I have a 100W 8 ohm amp using a single MJ21193/4 pair, but the TO-3 cans are rated at 250W. It was a close call whether to parallel them or not, but the amp has been running without problems, with a rail voltage of +/-56 while idling.
The MJL version at 200W really need to be parallelled.
Drivers I used were 2SC2238/A968 but these are long gone. MJE15032/33 are ideal replacements as they have similar gain (70 min) and voltage (250 vs 160 in the C2238).
MJE15028/29 have a lower gain (40 min) and that will push the VAS current requirement up unless additional pre-drivers are used (3EF configuration).
Depending on the load requirement at least 5A peak is needed (100W/8 ohms) and MJL gains will be at least 30, so the base currents will need to be 167mA. Allowing for some margin (lower load, higher frequency) perhaps 250mA is suitable. That would need 6.25mA, so the VAS would have to run well above the usual 6mA or so.
I used 2SC2238/A968 at 20mA in cascode for the VAS stage in the 100 W amp mentioned.
You might use KSC3503/A1381 for these, cascode optional, but will need heatsinking well.
IF you can get MJE15032/3 the VAS current can be relaxed a bit.
If you run a single pair MJL outputs, 75W at 45V rails is a real possibility. Double up for 4 ohms and could then get 150W, but again the drive currents needed will be higher.
MJE15028/9 should be OK with 55V rails but you would need to check that they do not go higher on light loading. You only need 40V for 8 ohms so 50 V on idling should still be possible for 100W. Include reverse bias diodes for inductive spike protection across the output devices. I used UF4004's for 50W amp or UF5406 for higher powers.
The MJL version at 200W really need to be parallelled.
Drivers I used were 2SC2238/A968 but these are long gone. MJE15032/33 are ideal replacements as they have similar gain (70 min) and voltage (250 vs 160 in the C2238).
MJE15028/29 have a lower gain (40 min) and that will push the VAS current requirement up unless additional pre-drivers are used (3EF configuration).
Depending on the load requirement at least 5A peak is needed (100W/8 ohms) and MJL gains will be at least 30, so the base currents will need to be 167mA. Allowing for some margin (lower load, higher frequency) perhaps 250mA is suitable. That would need 6.25mA, so the VAS would have to run well above the usual 6mA or so.
I used 2SC2238/A968 at 20mA in cascode for the VAS stage in the 100 W amp mentioned.
You might use KSC3503/A1381 for these, cascode optional, but will need heatsinking well.
IF you can get MJE15032/3 the VAS current can be relaxed a bit.
If you run a single pair MJL outputs, 75W at 45V rails is a real possibility. Double up for 4 ohms and could then get 150W, but again the drive currents needed will be higher.
MJE15028/9 should be OK with 55V rails but you would need to check that they do not go higher on light loading. You only need 40V for 8 ohms so 50 V on idling should still be possible for 100W. Include reverse bias diodes for inductive spike protection across the output devices. I used UF4004's for 50W amp or UF5406 for higher powers.
Last edited:
Whether or not 15028/9 are suitable for use on a 40/0/40 toroid really comes down to the regulation of the transformer. You might want to see what your options are there first - and maybe in the meantime some 15030/1 will present themselves and solve that dilemma for you.
Of course you could construct a low current 150 volt supply, put a 10k resistor in series with the transistor, and make sure it doesn’t leak (when heated or cooled).
A 400 VA trafo might put out as much as 44/0/44 unloaded. I had some 500 VA 80VCT EI’s that did. The more oversized it is, the better it will be. You could get away with a 400 VA in this application, but more is better. Definitely oversize if you’re planning to drop it to say 35/0/35, regulation may be good enough to get your power if oversized.
Of course you could construct a low current 150 volt supply, put a 10k resistor in series with the transistor, and make sure it doesn’t leak (when heated or cooled).
A 400 VA trafo might put out as much as 44/0/44 unloaded. I had some 500 VA 80VCT EI’s that did. The more oversized it is, the better it will be. You could get away with a 400 VA in this application, but more is better. Definitely oversize if you’re planning to drop it to say 35/0/35, regulation may be good enough to get your power if oversized.
MJE 15034/35G
I'm overwhelmed by the response i'm getting here
Actually i already invested heavily in this hobby, so much so that for the rest of my life i no longer have to worry about op-amp & transistors! That's why i was reluctant to invest more but after reading the post by John Ellis i'm seriously considering mje15034/35 This is probably my last big purchase(at least for me) of semiconductors(5p for each sex, total 10unit). Btw here in india RS comp is the only reliable online parts supplier with free shipping.
I'm overwhelmed by the response i'm getting here
Actually i already invested heavily in this hobby, so much so that for the rest of my life i no longer have to worry about op-amp & transistors! That's why i was reluctant to invest more but after reading the post by John Ellis i'm seriously considering mje15034/35
This is probably my last big purchase(at least for me) of semiconductors(5p for each sex, total 10unit). Btw here in india RS comp is the only reliable online parts supplier with free shipping.Interesting!
Hi, i was digging a bit...after reading some thread i want to know something. Do you recommend mje15034/35 as a driver tr for an amp with +/-30v rails? What if i use them for VAS, i know they're overkill but just saying. According to datasheet they have excellent gain at low collector current, aren't they better than Mje32/33?? I'm assuming they're simply different animal. How to read their SOA Graph? I see DC SOA knee at 50volt-1amp &
at 100volt-200ma, for mje15034/35. Very robust, huh?!
Hi, i was digging a bit...after reading some thread i want to know something. Do you recommend mje15034/35 as a driver tr for an amp with +/-30v rails? What if i use them for VAS, i know they're overkill but just saying. According to datasheet they have excellent gain at low collector current, aren't they better than Mje32/33?? I'm assuming they're simply different animal. How to read their SOA Graph? I see DC SOA knee at 50volt-1amp &
at 100volt-200ma, for mje15034/35. Very robust, huh?!
Attachments
Last edited:
I'd probably say not for +/-30V rails.
The reason is that while many parameters are quite good, the high voltage capability means a lightly doped collector has been used. You can see in the data that the gain with 5V collector is a fair way below the gain at 20V. That says the quasi-saturation is probably quite severe. I've not looked at that yet for the 15034/5, but it suggests that these devices are intended to drive high gain output transistors, perhaps of the 3281A type etc. and for higher supply voltages.
The older MJE15028/29 have a better performance at lower voltages.
The popular BD130/BD140 can also be used for amplifiers with 30 to 40V rails.
But this is only part of the equation. Another question is what gain you need at what current. That depends once again on the output transistors and output power you are aiming at.
BD139's would probably drive MJL21193/4 but may struggle if you intend to use 4 ohm speakers needing higher currents.The 28/29 have gains of 40 maintained to 2A at 2V collector. The downside is that you will need a suitable VAS current. A BD139/140 VAS could work at 20mA.
You have to look at all the specifications to really check the suitability of a given set of transistors for a particular requirement.
The reason is that while many parameters are quite good, the high voltage capability means a lightly doped collector has been used. You can see in the data that the gain with 5V collector is a fair way below the gain at 20V. That says the quasi-saturation is probably quite severe. I've not looked at that yet for the 15034/5, but it suggests that these devices are intended to drive high gain output transistors, perhaps of the 3281A type etc. and for higher supply voltages.
The older MJE15028/29 have a better performance at lower voltages.
The popular BD130/BD140 can also be used for amplifiers with 30 to 40V rails.
But this is only part of the equation. Another question is what gain you need at what current. That depends once again on the output transistors and output power you are aiming at.
BD139's would probably drive MJL21193/4 but may struggle if you intend to use 4 ohm speakers needing higher currents.The 28/29 have gains of 40 maintained to 2A at 2V collector. The downside is that you will need a suitable VAS current. A BD139/140 VAS could work at 20mA.
You have to look at all the specifications to really check the suitability of a given set of transistors for a particular requirement.
OK, but that won't be possible with 30V rails unless you use a bridge configuration.
You need 40V plus a tolerance for a minimum transistor voltage and power supply rails, so looking at +/-45 V minimum, probably +/-50V. And a peak output current capability of 9A.
And, the worst case power dissipation could be in the region of 75W for each transistor (when on) so a mean dissipation of 37.5W each assuming +/- output, not DC. Total power 150W dissipation per amplifier. BUt that may be reduced in practice because musing is not continuous and the power supply droop will cause the voltage to reduce at higher currents.
With a pair of MJL21193/4 you will need a base drive of 300mA as a rough guess.
The MJE15034 will be fine up to 500mA (beyond this the gain falls but so does the fT). So, they are usable, and though the 32/33 have a lower gain, it is maintained at a higher current (1A).
I think that the 32/33 would be my preference for the output devices you have.
You need 40V plus a tolerance for a minimum transistor voltage and power supply rails, so looking at +/-45 V minimum, probably +/-50V. And a peak output current capability of 9A.
And, the worst case power dissipation could be in the region of 75W for each transistor (when on) so a mean dissipation of 37.5W each assuming +/- output, not DC. Total power 150W dissipation per amplifier. BUt that may be reduced in practice because musing is not continuous and the power supply droop will cause the voltage to reduce at higher currents.
With a pair of MJL21193/4 you will need a base drive of 300mA as a rough guess.
The MJE15034 will be fine up to 500mA (beyond this the gain falls but so does the fT). So, they are usable, and though the 32/33 have a lower gain, it is maintained at a higher current (1A).
I think that the 32/33 would be my preference for the output devices you have.
The lead time for mje15032/33 is unpredictable, i checked several merchant website including mouser & digikey. Due to current situation worldwide i placed an order for mje15034/35 at rs components. I hope for the bestThanks everyone.
For VAS use very low Cob like 2SC3503/2SA1381. If you force to use high Cob like mje15032/33, than cascode them. With low Cob VAS, you can achieve very low distortion at high frequency. For good performance in high frequency, always use low Cob and high fT transistor every where. You can not achieve very high slew rate with MJL21193/94.
But not all people want to high quality at high frequency. Sometime, they can not know the (sound ) different.
Hi Bimo, it's nice to have you here. I've no concern about slew rate(usually not a problem) & 4mhz(mjl93/94) ft is still pretty good if not best. Yeah, i've my eye on those, excellent transistors for vas, hard to get but not rare (2sa1381/c3503). Good VAS transistor is difficult to obtain today. Btw what do you think about mje15034/35 driving two pair of mjl93/94? Regards.
Last edited:
rs components have KSC3503/A1381.
But not matched gains. D group for NPN E group for PNP.
Seems difficult to get matched components. Distortion is lower with matched components even if matching less important than it was (if using these as predrivers for a 3EF stage).
But as you probably want these for the VAS, should be no problem.
But not matched gains. D group for NPN E group for PNP.
Seems difficult to get matched components. Distortion is lower with matched components even if matching less important than it was (if using these as predrivers for a 3EF stage).
But as you probably want these for the VAS, should be no problem.
Last edited:
Try Farnell. Packs of 5. Still only D and E grades though.
re ZTX -I've not tried these but they look suitable. 5pF Cjc max at 20V comparable to 3503 3.5pF typ. at 30V; ft 50MHz min compared to 150MHz typ. (both at 10mA).
I would advise mounting on a heatsink, a little more difficult with the Zetex/Diodes "thin TO-92" case than a TO-126, but worth doing so even if the VAS current is 10mA. A relatively small heatsink would be sufficient unless you run the device at 20mA.
Datasheet does not seem to allow for heatsink power increase, but most transistors will run more power when heatsink attached.
re ZTX -I've not tried these but they look suitable. 5pF Cjc max at 20V comparable to 3503 3.5pF typ. at 30V; ft 50MHz min compared to 150MHz typ. (both at 10mA).
I would advise mounting on a heatsink, a little more difficult with the Zetex/Diodes "thin TO-92" case than a TO-126, but worth doing so even if the VAS current is 10mA. A relatively small heatsink would be sufficient unless you run the device at 20mA.
Datasheet does not seem to allow for heatsink power increase, but most transistors will run more power when heatsink attached.
Last edited:
Yes, available there but i have to buy 60unit each!! This is the only thing i don't like about RS. They're business to business supplier
This is what we were ALL up against back in the 70’s and 80’s when I got started. Regular “mail order” places didn’t stock anything over 80 volts, and Digikey was just another one of them. The big boys that carried the whole Motorola line wouldn’t even do business with you unless you had a tax ID on file, and minimum orders were too much for a kid on a $200 a YEAR budget. I remember seeing the Leach double barreled amp and drooling - but where the |-*(% was I supposed to get MJ15003’s??????? It is/was VERY frustrating. We are spoiled in the US now when it comes to electronics - even now with the shortages and obsolescence in two years - we really need to count up our blessings. There isn’t the conspiracy to prevent you from buying the stuff like there WAS 40 years ago. Thank GOD MCM came along when it did or I might have just thrown myself off a bridge with a dead Phase Linear strapped to my leg.
If you can afford 60 of each at all, just buy them. I did lifetime buys (200 each) years ago before it got this bad. Including those and several of my favorite drivers which are now NLA. I have the proper solution, it’s called the C4793.
If you can’t get the VAS transistor you *want*, use a compound VAS so that the power device can be a bigger high capacitance one, or use an EF3 output stage and go down to TO-92s (or SOT 223) for the VAS. Again, we’re soiled. We didn’t have anything like the C3503 back then. Had to choose between MJE340, 2N3440, or the video output transistor of a dead CRT, depending on what you could salvage off a board and good luck with a PNP without a tax ID and a lot of money.
mje15034/35 have the best SOA for driving output transistor. Those are able to drive more than two pair output transistor. But those are not the best in sound quality. I will choose driver from Toshiba.
We have our own Element14 here but i need TAX ID, as already mentioned by wg_ski. I think i should borrow someone's
Hmmm...Heat is the best enemy of semiconductors!
But then there was no problem of counterfeit components like today. Anyway don't make me nostalgic or i'll cry.
From RS? Never! But i understand your point. I'll try at local stores! They're real treasure with unbelievable offering! I'll try for 50 each
Very much available from Toshiba & KEC but big to-220 insulated package. Good spec though.
Last edited: