#97 Simple function (signal, waveform) generator

[@youtuuba]

Ralph Bacon, a few observation with this video: - The XR chip, even when it is not a counterfeit one, cannot produce an actual (proper) sine wave. The best it can do is an approximation of a sine wave. On a quick glance it kinda resembles a sine wave, but on closer inspection, or if measuring the distortion, it is pretty corrupt. If somebody just wants a low-harmonics waveform for signal injection, it is fine, but if using to design or evaluate filter circuits, etc; it simply won't do. So, only useful for "experimenters" if they already have a good understanding of the limitations of this marginal sine waveform, and most people who care will already have a much better function generator and won't touch THIS one with a 10 foot battle lance. - If you mentioned it in your video, I missed it, but this kit is missing the critical AC coupling output capacitor, thus a high DC offset, so immediately will cause all sorts of issues for "newbies" trying to do experiments with it. - When you did your diagram to explain the function of a Schmitt Trigger, you were in fact describing a normal comparator, quite different from the hysteresis that a Schmitt Trigger provides. In other words, you showed it switching at the same voltage level, while in fact a Schmitt would turn on its output at a higher voltage and then turn off its output at a lower voltage. - This kit has so many blunders in its design and parts supply that it probably causes more newbies to quit out of frustration that it encourages anyone to continue in the hobby. Even with all the 'fixes' promoted on various KZbin videos, it remains a pretty awful and awkward signal generator. A 555 astable can give a nice square wave, or a Wien oscillator built around an op-amp can do a better job (albeit still imperfect) of sine wave generation, and triangle waves can be easily derived from a square wave using another op-amp. One 555 and a dual op-amp can give better results than this kit.

[@RalphBacon]

I hear you, although the kit with the original chip works very well indeed. It's that fake chip that causes a lot of problems. For actual function generation I use a dedicated Kuman FY6600-60M dual channel device, nice clean signals and you can even program your own, which is useful to emulate switch bounce! That said, for simple tone generation this kit here still works "OK" although you do have to have your wits about you. But beginners are loathe to spend the money (on that Kuman, for example) so I show what is available. Yes, the ubiquitous 555 does give a nice signal, too. Good to hear from you, thanks for all the info and heads-up. 👍

[@Ed19601]

Xr2206 that's really an oldie. I think I built a signal generator with it some 35 years ago. Elektor project

[@RalphBacon]

Yes, it had a good run, Ed19601, about 4 decades or more! Now retired, except that they are now producing fake chips. Makes you wonder why they went to the effort really. Oh well, it still works (sort of) at 9 volts!

[@Ed19601]

Ralph S Bacon good to know that. I am afraid the old XR isn't the only chip that's being faked

[@RalphBacon]

Indeed, just ask FTDI. (And the poor Arduino owners who had their USB to Serial converters bricked by FTDI as a result).

[@Ed19601]

Fortunately that didnt happen to me yet, but I did get some questionable components from time to time, like an NRF24L01 and I learned to avoid 1 store on Aliexpress like the plague. I think it was 1977 that I built it, but there was a 1984 follow up ketel.home.xs4all.nl/ham/84111NL.pdf (dutch, but the circuit is clear)

[@RalphBacon]

My, my, how time flies. That article was published around the time I was getting stuff published in Practical Electronics (a digital model railway signal controller, something that could be done today with a $2 Nano in a very short time!). I struggled a bit with the Dutch (even with my German background) but I suspect the Elektor design was considerably better than the kit I demoed. I did laugh at the last paragraph where it is suggested that perfectionists could experiment a bit with various capacitors - a bit like "we know it could be better but we've spent too long on this already"! Anyway, so you now avoid a particular Aliexpress seller? Be brave, name names! Or partly anyway. Instructables have just sent me an update to avoid a seller too, initials J.O. Coincidence? If it [the price, usually] seems too good to be true...

[@t1d100]

The duty cycle from the XR2206 Triangle Wave is different from the duty cycle of the Schmitt Trigger Square Wave, because the Schmitt Trigger 50/50 duty cycle trigger voltage point is different than the 50/50 timing point. In other words, the duty cycle trigger point needs to be adjusted.

[@RalphBacon]

More than likely. For such a simple device I'm not surprised if it's now exactly 100% but is it good enough for hobbyist use?

[@t1d100]

No worries, as I know there is a language barrier for this particular term, but... In the USA to "knock up," is an impolite term for "to make pregnant." No kidding, here. We might say "knock together," to mean cobble together. Also, we would not say "I'll knock you up in the morning." Thought you might like to know.

[@RalphBacon]

Hmm... UK also has the same "knock someone up" but it sounds as though it's from the second world war years to my ears. That is, somewhat old fashioned. But we can also knock up a quick prototype without getting schoolboy sniggers so I guess it is a bit different here. Good to know!

[@HazeAnderson]

I am currently in the process of rebuilding my kit according to KainkaLabs's suggestions. I built the kit in November 2018, desoldered all the pieces off about 1 month later, and I am now messing about with the clone XR-2206 chip on my breadboard, hoping to solder it all to a new PCB within a few days.

[@RalphBacon]

Good luck with your desoldering project, Haze Anderson and thanks for posting.

[@jstro-hobbytech]

Best way to learn to solder is to make led panels with blank perf boards and simple logic circuits. Like a panel of leds where some blink depending on the 555 timer and a decade counter. That's how I learned. The first panel I made was a 5x7cm board using only random blinking rgb 3mm 2 pin leds. My bench supply cannot kill it and it can put 10 amps through it hahahha Pro tip. Buy cheap Chinese led cube kits for 9 dollars and get 1000 leds and 8 74hc573 ics and other bits and bobs for real projects. Where I live on east coast of Canada soldering kits are expensive along with leds. I wait for a sale and for 20 bucks I get 2000 blue leds and lots of other cool switches and ics. I throw away the project pcb and stc branded mcu and keep the rest haha. There's lots of hacks to get things for far less if you're on a budget

[@RalphBacon]

Sounds like a cool (and dollar-stretching) tip, Joey, thanks for sharing!

[@jstro-hobbytech]

I just ordered a 2 Chan 100mhz function generator which I got at cost. Totally overkill but I was going to buy a 20mhz model and this was offered for less. I'm going to use it to a-b opamps using high and low pass filters. I'm splitting my time between creating circuits that use code and trying to mimic them using only passive components. You know, simple stuff like pir sensors for lights or for when my cat Ralph is into things he shouldn't be while I'm in bed haha

[@RalphBacon]

What a great name for a cat! 😸

[@raymondheath7668]

Back in the early 80's we used 74LS14 schmidt for logic level voltage. I built this signal generator and will try the mods to improve it

[@RalphBacon]

I have an (allegedly) better one on that slow boat from China for about £1 more, mentioned in these video comments by Javier; it's based on a CL8038 with an op-amp, from AliExpress. Now, Raymond, If you go ahead with the improvements with your existing one it will be interesting to see how the two compare. If the improvements are really worthwhile (I have no reason to think they are not) then I'll probably implement them too. Look forward to see the results!

[@elektron2kim666]

Functions are the sub formulas around the thing and the components.

[@RalphBacon]

Quite so, a good way of describing them.

[@mikebarton3218]

Hi Ralph, thanks for taking the time to answer my comment. I bought one of these kits a few weeks ago but haven't got round to assembling it yet - it's in a drawer somewhere. I wonder which chipset it uses. Thank you for the well produced and very informative videos. I look forward to every one. Mike.

[@RalphBacon]

Thanks you for your continued support, Mike! I'll lay odds yours too is a fake chip (try running it at above 12v and you will see that jitter), as the chip hasn't been made since about 2013. Not by the original manufacturer anyway. But running at 9v it's not too bad with those enhancements suggested by Kainka Labs (and the other one you mentioned). However, just like in Star Wars, this was not our only hope. There is another (kit). Stay tuned.

[@Tarbard]

I learned something and enjoyed it too.

[@RalphBacon]

I find that if I enjoy doing something then I don't even realise that I am learning, so it's great to hear that you enjoyed this video (and learned something too, win-win). Thanks for the feedback.

[@boldford]

Have you compared the component values in this kit to the application sheet. I think they used whatever was laying around in Shenzhen. If it works, it's more by luck than judgment.

[@RalphBacon]

Ha ha, you old cynic Brian, they wouldn't do that! Er, would they? Er, oh, I see what you mean. Hmm. Well at least you get a nice acrylic case! Actually I've had a different one delivered now, for a similar price which is *supposed* to be much better. Uses a totally different (fake?) chip to generate the signals. All will be revealed in a future video, let's hope it's better than this one.

[@mikebarton3218]

Hi again Ralph. There are a couple of videos from Kainka Labs and Electronic Stuff about the problems with these kits including fake chips. Mike.

[@RalphBacon]

Yes, indeed, I put a link to one of them in the video description. But there is (allegedly) a better signal generator kit available for about the same price, it's on order, we shall see in a future video!

[@binarybox.binarybox]

Several people have pointed out the problems. Here are the two videos from KainkaLabs... . kzbin.info/www/bejne/ZmO7paGbeK2heqM kzbin.info/www/bejne/nJDKn6KLZ5d3oJI If you have a fake chip, using under 12 volts would cure the instability and use capacitors on the outputs to avoid the dc levels. Best to check all the probs b4 connecting the kit to anything.

[@boblewis5558]

OMG! That is the exact same chip used in an Elektor project back in the late 70's and used in my mini function generator from that time! Still have it, still use it, still works! I MUST be getting old since I even remembered the precise chip number - well without the "cp" (no spoilers!) 😁😁

[@RalphBacon]

Well, some of us are not convinced it is _exactly_ the same chip, as it has issues at the higher frequencies. But it's OK for occasional use of the type you mention (I suspect you have a kosher chip if you built that in the late 70's - did they even have computers then?). Can you remember how much you paid? Significantly more, I suspect...

[@boblewis5558]

@@RalphBaconOops! Whilst I knew the XR2206 name well and I found my original photo positive for the Elektor board, originally from the January 1978 issue, I screwed up. Here are a bunch of pics of my function generator, which as you can see quite clearly is NOT based in the XR2206, despite my having a photo positive of the board ready to print/etch!! photos.app.goo.gl/FfGuqLxY5dZAqhqt7 Clearly, I never got around to making it (!!) and used another design. BTW, there's a great archive online of almost the whole series of Elektor from a guy in India, which seem to be exactly the same as the UK issues. Many of their circuits are still useful and relevant to electronics enthusiasts and worth a peruse. Some of the most useful quick projects were in their bumper double issue summer editions: archive.org/details/ElektorMagazine I used to have an almost complete set of my own from Number 1 right up to around 1997/1998 which I was forced to junk at the insistence of the dragon indoors when we were moving back up to Yorkshire two years ago! DAMN her! 😉 It totally threw me when I opened up the unit earlier today (first time since it was built!) I can't remember now WHERE I bought the board, but suspect from Maplin - sadly also no longer with us. It wouldn't have been as a full kit as I recognise many of the components as ones I had, and still have, in my boxes of parts. The chips and probably switches and pots, would have been from Farnell's in Leeds as after I'd left Ferranti (Manchester) and joined DEC (a name you previously mentioned) in Leeds in 1979, Farnell's were just down the road. Yes there WERE computers back then 😝, REAL ones, with lamps and switches on the front(!), not PCs! I know because I had to repair them and the associated dishwasher sized disk drives with the magnificent capacities of 5Mb to 26Mb! They were also around in 1968 when I was using an ELLIOTT 903 with the ENORMOUS memory of 4k (yes 4096) bytes of memory which soon doubled to 8k! 😎. So less of the lip "youngster" LOL 😉 The chip in my unit can be seen to be an Intersil ICL8038, also now obsolete but still readily available (more clones!!??), that does pretty much the same job as the XR2206, and makes for a pretty neat and compact unit. Unlike the XR2206 though, the chip handles from a very low 0.001Hz to 300kHz. Datasheet here: pdf1.alldatasheet.com/datasheet-pdf/view/67443/INTERSIL/ICL8038.html It is certainly a kosher chip, but I'm not sure about today's. I honestly can't recall exactly how much the unit cost but it would not have been too expensive ... I'm tight! Probably less than £15-£20 including the case. The board was almost certainly a Maplin board - also sadly no longer around.

[@RalphBacon]

Well, at least you were _about_ to use that chip at some time in your dim and distant past - but I guess the kit you actually used worked better (or was less effort, or was cheaper). But I'm amazed you even remembered this chip from the 1970s!

[@dragan3290]

The dso 150 looks identical to the dso 138? I have the 138 and I love it but I guess your dso 150 is better looking. Cheers from Australia

[@RalphBacon]

Glad you're happy with the 138! I still use the 150 on occasion (for low frequency work) and for the money it is well worth it.

[@pekkagronfors7304]

Interesting and informative as usual. Cheap stuff like this is perfect to learn the basics with. I also have the DSO oscilloscope like yours.

[@RalphBacon]

In which case, Pekka, you'll like a couple of other cheap things I have in the pipeline. As you say, they are ideal to learn the basics. When and if someone outgrows the JYE oscilloscope at least they will know *why* they want or need a "proper" one! Thanks for posting, good to hear from you again.

[@MrBobWareham]

Hi, Ralph, I think you will find that as the chip is a sad copy of the original that it is not very good and providing a true signal if I remember it was something to do with voltage I do remember a review some time ago, Bob

[@RalphBacon]

Yes, it is a copy chip which is not as good as the original. Pity. I still have another signal generator kit which has a better spec but have never built it. One day, one day...

[@MrBobWareham]

Hi, Ralph, I think you need to look on youtube to see reviews on this kit to find out the chip is not genuine and the output is not true, have you now got a dry cat? all the best

[@RalphBacon]

Yes, I came to the conclusion, Bob, during the video that the chip was not genuine (my 12v test proved it really) but that it still worked at 9v, so for the money it sort of still has value. Maybe. But the video link I posted does detail all the changes required to get this to work much better. Benny is most certainly dry now, but as it is getting colder he comes indoors more often now anyway!

[@markhall3323]

…for me biggest problem using an Arduino Due etc is the very limited bandwidth they just don’t run fast enough to generate high frequency signals

[@RalphBacon]

Yes, you need a faster processor such as the STM32 or even an ESP32.

[@javierpallalorden]

Hi Ralph, very nice video. I bought the ICL8038 version with the 7660 voltage converter and op-amp, far better version for under €8 (on AliExpress).

[@RalphBacon]

I will certainly take a look at this other version, Javier. The kit I used can be had for about £5 on AliExpress so it will be interesting to compare! Thanks for the heads-up.

[@boblewis5558]

@@RalphBacon after you've read my latest long reply to you, if you DO try the ICL 8038 version, it may be interesting to compare the then and now versions since mine is 36 plus years old and a kosher original chip! It's socketed too so it'd be easy to do a direct comparison in the exact same circuit. No other variables other than the chip change!

[@KJW648]

Have you hacked your Rigol DS1052E Oscilloscope into a 100MHz DS1102E ?

[@RalphBacon]

It wasn't me that did it, the seller offered this simple 100MHz upgrade for about £5 extra so I took him up on it (better than me breaking open the case and invalidating the warranty). You've reminded me that I never did check how well it could display a (near) 100MHz waveform (mainly because I don't need ones that high) but now I might just do that. You got one of these too, Kuk?

[@KJW648]

Yes, I have DS1052E badge running 100MHz, the DS1102E and DS1052E are the same model, the only difference is the badge and software. I modified mine myself and see how it is done EEVblog #70 kzbin.info/www/bejne/gp_LiZmMjryMjqc and #77 for update :)

[@DrexProjects]

I bid low on these on eBay but never won. Good to know they are not that great. Thanks for the Vid.

[@RalphBacon]

Another viewer @Javier says "I bought the ICL8038 version with the 7660 voltage converter and op-amp, far better version for under €8 (on AliExpress)." so maybe that would be a better choice anyway (I will take a look later).

[@MrJanulis]

1. Duty cycle SHOULD act like it does with different amplitude signal you feed it as it is a comparator circuit (just with hysteresis). 2. Rotary Encoder will bounce at both fronts (rising and falling) and I can hardly understand why you are trying to use Schmitt trigger for debouncing?

[@RalphBacon]

Ah, you mean when converting the triangle wave into a square wave, as the signal gets larger in amplitude the pulse stays high for longer due to the hysteresis of the trigger - the signal does not fall to the point the Schmitt trigger switches off again until a longer period has elapsed. Yup, you are spot on there. Hmm, something I should mention in my next video on this subject, thanks for the clarification. Now, I'm using the Schmitt trigger with a capacitor; the capacitor filters out the quick changes by stopping the voltage reaching VDD immediately thus preventing tiny (as in short) bounce spikes from triggering the Schmitt trigger due to the hysteresis. The Schmitt trigger thus only fires on longer pulses that we can reasonably assume to be the final state of the switch. It's a classic design and we'll see it better when I connect up my other signal generator in a future video. The capacitor value is important to filter 'spikes' from bounces but not slow down the edge too much - but then the Schmitt trigger cleans all that up rather nicely. Thanks for posting Aidas, nice to hear from you.

[@MrJanulis]

I agree that it is a classic design :) but I guess it was designed way before cheap microcontrollers were born :) and worked really well. But for now, we have different possibilities to process an incoming signal. Why am I arguing? Each RC integrator is designed to filter off certain frequencies and can do a decent job. But what exact frequencies we would like to filter? It greatly depends on the switch / encoder we are using - the different materials used for contacts, the different lengths of contacts, and even the age of the encoder - all that count. Not to mention rotation speed. We need a universal solution that can work on different encoders from different manufacturers with different weary levels and different users patterns and moods. I just found an amazing document on this subject - hope it will help you, Ralph, to perfect your solution we all want. There it is: www.eng.utah.edu/~cs5780/debouncing.pdf My pleasure to help as much as I can (would like to have the similar equipment you had and experiment a lot, but due to our way of living - a lot of traveling and volunteering around it's impossible). With warmest regards, Aidas

[@RalphBacon]

That document is a real eye-opening, Aidas, thank you so much for posting. I'm very tempted to create the enhanced RC debounce circuit on page 15 as he indicates it is almost foolproof. Hmm, will add this to my list, thanks again.

[@MrJanulis]

I was hoping you will do SW debouncer as on Page 20 or at least page 19 :) Why you are trying to ADD conponents instead of removing them? Anyway, happy you found it helpfull

[@RalphBacon]

It's a funny old world, Aidas. Last video I was reprimanded for NOT doing hardware debouncing, and this video I'm being sternly told not to add yet more hardware, do it in software! Just goes to show you can't please all the KZbin viewers all the time! But I will certainly look at the software side of things (again) to see what I might learn. It's a very good article, I must say.

[@sickvic3909]

Nice. Wish I had this stuff as a kid. I wouldn't have to ask this now. What the hell do you do with it? :)

[@RalphBacon]

As an example then, Vic, if you were driving a MOSFET for PWM purposes (a robot motor, say) you could inject a square wave into that MOSFET-based hardware to prove your circuit was working correctly whilst at the same time determining the best frequency and duty cycle before actually programming an Arduino to do that (it's quicker by far to use a function generator than constantly modifying Arduino code). There we are, a Real World example!

[@sickvic3909]

Ralph S Bacon IC Thanks

[@MrBobWareham]

Hi Ralph have a look at this @

[@RalphBacon]

Bob, your link is not there, it is an ex-link, pining for the fjords.

[@avejst]

Thanks for sharing :-)

[@RalphBacon]

Glad you liked it Asger, it's caused a fair few comments - and guess what? A better kit is available which I will have to share in a future video!

[@mikebarton3218]

Ralph, that's a rubbish square wave! A 555 could do better than that. :-)

[@RalphBacon]

Well, with a bit of tweaking it sort of turned out OK (rubbish at the [near] 1MHz, admittedly) but there are alternatives apparently, Mike, see the post from Javier here in these comments; I'll investigate that too to see if it's worth looking into. And yes, the 555 was (and still is) a good chip for this sort of thing (not at the 1MHz mark either though).

[@Ed19601]

unless you get a dodgy 555 as well ;-)