Episode 16 - Review of a $50 Chinese signal generator

Jack reviews a $50 signal generator. There are a lot of these on eBay. It has flaws but offers some decent value. For lots more about embedded systems (including over 1000 articles), visit www.ganssle.com. And subscribe to my twice-monthly newsletter that has more reviews like this! Go to www.ganssle.com/tem-subunsub.html

[@HitAndMissLab]

I have this one and I am very happy with it. It was not mentioned in the spec and in this video, but this signal generator can do Sawtooth waveforms as well. Both rump up and rump down. You just set it to triangle and than when it asks for a duty cycle, you move it to 49% or 51% for ramp downs or rump ups.

[@davidprice2861]

Very well presented review, concise, yet all the important and relevant information. Thanks.

[@david203]

Very nice review for someone who needs a signal generator that is not of laboratory quality. Good review of most of the features of interest.

[@injoelsgarage3934]

Excellent review. You should make more of them. Cheers, Joel

[@johncunningham5435]

Interesting, a signal generator at $50. not bad. Thanks Jack! a very informative presentation, well done.

[@GONZOFAM7]

Pretty nice. It will work for me in the audio range. I'm going to use it as an input to a synthesizer.

[@hitechespresso]

I got one. Works perfectly. Thanks, Jack!

[@stephenlee2284]

I can't figure out the sweep function. Pushing the rotary encoder adjust button (ADJ) seems to just shut it off, and I can't enter the sweep range. Oddly, if I switch from LINear to LOGrithmic, it sweeps from some low freq(I can't read the numbers when they are changing that fast) up past 42Mhz. It's only the 2Mhz model. SGP1002S. Help

[@billybbob18]

Thank you. You saved me from wasting 50 on something that would be totally useless for my application.

[@leonardmilcin7798]

"Things get interesting as frequency goes up" Rule #1. Things get always interesting as frequency goes up.

[@alperenalperen2458]

Wow you got a youtube channel that's nice :D Looking forward to see more videos of yours

[@dandwrasan2342]

Fantastic vid 👍🏻😊🇬🇧 please can you tell me what the term “up conversion “ means or is it up mixing?

[@davidhaney1394]

for setting up reel to reel tapes i find this does the job ok , im only using sine waves at 1 and 10 khz at .75v and for that its just the job its small, and it just works , but you need a scope .

[@mikehawk6873]

I want one so I can practice at home but the ones at school are different I wonder if I will still get the same values when working with circuits?

[@JasperJanssen]

The back and forth at higher frequencies/odder duty cycles looks like quantization noise - i.e. You're seeing 1 sample worth of the underlying dac frequency, when the numerical result is right in between the two samples.

[@a.lithia5664]

I've tested the Rigol DDS generators in our school's lab, and it also has this funny effect, even at 10MHz (trigged on the sync output). I think this is the artifact of digital non-integer frequency synthesis, since my own HP function generator that runs on a VCO doesn't have that weird jitter at all.

[@mrkv4k]

Yes, it's quantization noise. I've build an DDS generator as my bachelor thesis and this is common thing. When square wave is generated, the DAC basically receives the highest possible value (for example 0xFFF for 12bit DAC) for fist half of the period and then the lowest possible value for the other half (0x000). When your main clock runs at 20Mhz, and you are trying to generate something like 1.6Mhz, it's not evenly divisible and you are gonna get 12.5 samples per one period (6.25 per half), which means that your internal counter will jump one sample every other period - the patern will be 7x0xFFF - 6x0x000 - 6x0xFFF - 6x0x000 - 7x0xFFF -6x0x000, etc. You can think of it as adding 6.25 over and over to itself and changing the level every time the internal counter is higher or even that value, like 6.25 (change on 7th sample), 12.5 (change on 13th sample), 18.75 (19th sample), 25 (25th sample), 31.25 (32th) and so on. The only way around this is to either change the main clock depending on frequency generated (almost impossible) or to go significanly higher with the clock speed.

[@cidpusa700]

Thanks awesome review and the prices have fallen on e-bay to under $20

[@BOROZ28]

What is your take on the Feel Tech and similar signal generators in the range of 100 USD and plus. Seen few reviews but undecided.

[@WillemExist]

The jitter makes complete sense. They have implemented a numerical controlled oscillator. Basically a feedback accumulator. If the fpga clock frequency is too low, you will have that kind of jitter.

[@Graham_Wideman]

I think you are right. But that's a very undesirable design decision, as it results in this jitter. It would be far better to simply use (and specify) a coarser PWM resolution.

[@cheewurz]

What would you Recommend for a Home Hobbist Oscilloscope?

[@mattcero1]

Great video sir! Would this be a good product to test speakers like tweeters, mids and woofers? I have about a thousand I need to check for sound. If so, how loud would the test be? Thank you.

[@redicebluefire2552]

Know of any affordable duel or more channel analog output AWGs ($400 range)?

[@msdesignru]

Thanks for this review

[@HillOrStream]

Nice review, didn't know these existed, last time I looked at low cost signal generators they were all pretty poor, Bought an ATiny based one that only uses six bits, so the sine wave is poor. Would of like to have seen how well the sine looks at the bottom of the frequency range and if it goes down to 20Hz or below, but I guess I'll find that out when I search for it

[@MarkSeve]

Its Sept, 3rd 2019. Seems I'm a little over 4 years to late to enter the drawing for this one. :( lol Thanks for the info though Jack. Keep up the great work you do sir.

[@Shadi2]

the STC micro is probably from their very popular 8051 line.

[@HashanGayasri]

For that amount, you can add a tiny bit more and get a 40M samples /sec USB oscilloscope with a way better function generator (with regards to the output waveform - not features) .

[@wither8]

Found you via Dave's video blog, fantastic quality! Gonna take me a while to get through 1000 articles! Can you dump the AHDL from the Altera to see why the square jitter/deformation is happening? You'd think from all the 3 forms, the square would be the easiest to synthesize from a digital component...

[@1pcfred]

+wither8 I guess for the cheap price they had to screw something up.

[@markrhine5809]

I am look for a rf signal generator for AM radio alignments, would this work?? Very interesting video..

[@cajuncoinhunter]

What cable were you using from that unit to your scope ??? If it was a supplied cable , it will show that on a scope, the better the cable the better the signal... I've seen it happen before.....

[@pirateman1966]

No link on the generator?

[@NatureAndTech]

Does the amplitude change with frequency? Looks like it.

[@Graham_Wideman]

As Jack said in the video, he was displaying what happens at the upper limit of operation, where of course the amplitude will decrease when exceeding that frequency.

[@themainproblem]

Only 3 comments? Great video! How do you like your HP 8640B? I have one and think it's a wonderful piece of equipment.

[@azzym312]

If you can make brass gears for the cavity unit. They are the best in the world but age is showing.

[@tomstern1681]

These waves are when terminated with 50 Ohm? I have a similar device which I bought for $100 with a case and all and I can tell you, if you have $50 you can also get $100 - $150 and get a decent generator.

[@svampebob007]

Dave sent me here too, though you should fix your hyperlink in the description it's not so hyper ;) subed and looking forward to see more content from you :)

[@garyha2650]

Is the increment for frequency ADJ knob based on where the cursor is situated?

[@Graham_Wideman]

Yes

[@Pillowcase]

got one of these after watching this video - really neat gear for the price

[@root6683]

Pillowcase what do u use it for ?

[@TheZooman22]

I looked at those a while back, they seemed to be too good to be true. I decided on a used (2 MHz) Kenwood function generator, which is fine for school atm.

[@Toturials]

hi, Sine Wave Output Volt AC Or DC . thanks

[@root6683]

What is usage of that thingy ?

[@kardeef33317]

just found your channel and subscribed. I guess I'm a little late for the contest. On a fixed income and could use stuff like that .It would work great with my DSO 138. You work with what ya have. Nice channel, glad I found it. Thanks for the Vids :)

[@fanwlkr]

The rounded square seems to be common on the cheap ebay ones. What I still wonder is how it sounds ;)

[@AlienRelics]

That is bandwidth limitations of the output amplifier/buffer. The question is, what do you require, not what looks good.

[@willynebula6193]

just found your channel' great vid! Keep up the great work new sub bang for buck you get what you pay for

[@TheAq93]

so if i got it right.. with a very low frequency lets say 150 hz will the square wave create almost perfect pulses?

[@AlienRelics]

No, of course not. "Perfect" is an ideal. "Almost perfect" has no more descriptive power than "perfect". Define the pulse you need, otherwise it is like someone asking for a signal generator that goes from 0Hz to someothernumber, or a power supply that goes from 0V to someothernumber. Why would you need a signal generator that puts out 0Hz (no signal) or a power supply that puts out 0V (no poweer)?

[@animarkzero]

Some metal shielding on the back may help signal quality a bit!?

[@AlienRelics]

No, that just helps keep it from generating RFI. It does nothing for signal quality.

[@Graham_Wideman]

@@AlienRelics Yay, someone talking sense.

[@alexk2490]

can the frequency chenge without turning off signal? online, so to speak

[@nor4277]

It would have been nice if it includes a case for it even if they changed for it ,and not everyone.needs a top of the line signal generator ,I work on tube radios ,I bought a pocket oscilloscope and I ordered a signal generator kit for under 5.00 ,a Idea I got from.great scott.

[@3dmaxuser]

Great video :)

[@freequest]

I know this is a Chinese signal generator but could it be used to generate Australian signals?

[@ostheer]

Support for Australian signals was removed in version 1.4

[@freequest]

Darn it. Figures as we need stuff suited for use upside down.

[@ChristopherSibert]

I tested mine here in the US, I flipped it upside down, and it worked fine, but all of the signals were inverted, so you have to adjust for that.

[@ericlum9859]

I p

[@mariaforzen7836]

freequest jnbl

[@tHaH4x0r]

Very nice video! That jitter seems pretty weird to me as well... Also, the Altera part is actually a CPLD, not an FPGA ;)

[@frankrose9952]

Very nice video. The jitter is a something that all DDS synthesizers do. The DDS output is only jitter free when the output frequency divides evenly into the DDS clocking frequency. For example, if the output frequency is 10 Mhz, and the clock frequency is 100 MHz. then there are 10 samples per output cycle (100 Mhz / 10 MHz) and the output waveform starts and stops at exactly the same spot int he sine wave for each cycle. but consider what happens when the output is 9 Mhz. Now, there are 11.111 sample points per cycle, and the DDS starts and ends each cycle at a different point for each cycle. At first, this doesn't seem like it should be a problem, but remember that the register that controls the output frequency of the DDS is an INTEGER register, and the DDS can not do floating type divides, so each output from the DDS is very slightly different than the last. This jitter is on ALL non-integer output frequencies. It is just not noticeable at lower frequencies. Here is an experiment you can do to see this effect: Set the generator to 10 Mhz and set the scope so that one or two cycles are visible (so that the jitter can be easily seen). Now, slowly reduce the generator output, decreasing the frequency by very small steps (10 Hz, for example). As you decrease the frequency, you will notice that the jitter will be better at some frequencies, and worse at other frequencies. And some frequencies will have no jitter at all. the frequencies that have no jitter are the ones that where the output frequency is an integer multiple of the clock frequency. If you know the frequency feeding the DDS circuit, you can compute these frequencies. DDS synthesizers are wonderful devices. You can generate frequencies with very high frequency resolution - down to .001 Hz per step in some cases. But nothing is for free, so to get high frequency resolution, you pay for it with output jitter. Also, that output jitter on the scope shows up on the spectrum analyzer as side spurs, like the ones you show on your video. But for $50 bucks, what would you expect? I'm amazed that what you have sold for as cheap as $50. The cost for parts and labor alone would be more than that.

[@tHaH4x0r]

+Frank Rose Of course, i should've known and makes a lot of sense. Thanks anyways for the reply.

[@azzym312]

You got it right, but NO not in this case, in this case it is side bands due to noise on the power rails to LPF circuit. Can be fixed pronto

[@Graham_Wideman]

@@azzym312 And again you made this comment which is simply incorrect. How would a 10MHz signal acquire a sideband at +/- 5MHz? You would have to modulate it with a 5MHz signal. Where is that 5MHz signal going to come from? You think the power rails? From a switching power supply? Since when does a cheapo switcher operate anywhere near that frequency. And yes, obviously 5MHz could be a distant harmonic of a 100kHz switcher, but if it was really leaking that much onto the rails, there would be all kinds of other lower harmonics on there too. So no, this is an artifact of the DDS algorithm, not noise on the rails.

[@crazyham]

Excellent Video, Thank You. I had just ordered one prior to watching. This will definately Suit my Purposes ⚡🙏⚡

[@shana_dmr]

I've seen many iterations of these el cheapo Chinese DDS things and it's baffling for me that most of them have the output clipping issue - has nobody at the manufacturer's site bothered to connect this thing to the scope and play with the knobs? :)

[@JasperJanssen]

They just expect you to know what you're doing, and aren't expecting to save you from yourself ;)

[@rohiniesakki3592]

Can anyone tell the concept of this

[@andreasparasxos5372]

Can you please explain me what exacly offset and ampitube does??which are the better settings?? Thank you

[@AlienRelics]

A sine, square, or triangle wave with no offset is equally above and below 0V. That is often just what you need for a signal for testing amplifiers. But for a square wave used as a clock source, or a triangle wave for some other testing purpose, you often need all of the signal to be positive with respect to common. That is what "offset" does. Amplitude is to adjust the peak to peak voltage of the waveform. One is not a "bettter" setting. It would be like asking which is better, an accelerator or steering.

[@andreasparasxos5372]

@@AlienRelics thank you my friend

[@AlienRelics]

@@andreasparasxos5372 Glad to help.

[@MilanKarakas]

At 7:10 on your video, those are not 'harmonics', it is rather product of some spurious modulation where you get main frequency and products - sum and difference. Else, great video!

[@Graham_Wideman]

Actually, I think those really are harmonics. To the left you have the large spike at DC. In the middle you have the intended fundamental at 10 MHz. At the right edge you have the first harmonic of 10MHz, at 20MHz. In the middle of the left half of the screen you have a small peak at 5MHz.. that indicates a 2-cycle pattern of irregularity, possibly to get an exact cumulative frequency given that the sample output rate runs at some fixed clock rate that might not produce exactly 10MHz if every cycle was identical. And then a similar peak at 15MHz. Of course, you could describe that as the 10MHz "modulated" by "spurious modulation" of 5MHz, I guess :-).

[@soulrobotics]

the only question i have is what is the load.... seems to me that could be a mismatch in the line... but... well...is cheap. for sure need a filter

[@blackoakcbshop6371]

Could you use this for CB Radio?

[@Graham_Wideman]

Unlikely... it doesn't have a microphone or antenna. :-)

[@PapasDino]

I'm reposting my comment as it doesn't appear that it will appear as a public comment because I included a URL; let's see if this works: Thanks for the review Jack, a nice adjunct to the KZbin review and simple upgrade of a similar more capable DDS generator by "devtty0". BTW, are you a ham radio operator? The HP8640B is a hint that you might be. 73 - Dino KL0S

[@azzym312]

What do you mean by harmonics on both sides at 7:03 ?

[@tanukicyber]

It is referring to the two spikes, one left and one right, of the large spike in the center

[@azzym312]

These two spikes are NOT harmonics as he calls them, they are side bands produced because of the amplitude modulation probably caused by a noisy power supply rail. The unit will benefit from a little bit of ground rerouting and filtering.

[@tanukicyber]

+Azzy M You may very well be correct! I was just reiterating his description. I wouldn't be surprised at all that it has poor grounding and shielding, after all, it doesn't even have a power s/w and it has terrible jitter. Not a useful too,l except to show a child what a basic waveform looks like. I did enjoy it though.

[@Graham_Wideman]

The display he shows has the fundamental (10MHz) in the middle, the DC spike on the extreme left, and a 20MHz (harmonic of 10MHz) a the extreme right. The peaks at 5MHz and 15MHz are not harmonics per se of the 10MHz, but are rather due to some irregularity in the lengths of single cycles played out by the digital synthesizer, in order for it to produce a cumulatively precise 10MHz while operating with a fixed time per DAC sample. Apparently in this case, the strategy leads to half the cycles being one time length, and half the cycles being slightly longer, ie: an approximately 2-cycle pattern that repeats at 5MHz. These are not cause by a noisy power supply rail, or a grounding or shield problem -- those would produce very different artifacts.

[@danmyself5341]

thanks

[@trentjackson4816]

It is a toy. Bit of Sunday morning fun for fifty bucks. I still prefer CROs rather than 2K USD DSOs. Is there something wrong with me?

[@to-tt7fc]

Sine wave has harmonics ?

[@azzym312]

He is wrongly calling side bands ( 7:00 ) harmonics. This is AM of the signal likely caused by by PSU/ground defect. Easy to fix.

[@TheCarlos206]

Yes it has harmonics because it is not a pure sine it is a DDS sine...

[@Graham_Wideman]

@@TheCarlos206 Correct.

[@jason-tj6kj]

Stupid question: can you generate sound and possibly make music with this device? I don't have any technical background in electronics or science, just love to experiment with noises and such for "art." Thanks!

[@Graham_Wideman]

Yes you can, but it's a rather clumsy and frustrating way to do it.

[@yoramstein]

Thanks for intresting review. For us$300 one can get a Siglent or a Rigol double channel boxed with much better quality -don't wate money on those!

[@joeambly6807]

Duh performance will be better at 6 times the price!

[@ronson232]

I think his point isn't that a $300 one is better, it is that a $50 one isn't good enough.

[@octapc]

stating the obvious we all know, so, makes his comment frivolous.

[@y_x2]

They use a very crude D/A you can probably do better today with the right stuff.

[@MrTruth0teller]

Hi, Nice review. Smartphones can be used as signal generator. I use my android phone with free signal generator app, which gives output at audio 3.5mm jack. Signal generator for free.

[@underverser]

Does it ecxeed output beyond 20-20.000Hz? I don't believe that...

[@computernuts69]

that is one sexy Oscilloscope :)

[@kishoresaldanha815]

Touching those QFNs by hand might probably zap them !!!!!

[@joeambly6807]

Nope

[@jamesraykenney]

Not while they are 'in circuit'.

[@garyha2650]

Careful not to hook up the power backwards, it will burn out

[@rosalvacoldwell215]

goood

[@researchandbuild1751]

That square wave looks like complete garbage even on the low frequency

[@Graham_Wideman]

Apparently this is garbage quantized with a one-bit ADC.

[@valmordumke5160]

Generator of Functions everyone put the generated direct connected in the ociloscope more nobody put or knows how to use the generator in test in the circuit ??????? just showing the generator on the oscilloscope is of no use as this anyone knows how to do

[@mrkv4k]

It's clearly described as review of function generator, not a video on how to use it. What 's your problem?

[@Graeme_Lastname]

Junk is as junk does. ;)

[@SatanIsTheLord]

In which forced labor camp it was made?

[@teds9896]

What do you want for $50? Actually I've seen for $30. You've got to factor in: 1) How much demand is there for a product like this worldwide? 2) Their board manufacture is going to be cheap meaning you're not going to have the expensive discrete components necessary to really clean things up. 3) Yes there's still some amazing used equipment to be had on Ebay; it often still costs hundreds and are you getting anything newer, for example HP, than 20 years old? In its day (and still similar units) often cost thousands.

[@azzym8794]

*I have some B&K 10M and 20M sig gens. Their output is the same as this Chinese device, or worse. I regret having bought them. I have repaired one of the Chinese units for a friend. I was able to get rid of the Jitter by adding some bypassing and re-routing the ground. The Chinese will do that soon too, I am sure. The B&K fail to produce sine at half the rated max frequency.* *Why did someone not produce the same in say US or Austrlia, say. Jack Ganssle should also have shown comparison with a better known brand. This think is a remarkably well made device.*

[@1pcfred]

Even for the money I can still complain about that poor square wave performance. It is a shame really. Why did they goof it up like that?

[@spoderman15]

China.

[@andytan4993]

Absolutely,cheap goods with poor performance. It's unhelpful for a Spoder man be biased against China.

[@azzym312]

What are you expecting for 50$? Agilent quality. A similar Agilent unit goes for 6k$ to 8K$ used. In the hands of someone who UNDERSTANDS what he is doing, this is a perfect device. A little re-routing of ground and filtering of the power supply will do absolute wonders. But you have to KNOW. Your knocking of the products exposes your ignorance. Go learn some electronics. You tell me where it fails to do what it is supposed to do and I will tell you why.

[@Michael-w8v]

You can get old HP too. They're better made. There is bigger oscillator on some signal generator.

[@1pcfred]

Michael Marr but can everyone get an old HP now? This thing, I can just order one. Bang, done.

[@bazzaar1869]

Pretty clueless review, you can not measure the rise and fall times of square waves at 10MHz with x1 input of the scope! Notice the way changing the frequency resulted in changing amplitude, clear sign that input capacitance of the scope is having an effect.

[@Graham_Wideman]

Are you referring to the amplitude variation with frequency at around 2:50? That's due to operating near the signal generator's specified bandwidth limit of 10MHz, so of course the amplitude drops off with frequency. Also, I think you may be confusing bandwidth issues regarding 1x vs 10x _probes_ contrasted with the actual _scope input_, which is always "1x". As you can see at 6:19, Jack is not using a probe. He's using a BNC cable. I didn't catch whether he mentioned it, but he may also have the scope set to 50 ohm input impedance, considering this is one of the few function generator reviews which shows no ringing on the square waves. If true, the scope input capacitance will have negligible impact.

[@gacherumburu9958]

Poor man's equipment

[@stevenbliss989]

What is the MAX II chip??????