Hmm, it says Texas Instruments, but the one in the photo seems to have the logo of Signetics, which looks like it stopped chip production by 1995. Vintage!
3 Likes
Signetics.
Stopped producing around the 80s and then became Phillips, which is now NXP.
It’s been a long time since I’ve seen this brand. Now i’m scared…
Wikipedia says:
“In 1995, Philips spun off the assembly and test operation in South Korea, which was started by Signetics in 1966, as an independent subcontract service provider. They continue to use the name Signetics . Since 2000, Signetics is primarily owned by the Young Poong Group.”
1 Like
Signetics released a catalog in 1991 (which included the LF398)
The South Korean Signetics apparently used to assemble PDIPs, though not any more (now just SMD):
https://web.archive.org/web/20010420142236/http://www.signetics.com/st_assc.html
By “assembly” I infer they mean they package dies produced elsewhere. If that’s the case I presume the product would bear the logo of the semiconductor manufacturer rather than Signetics.
2 Likes
I think this makes sense.
In the catalog you posted says: LF398 SO-14 Package.
Unless this ebay store sells oldstock components, but I think it’s suspicious they put Texas Instruments in the description and Signetics in the product photo, as noticed by you.
1 Like
Since this topic is about S&H, I really like this one by Ray Wilson:
PS: If it’s not the right place to post this, please let me know.
The big advantage from Ray’s Micro S&H is the use of common components that you have on hand and some extra features like Glide Control (very interesting) and Trigger Out.
Okay, there are JFETs and those are hard to get, most I see out there are fake or out of spec.
But, since JFET acts like a switch in this design, it won’t require as much as a buffer. I tested several types and worked ok.
I saved some components from the time I was building a lot of Ray’s stuff.
Some of them:
Maybe modern components work better, who knows.
There’s also this version, I remembered this one when watch the Moritz’s project:
Ray’s Single Chip Simple Sample and Hold
In this one Ray uses TL084, and does not have polystyrene capacitors.
Both versions are easy to build and can fit on a small piece of stripboard.
On the MFOS website there are other different types.
Anyway, just to kill time while the LF398 (the real ones) aren’t delivered.
4 Likes
Ah many thanks, I’ll try that too 
It seems the MPF 102 is just a simple N-channel JFET and I guess it’s replaceable with J112 or J113 (I have tons of those).
Meanwhile I also also built @moritzklein’s S&H and added a 40106 oscillator circuit to it as internal trigger. It also works nicely and is very simple. Here is the base circuit from Moritz’s Sample and Hold YouTube video
6 Likes
The LF398 from the other seller have arrived, they work like a charm! The test circuit, the breadboard and also the strip board 
9 Likes
Excellent!
And not the same vintage Signetics chip as in the listing photo — I can’t quite make out what it is though.
3 Likes
Analog electronics are already hard enough without having to debug faulty/fake components. In my experience paying a little extra for the peace of mind that reputable sellers bring is well worth it in the long run.
In any case I’m glad you got it working! My S&H circuit is still sitting half finished on the breadboard…
5 Likes
I fully agree! I usually don’t use cheap components but the LF398 was hard to source and I did not want to wait weeks. 
Now I have learnt another lesson
5 Likes
Nice job! 
1 Like
Looks like National Semiconductor.
Sorry, forgot to post a good picture! Yes as @peternoya spotted, the logo is from National Semiconductor. Whatever is inside, the stripboard is still buggy but the test circuit works nicely and it holds 90% of the voltage for a couple of minutes. Maybe I’ll try the DAQ features of my new scope 
Anyways, I have also rebuilt the René Schmitz circuit on breadboard and it works. Here is triangle signal sampling:
Later I’ll debug the stripboard… The panel is already done but the signal is acting weird 
7 Likes
Reanimating this thread!
I completed my S&H pcb today (Rene Schmitz YASH) and it works, but I have pretty high droop rates 
like more than a semitone in 10 seconds. Do you observe the same in your builds? From the datasheet I read that the droop rate depends on the hold capacitor and 1nF (I used PP, btw.) corresponds to something between 1e-1 and 1e-2 V/sec, so let’s say it’s 1e-2 that would mean 10mV per second right? That’s a lot! So that means that my S&H works as it should but I somehow expected it to be more stable…
1 Like
Yeah, I get a considerable droop rate too. From reading it appears droop rate is lower with higher capacitance, of course that also increases charging time but as built charging time is 3 µs so there’s definitely some room for slowing that down. Haven’t tried it though.
2 Likes
You mean when I increase the holding cap I also need to increase charging time? Makes sense! I will check the datasheet later to see if I can compute the values
1 Like
For comparison MFOS’s (discrete) S&H uses a sample time of 100 µs and a 10 nF hold cap (http://musicfromouterspace.com/analogsynth_new/SMPLANDHOLD200607/SMPLANDHOLD200607.php).
2 Likes
Hm… okay, I think the relation should be linear, because the time constant of capacitor charging is
$ \tau = R C $ (oh, latex not working )
So, if I want to use 100nF as hold cap, I would need to increase charging time by 100 as well. 300 us sounds okai… that would mean I need to either use a 47nF cap or a 1M resistor in the timing circuit, I guess. I should have used sockets for these on my board 
2 Likes
Breadboard it! I have no idea how short the duration needs to be for it to be non audible, other than that MFOS used 100 µs so I’d guess that’s safe.
3 Likes