Sanken 2SC1831 substitute - MJ21194?

[Tone Meister]

Sanken 2SC1831 is an obsolete bipolar output transistors for an early model (1987) Sumo Nine Plus home audio amplifier. The Sanken site doesn't list an alternative. Owner of the amp is concerned that any replacement might not sound as good as the original part. It seems to me that the MJ21194 would be a suitable replacement, but I wanted to run it by the pros first because I think we're gonna need 20 of these.

I haven't located a schematic for the early version of the amp, but if I do find one I'll post it.

[spamunkulus]

my old Databook from 1981 says:

2SC1831

SANKEN NPN-Darlington Powertransistor

Uce=70V
Imax=10A
P(25)=100W
hfe=500
Ft=10MHz

A darlington, very strange.
If that is true, the MJ21194 would not work.
In my schematic (later Version) a 2SD555 is used.
This might be replaced by the MJ21194, though the SD555 was a bit faster.

[J M Fahey]

Please search and post the datashet here.
Specially one which shows case and pinout.
I´m certain there are modern replacements available but besides electrical compatibility they must also fit mechanically.

[Tone Meister]

my old Databook from 1981 says:

2SC1831

SANKEN NPN-Darlington Powertransistor

Uce=70V
Imax=10A
P(25)=100W
hfe=500
Ft=10MHz

A darlington, very strange.
If that is true, the MJ21194 would not work.
In my schematic (later Version) a 2SD555 is used.
This might be replaced by the MJ21194, though the SD555 was a bit faster.

Quoted from llwhtt (Craig) over at diyaudio forum:

"No, the 2SC1831 is not a Darlington, the subs come up as Darlingtons but it is a Super Beta, current gain is similar to a Darlington.

The original SUMO NINE used the 2SC1831s driven by opamp only, the 2SC1831s had enough gain to be able to do that.

The SUMO NINE PLUS has an opamp, drivers and outputs. The opamp (LF412) does not have enough current gain for normal transistors (2SD555 or MJ21194) so the driver (TIP31C) was added. The driver and outputs are in a Darlington configuration."

[Tone Meister]

Please search and post the datashet here.
Specially one which shows case and pinout.
I´m certain there are modern replacements available but besides electrical compatibility they must also fit mechanically.

Juan,

I can't locate a datasheet, but I did find the electrical specs. It's a TO-3 case but I don't know the pinout for certain.



2SC1831 Transistor Datasheet. Parameters and Characteristics

Type Designator: 2SC1831
Material of transistor: Si
Polarity: NPN
Maximum collector power dissipation (Pc), W: 100
Maximum collector-base voltage |Ucb|, V: 90
Maximum collector-emitter voltage |Uce|, V: 70
Maximum emitter-base voltage |Ueb|, V: 6
Maximum collector current |Ic max|, A: 8
Maksimalna temperatura (Tj), °C: 175
Transition frequency (ft), MHz:
Collector capacitance (Cc), pF:
Forward current transfer ratio (hFE), min: 500
Noise Figure, dB: -
Package of 2SC1831 transistor: TO3

[The Dude]

2SC1831 Generic Sumitomo Transistor C1831 | eBay

Not Sanken, but maybe?

[Jazz P Bass]

Sumo Nine Plus output transistor replacement - Sanken 2SC1831 - diyAudio
The good news is that you have the Nine +.

[Tone Meister]

Sumo Nine Plus output transistor replacement - Sanken 2SC1831 - diyAudio
The good news is that you have the Nine +.

Yeah, he has a Nine Plus, but that's a schematic for the later version that uses NEC 2SD555 chips. I have a strong feeling I am fixin' to learn something before this is over ...

[J M Fahey]

Messing with that amp is opening a can of worms , but here´s a couple side notes:
1) D.A.T.A Book lists them as Darlingtons.
A DIY Audio member claims they are not so, but "super beta" , whatever that means.
I´d rather trust the World´s largest transistor data book.

2) Motorola claims their 2SC1831 direct replacement (not ifs or buts) is 2N6056 , (which is a Darlington) and meets or exceeds 2SC1831 parameters ... and is widely available.
2SC1831 ... - Datasheet Search Engine Download
I´d also trust Motorola.
Futurlec has all you want for U$2.60 each.

3) the 2nd source Sumitomo may be equivalent to the original one, but besides being old and rusty its mechanical construction is *horrible*.
It uses the cheaper than cheap "double thickness" iron base, without an internal copper heat spreader "coin" and just a small copper plug in the middle (easily visible, it´s the stamped disk in the middle of the base) which means that in practice this TO3 has same or less copper surface than a TO220 transistor.
I would avoid them like the plague.

EDIT: that said, I would replace dead Sankens with On/Motorola and all transistors in parallel so each pack is same brand/batch .

You might end with a couple spare pulls, not bad for future servicing.

And of course, "sound will be exactly the same", they are being used as unity gain followers which is the great equalizer.

[km6xz]

Yes, they are Darlingtons and the On Semi will be a better part and cheaper. If you did not rebuild the supply, it should sound identical so just tell the owner that you scraped out all the fairy dust from the old ones and re-installed in the new modern transistors so they have the "vintage silicon" sound. High end hi-fi'ers are very sensitive to suggestion so tell him it does sound better as he is test listening to it.

[Tone Meister]

More information and schematics are attached, one with the relevant section marked up.

"I guess there are some disagreements on the SanKen 2SC1831 as being a Darlington or not a Darlington. In my Japanese substitution manual SanKen's replacement for the 2SC1831 is a 2SC4131. It's NOT a Darlington either but it's also not a TO-3 case it's a TO-3P, won't work without some mods."

"OK, here's the story of this particular version of the NINE+. The NINE+ was originally designed to use the NEC 2SD555, an expensive transistor. AFTER James Bongiorno sold the company the beancounters didn't want to purchase those expensive xsistors so the circuit was "bastardized", James' word, to use the 2SC1831 of which they had 1000s.

"James Bongiorno used to hang out on the SAETalk forum, there he answered some questions about the NINE and the 2SC1831. James never had a problem suggesting replacement components for obsolete devices but the 2SC1831 was one where he very adamant about, there are NO replacements."

"So for this amp to live on it needs to be un"bastardized" and the MJ21194s fitted. There are several components in the bias circuit that need to be removed and replaced with the much simpler bias circuit. Also the output emitter resistors need to be replaced, from .51 Ohms to .2 Ohms. The peak current resistor, R38 looks to 1K in this version, needs to be 120 Ohms in the normal PLUS. The fan appears to be a DC fan instead of the usual AC fan, that shouldn't matter a bit."

"Not knowing how they did the mods physically is my only concern. Was the PCB hacked or not. Comparing the two schematics it looks as though no traces had to be cut. The normal bias circuit is just two resistors and two 5K trimmers per channel, the "B" version has two resistors, two 100K trimmers, four xsistors, and two diodes. There is not much space in a NINE so I'm not exactly sure where these parts were installed and if the transistors were for thermal tracking. I guess it really doesn't matter as all of it will be gone ..."

[km6xz]

So it looks like the design was a work in progress, which was version is yours? It appears to be neither of the prints
Does yours have the TIP31 drivers? The 2SD555 is a conventional moderate beta Bipolar 200 volt power transistor. The devices yours had in it were high Hfe devices. Do you have a curve tracer available to test the good channel's transistors to end the speculation of what they had in it? If not you can set up a test jig with a bench supply and a few resistors to see what current gain the good channel's transistors are. If it does not have the driver transistors, it has to be a Darlington or it is not going to work. The passive offset circuit that creates a DC servo with each input opamp is pretty simple in both versions of the schematics presented but neither of these are what you have.

One of the problems with working on low production volume gear is that every unit off the hand assembled production like is just as likely to be different as similar. Why would they worry about transistor costs when their retail price was very high? It is entirely possible, being a shoe-string operation, they just did not have the money to place a minimum manufacturing lot quality so changed the circuit with every batch as to which parts they could get. There is no rocket science with this circuit, the only novel aspect is the bias circuit,that is just a DC servo with built in offset from the output of the opamps.

[Tone Meister]

The amp is not mine; it belongs to a friend who indicated it has the Sanken 2SC1831. I have not seen the amp for myself and have asked him to open the amp back up and to look for the TIP31 drivers so we can ID exactly which version he has. I've posted this here because I wanted to learn all I could and be reasonably sure it is something I want to take on before he orders a bunch of parts that won't work. Without a solid plan in place, I don't want to disassemble an amp that appears to need a partial circuit re-design to fit currently available chips.

If I have this right, the amp was not a work in progress from the standpoint of the original designer, James Bongiorno. He designed the original Nine Plus (8-25-82) amp around the NEC SD555 and was satisfied with its performance with that transistor. It was later after he sold the company that the circuit was altered and the Sanken 2SC1831 transistors fitted, likely to cut costs. I simply don't know the full history of the circuit, but the circuit was modified at some point. The schematic SUMO NINE PLUS_+ is the target circuit and the schematic SUMO PLUS NOT is the amp we believe my friend has. I should not have attached the Sumo_Nine schematic.

I think the goal is to return the amp to one of the earlier circuits by redesigning the bias circuit to accommodate the MJ21194 outputs. If I am interpreting my source correctly, that would entail removing some parts and changing values here and there. Juan may have this one right, it may be opening a can of worms. But, if I can get a consensus on the redesign I'd be willing to take it on.

[Jazz P Bass]

Can-O-Worms.

[Tone Meister]

Can-O-Worms.

Does that mean it's time to go fishin'?