I see we have an ATF20B Function Generator in the lab. I have been considering buying a FG for my own lab and was considering this one until I saw this YouTube video about a serious bug in the firmware for the DC Offset.
Has anyone encountered this issue with the DMS ATF20B? I haven’t had the time to come in and hook it up and try to repeat the problem so I was hoping someone on the wiki would either know or try this out next time they get a chance.
If this is true, it’s a deal-breaker IMHO, so I’m also looking at the Rigol DG1022U.
I guess I’ll go with the Rigol DG1022U for my lab when I can afford one.
NOTE: This post in no way means to infer anything about the committees decision to put the ATF20B in the lab. Sometimes you take what you can get based on info you have at the time. I’m sure ATF20B meets 90% of the members needs. Thank you committee members for all you do to improve the lab. I’ve seen some very nice improvements.
I have a DG1022. It’s decent for what it does. It has a neat feature where you can set the input impedance of the load and it will calculate in the divider with the internal 50 Ohm output impedance when setting voltages.
There are a few things that are pretty lacking though:
Lousy THD, ie not good enough for any real audio testing
PWM output will glitch and ‘runt’ if you dial through the duty cycle. This makes it unsuitable to control the PWM input of something you want to fiddle with by hand.
The backlight inverter has an audible squeal. It’s not super loud, but noticeable in a quiet room.
The label for CH1 is exactly between the two BNC jacks, so I regularly plug cables into CH2 unintentionally when not paying attention.
If that doesn’t bother you, it’s a decent low-cost option!
Yes I’ve seen this issue with the DMS ATF20B. I couldn’t make sense of what the DC offset was doing when I tried to use it so it’s nice to see that it wasn’t just user error.
Some long while back I got into the habit of putting a scope/meter on the outputs of pwr splys and sig gens to verify what their readouts indicated is actually what’s at the output. Adjust accordingly.
Problem with that view is that all equipment from the mid-80s onward is essentially driven by software.
This also means that bugs are easy to fix with a firmware update, usually. Has anyone checked if this is the case with the aforementioned signal generator?
I asked that question in the comments on the YouTube post that identified the bug. So far, no answer. Given the stories I’ve seen on how unresponsive the Chinese manufactures are (Rigol, Atten, etc.) I wouldn’t hold my breath for any new firmware in the immediate future.
Just another reason to go with an HP/Agilient box “if” you can find/afford one.
Correct. The adage of “you get what you pay for” is fairly true on the steep part of the price/performance curve. Once it flattens out, you might just be paying for a name.
The EEVblog guy tears apart a lot of stuff, and the Chinese-branded stuff is sometimes pretty horrible (loose parts, shoddy layout, DOA, etc.). However, Rigol is one of the few exceptions.
“Easy to fix” IF the vendor provides that kind of support.
And therein lies the rub. If they’ve shipped it out the door with a serious flaw, what odds they’ll bother to make improvements, when people are buying regardless?
I’ll concede that a really good software-driven instrument can be better than one which is pure hardware (or logic-driven, but without software as such, although that tends to appear to be the sweet spot where I’m concerned). But that’s not what we’re talking about, now is it?
Shipping broken hardware out the door with no software fix possible is better? I call BS on that argument. At least with software, you might get a 3rd party fix if the company doesn’t do it.
The “offset problem” mentioned above is a simple user interface bug (which requires fat-fingering a softkey), not a problem with the actual output. I’d consider it a minor irritation, at worst.
There are two problems shown in the video. One is a user interface problem, the other is the offset problem.
And I’d suggest that “broken hardware with no fix possible” is unlikely to get out the door in the first place, because it’s obviously unfit for purpose. Once upon a time, there used to be bulletins published with information on how to fix such problems as did crop up. Whether these were more likely than a software fix from a Chinese vendor? I can’t say for certain. There have always been good & bad vendors. But you didn’t have the DMCA standing in the way of taking a soldering iron to your instrument, & if you found a fix, you were free to share it.
This is kind of a moot point. Without the capabilities provided by firmware and signal processing, these instruments wouldn’t exist at these price points.
It’s true that a lot of products are shipped early due to updatability, but that also contributes to its price point.
Unfortunately I think a big part of the initial quality difference is also in the way that software is treated/handled by businesses and engineering managers. ie you wouldn’t produce a big piece of gear without a talented engineer at the helm, but software is notorious for being a crap shoot, having a reasonable likelihood that shoddy and ‘barely good enough’ engineering practices have caused major structural problems that can’t be seen with the naked eye.
