Can anyone with a LM301AN on hand relay a quick measurement for me?
Quickest way would be diode test on meter, negative lead to pin 7, positive lead to pin 2 or 3. These are the respected input transistors tied to V+.
I just want to know if what I have looks per usual. Comparing with other BJT input opamps only gets me so far, as I can only verify what they aren't.
I've got 800mv on both inputs on all of a few dozen.
I would normally give the benefit of the doubt, but the closest similar images I can find are of known fakers but with the brand logo rubbed off.
Plus, I thought that genuine devices also have the dot impression, which these do not. In fact, it seems that all Nat-Semi ICs I have on hand until these, have the dot impression.
I know I'm a pain.
You are lucky, I have a 1970s nat semi LM301A around my desk.
Old school digital 3.5 digit DMM. When measuring a diodes with my DMM, the positive lead on the anode of a diode gives a normal reading.
I'm getting no diode voltage in either direction; pins 7 to 2 or 3.
Ohms: Negative lead to 7, positive lead to 2 or 3 gives 18MEG ohm. Other way is > 20M.
It's possible the 800mV comes from protection diodes to +V that are part of their production process.
See which direction they point.
Well... according to the datasheets with schematics it shows that both inputs go through an NPN base with collector tied to V+, so my go-to was the simple diode test - as I do when verifying transistor polarity. So I'm confused on why this isn't working for you?
I've got 8M resistance with negative lead on V+, positive on 2 or 3. Shows open on my meter flipping them around.
Trying them out anyhow. Fingers crossed.
QuoteWell... according to the datasheets with schematics it shows that both inputs go through an NPN with collector tied to V+, so my go-to was the simple diode test - as I do when verifying transistor polarity. So I'm confused on why this isn't working for you?
That makes a lot of sense!
Dammit, tried another meter.
Diode range = ohm range 2k: -ve lead on pin 7, and +ve lead on 2 or 3: 810mV
Reverse leads and no reading.
Ohms range 20M: -ve lead on pin 7, and +ve lead on 2,3: 500 kohm (some meaningless measurement of 0.5V to 0.8V)
Reverse leads and get > 20M.
Try previous meter, now diode range: -ve lead on pin 7, and +ve lead on 2 or 3 : 0.79V !!!
Ohm range 20M: readings as before.
Grrr ... WTF ... surely a glitch caused by one of the UFO flyovers.
A third multimeter:
Diode range: -ve lead on pin 7, and +ve lead on 2,3: 0.76V
Open in reverse.
Resistance range only goes upto 2M but reads open in each direction.
When I ratched the ranges to 2k, I get :
-ve lead on pin 7, and +ve lead on 2,3: 990 ohms (again a meaningless equivalent of 0.5V to 0.8V)
Open in reverse.
In reality I think my first diode measurement was dodgy because the chip had oxidized pins. I plugged the chip into the breadboard to hold the part but measured from the top by placing the probes on the fat part of the leads. It could be 10 to 20 years since that part of the pins has been touched.
Thank you, Rob!
Are you getting a resistance reading between pins 6 & 8? Two nat-semi sheets and the Alfa AS301AN show a few resistors in between the comp & output connections. Expected ~7K resistance.
The on-semi sheet does not and the chips I have show open there as well.
I got 25.3k/25.9k depending on polarity. I checked two ranges 200k and 2M to make sure the readings were resistive and not due to VBE drops. (Also checked with two meters.)
I've extracted internal schematics before and the datasheets vary. Sometimes I don't believe them. Other times I merge info from more than one.
So your 7k5 comes from the National Semiconductor data (1999, 2000),
(https://i.postimg.cc/hJdLT6z6/ns2000-LM301-A-schematic.png) (https://postimg.cc/hJdLT6z6)
The ST 1995 datasheet shows 25k, (agrees with my measurements but my chips are National Semiconductor 1976)
(https://i.postimg.cc/v49Tzshr/st1995-LM301-A-schematic.png) (https://postimg.cc/v49Tzshr)
And when I see these variations I often go back to early Fairchild datasheets, this one from Fairchild 1973 and it has no such resistor,
(https://i.postimg.cc/cK0xB8rW/fch1973-LM301-A-schematic.png) (https://postimg.cc/cK0xB8rW)
Look closely and you can see different connections due to drawing errors!
Try to find national semiconductor 1994, 1995:
got 1994, it has the 7k5,
https://pdf.datasheetcatalog.com/datasheet/nationalsemiconductor/DS007752.PDF
I figured there could be a typo in the schematics. So far they look... okay. But the considerable difference in resistance measurements is a flag.
I was already prepared to cut my losses and snag the new Alpha ones, but was hoping this "reputable vendor" would pull through - even after sending me some crap by mistake instead. Being in the NOS market for so long, I'm not surprised at all. They look very convincing, just not what I've been seeing after sourcing obsolete stuff for the better part of a decade.
Plugged 32 of them into a massive build last night. LEDs are lit, LFO flashing but no output.
Not blaming the opamps just yet. I'm actually kind of happy that it didn't work right away as I've been too lucky lately and I can keep working on it :icon_lol:
I can only hope that the problem is not these opamps. On to debug & calibrate this beast which should make it more clear if the chips are legit or not.
Quote from: Kevin Mitchell on February 13, 2023, 06:17:57 PM
Can anyone with a LM301AN on hand relay a quick measurement for me?
Quickest way would be diode test on meter, negative lead to pin 7, positive lead to pin 2 or 3. These are the respected input transistors tied to V+.
I'm a pain.
I have three parts, each LM301AN. #3 is early, natsemi waves, flat divots, batch /749.
#2 is squiggle [late logo] natsemi, curved divots, wide white band, code M53AD.
#1 is waves natsemi, curved, wide white, code :M8930.
ohms test between pins 6 & 8 - 1 and 2 show ~5k each way. 3 shows ~6k each way.
diode test same pins - 1 and 2 show 1V35 one way, 1V11 the other. 3 shows 1V41 and 1V21.
diode test pin 7 to pins 2/3 - all three show ~~800mV one way, open the other.
same test with either hi or auto ohms - all show open, no reading.
meter is a piece of shit autorange. doesn't know where 0 ohms is.
I dunno what any this means.
Thanks, duck!
Quote from: duck_arse on February 14, 2023, 09:05:52 AM
I dunno what any this means.
We're checking the forward voltage drop on the input transistors and internal resistance through pins that are tied internally through a few series resistors. It seems ~7K to 25K is expected for comp to output resistance - depending on the flavor (maker, year, legitimacy). Mine do not have this resistance and shows open. 800mv seems to be the common ticket and shows that the inputs are likely single NPN transistors with the collectors tied to V+. So we know they're similar and correct in that respect.
QuoteMine do not have this resistance and shows open.
I wouldn't be overly concerned about that at the moment. Interesting that the old Fairchild LM307 datasheet (similar to LM301A) has the resistor but no value.
Something the LM301A does that many opamps can't is feedforward compensation. If you look at the LT datasheet there is a waveform at the bottom right of the (third column) on page 3 (2-299) and the middle circuit on p5 (2-301). If you can get that fast waveform from the inverting circuit there's a good chance you have something at least representative of an LM301A. Common opamps will show waveforms like that at the bottom left of page 3, more than 10 times slower. Unfortunately the breadboard capacitance will mess with it a little.
https://www.analog.com/media/en/technical-documentation/data-sheets/lm101a.pdf
The voltage drop to +V your are measuring is another quirk of the LM301A. The LM301A common mode range goes up to the +V, not many opamps did that.
[Another very old chip was the Lm709, it had some similarities with the LM301A. I haven't looked at the schematic.]
I just remembered I had a board with 2x LM301AN's which was easy to get to.
Parts are National Semiconductor 1979.
*In circuit* I measured pin 6 to 8 at 4.9k ohm, both parts are same batch and measured the same. Not far off Duck's 1977 devices.
The feedback resistor on pin 6 is 470k. I didn't want to pull the parts but I tried shorting some of the resistors on the board and it didn't change the measurement.
Interesting... another one with resistance between comp and output unlike mine (I know, don't think too much on it). That last datasheet you've shared doesn't have that as well, another typo or perhaps minor discrepancies that didn't warrant a part # change? Anyways, you're right that I should plug it into one of the suggest circuit applications and scope out the transition speed.
I'm not entirely convinced that they are 100% real, but I've yet to have a firm reason to doubt besides esthetics.
Fixed my big project up, the one that uses 32 of these chips. First and last time I use spade lugs for speaker connections lol. We're bleep-blooping away. Still need to calibrate which will be the real tell about their authenticity, but it's looking good so far!
For the archive;
(https://i.postimg.cc/x0YLp513/20230215-073220.jpg)
Just did a google search on the # on top.
Someone over at DIYAUDIO had a chip with that same #, but it was an LF356 - a different single opamp.
While I barely have a grasp on mf codes, that's definitely suspicious. I'm thinking rebrand, though typically it's printed and not etched.
Edit: They're not the same as amongst other things, the 356 has p-channel jfets on the inputs and our measurement methods would be of opposite polarity.
Any thoughts?
Quote from: Kevin Mitchell on February 15, 2023, 08:20:04 AM
Interesting... another one with resistance between comp and output unlike mine (I know, don't think too much on it). That last datasheet you've shared doesn't have that as well, another typo or perhaps minor discrepancies that didn't warrant a part # change? Anyways, you're right that I should plug it into one of the suggest circuit applications and scope out the transition speed.
I'm not entirely convinced that they are 100% real, but I've yet to have a firm reason to doubt besides esthetics.
Fixed my big project up, the one that uses 32 of these chips. First and last time I use spade lugs for speaker connections lol. We're bleep-blooping away. Still need to calibrate which will be the real tell about their authenticity, but it's looking good so far!
For the archive;
(https://i.postimg.cc/x0YLp513/20230215-073220.jpg)
Thanks for giving us all such clear images.
From the marking on the chip, it corresponds to that for a chip made by National Semiconductors with the following characteristics:
Date of manufacture - Jan/Feb 2002
Die manufacturing fab - Greenock, UK
Assembly plant - Malaysia
Lead finish - Pb- Free Tin
But looking at the datasheet it seems to show that National used a dot to mark pin 1 on the package as well as the notch at the end as shown in my picture from the datasheet.
So, although the marking code is consistent with National Semiconductors for chips made in 2002, the package looks inconsistent with the datasheet. Also the fine hatching of the laser engraving of the National logo looks a bit suspicious to me.
(https://i.postimg.cc/D4gcCMhg/LM301-A-package.png) (https://postimg.cc/D4gcCMhg)
Thanks for chiming in!
And thank you for the breakdown on the code. That explains how two different ICs can have the same info stamped.
I totally agree with the dot observation. I had noted this in my top post as all nat-semi ICs in my inventory have this dot. This was the first thing I had noticed that made me suspicious. My lack of familiarity makes me unsure. As in, does this imply that ALL genuine nat-semi ICs have this dot? Also, are there other brands of LM301AN or AP chips that don't have the dot impression?
My experience with resurfaced ICs is that it's probably cheaper to print or stamp the label, I've even seen real chips get resurfaced - likely for a bait n' switch effort. In these strange times where practically all components are at slow turnout and high demand, they (counterfeiters) could certainly make the line more blurry by investing in machines to make it more convincing.
FWIW, I couldn't tell that the markings were etched until I took the photo this morning.
Anyways, as I've said I have a big project that takes many of these. If I can't calibrate it properly that will be my sign as there's many places where the design will not function properly with anything but a LM301A
Edit:
Shout out to my other thread, "Old or Fake Chips? (https://www.diystompboxes.com/smfforum/index.php?topic=121410.0)"
Maybe I should change this one to "New or Fake Chips?" :icon_lol:
QuoteInteresting... another one with resistance between comp and output unlike mine (I know, don't think too much on it). That last datasheet you've shared doesn't have that as well, another typo or perhaps minor discrepancies that didn't warrant a part # change?
FYI, the 1996 Motorola (pre Onsemi days) datasheet doesn't have the resistor either.
It's very hard to workout when those schematics represent reality or deviate from reality (or have errors).
QuoteAnyways, you're right that I should plug it into one of the suggest circuit applications and scope out the transition speed
It's something concrete at least that be confirmed by measurement. Another set-up is a gain of 10 (not unit) with a smaller compensation cap for more bandwidth but it's a less convincing test than the feedforward compensation test.
QuoteI'm not entirely convinced that they are 100% real, but I've yet to have a firm reason to doubt besides esthetics.
You can spend a lot of time doing forensic research to work out if the finish, makings, pin dimpling are all consistent. To do that you have to get the year correct. For example,
QuoteDate of manufacture - Jan/Feb 2002
Die manufacturing fab - Greenock, UK
Assembly plant - Malaysia
Lead finish - Pb- Free Tin
For example is it possible that date code is 2012? and are those plant locations consistent with 2002 or 2012? Companies close down plants and move manufacturing locations over time - sometimes this info isn't easy to find although old databooks will give lists of locations, then you have to go over many years of databooks. TI took over NS in 2011 so 2012 would be a fuzzy year for getting info.
QuoteBut looking at the datasheet it seems to show that National used a dot to mark pin 1 on the package as well as the notch at the end as shown in my picture from the datasheet.
So, although the marking code is consistent with National Semiconductors for chips made in 2002, the package looks inconsistent with the datasheet. Also the fine hatching of the laser engraving of the National logo looks a bit suspicious to me.
Without doing forensics the lack of the pin 1 dimple is suspicious to me.
I think SI and TI has the top half-moon dimple but NS has the extra pin 1 dot. Another feature worth checking is if NS had chamfering around the edge of the chip in either 2002 or 2012. Font size and style is another one. Mark Hammer posted a link to an excellent video in the Lounge section a few years ago and it mentions stuff like rubbing acetone (?) on the marking to working out if the chips have been remarked.
QuoteAs in, does this imply that ALL genuine nat-semi ICs have this dot?
You would have to do a lot of digging to confirm the "ALL". When you start gather info on this stuff it's frightening how many variations have occurred over time.
I found this document but the only useful info is when changes occur after the TI takeover, end of 2012. So that leaves a gap for most of 2012.
https://www.mouser.com/pcn/nationalsemi_20126102.pdf
The doc talks about the year marking being YY format before the change, so maybe the 2012 devices are marked 12 and your devices are in fact 2002.
See,
http://www.elenota.pl/datasheet-pdf/147885/National-Semiconductor/MARKING?sid=67b5a593918b05787cce7f253e6ffd1a
This one has a table,
https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/196/marking_5F00_conventions.pdf
Go to the table 2002: 0206 206 21 A
If I'm reading that correctly, presumably a combination of 1A is valid.
That's not to say it is legit, just consistent.
This looks like a 2003 chip with JR31AB marking. Cannot vouch for authenticity even here but it's interesting it has the pin 1 dot. Notice how the part number in closer to the date code and LM is on the same line as the 1458.
(https://i.postimg.cc/bdJNRSgx/NS-LM1458-N-Op-Amp-IC-JR31-AB.png) (https://postimg.cc/bdJNRSgx)
I've come to the conclusion I don't trust any of the 2000 and after markings!
Thanks for digging in, Rob! I really do appreciate the breakdown, references and your opinion - all of which are valuable to myself and future visitors.
I've just purchased an LM301 (AH, the metal can variety) to compare. Should be legit, it's smallbear stock. I'll side-by-side these with the feed-forward comp circuit and share the plots. Will post that some time next week.
Calibrating the beast tonight, been busy playing with my 3d printer after hours.
After sifting through the schematic of said beast, most of these opamps are using the compensation cap. Considering it's current functioning status being untuned & uncalibrated, I think is good sign. But we'll see!
The absolute prof will be in the plots. I'll throw whatever other BJT input opamps I have in the mix as well.
QuoteThanks for digging in, Rob! I really do appreciate the breakdown, references and your opinion - all of which are valuable to myself and future visitors.
No problem. I had a few attempts at doing stuff like this in the past (ie. building a valid marking library) and to be honest it ends up being an overwhelming task.
QuoteI've just purchased an LM301 (AH, the metal can variety) to compare. Should be legit, it's smallbear stock. I'll side-by-side these with the feed-forward comp circuit and share the plots. Will post that some time next week.
Good idea to have some trusty reference devices.
QuoteCalibrating the beast tonight, been busy playing with my 3d printer after hours.
After sifting through the schematic of said beast, most of these opamps are using the compensation cap. Considering it's current functioning status being untuned & uncalibrated, I think is good sign. But we'll see!
The absolute prof will be in the plots. I'll throw whatever other BJT input opamps I have in the mix as well
Good luck with it. Many circuits which used the LM301A and the basic 30pF compensation cap will work fine with an LM741 or even a TL071. All you need to do is pull the compensation cap.
The tricky ones are:
- The compensation cap isn't 30pF. Usually a faster (or slower) circuit, here try a TL071.
- when the circuit uses feedforward compensation to speed up the response. Not a lot of options here
without considering the design details. Pulling the compensation network and using a TL071 can sometimes work.
- input DC levels close to the +V rail. This was unique feature of the LM301A. You would need to
use a modern opamp with common mode to +V rail.
- I knew I forgot one, it just wouldn't come back in my head.
Offset/balance adjustments. The connection is specific to the LM301A you would need
to rejig those for any replacement opamps.
I've been running on negative sleep the last few days and I forgot to post these. I found the 1989 Raytheon schematic has the 5k resistor. A lot of variations on the schematics (I've found this to be the case for LM741's as well, so no surprise).
Raytheon 1989, 5k
(https://i.postimg.cc/rKxnPDSQ/ray1989-LM101-A-schematic.png) (https://postimg.cc/rKxnPDSQ)
RCA early 70's, 7k5,
(https://i.postimg.cc/0zCt9hXS/RCA197x-CA101-LM301-A-schematic.png) (https://postimg.cc/0zCt9hXS)
editted
Quote from: Rob Strand on February 16, 2023, 04:03:19 PM
Good luck with it. Many circuits which used the LM301A and the basic 30pF compensation cap will work fine with an LM741 or even a TL071. All you need to do is pull the compensation cap.
You know Rob, I figured if there was going to be any trouble it would be where the comp cap value differs the most and go figure, there's trouble afoot.
I won't drag you into the nitty-gritty of the project's circuit, but here's what's I'm seeing so far.
I've gone through 5 sections of calibration and decided that I should stop.
I have 3 sawtooth sources that are provided from circuits with no LM301. They look correct.
-Sine wave shaping, set up a nice sine wave at the proper peak-peak voltage (2 LM301, each have a 30pF comp cap)
-Clock (the flashing LED of optimism), set frequency & duty cycle no problem (1 LM301, no comp cap)
-Preamp, This has high gain at minimum actuation of the gain control and clips far too soon making it very difficult to set properly - if you breath on the knob it'll probably move a couple of volts. (1 LM301, 5pF comp cap)
-Envelope Follower - peak-peak voltage is a few volts above what it should be. (3 LM301, 2 30pF comp caps and one 330pF)
-Ring Modulator - output is not at all what is expected, should resemble a modulated mix of previously verified sine and saw-tooth, shows only a quick spike. Also, Audio mode should be half of the peak-peak voltage than DC mode, which it isn't - it's a few volts above DC mode (2 LM301, 1 10pF comp cap and 1 220pF from comp (pin1) to inverted input)
Of course, I'm no pick n' place soldering robot. There could be builder errors though being this far into the game I know that my suspicions are justifiable.
I'm going to put an order in for AS301AN chips today to get ahead of the impending doom - been itching on it too much. I'll take a hit for peace of mind. But I must know what the deal is with these chips.
Edited: forgot about the verified sawtooths and sine calibration
QuoteOf course, I'm no pick n' place soldering robot. There could be builder errors though being this far into the game I know that my suspicions are justifiable.
I'm going to put an order in for AS301AN chips today to get ahead of the impending doom - been itching on it too much. I'll take a hit for peace of mind. But I must know what the deal is with these chips.
I'm on the same page, looks suspicious. The good thing is you are onto it! A bit of circuit weeding and I'm sure you will push it over the line.
No plots yet, have to rethink my 200KHz pulse source to simulate the Feedforward comp inverter for it to be 1:1 to the datasheet circuit.
While I wait for my LM301AH to come in, I had received the AS301AN chips today. I immediately installed them into the big project and right away I'm hearing improvements.
First off, the basic stuff we've been doing.
Diode test shows 797mv - to be expected from what we know.
Resistance test shows ~6.7K
Scoping the feedforward circuit with a 0-5v 200KHz pulse there are some very obvious differences. While not a perfect representation of the pulse wave shows on the output, the difference between the two chips is far from negligible. My immediate observation is that the output signal of the Nat-semi chips show 1.5v p-p and the Alfas show ~3.8v p-p. That's a hell of a difference.
Edit: trying the test again tonight just to have the plots drawn out. Would be nice to actually use the record function on my scope for once lol. Was having trouble level shifting a 5v clock pulse for ±5v. My workaround will be a transistor switch into a cmos buffer.
QuoteNo plots yet, have to rethink my 200KHz pulse source to simulate the Feedforward comp inverter for it to be 1:1 to the datasheet circuit.
When you check the slew-rate it's good idea to stick to the same supply and input voltages as the datasheet, also the same load if one is shown or implied. Otherwise you can get quite different looking waveforms.
QuoteScoping the feedforward circuit with a 0-5v 200KHz pulse there are some very obvious differences. While not a perfect representation of the pulse wave shows on the output, the difference between the two chips is far from negligible. My immediate observation is that the output signal of the Nat-semi chips show 1.5v p-p and the Alfas show ~3.8v p-p. That's a hell of a difference.
The original NatSemi devices are looking at bit fishy aren't they. Maybe confirms you original suspicions.
Well really, this absolutely confirms my suspicions;
±5v 200KHz pulse signal, opamp is powered by ±15v
AS301AN
(https://i.postimg.cc/JGcKn4HN/chart-1.png) (https://postimg.cc/JGcKn4HN)
"LM301AN"
(https://i.postimg.cc/xXBWG1gX/chart.png) (https://postimg.cc/xXBWG1gX)
Ignore noise from input signal (noisy probe)
QuoteWell really, this absolutely confirms my suspicions;
That LM301AN waveform is totally whacked.
What's really weird is there's no negative swing.
You might perceive it as a signal inversion. You might be able to confirm that with a low frequency sine wave overdriving a non-inverting buffer stage.
An other alternative is the device is very slow and cannot recover at the 200kHz rate, you would have to drop it to 20kHz and see what happens.
Another effect is the output stage cannot pull down much current and it's kind of rectifying some capacitive load. Try the probes on x10.
Bottom line though: AN301AN looks respectable and LM301A looks whacked!
I wonder if there's any YT videos of someone busting fake LM301A's with an oscilloscope?