Anyone using Instek AFG-2000 or AFG-2100 arbitrary function generators?

[sanka]

I was looking for a budget function generator and found that there was a lot of eevblog love for the Instek SFG-1003. I was about to buy it, but while checking out Instek's website, I found that they had a newer 5MHz arbitrary waveform generator  AFG-2005 (or AFG-2105 if you wanted modulation & sweep capability and also frequency counter).

I don't really need modulation or frequency counter from the function generator. I could occasionally use arbitrary waveforms. So I bought an AFG-2005 yesterday for $250. A bit more than the SFG-1003, for sure, but still seems reasonable.

I haven't received it yet. I was wondering if anybody here is using the AFG-2000 or AFG-2100 series, and had any reviews or words of wisdom to share.

Thanks

[saturation]

I have the SFG-1003, basically is puts out a clean signal, jitter free to under 5ns and provides all the function in its spec sheet as written.  There is no reason to expect your better unit performs less well.

The bigger issue is can you get the same performance with a cheaper unit?

If you read the Siglent thread, you will find Instek is money well spent. 

https://www.eevblog.com/forum/reviews/siglent-sdg1020-arbitrary-fungen-first-impressions-and-a-possible-signal-glitch/?topicseen

.. but for higher frequency capacity, but with more jitter and inconsistencies see the thread on the Hantek 3x25.  There are gotchas to that unit but for $150, there are few FG that run to 100 MHz.

[sanka]

Got the AFG-2005 function generator today. Haven't played with it much. But I thought I'd post a screenshot of what its square wave looks like at 5MHz – the rated max frequency. See attachment. What do you think?

The spec sheet claims rise time of less than 25ns at 50? load. My oscilloscope hasn't shown me anything less than 29-30ns at 50? load. Rise times are higher (about 40ns) with no load – i.e., just connected straight to the scope. Rise times are also higher (about 36-38ns) with higher amplitude, such as at 5V or higher.

I'll play with it more.  I might open it up and take some pictures as well. In the mean time, if anyone is interested in any specific test case you want me to run, please let me know.

[saturation]

At the output of a typical DDS FG is a sharp bandpass filter; this is what you are seeing;  the edge roll offs at the higher i.e.,  5->9th harmonics, 25 MHz and up.  For square waves, check the output from the TTL port, which in typical units is not filtered, and you'll get sharper rise times.

[Salas]

I am using AFG-2012. When I had Rigol 1052E (100MHZ mod) it was showing <21nS rise time on square waves. You should expect your best square only up to 10 to 20 times less frequency than your full generator bandwidth. Same thing happens with your scope's bandwidth. Good thing is that the Instek actually outputs the best it can up to max frequency when many others stop at 1MHz no matter if their sinewave output is 25MHZ or more. Mine was rather acceptable up to 5-6MHz on the DSO although naturally rounding as it went up in frequency. Also its stable enough for square wave jitter at low frequencies and that is the reason I opted for it over the much fancier Siglent dual output (thanking member Drieg of Silcon Electronics for checking that aspect on the Siglent for me at specific frequencies and amplitudes I asked him for). The arb editor is easy to use also. The gen performs well and its ergonomic. Its calibration for level spec was marvelous when I checked it with my Fluke 87V down to single digit mV level. Comes with a cal certificate as you know, and its actually honored. Higher MHz model with modulations and sweep I avoided because of price since I don't care much about those functions or range especially when  the economy here sucks big time. Here is a picture of it:

Unfortunately I have sold my Rigol 1000 series recently so to upgrade into 2000 series which is not here yet and I can't post you screenshots with rise time metric right now, but here is a 2VPP 1MHz photo from my HAMEG HM204-2 analog 20MHz scope. It shows the same leading edge roundness on its 1MHz 2VPP probe cal output too. There is a wee step on leading and trailing edge if you can see it. It should be some artifact of the digital synthesizer and filter, or something else that it may improve on the DS2000. It's ironing out from 3VPP upwards though. I don't remember it as a clear step on the 1000 but I better see it again on the new DSO soon so to be fair. Picture is taken with 50 Ohm inline termination at the scope by a Trompeter Electronics adapter through 3 feet LMR195 coax BNC lead.

[saturation]

https://www.eevblog.com/forum/reviews/cheap-oscope-probe-quality/msg31749/#msg31749

On this old test, the SFG-1003 can under 10ns rise times easily through the TTL, aka sync, port.  Its probably much faster, but ~5ns is limited by the DSO.

[Salas]

Got the AFG-2005 function generator today. Haven't played with it much. But I thought I'd post a screenshot of what its square wave looks like at 5MHz – the rated max frequency. See attachment. What do you think?

The spec sheet claims rise time of less than 25ns at 50? load. My oscilloscope hasn't shown me anything less than 29-30ns at 50? load. Rise times are higher (about 40ns) with no load – i.e., just connected straight to the scope. Rise times are also higher (about 36-38ns) with higher amplitude, such as at 5V or higher.

I'll play with it more.  I might open it up and take some pictures as well. In the mean time, if anyone is interested in any specific test case you want me to run, please let me know.

Screenshots taken with AFG-2012, 1m LMR195 coax on main output BNC, 50 Ohm feed-through at the scope's end. 1MHz few periods, 1MHz rising edge, 5MHz few periods. Hi Res and 56Mpoints were the acquire parameters on DS2000.

[sanka]

Screenshots taken with AFG-2012, 1m LMR195 coax on main output BNC, 50 Ohm feed-through at the scope's end. 1MHz few periods, 1MHz rising edge, 5MHz few periods. Hi Res and 56Mpoints were the acquire parameters on DS2000.

Thanks, Salas. I suspect my higher rise times are contributed by the scope probe in the middle. I'll also hook up a 50 ohm feed through terminator on a coax from the FG to the scope, eliminate the scope's probe and see if things look different. I don't have LMR195, but I'd imagine that a regular RG-58 for such a short distance should be fine at this frequency.

I see you got your Rigol DS2072. I have had one for a couple of weeks. There is a lot in it for me to learn.

[sanka]

Screenshots taken with AFG-2012, 1m LMR195 coax on main output BNC, 50 Ohm feed-through at the scope's end. 1MHz few periods, 1MHz rising edge, 5MHz few periods. Hi Res and 56Mpoints were the acquire parameters on DS2000.

Hi Salas,
I eliminated the scope probe and connected in the manner you took your measurements (except RG58 instead of LMR195) and my values are almost identical to yours. Thanks.

For square waves, check the output from the TTL port, which in typical units is not filtered, and you'll get sharper rise times.

Thanks, Saturation. Yes, I now see 2-3ns rise times on the TTL / SYNC port. The less jittery Main port does 8-22ns under different conditions. Certainly less than the manufacturer claimed <24ns. Pretty good!