February 15, 2013 - filed under: Shop
Date: Revised on 1st Nov 2013
Toshiba 2SJ74BL and 2SK170BL match pair for sales: (NEW PRICES from 1st Nov, 2013)
- 2sk170BL/J74BL match pair HK$120/pair (Pair = K170 x 1pc and J74 x 1 pc) (about US$16). Idss 6.4 to 9mA only.
- 2SJ74BL : HK$160/pair (about US$22 – 2 pcs) or HK$320/quad (about US$44 – 4 pcs). Idss 6.4 to 11mA.
- 2SK170BL match quad : HK$100 (about US$13 – 4 pcs). Idss 6.4 to 9mA only.
- Note that Idss of K170 at 10mA and above is basically not available. If found, the price is HKD$100 each.
- Unmatch will be same price as match above. Only BL grade of above jfets are available.
Shipping is HK$35 (US$4.5) to worldwide by registered airmail and paypal fee 5.5% (4.5% for purchase amount >= HK$500). All matched within 0.2mA Idss @ 10Vdc. Maximum match is 4 pcs K170 and 4 pcs J74 with same idss match together in one group. Note that the temperature at measurement is about 28 to 30 degC in Singapore.
There is further delay for the sales of other kits and parts. Tentatively it will be by end of 2013.
Thank you for all the support.
Many diyers have been buying some fake 2SJ74 from China sources or even from US source and thus some of them ask me for the measurement curve of this genuine Toshiba part. Here is the file which you can enlarge to see the curve and estimated Yfs (gain) of the part. The jfet selected is at about 10.8mA Idss (top curve). The setting of Tek 577 is highlighted for easy reference but you can zoom in the pdf to see actual setting of equipment.
Note that the part is laser marked as it is new part production in 2008, probably the last production batch in Japan. For ink mark, those are very old part that I have sold before – 2SJ74V.
You may notice that I have increased the price from 2013 because there are not too many stock left for sales.
November 23, 2012 - filed under: Pass Amplifiers
Those who had brought my F4 kits few years ago can purchase this kit or boards for a quick conversion from F4 into F5T because all the holes and mosfet locations will be same as the F4 pcb. The optional MUR3020W can be obmitted if not use. If F4 buyer has the original IRF9240/240 power mosfet, it can be re-used.
The To-92D transistor are Toshiba 2SC2705Y/2SA1145Y rated at 800mW and thus it is very safe to use at 60 deg C environment for less than 200mW dissipation. It would survive safely under a maximum supply voltage of +/-32V. The Panasonic capacitor is rated at 35V and thus a supply voltage of no more than +/-32V is recommended for this kit.
I build a version using Toshiba Mosfet 2SK1530/2SJ201. The supply voltage is +/-22V and thus I do not use the cascade transistors (2SC2705Y and 2SA1145Y). I also change 3 resistors and details refer to the “F5T-BOM-cost” document. The change on R11/R12 is to increase the thermal stability and R27/R28 is for increasing the supply voltage to jfets to be minimum 10V or above for a supply voltage of 22V. 49.9R to replace 47R do to stock issue.
Sound wise, it is more powerful than the F5 original due to higher current output and more pairs of mosfet is used. Overall the sound is smoother and warmer than F5. Nelson said this amp sound more relax!
Please look at the F5T-BOM-cost documents for the detail kits offer – Price & BOM options.
Details: F5T-BOM-Cost2 (updated shipping cost)
Manual: F5 Turbo Assembly Instruction v3 (NEW)
Power supply ripple comparison CLC vs CRC: It can be seen that an inductor of 0.47mH air-coil has improved the ripple peak to peak level to 20 times smaller. See attached: F5T ripple
After building another set with Cascade transistors, the noise level with grounded input is about half of that without cascade transistors. The reason is that the filtering of the transistors at the base is much better and thus the jfets is supplied with a more stable supply with lower noise level. Here is the measurement comparison.
Notes for F4 up-graders: For the MUR3020W, I have no pre-tap holes for them and thus I use a L-Bar to fix the 3 power mosfet and MUR3020W as shown.
I have built the third set with 32V supply as suggested by Nelson Pass. Here is the measurement of distortion vs Power Output at 8.2 ohm load. With higher supply voltage, the distortion is lower at 0.004% 1W. The non-cliped output is about 50W. For this measurement, the bias is about 1.6A to 1.7A for each channel. The conclusion is that the distortion is lower at 32V and higher at 22V supply. With no surprise, this set sound the best after adding the Solen 1mH coil in the power supply (CLC) filter.