[sdiy] [OT] LM13700 spotted in commercial product of big Manufacturer

Brian Willoughby brianw at audiobanshee.com
Sun May 15 22:19:25 CEST 2022


What about the Analog Devices SSM2018 and SSM2164?

... or is a VCA (OVCE) not always interchangeable with an OTA? I realize that an OTA is flexible enough to be used in many ways, but it seems that a VCA is the most useful.


One promising company is THAT Corp, who obtained military chip fabrication that's capable of very useful analog silicon, and they seem to know how to use the equipment. Their prices are high, but considering the limited volume, it seems a reasonable price to pay.

http://www.thatcorp.com/THAT_Semiconductor_Fabrication.shtml

They have three VCA chips in production:

http://www.thatcorp.com/Datasheets.shtml


p.s. Thanks for the Twitch.tv handle.

On May 15, 2022, at 04:04, Matthew Skala via Synth-diy <synth-diy at synth-diy.org> wrote:
> On Sun, 15 May 2022, Roman Sowa wrote:
>> Why would ever LM13700 disappear?
>> It's the cheapest OTA with thousands of uses.
> 
> There are two different issues being conflated in this thread.
> 
> The LM13700N, meaning the DIP version, certainly can be discontinued
> because it already has been.  Texas Instruments PDN number 20210505000.3
> announcing this was dated a year ago Thursday - May 12, 2021, with the
> last order date being one year later.  I have my "last time buy" box right
> here, next to my similar boxes of TO-92 transistors.  "Why" they would do
> this is no longer a high-priority question given that it did happen, but
> the PDN says "Reason for Widthdrawal/Discontinuance:  The die attach for
> these products is being discontinued."
> 
> That is probably what the original poster had in mind when describing the
> use of an LM13700 in a new design as "unwise."  However, it was just
> specifically the DIP version that was discontinued by TI, and I think it's
> unlikely that Fender is actually using the DIP version.
> 
> All DIPs in general are on borrowed time, a fact which concerns me given
> that I really like DIPs and I sell products that use them.  I've talked
> about it several times in my video stream (Mondays, 3pm Eastern,
> "matthewskala" on Twitch).  There are DIP LM13700 clones available from a
> couple of Chinese manufacturers.  CoolAudio offers a 13700, and their data
> sheet mentions a DIP version, but it's not clear that they really make and
> sell anything but the surface mount version.
> 
> The LM13700 - no suffix, meaning the silicon die that might be packaged
> some other way than DIP - is a different question.  That is a popular IC
> with a big market, and it would be harder for TI to discontinue all
> versions of it than to discontinue just the DIP version.
> 
> But they still could!  Remember that one important reason the LM13700 is
> so popular now is because it is pretty much the last OTA available.  Many
> people using the LM13700 would really prefer to use the CA3080, CA3280, a
> Roland custom chip, or similar... but those are all gone.  Being the last
> one available improves its marketability, but the reasons why the others
> were discontinued do also apply to the LM13700.
> 
> The only remaining TI LM13700 is the LM13700M, the -M as opposed to -N
> denoting the SOIC16 package; and SOIC is a "dated" package too.  It's
> surface-mount and seen as more modern than DIP, but it's basically the
> oldest generation of surface-mount.  Most mass-produced designs are aiming
> for smaller packages and the fact the LM13700 is not offered in a smaller
> package does not bode well for it.  If they really thought this was a chip
> with a strong market for the long term, they would be introducing newer
> packages for it.
> 
> Consider the CA-series chips, like the CA3080 and CA3280.  All of the
> CA-series chips are gone now because they had to be made on a special
> process that offered physically large bipolar transistors for analog
> circuitry - totally different equipment from what's used for making modern
> digital CMOS.  Eventually the company that ended up owning that production
> line after however many mergers (I don't even remember who it was), didn't
> see the business case for running the special CA-series semiconductor
> process anymore.  Quite possibly, there was only one technician who knew
> how to run those machines, and that person retired.  All analog bipolar
> processes will go that way some day; and new ones are not being created to
> replace them.  New analog processes, to the extent they exist at all, are
> CMOS processes.
> 
> So I don't think the LM13700 is secure forever, even though it's probably
> secure enough for the next few years that Fender can reasonably use it in
> a new design; provided they're not using the DIP version, which is
> already discontinued.
> 
>> Just like TL0x, LM324, NE5532, LM2902 and other ancient designs it's not going
>> anywhere simply because it's cheap, and for decades used by large companies in
> 
> TL07x is an interesting case.  Go to the TI Web page for the TL074 and
> there's a big notice on it saying "A newer version of this product is
> available," pointing at the TL074H.  Read the data sheet for the TL07x
> chips and it is conspicuously two data sheets.  Nearly all the tables of
> actual data are split into "TL07xH" and "all others."  The top bullet
> points on the front page, and the first paragraphs of the main text
> sections of the data sheet (sections 3 and 8) are devoted to telling
> readers that the TL07xH is better than the other versions.
> 
> Notwithstanding the similar model number, it seems clear that the TL07xH
> is a completely different product, with a much newer design, and TI is
> strongly encouraging you to use it instead of the traditional TL07x.
> 
> The TL07x "all others" chips are JFET-input bipolar chips.  They would
> have to be made on a process specifically designed for such, and not
> useful for much else.  The TL07xH is analog CMOS, presumably made on a
> different process much more similar to processes used for modern digital
> CMOS.  It doesn't mention CMOS in the data sheet, but does on the TI Web
> site.  References to JFETs have been removed from the TL07x data sheet
> text, although the schematic in the data sheet still shows JFETs and BJTs
> and there is no schematic showing MOSFETs.  The TL07xH is also surface
> mount only.
> 
> We don't need a weatherman to tell which way this wind is blowing.
> 
> -- 
> Matthew Skala
> North Coast Synthesis Ltd.




More information about the Synth-diy mailing list