Don't buy any kit with XR2206 WFG

[dbctronic]

Every once in a while, you need to bang up some little analog project and don't want to get stuck programming to drive a digitally controlled thingy. Sure, digital waveform generators there are, aplenty, but...

Anybody remember the venerable XR2206? Easy to use, full analog square/triangle/sine generator. Well, after they went out of production, one or more Chinese firms made knockoffs. There are still a few floating around, always with the date code 0489. You may see a little bright red board WFG kit, in particular. DON'T BUY!!!

This run of chips has a sine output that is only about 20 mV, with huge switching spikes that are impossible to filter out. I found out that this was a generally known problem.

I went on a tear last year looking for XR2206s, and in desperation bought a couple of such kits from several firms just to snag the chip. All had the bogus 0489 b******, and they all had the defective sine output. And the kit was perfectly useless without the 2206, of course. The kits were going for US$4 to 6, but when you add up several of these plus individual shipping and wasted time ... 

Time marches on...

[Seekonk]

Like how hard is to copy a chip. I don't mind fakes as long as they work.  I use a number of IR2153 self oscillating half bridge drivers.  A feature I use is shorting the timing cap to turn both outputs off. This is a stated feature in the pdf.  These fakes turn off the hi side driver, but turn on the low side. If just generating AC that isn't a problem. My circuit becomes non functional. Now I never know what I will get when I order.

[dbctronic]

Righto, Seekonk. If you stray from buying trustworthy parts, strongly consider playing it as plain vanilla as possible. A knockoff 2N3904 most likely has the standard base-collector-emitter pin order, but don't be surprised if some have a little below the minimum hFE, or burn up a little below VCE(max). These are minor fudges, but a blatant violation of specs such as you experienced is not, and more than once I've had similar designs in mind, especially trick enable/disable techniques such as yours, that would not have worked and would have required more redesign than it's worth.

I suspect that trick enable/disable designs are rather often subject to rude surprises. Annoying when making a one-off (    ), VERY annoying when in the middle of a production run (    ).

[Bud]

 If you need just one IC, drop me a PM. I have a NOS one that i can send.

[ledtester]

When did Exar stop production? I see a lot of photos with a "F0919" date code, eg:



And some more variations:





And are the ones sold by Jameco fakes?

https://www.jameco.com/z/XR2206CP-EXAR-Corporation-IC-XR2206CP-Monolithic-Function-Generator_34972.html

[Bud]

Based on this

Wide Supply Range, 10V to 26V (18V recommended for optimal operation)

my bet is Jameco is selling chinese fakes. The Exar datasheet specifies no "optimal operation" supply voltage. On the other hand chinese fakes are known for lower max operational voltage.

[Bud]

There you go, a post from 3 years back:

The original Exar is a good part, but long gone. A few years ago I got some from Jameco that obviously weren't good copies of the original. They generally work OK, but won't meet the sweep specs for a real Exar part, and only come close for one specific supply voltage. I'd be wary of using them in any circuit where a wide range DC signal is used for frequency control.

[ledtester]

Here's some more info on the operating voltage story:

https://youtu.be/ZCdTxRASqiQ?t=4m6s

The Elenco Analog/Digital Trainers (XK-550, XK-700) also use an XR2206 with +/- 12V supplies.

[dbctronic]

@Bud: Thanks much for the offer, but the one-off project got done another way, in fact without the oscillator altogether.

Yes, I ordered 2206s from Jameco, Mouser, and several other suppliers, all were bad 0489s. Decided not to try my luck on eBay!!! Exar stopped production in the 80s. Surprised to hear of any knockoffs with later date codes, since nobody sold me any, but I'd be wary of using them, based on my experience and the stories above.

Another chip I wanted, but which is long gone, is the Maxim MAX038, a 30 MHz analogue of the 8038/2206 family. I actually talked to a support rep at Exar who had worked for Maxim at the time that one went out of production. The week the discontinuation announcement went out, her phone lit up with "Can't find them! Where are there any???" In less than two weeks they were totally gone except for the occasional popup on eBay or the like. I really wanted that one, and may again next year, time permitting. I have actually tried to design my own discrete knockoff, complete with UHF transistors pressed into service as a high speed bilevel Schmitt trigger, but I suspect it's going to be too much work.

[wraper]

Yes, I ordered 2206s from Jameco, Mouser, and several other suppliers, all were bad 0489s. Decided not to try my luck on eBay!!!

Sounds like bullshit. Mind to explain how you managed to order it from Mouser when they are displayed as obsolete for a quite some time?

[srb1954]

I designed an XR2206 into a product in 1978 and that product stayed in production for 13 yrs to 1991. At that time the factory producing this product would have never used anything from China so I guess the purchasing dept were still able to source the genuine part until 1991.

[dbctronic]

Perhaps memory mis-serves me on Mouser. I know I ordered three chips, and two of the stupid WFG kits from, of all places, WalMart online. Tayda was one supplier, my only unhappy purchase from them, and it's the only thing Jameco has sold me that was a dud.

Somebody here may want to rein in their trollish posting manner a bit.

If they were still being sourced in '91, well, yeah, you could probably still scare them up if they went out in the very late 80s, or in '90. I had been told they went out in the late 80s. Anyway, long ago.

[kripton2035]

I bought this xr2206 kit on aliexpress
https://www.aliexpress.com/item/32963105557.html
and the date code is F0918
will try to find it in my boxes and test the sinus output voltage.


Edit : it outputs a 5v sinus, that is quite distorded, but it is ok until 3v pp
I measured @ 700Hz

[vk6zgo]

Its not always just Chinese "clones" that have problems.

Back in the 1990s, I had to pull apart an "in house"  made TV transmitter remote controller to change the operating method of one of the emergency programme sources.

Part of this device used a couple of National Semiconductor monostable ICs to provide longish delay times.
Whilst playing with the unit, I dropped, & lost, one of the monostables-------------bummer!

No problems, though, plenty in the store.

I grabbed one, checked the operation---------suddenly, it kept timing out early, so the controller function didn't work any more.

I borrowed the IC from one of the other functions, & all was well, but when I tried the function the new one was swapped to, it too, failed to work as designed.

Must be a bad IC, let's just grab another!
After 4, it became obvious that the new devices had a problem the originals did not have.

Much research followed----no Google back then, just digging through books.

It turned out that NS had been getting complaints that the mono wasn't reliable for short time intervals, so they redesigned it.
The problem was, it now didn't work properly for long intervals, so they re-released the original IC under a new part number.

We had made just two of these controllers, so it wasn't a major drama ---------we just bought a stock of the new part number ICs, changed the documentation to reflect this, & were good to go.
Imagine if we had been a manufacturer with tens of thousands such units out "in the wild"!

[Alex Eisenhut]

But does the kit itself work with a real XR2206? Asking for a friend... who buys lots of junk on eBay....

[dbctronic]

Alex:

I didn't try it. It's based on a junk chip (hard to believe they didn't know), so I junked it all. It has a bright red circuit board, about 1-1/2" square, and a screw on lexan front panel, actually pretty attractive. But untrustworthy. Make purchasing decision accordingly, sez I. As has been pointed out here before, eBay and PayPal don't really have your back for these small dodgy purchases.

See Kripton2035's post up this thread a few places, purchased on AlieExpress. That's the kit I bought, better sine output but still distorted. Somebody else further up the thread offered to sell a known good 2206, but evidence suggests your friend would best not bother installing it in this kit. Hmm.... keep looking.

[rsjsouza]

Just a warning about the XR2206... Even the original. It suffers from spikes on the sinewave and the fix is somewhat obscure.

The discussion below mentions a few details
https://www.diyaudio.com/forums/solid-state/129493-xr2206.html

[Alex Eisenhut]

Hey thanks, I might have to update my sig again. 
I am waiting for two of these little function generators, the other type is based what they claim is a ICL8038.

http://www.mit.edu/~6.331/icl8038data.pdf

https://www.ebay.ca/itm/Professional-ICL8038-Function-Signal-Generator-Sine-Triangle-Kit-NEW-Wave-K7S9/174306174318?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

I mean this isn't some high accuracy thing, but it would be nice if it worked about as well as a late 1960s shoe-box like a Wavetek.

Just a warning about the XR2206... Even the original. It suffers from spikes on the sinewave and the fix is somewhat obscure.

The discussion below mentions a few details
https://www.diyaudio.com/forums/solid-state/129493-xr2206.html

It's the PWL method of warping a triangle wave into a pseudo-sine. That tip is hard to get rid of, but the ICL8038 seems to have a more complex PWL implementation. (That pile of transistors between pin 3 and 2 in the 8038 vs the much simpler 2206 structure).

I hope it'll make a decent waveform.

[rsjsouza]

Just a warning about the XR2206... Even the original. It suffers from spikes on the sinewave and the fix is somewhat obscure.

The discussion below mentions a few details
https://www.diyaudio.com/forums/solid-state/129493-xr2206.html

It's the PWL method of warping a triangle wave into a pseudo-sine. That tip is hard to get rid of, but the ICL8038 seems to have a more complex PWL implementation. (That pile of transistors between pin 3 and 2 in the 8083 vs the much simpler 2206 structure).

I hope it'll make a decent waveform.

Yes, that is quite difficult to "purify": I have a signal generator that operates in the same way (BK Precision 3030) and it works well, although the spikes show up depending on the load.

[AdrianFu]

I bought one of the Chinese XR2206 function generators. It currently sitting in the junque-box. Might use the pots one day. Ended up putting together an Arduino based keyboard input with digital display function generators.
Works much better.......well just works unlike the Chinese version

Adrian

[jewelie]

Quick update on the cheap Chinese XR2206 signal generator kits.

I got one recently, cheap as I could get from AliExpress, for the hell of it, knowing that they were a bit iffy.

Mine does have the high frequency spike issues on the sine wave, which I'm not particularly bothered about.

Curiously though, unlike all the videos and reports I've seen before, and unlike what the supplied paperwork said, it seems to be stable above 12V!?

Additionally, with a little bit of modding using the example circuits from the original chip's data sheet-

• a 10uF cap between GND and the amplitude pot centre, plus
• changing the 330R resistor to a 500R trim pot (the resistor that's in series with the jumper to change between triangle vs sine wave)

then I could can get a sine wave that seems to measure less than 1% THD (excepting those little spikes) and is stable and otherwise clean down to 20mA p2p.

I get the best sine wave shaping at approx 215R on that 500R trim pot, which makes me wonder if the 330R resistor used by default is a mistake and should actually have been 220R?

- Julie

[Kleinstein]

The best resistance for the sine trim can vary from chip to chip. So some may work better with 220 ohm and others may want 390.
The 330 ohm may be taken from a datasheet or testing some prototypes. There is a reason they had the extra timmer in the original and not just a fixed resistance.

[jewelie]

Fair enough!

BTW For reference the chip in mine that's stable above 12V appears to be this one marked XR2206CP HLF A8  http://hlf-ic.cn/en/products/show-134.html

[jonpaul]

in 1973, I used the  very similar Intersil 8038 function generator in my Popular Electronics Super Audio Sweeper.

50 years pass, I still have the hand wired prototype. still works, but needed new lytics and contact cleaning  of Allen Bradley AB pots and rotary  switches.

Amazingly it's  in cal, but Frequency at high end off 4%, needed a tweak of trim pot.
Was an MITS kit, and PE cover.

Found spare parts and discovered the super rare orignal ceramic and gold Intersil ICL 8038s as well as vintage Exar 2206s, (second design choices). The bipolar voltage regulator was a Raytheon 4194 in T0-66 métal can.  these rare ICS are seen in the photo

Enjoy,

Jon
the ramblings of an old retired EE

[Mark0x01]

An old thread, but finally inventoring by stuff and I have 17 new old stock XR2066 dated 8327 if anyone needs one.