Video of Uni-T Waveform Gen. Failure, need advice

[Seeandhear]

https://youtu.be/n9WYOPXJtfI?si=ECNi5pZ4FIav0tTs

Please let me know what you guys think. The Uni-T UTG962E waveform generator’s performance was pretty bad at school when I compared it to the other generators. I was trying to convince my classmates to buy Uni-T’s high value gear to be able to do the assignments from home like I do, but it backfired when everyone saw how bad my waveform gen was making the 1kHz sinewave at 20mVpp thereabouts, when compared to the school’s waveform generators.

So, later in my current lab we have to run the ground from our 15V to -15V dual-power supply in series with the waveform generator’s channel 2 with a 1MHz sinewave at 20mV to introduce noise to the power going to the 741 op-amp. Meanwhile, we use channel 1 from the waveform generator to produce the 1kHz sinewave at 20mVpp. The amplitudes might be slightly different in the lab but they’re all around 20mV.

I need a waveform generator I can trust to get my labs done at home, needs to have 2 channels and let run power through the different channels and such as used in our lab. What do you guys recommend I do? Should I get a replacement Uni-T UTG962E? Or should I go with a different piece of gear? Also, maybe I can’t hook 1 channel in series with the +-15V while using the other channel with a different ground reference. Our school uses two 1 ch waveform gens.

1.) Replace the UTG962E, maybe mine was a lemon.
2.) Your recommendation.
3.) I go balls to the wall and just buy a Keysight EDU33212A, unless I need two 1 ch AWGs.

Any assistance, advice, etc is welcome. Who knows, maybe I’m just doing something wrong. Thank you.

[jwet]

I watched your video.  Your setup is pretty messy and I see a few problems.

1. You don't specify your supplies but it looks like you're using a single supply.  The 741 is generally run at +-15v and is generally expected to provide +-10 V on outputs.  The UNI-T output is bipolar, unless you're capacitively coupled- this is not good.
2. As above, look at common mode voltage range and where you're feeding the input DC wise.  Op-Amps have a range of CMVR they work over.
3. Those big fat long power leads are like inductors, shorten these and use .1 uF bypass caps at the amp. 
4. Consider limiting the BW on your scope to 20 MHz, you're picking up a lot of noise and you don't need much BW.
5. Shorten scope test probe leads, twist ground lead of probe around probe and ground close in.  Tighten things up.
6. What capacitive loading?- use the 10x setting on your probes if you're not.  The 1x setting has a lot of capacitance and could be causing your amp to go unstable driving it.  Op-amps have a hard time driving much C load and the 741 isn't stellar.  The 10x setting has <20 pF which is a safe load.
7. The 741 is a prehistoric part that goes back to the 60's.  Try something like an LM358, a slightly newer but much better behaved part.
You're equipment is fine- you need more experience putting test stuff together.
Regards,

[Seeandhear]

Thanks Jwet for your reply. This is the start of my second semester working with electronics, so half of what you said is foreign. I’ll try researching each bit of advice you gave me.

The power supply is the Uni-T UDP3305S. I had Channel 1 and Channel 2 connected in series with the ground reference in between for the 741 op-amp. I didn’t know that had potential for mistake. Should I be using this “SER” mode for Channels 1 and 2 instead of manually connecting the channels in series the way I was doing it before?

Here is a video of just the Waveform Gen hooked up to the Oscope. I didn’t run it very long, but it seemed stable when changing settings. https://youtu.be/wC6DT2yQ0vQ?si=3GD-1zDkNe-jIi_G

I am in a rush to head to campus and I won’t get out until later but, this is good news if my only problem is my inability to hook stuff up. I can work on that. Your recommendations for shortening leads and such probably require me to buy shorter leads off amazon right? Maybe there is an easy way to get a capacitor in series with the lead at the connector like you suggested. I’ll look into it later. I suspect my main issue, currently, is the way I’m using my power supply.

I’m having problems getting user manuals from the Uni-T website. Weird.

[jwet]

The PS hookup shown in the PIC is good.  The idea is to have +-15v go to Vcc and Vee.  Op-Amps don't have ground.

Bring you Uni-T input into the circuit (show circuit- hand sketch with phone pic is fine).

You don't need new leads just put a bypass on the bread board at Vcc and Vee near the amp.

Probe the circuit and take ground in close.  Don't have long leads hanging around- they're antennas.

Post again and I'll try to help.

[Seeandhear]

I brought this problem up with my instructor and he reminded me that the reason we were using a 20mVpp amplitude was to have the best playground to see and test noise. So we were adding a noisy rail and then decoupling the chip from the power with capacitors.

The section of my lab assignment that was causing problems with my AWG was when purposefully using a 5Ω Ri with a 25Ω Rf with the 741 op-amp in a basic inverting configuration. This must have been causing my UNI-T AWG to struggle with the higher current draw or something. The lab meters didn’t struggle here but mine did. I haven’t had mine fail outside of this specific use-case. So, when using a 5k Ri, 25k Rf; there is no failure even with extended run-time and setting adjustments.

I almost feel like returning mine via Amazon, but maybe I shouldn’t. The $500 range AWGs in our college lab are so much cleaner in mV range and they can handle the use-case. If I was to return mine, then I could grab a Siglent SDG2042x. Let me know if you have an opinion. The instructor said that having a dirtier AWG isn’t a bad thing to have. I noticed mine doesn’t have a 50Ω mode, just “HighZ.” Too bad the product manuals require a password to download on the UNI-T website.

Thank you

[jwet]

A 50 ohm output impedance is nice- it shouldn't come up a lot, most input Z are >1K and you won't see it.  It sounds like you have a good instructor.  You have these Op-Amp equations in your head from class... Gain = RF/RI but there are limits, its good to see this in action.

I have one old HP audio generator that has a 600 ohm output and it trips me up now and then.  If you're in the return window, you might return and get something with 50 ohms output, its bit more versatile.  Test equipment tends to hang around once you own it.

General purpose function generators don't have great level control, especially cheap ones, usually they use a pot and 10 x step attenuator before the final power amp.  This can make for a little signal going into the final amp that tends to be noisy.  Better generators generate a nice 1V or more leveled waveform, go through the power amp and then put a 50 ohm attenuator after the power amp.  They are precision attenuating a clean signal and get a clean output.  Function generators can usually add DC offset to outputs that can make this difficult, you don't want to put a 5v DC offset into a 50 ohms attenuator so they have a two stage output- AC path and DC path.  Adds complexity= expense.

As far as brands and models, Uni-T is a lower tier supplier compared to Siglent.  Siglent is kind of mid Tier but generally great value- some of the high end stuff is top tier- they make scopes that are sold by LeCroy that are first class (I own one).  It all depends on what you want to do, its hard to know.  There are so many goodies to buy, its hard to choose.  After you buy all the equipment, buying cables and adapters is another expense and nice interconnects make lab time much easier.  I do mostly medium frequency analog to low frequency RF.  Everything needs a microcontroller so I have tools for that too- mainly a PC and some programmers.

I looked up the Uni-T you mention- here is a link to the manual I found-
https://res.cloudinary.com/iwh/image/upload/q_auto,g_center/assets/1/7/UNI-T_UTG900E_manual.pdf

It does have a 50 ohm output and says it can do a 10V signal into 50 ohm which is very good.  Any generator will have a hard time driving a 5 ohm load when its not setup properly. There is something in the Utility menu that lets you select the output Z as 50, 75 or high Z- I suspect you have it set at the latter.  The size is really nice and I think the quality is quite good really.  Take a long bath and read the manual cover to cover.  I think its a very capable instrument.  I really like the 2 channels.  This can be very nice for making modulated signals, exotic quadrature signals, etc.  Distortion is good. freq counter input is nice feature.  Great learning tool.  3 year warranty is great.  Good manual.  I think its a keeper.

When you're using a signal generator, think about what you're asking it do.  If you really want to drive 10V into a 50 ohm load, you will have power peaks (V^2/R) of 2W- RMS is 500 mW.  This is a sig gen not an audio amp.  I use to stock a lab in a physics lab as a job in college.  I also repaired equipment.  My boss put internal  fuses on the outputs of all the equipment and similar things on inputs.  We spent a lot of time repairing equipment.

Edit- I notice there are some good reviews and discussion of your UNI-T here on EEVblog, search, conclusion thumbs up.  One reviewer mentions that using a good USB power input is important and makes a difference.

Good luck in your schooling.

[Kleinstein]

It is possible that the generator has issues at very small amplitude. One could get around that with an external attenuator (e.g. 20 dB or 40 dB).
I would not so much judge the generator from the low signal. The  generator may well be good enough.

For the OP-amp the 741 should be fine with +-10 V supply, no issue there. It's an old type, but still not that bad.
The LM358 is in may aspects even worse (noise, output cross over) - it can still be OK to have a few for the bread board. I would more use RC4558 as a moderate speed ( 3 MHz typ), cheap type for tests on the bread-board. 

[electr_peter]

What is the power source of AWG and how it is referenced to scope?

UTG962E is capable of few mVpp sine output (that still looks reasonable(-ish), from 10-20mVpp and up even better - based on real tests). There shouldn't be any issues with abrupt level changes or excessive noise from AWG (excluding possibility of broken unit).
At such low levels noise has bigger impact - use proper probing techniques and limit BW if needed.
UTG962E output impedance is always 50 Ohm. High Z/50 Ohm setting just rescales Vpp by 2x in High Z, actual output impedance is not affected.

[jwet]

>> 741 vs the world...

Dave Fullagar, the original designer of the 741 at Fairchild, hired me out of industry and was my boss for a few years in early 90's.  He was one of the founders of Maxim.  He would be the first to admit that the 741 had LONG lived out its usefulness 20+ years ago (and that was in the 90's!).  We figured it hung on because so many students dissect it in EE courses. The context that I was making that comment referred to single supply operation which is what it looked like the OP was doing. 

[Seeandhear]

Thanks Jwet for your awesome information. This was a fun learning experience. I feel bad about the generator. I did choose to return it for the Siglent. The manuals being password blocked on the official Uni-T website was the last straw off my conscience.