Renesas (Hitachi) 2SJ162 and 2SK1058, as well as Exicon ECW20N20 & ECW20P20 and the likes, have integrated gate zeners.
Earlier Hitachi 2SJ50 and 2SK135 in TO3 packages do not have it, nor do the Semelab Alfet ALF08N16V, ALF08N20V and ALF08P16V, ALF08P20V.
Renesas
https://www.renesas.com/us/en/produ...power-mosfets/2sj162-silicin-p-channel-mosfet
https://web.archive.org/web/2024000...n/document/dst/2sj160-2sj161-2sj162-datasheet
https://www.renesas.com/us/en/produ...evices/power-mosfets/2sk1058-n-channel-mosfet
https://web.archive.org/web/2024000...ocument/dst/2sk1056-2sk1057-2sk1058-datasheet
Semelab Alfet
https://www.farnell.com/datasheets/866120.pdf
https://web.archive.org/web/20240225000131/https://www.farnell.com/datasheets/866120.pdf
https://www.farnell.com/datasheets/866122.pdf
https://web.archive.org/web/20240225000122/https://www.farnell.com/datasheets/866122.pdf
Yes
Why would I short the outputs...You have to remember when this was put together it was before I even had a PC at home, I had not got access to all the info we have today and I certainly never dreamt it would ever be seen by anyone other than me.
Good point. I'd forgotten that but checking old data sheets show they do. Whether I had appreciated that all those years ago I can't remember.
There is some confusion about exicon's protection - nothing is mentioned in the datasheet, except the gate-source breakdown voltage of +/-14V. However this matches the Hitachi 2SJ50/2SK135 range which do claim in the datasheet to have gate protection diodes (no further info provided). Someone has reported Profusion has claimed the Exicons have 1W 14V zeners. Presumably TVS diodes with the +/-14V specification?
So I did a more thorough test of an 2023 ECW20P20 p-channel device. Biasing the gate positive via a 560R resistor seemed to limit at +17V with the supply going upto 20V (didn't want to risk higher). Biasing the gate negative had a hysteresis behaviour, at about -16V supply the gate dropped to -10V or so, but recovered if the supply was brought back to -14V.
If it can be established with certainty which devices need external zener diodes, to avoid possible failures, how hard would it be to catch these at the stage of ordering a pcb for a build, by providing space to accomdate these, albeit with suitable legend, on a modified version of the present pcb and with notes in whatever accompanying paperwork ?
oops, I was too quick typing here from the top of my head, don't know where I got that from, but upon rechecking the datasheet it's true 2SJ50/2SK135 do have gate protection diodes as well.
While on alldatasheet.com I came by 2SK135 datasheet (attached below) from ISC (Inchange Semiconductor Company Limited), and checking their web page
http://iscsemi.cn/en/supply/4.html
they are listing 2SJ50/2SK135, 2SJ162/2SK1058, and the lower voltage variants of aforementioned ones.
Does anyone know anything about ISC's supply?
I have four pairs of Hitachi 2SJ50 /2SK135 that I bought over 20 years ago. I built the JLH Wireless World Mosfet 80-100watt Mosfet amp using a single set of these. I did have a failure of one device possibly as we used to get a lot of condensation in our home and failure coincided with switch on. I don't remember if this design had gate protection zeners or not and I have not kept copies of the articles in June, July, and August 1982. Later Mosfet designs use the lower rated 2SJ49/2SK134 as did his conversion of the CFP design that used inverted Darlingtons in combination with transistors -this update having been made in Hi-Fi News and Record Review in December1980.
If there are intending builders with new old stock Hitachi examples, the gate protection zener diode value is 4.7 volts
@Mooly forgive me if this has been asked already, but in place of 2N5551/2N5401 can we use MPSW06/56 or even ZTX751/651? They have the same pinout but the higher output capacitance might be a problem, I have quite a few quite well matched that need a home, and to save a few pennies
Someone may need to sim this.
MPSW - https://pdf1.alldatasheet.com/datasheet-pdf/view/12558/ONSEMI/MPSW06.html
ZTX - https://mm.digikey.com/Volume0/opasdata/d220001/medias/docus/1047/ZTX750.pdf
Someone may need to sim this.
MPSW - https://pdf1.alldatasheet.com/datasheet-pdf/view/12558/ONSEMI/MPSW06.html
ZTX - https://mm.digikey.com/Volume0/opasdata/d220001/medias/docus/1047/ZTX750.pdf
You have had your answer which is question of a little beer money spent on 2N5551 and 2N5401.@Mooly forgive me if this has been asked already, but in place of 2N5551/2N5401 can we use MPSW06/56 or even ZTX751/651? They have the same pinout but the higher output capacitance might be a problem, I have quite a few quite well matched that need a home, and to save a few pennies
Someone may need to sim this.
MPSW - https://pdf1.alldatasheet.com/datasheet-pdf/view/12558/ONSEMI/MPSW06.html
ZTX - https://mm.digikey.com/Volume0/opasdata/d220001/medias/docus/1047/ZTX750.pdft
If you are concerned with the costs of the build the hardware will take up a large percentage of the budget - items such as the case and the transformer. You could save money if you are prepared to compromise on power output of 70 watts.
A transformer with lower secondary voltage output should be less expensive than the unit Mooly has used. You should be able to find one with decent voltage regulation so the rectified output voltage holds under low impedance loads. That is probably an important compensation when power output is reduced so the transformer can boost the magnetic field for more current output on demand.
Last edited:
Just backing up the statement Semelab Alfet don't have gate protection diodes (concerns specifically the list of items below), from Semelab's app note:
ALF08N16V/ALF08P16V N/P 160V 8A TO-247
ALF08N16K/ALF08P16K N/P 160V 8A TO-3
ALF08N20V/ALF08P20V N/P 200V 8A TO-247
ALF08N20K/ALF08P20K N/P 200V 8A TO-3
ALF16N16W/ALF16P16W N/P 160V 16A TO-264
ALF16N16K/ALF16P16K N/P 160V 16A TO-3
ALF16N20W/ALF16P20W N/P 200V 16A TO-264
ALF16N20K/ALF16P20K N/P 200V 16A TO-3
ALF08NP16V5 N&P 160V 8A TO-247-5
ALF08NP20V5 N&P 200V 8A TO-247-5
Page 11:
Gate Protection & Safe Operating Area Limiting
Semelab’s previous family of lateral mosfets
included internal gate-source zener diodes to
protect the junction against overvoltage
situations. On the new family of Alfets these
zeners have deliberately been omitted for a
number of reasons to enhance the performance
and as a result some protection does need to be
added externally to the devices (in addition care
should be taken in handling the devices because
of the high impedance sensitive gate and static
handling precautions should always be
observed).
Semelab app note "Lateral MOSFET Design Recommendations for Audio Amplifiers" PDF
Last edited:
@mjona - I understand my friend, in all cases the designer knows best, I wrote saving a few pennies as a joke really, but - I have a serious amount of MPSWs bought to match perfectly for my Dynalo and Dynahi headphone amp builds and kits. I had no choice but to buy 500 pcs each from Rochester a while ago. They are obsolete. If you need any, get in touch pal
In any event they can be used - I will always ask, I could indeed use them here at some risk as dear Mooley pointed out - to be clear I will be building a 50w version with 35V rails, they could be used at a pinch, they are very decent, reliable 1W transistors.
So - yet again I will buy more transistors. I want to match as best as possible the 5551/5401 for this, so, I will have left overs. My stash, like yours most likely, is full of parts bought in excess, however I can pass them on to fellow builders here no doubt. It is very likely unnecessary to match anything for this amp, but it really makes me feel better to build to the best of my abilities and I enjoy the diligent process (the servo mitigates any imbalance in the grouping I guess? )
My transformer will be one of the last made by Terry at Canterbury Windings here in the UK, an absolute legend in the world of Toroidal Audio transformers. He will be retiring very soon and will only take orders that he can fulfil with his remaining stock of cores. I might just order another one for a future project.
It strikes me that Mooly has base and gate stoppers respectively for the output transistor drivers and FETs so there will be some two element combined curve in frequency response with the FETs no doubt well researched and proven to work. My advice is to stick to the script and don't branch out on another course.@mjona - I understand my friend, in all cases the designer knows best, I wrote saving a few pennies as a joke really, but - I have a serious amount of MPSWs bought to match perfectly for my Dynalo and Dynahi headphone amp builds and kits. I had no choice but to buy 500 pcs each from Rochester a while ago. They are obsolete. If you need any, get in touch pal
In any event they can be used - I will always ask, I could indeed use them here at some risk as dear Mooley pointed out - to be clear I will be building a 50w version with 35V rails, they could be used at a pinch, they are very decent, reliable 1W transistors.
So - yet again I will buy more transistors. I want to match as best as possible the 5551/5401 for this, so, I will have left overs. My stash, like yours most likely, is full of parts bought in excess, however I can pass them on to fellow builders here no doubt. It is very likely unnecessary to match anything for this amp, but it really makes me feel better to build to the best of my abilities and I enjoy the diligent process (the servo mitigates any imbalance in the grouping I guess? )
My transformer will be one of the last made by Terry at Canterbury Windings here in the UK, an absolute legend in the world of Toroidal Audio transformers. He will be retiring very soon and will only take orders that he can fulfil with his remaining stock of cores. I might just order another one for a future project.
Last edited:
The TL071 might struggle at low supply voltages unless you drop the value of the series supply feed resistor.
The resistor is adjusted accordingly (together with some other parts according to what I had available in the parts bin) and it does work fine at that low voltage already. I measured -8V at the IC, with the outputs around -2.1V each, which is correct for the lowered feedback resistor of 6k8 instead of 22k that I used. The IC is a TL052, which is the better-spec'd and lower-offset dual version of the TL071.
That's looking good
obviously feel free to experiment.I think this is a very important point.
Whatever we might want to alter or do differently today and whatever criticisms there are of the original design, the fact is the original does work and makes great music.
Running the original amp with low global feedback was a very conscious choice... it just sounds so good,
There's so much gain in front of the amp already, that I would probably get away with a unity gain buffer
! Pretty sure though that it will still sound good, due to the singleton input stage and lateral output stage. And in case it doesn't, I know which parts to change first
That's good enough for me Mooly ol' bean
Correct transistors will be used mate.@mjona thanks for the advice and knowledge. I am a total novice in the electronic engineering arts. I'm a mechanical engineer originally, self taught mostly in this game.