDS scope. (See below.)
Repairs
You might be interested in the following slightly edited post from the Yahoo TekScopes forum (free registration required):
Jan 07, 2009 7:13 am Hi all, That question reminded me of a scope I purchased a couple of months ago, an Instec GDS-830, that came in with a burned out PSU. So I mailed the south Korean manufacturer, GoodWill Instruments, and asked them to send me a brand new switched power supply. "- That model is over 5 years old, and we no longer have any spare parts." I did not answer that. Instead I turned to the distributor in Sweden and asked the same thing. The man I spoke with did not think they would have any PSU as it was over 5 years old, too. Our distributor is maybe northern Europe's finest, so I asked what sort of oscilloscope manufacturer thinks that an instrument should have unobtainium PSU's after only 5 years of use? That's real crap, if so, I said. "-Well I can send them a request, he said." After about 24 hours he returned with: "-Your lucky! They got it in spare! But it will take about 4 months" No problem, I said. Take your time, I can fix things with my intuitiveness. (as I am a pro service tech ;) Well it took 1 month, and now it is sold to another lucky guy in Sweden. And I did not have to troubleshoot a loose-firebombish PSU. Never let them fool you! Best regards, L.A.