Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

June 13, 2024, 06:34:57 PM

Login with username, password and session length

Recent Posts


FET Self Biasing Method

Started by Littlewyan, March 23, 2021, 08:39:54 AM

Previous topic - Next topic


I've been building myself a simple amplifier (Of which I will post more details about in a separate thread) which uses a TDA2050 chip of the power amp and a few JFETs and NPN transistors in the pre amp. I've breadboarded it and it works fine, but I don't believe the bias of my FETs is correct. I've googled and googled, read various websites and have gotten nowhere, everyone seems to have different methods of finding the correct bias. Can someone please help? I'm using J202 JFETs, datasheet is here:

My supply voltage is 35VDC. I'm currently using 33K Drain and 1K Source. My VDS is 20V which is roughly ok, but the JFET is only drawing 0.46mA. The datasheet states it should be between 0.9 and 4.5mA.


I realise now that IDSS is actually the maximum current the JFET will draw.


JFETs are so inexact in their specs, there's really no way to simply slap on a few pre-determined components and have it come out working exactly as it should.  MOSFETs are a little better, but there's still some voodoo in there.  I'm afraid that as long as you want to design with FETs, you're going to have to live with the reality of trim pots as biasing devices, and manually adjusting each one by ear.  Some think the effort is worth it, others not so much.  I remember someone came up with the brilliant idea of using a diode instead of a resistor at Source.  It worked, but not always optimally.  Sorry I don't have a better answer.


I think you're right. Unfortunately in my case I designed my preamp around the FETs have 18K Drain Resistors, however I had to bump up to 33K to get the bias right in the actual circuit which increased the AC gain too much. I tried adjusting by Source Resistor but this didn't help much either. In the end I had to lower the supply voltage by 10V so I could stick with the 18K Drain Resistor. Luckily the tone wasn't affected by this (It's a completely clean preamp anyway).


Hi. Edvard is correct about FETs having a wide range of specs... that is why they are not favoured in mass-produced equipment. However it is possible to allow / design for this.
There are simple circuits that will allow you to measure Idss and Vgs cut-off of FETs using a few components and a multi-meter. Once measured a precise design can then be done... and component values can be calculated. The Drain resistance value you have mentioned seems way high.
There are multiple articles on the net re biasing calculations. There is one in particular by Dimitri Danyuk to set up bias so that an FET works like a triode vacuum tube... Google it. I have built many circuits using this principle. They work very well.
For your (and others) info I have attached something that I wrote for a magazine a couple of years back. It is copyright... but I would welcome your feedback and comments.


Yep, as the guys said, fets can be all over the place. It's very common to buy largish batches and select/match parts with similar characteristics. Many parts also come in different grades, meaning that you might be able to buy parts closer to what you need in your design.


I buy J113's a hundred at a time from Mouser ($0.13 a piece) and use a FET tester I built (the simple one from runoffgroove) then spend a rainy Saturday sorting them. It takes less time than you'd think (and I live in South Florida so we have many rainy days...). The J113 works great for Danyuk's circuit.


I have been fortunate to have got a supply of original 2SK and 2SJ FETs... like the 2SK30, 2SK369, 2SK170... etc... which has given me a wider choice of options to vary and optimise designs... still using (primarily) the Danuk process.
I have even developed a couple of XL spreadsheets which integrate the key formulas... and then expand the calculations to other areas e.g. Voltages, Currents, Gain, etc...  If you would like the xls file send me an email address(s) and I will send it that.