I got two batches (100 pcs. each) of BF862 low noise n-channel jfets, from two different providers (DigiKey Corp. | Electronic Components Distributor | United Kingdom Home Page and Mouser Electronics - Electronic Components Distributor). Both are 100% genuine NXP parts, the only difference being the origin. One batch is marked 2Aw (w stands for China) and the other 2Ap (p stands for Hong Kong).
While they both have the noise floor @100KHz as per the datasheet (~0.8nV/rtHz), the Chinese parts have significantly larger noise corner frenquency (~2KHz) compared to the HK parts (<500Hz). Therefore, the Cninese parts have about twice the noise of the HK parts, @1KHz. Devices in both batches have essentially the same average Idss of about 16mA.
Not a huge deal for audio, but could be an issue for low and very low frequency applications. BTW, the datasheet does not specify anything about the low frequency noise.
While they both have the noise floor @100KHz as per the datasheet (~0.8nV/rtHz), the Chinese parts have significantly larger noise corner frenquency (~2KHz) compared to the HK parts (<500Hz). Therefore, the Cninese parts have about twice the noise of the HK parts, @1KHz. Devices in both batches have essentially the same average Idss of about 16mA.
Not a huge deal for audio, but could be an issue for low and very low frequency applications. BTW, the datasheet does not specify anything about the low frequency noise.
Last edited:
The w's were from Digikey and the p's from Mouser. The Digikey silver bag says "Origin: Taiwan".
The BF816 never claimed to be anything but an amplifier in the MF/HF frequency range. Even in that application the significant capacitance, noise and cutoff frequency parameters only show "Typical" values. This was noticed in the "Discrete Opamp Open Design" thread ( http://www.diyaudio.com/forums/analog-line-level/218373-discrete-opamp-open-design.html ), but there are very few suitable substitutes so we are forced to cross our fingers and hope for good luck when we use the BF862 in audio (or nearly-DC) applications.
Those noise corner frequencies are significantly higher than the Data Sheet values for "audio" JFETS like the 2SK170, LSk389, 2N5911, etc - which seem to be on the order of 100 Hz.
The really unsettling observation is the variation in noise corner. It suggests that it may be desirable to select an "audio-grade" version of the BF862 based on noise performance, but your comments suggest that entire batches of the part may not pass the screening parameter.
It's too soon to conclude that "Hong Kong" parts are preferred to "China" parts. That's the way your two batches measured, but a different production lot from China may have equal or better performance than your Hong Kong batch. Perhaps there's a Forum member with specialized knowledge of JFET fabrication, who can tell us whether noise performance like this is likely to correlate with the original fab, or the packaging location, or some other identifiable characteristic - or is simply typical of batch-to-batch variation regardless of other factors.
Are you willing to share your measured values here? The average Idss you reported is in the center of the Data Sheet range (which is good!) - is there much spread among your parts?
Dale
The noise corner frequency is essentially determined by the process cleanliness. While some variations are to be expected, this is more like a constant of the manufacturing facility.
Idss was 16mA, pretty much a gausian distribution slightly skewed toward higher values, with a dispersion of +2/-3mA.
I'll check it myself, but it is hard to believe that K170 has only 100Hz noise corner frequency. That sounds more like a value for bjt's. And with 4-20nV/rtHz @10KHz, 2N5911 is anything you want but a low noise device.
Idss was 16mA, pretty much a gausian distribution slightly skewed toward higher values, with a dispersion of +2/-3mA.
I'll check it myself, but it is hard to believe that K170 has only 100Hz noise corner frequency. That sounds more like a value for bjt's. And with 4-20nV/rtHz @10KHz, 2N5911 is anything you want but a low noise device.
Last edited:
I've found bad ones too. I don't know of any devices with a guaranteed corner. Did you see Dimitri's data?
7" or 11" reel, in principle same part. Actually first time I heard about a 11" reel, normally they're 7" or 13"....
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.