Someone had to invest and sacrifice some gyrator boards to test various lower FETs (either depletion or enhancement devices as well as TO-92 or SMD options). That was me.
Why? Because I want to push this circuit further and find the best options as well as provide to the builders out there some other device alternatives when they can’t solder SMD components.
So let me present you the abused test mule and the various boards under the mercy of my tests:
So What FETs where analysed? Here is the list:
- BSSH111BK: N-Channel, Enhancement, 55V, 335mA, SMD device (needs to be mounted in 45 degrees or with an adaptor)
- 2N7002: N-Channel, Enhancement, 60V, 115mA, SMD device (needs to be mounted in 45 degrees or with an adaptor)
- 2N7000: N-Channel, Enhancement, 60V, 200mA, TO-92 – fits perfectly on the PCB board.
- BS107A: N-Channel, Enhancement, 250V, 250mA, TO-92 – fits perfectly on the PCB board.
- BSS123: N-Channel, Depletion, 100V, 150mA, SMD device (needs to be mounted in 45 degrees or with an adaptor)
The test circuit
The test circuit is very simple. The classic DHT preamp stage with a 4P1L in filament bias. The HT supply is regulated 300V with a decoupling 100μF ASC Oil cap with a 100nF film cap in parallel. To separate FR impact from the load, the load is the test set in parallel with the 510kΩ R8. The following components differ between boards as am planning to use them in different circuit from the 4P1L but shouldn’t affect the results of the below tests:
- R4: Either 300kΩ or 390kΩ (preferred for 4P1L)
- C1:100nF and R6 10MΩ
- R7: 300Ω or 470Ω
- Rf: 15Ω
All tests were conducted with If=650mA and the 4P1L biased to Ia=30mA
Below are the tests conducted.
So in summary, here’s my doggy bag take:
- BSSH111BK will give you the best performance. However, there’s isn’t any free lunch. You will have to solder the SMD device on 45 degrees. Which is tricky, see my previous post here.
- If you want to avoid the hassle of SMD soldering and need +30mA current in your stage, then go for either 2N7000 or BS107A. Both TO-92 devices fit on the board and perform really well up to 1MHz.
Ha! Ha! Ale, that’s great, what a big surprise! I’m sure many DIYers will read this post with interest. thanks!
I guess M3 is still DN2540 ?
Not always, see the details on each measurement
Hi Ale,
Not directly related to the question of different FET options, but have you ever tested options to allow them to work with some more “headroom” (I don’t know if that is the correct word here) to decrease capacitances?
One option I have been playing around (but not performed serious measurement) is a A23 12V battery in series with the gate of the upper device. http://www.ebay.co.uk/itm/Small-Single-A23-Battery-Holder-1-5V-DC-2-x-PCB-Pins-/291875339433
Wish you a nice weekend!
Erik
Hi Erik, not other than using IXTP08N100D which helps increasing the VDS of the JFET. Sounds really good as well as measures better. Many of the plots here are with this device
The 12V A23 battery is a good idea! I may hack the board to try it when I get the time.
Cheers, Ale
Hi Erik
I did the test in the end. I compared a 4P1L loaded with the gyrator using a BSH111BK and DN2540. Biased at 30mA and loaded with 100k resistor. The gyrator performs well up to 900kHz on this setup. Bear in mind the BSH111BK is best lower device I’ve found.
When I replaced the bias of top MOSFET with 12V battery the improvement is 100kHz. FR goes flat up to 1Mhz with the same load. Impressive, but long leads of the battery holder made it difficult to stabilize. Anyhow, got it working and is great. Not sure I’d go all the trouble for 100kHz of additional bandwidth considering the response is already really good.
Perhaps with other lower devices with not such great parameters as the BSH111BK, I’d consider hacking the board to include the 12V battery.
Great idea either way!
Cheers
Ale
Hi Ale – how do these new devices compare in sound quality with the BF862 which can be used at 25mA? Are any of them an improvement on BF862 at that current?
Andy
Hi Andy,
I’ve been travelling over the last 2 weeks and haven’t had the chance to listen to them yet….
cheers
ale