I’ve been down a little rabbit hole. Many good through hole parts have been EOL’d in the recent years and if you weren’t lucky enough to snag some or too late to the party you missed out.
Luckily sometimes you can get by with surface mount equivalents.
SK209 used for SK117
2SC3324 for a low noise high gain similar to 2SC2240
2SC2713 and 2SA1163 being similar to C2240 and A970.
There are some others I have as well, but can’t think of them off the top of my head.
Are there any good large surface mount transistors that would replace or be somewhat near as good for drivers?
Something like 2SC4793 and 2SA1837 or 2SC5171 and 2SA1930 (though I read these last two have poor SOA, so maybe similar specs but better SOA)
Or 2SC4883 and 2SA1859. Any devices that could be soldered to adaptor board and use in through hole designs?
Dan
Luckily sometimes you can get by with surface mount equivalents.
SK209 used for SK117
2SC3324 for a low noise high gain similar to 2SC2240
2SC2713 and 2SA1163 being similar to C2240 and A970.
There are some others I have as well, but can’t think of them off the top of my head.
Are there any good large surface mount transistors that would replace or be somewhat near as good for drivers?
Something like 2SC4793 and 2SA1837 or 2SC5171 and 2SA1930 (though I read these last two have poor SOA, so maybe similar specs but better SOA)
Or 2SC4883 and 2SA1859. Any devices that could be soldered to adaptor board and use in through hole designs?
Dan
Well, they killed most of good SMD drivers as well. If you're still interested, Toshibas TTA006B\TTC011B are direct replacements for 2SA1837\2SC4793. The problem is they come in TO-126 packages with emitter and base pinout changed. There are also good drivers from Sanken, 2SA1668\2SC4382. Btw, I think there are some SMD FZT series in SOT-223-4 package like FZT758\FZT658. High voltage, small Cob etc.
Toshiba TTA006B\TTC011B
Sanken 2SA1668\2SC4382
Diodes Incorporated FZT758\FZT658
FZT Series
Toshiba TTA006B\TTC011B
Sanken 2SA1668\2SC4382
Diodes Incorporated FZT758\FZT658
FZT Series
Last edited:
There's 2SA1552 / 2SC4027 in a dpak package, the design issue will be power dissipation, they'll be okay in a VAS stage with say 50v rails at 5ma where they'll be dissipating around 250-300mW but in a driver stage they will be dissipating somewhere in the order of a 1Watt or so, this will be difficult to manage in a smd package. You'll be needing large copper pours and smd heatsinks.
For driver stages, through hole packages are still preferable, there not many transistors left that are still being manufactured.
Sill need to be mindful of SOAR parameters for through hole mounted transistors as well.
For driver stages, through hole packages are still preferable, there not many transistors left that are still being manufactured.
Sill need to be mindful of SOAR parameters for through hole mounted transistors as well.
When Toshiba finally decides to quit the A006/C011 the world will probably stop spinning and at least a few audiophiles will fly off into space. At least with them you have the option of using one or two in parallel. Two C4027 with enough copper pour for two watts doesn’t save any space, does it? Defeats the main purpose of SMDs for sure.
The last survivor will be the MJE15032, but it’s oversized for many applications and not always considered fast enough.
The last survivor will be the MJE15032, but it’s oversized for many applications and not always considered fast enough.
It is true that SMD have the disadvantage of dissipation, but I figured that if I were to solder to an aluminum or copper substrate that would be acting as a heat sink with legs coming off for the through hole board. Similar to this.
Shoot, you can even make the silk screen and traces so that you can solder two of them and run them in parallel on the same board. I will check out those devices you recommended. Luckily I’ve been able to come across some C3298/A1306 and snag 50 pair. Out of curiosity if anyone has any thoughts on those I’d love to hear it. I was initially on the lookout for C4883 and its partner. Do you think mounting these SMD parts to copper plate would be sufficient? What’s the difference between the 2SC4027 “T” and “S”. I thought that it might’ve been packaging, but they seem to be the same packaging. Does it matter, are they the same?
Awesome, thank you so much for the recommendations. Luckily I picked up those Toshiba devices a while ago, but I’ll pick up more on my next order. I have 50 pair currently, but didn’t realize they were regarded so highly. Out of curiosity, what makes them so much better than the TTC004b and TTA004b? The C011b looks to be higher voltage and a little less current capability. What makes it special?
What makes the C4832/A1668 special? I see they’re 15 MHz/20 MHz devices. The MJE are 30 and are considered too slow by many.
I will definitely take a look at those devices suggested above. I’m not too worried about saving space, as long as I can get it to fit into the board that’s all that really matters lol. I’m just looking strictly for performance.
The MJE15032/33 are definitely decent drivers, they were some of the first I bought when I got into this hobby and I still have plenty of them. But after visiting here sometime, I found they’re more about brute force. I have been told on multiple occasions that there are superior drivers out there as long as you don’t need the current and dissipation capabilities of the MJE devices.
Dan
Last edited:
The 004 is the 160 volt class device, which may or may not be enough. Toshiba and not others made a LOT of those in the early days - you could choose from a dozen different pairs. The 230 volt version will work almost anywhere but as voltage goes up, quasi-saturation gets worse. It just depends on your rail voltage which one you want. The 230V version IS the same die as the original 2SC4793, just re packaged. They do not have super SOA, so it is common to see two in parallel as drivers when rail voltages get above +/-55V.
The MJE15032 is about brute force - useable to drive up to six C3281-class outputs in pro PA applications. But usually when I need to drive more than four I just use another of the same output, or at least a C5242, as it’s usually an EF3 by then anyway so the VAS will be loaded with something smaller. It’s too much brute force for a 50 watt amp, sort of on the fence at 100, probably about right over 200 (By hi-fi standards).
The MJE15032 is about brute force - useable to drive up to six C3281-class outputs in pro PA applications. But usually when I need to drive more than four I just use another of the same output, or at least a C5242, as it’s usually an EF3 by then anyway so the VAS will be loaded with something smaller. It’s too much brute force for a 50 watt amp, sort of on the fence at 100, probably about right over 200 (By hi-fi standards).
Look at the bright side. We still have ever lasting MJE340/350, MJD340/350. 😁
Edit: btw, https://www.onsemi.com/pdf/datasheet/mjd340-d.pdf
Last edited:
Different class of devices - they do have their uses but generally not a first or second choice for top end products and designs. Good compromise when things can be compromised like in PA amplifier VAS or predriver service. Been there done that bought the T shirt.
But there used to be many more (real) makers of that pair than there is now. MJE/MJD1503x will survive them.
But there used to be many more (real) makers of that pair than there is now. MJE/MJD1503x will survive them.
Okay, awesome info to have, I’ll definitely be picking up some more TTC011/TTA006. There have been multiple occasions where I have replaced faster TO220 devices with TO126. Generally older Marantz, C1913 and A913 to C2690 and A1220. I’ve done a couple that get every day use and have been functioning perfectly for at least a couple years. so I would think with proper heat sinking they should be able to do as well as their TO220 counterpart. And I do see them doubling up, just as in that NAD amp you were helping me with, doubled up the C4793 and A1837 to drive 4 pair of C5200 and A1943 on +/-74v.
What is your opinion of the C3298 and A1306? I was originally looking for some C3884 and A1859, but no luck in finding them in bulk. I am building a pair of wolverine amplifiers and getting a couple pair of the C3884/A1859 from member fireanimal. So I’ll have them for this build, but just wanted your opinion on them, a close second best? Is it considered a good driver, better than the C4793 combo?
Dan
Btw, 2SA1837 / 2SC4793 are still manufactured by Unisonic, Taiwan and distributed by Profusion Semiconductors. They don't sell dodgy / fake products.
Unisonic 2SA1837/ 2SC4793 Datasheet
Profusion 2SA1837
Profusion Bipolar Transistors
Unisonic 2SA1837/ 2SC4793 Datasheet
Profusion 2SA1837
Profusion Bipolar Transistors
Where are you getting C3298/A1306? Is unisonic making those too? They were the original driver pair for the C3281/A1302. Notice how close the part numbers are in the sequences. Those drivers were discontinued by Toshiba long ago.
They are Toshiba devices. I’m getting them from a forum member that looks to be trusted by the community here. He said he picked them up in Akihabara, Japan about 20 years or so ago. He has offered to send them to me without payment and I can pay him after having them checked out. So I’d be inclined to trust him.
What are your thoughts on those particular drivers? Quite close in numerical order to those outputs.
Dan
Last edited:
That’s awesome to know that those parts are still available. I would imagine that shipping from the UK to the US isn’t going to be killer either. I just wish that device was a little more robust. I’ll definitely grab some though.
Dan
Well if you find old stock that’s the way to get them. The few I still have are pulls. The PNPs are the green ones, and both types are the B’s (highest voltage grade).
They were more or less intended to be used as a set with the C3281’s. They are good parts, just not 200 or 230 volts, if that is needed. But as you’ve seen 230 volt vceo rating doesn’t mean C4793’s handle a lot of power at 74 volts.
Would you be so kind and take a picture of them and post it here ? I need to see something. I also have a bunch of 2SA1306B/2SC3298 but unfortunately they aren't a good match. One is a 200V device, other is 160V device. Also, the font used for lot number and the logo isn't same between devices. Meh.
Edit: ninja'd by wg_ski... 🙂
These are not my pictures but the transistors looks exactly same. They are tested in a real rig, +/- 78V, no problems.
Last edited:
Yeah exactly, green PNP and black NPN and type B. I don’t think I’ve ever been in a situation where 200 or 230v was needed, don’t have an example where that would be, meaning a known amp model?
Basically what I am wondering is that now I have these coming should I continue looking for C4883a and its partner? Or is there nothing really to be gained by the extra 20 MHz? I know that the 4883 has an extra 500 mA current rating, but assuming that we’re talking about a circuit where either could be used, any reason to strive for the others?
Dan
The "T" and "S" suffix is the hFE grade.
S - 140 to 280
T - 200 to 400
Accuphase use 2SA1552 / 2SC4027 in some of their amplifiers.
That was simply after the big change. Toshiba changed logos right around the time green PNP packages disappeared. Nothing wrong with either of them. I would probably breakdown test the non graded one if I was running 80 volt rails, as it is only guaranteed to 160V. I’d want to see some margin. They still could be a good gain match - hFE and vceo is roughly correlated in the long term average, but both are Y grade. If you want tighter you have to test them yourself anyway, and it is possible to be matched.