I got one of those cheap op amp testers from AE and decided to test all the op amps in my stock. And together with the tester, I ordered a few more op amps from AE that the prices looked too good to be true. I figured I could test them on the spot. Who knows, they could test fine. After all I got a few working LM3600 and LM13700 from AE that worked fine in the oscillator jig.
The tester puts the op amps in a single 12V supply voltage follower configuration, two anti-parallel LEDs at the output, and injects a (selectable 10Hz and 200kHz) quare wave at the input. If the LEDs do not light at all, light only at the lower frequency, are very dim, or only one of the LEDs lights up, then you get a rough indication that there might be something wrong and that you should probably run a few more tests.
As a reference I had LM741, TL072, TL074, LM324, LM358, NE5532, LF353 and MC34083 that I got from reputable physical shops. The LM741, LM324, LM358, LF353 and MC30483 were vintage NOS parts, while TL072, TL074, and NE5532 were newer parts the shop got from another reputable supplier (tme.eu).
Results:
-
All the TL074 and TL072, and LF353 from ebay and AE (supposedly Texas Instruments) were tested as fake. The could not light the LEDs at 200kHz. Genuine TL074, TL072 and LF353 did not have such problems. When powered by a single 12V supply and the inputs were grounded, the output stayed at 0V, as a LM258/LM324 should behave. A genuine TL074 gave 11.5V at the same configuration indicating the typical phase reversal problem in those ICs.
-
LM833 (marked HLF, a Chinese brand) from AE also tested as fake. At 200kHz only one of the LEDs lit up, indicating that the op amp could swing down to ground but was not fast enough in the other direction, a tell tale that this was a rebadged LM358. For reference, the similar NE5532 worked fine. Additionally, when powered by a single 12V supply and the inputs were grounded, quiescent current was measured at 1.2mA which is too low for a genuine LM833 (should be 4-8mA).
-
OP07 (supposedly Texas Instruments) from AE were fake beyond any doubt. Genuine OP07s are likely too slow to light the LEDs at 200kHz anyway, so I tested resistance between pins 1 and 8 (the offset trim) where according to the datasheet there should be some resistance. Instead, I got an open circuit, just like I got when measured between 1 and 7 (supply) and 8 and 7 where there should be resistance as well. I also measured the quiescent current at 0.3mA which is too low for an OP07 (I couldn’t find figure for 12V, but 4mA for +/-15V). I also noticed a huge offset at the output in the tester, where a genuine OP07 should have less than 0.1mV
-
OPA2134 from AE (supposedly Burr Brown) were also fake as well. First, they failed the op amp tester test at 200kHz, which a genuine OP2134 with its bandwidth should have passed. Second, measuing resistance between inputs and supply pins showed a resistance between 1-2M and a diode drop, indicating that the op amps were bipolar. If they they had JFET inputs like genuine OPA2134 they should have shown an open circuit.
-
LM741 (supposendly National Semiconductor) from AE were also fake. Genuine LM741s should be too slow for the 200kHz test, so I measured resistance between pins 1 and 4 and and 5 and 4 (offset null pins to - supply). The ICs showed an open circuit when the LM741 datasheet shows 1k resistors in between those pins. As a reference, the genuine NOS LM741 indicated 1.2k. Quiescent current was also measured around 0.3mA which is too low for an LM741, although the genuine LM741 also measured low (which could be due to being LM741J military variant in ceramic package).
I also got a few RC4558 from ebay and LM258 (LM358 with wider temperature range) from AE, which are likely fine although I need to do a few more tests.
Now I got my money back on those last few orders but spent a lot of time testing, etc. I did not mind the time spend as I treated it as a learning experience, but I will likely not do it again.
At this point, one might ask if it’s really worth buying any ICs from ebay and AE at all. The thing is that for most people outside the US buying from Mouser, Digikey or Jameco is not an option due to the huge cost of postage, added VAT and import fees. Buying from Tayda or shops from the UK (thanks to Brexit) also incurs VAT, import fees and a long hold up at the customs. Even shops within Europe will charge you more than 10 EUR for a small package with a couple of ICs. Considering these, ebay and AE sound attractive especially since you can combine your order with other components that are usually fine and sold at very competitive prices (IC sockets, headers, 1% resistors, ceramic capacitors, 4148 diodes, LEDs, etc). Unless you like the risk and enjoy spending time testing, it’s not worth it.