I was reading (if memory does not fail) that this transistors are in a "ring" configuration and are composed of many transistors in parallel, but how many ? does someone know just for curiosity ?
Thanks
Thanks
Yes, ring emitter transistors, and are made of many transistors in parallel, but do you know how many,just for my curiosity .
thanks
thanks
Perhaps this is too simplistic. There is actually 1 transistor with multiple emitters and though that confers the benefits of multiple transitors in parallel, it is much more by virtue of the extremely close thermal coupling.
However, you won't be able to simulate these devices with multiple small transistors just coupled together in free air or even on heatsinks. The close integration on a single die makes quite a different, unique device and the precise detail would be covered by patents. I'd imagine the number of emitters formed is not really exact or completely consistent in manufacture but there is some indication in docs like this and their citations:
https://www.google.com/patents/US5786622
Last edited:
I believe ring emitter transistors were a Fujitsu only dealie, and the transistors in question are from Toshiba, so no RET.
Here's a small blurb about it:
Here's a small blurb about it:
Last edited:
It may not be quite that clear-cut. Take a look at some post details in this old thread:
http://www.diyaudio.com/forums/solid-state/53508-skinny-sanken-2sa1216-epitaxials.html
Thanks, pretty interesting read. Strange though that many manufacturers are mentioned there - but not Fujitsu. Why would they claim it is their unique RET technology? I suppose there are many ways to go about it, or something.
Simply put, a ring-emitter device consists of a great number of transistors on 1 die.
So, as Jacco says in the old thread this transistors are composed of many transistors in parallel as i suspected, and thats why they are the fastest transistors available for audio aplications, but i could not find any more technical information about this subject in the net.
I see. But I assume 2SA1302/2SC3281 are not RETs, though? I've never seen them advertised as such, and they are relatively slow (30 MHz) compared to the RETs mentioned here and in that thread (60-80 MHz)
Jacco was describing the specific modern Sanken types of LAPT (Large Area Parallel Transistor) that indeed are claimed to have multiple transistors integrated in parallel on the die but that is not quite the same as the competing Ring Emitter or perforated emitter technologies which your OP referred to, even though they may be classed together in terms of improved performance.
These industry details have all been discussed here before, but in which thread?
Member ilimzn I think, has a better historical knowledge of semiconductor developments. Perhaps he can fill in some details and facts if interested.
Edit: This OP by Tiefbassubertrager might make interesting (exhausting) reading and discusses Toshiba types:
http://www.diyaudio.com/forums/soli...istor-families-audio-power-output-stages.html
Last edited:
Thank you for the link for the thread ! i will have a very interesting reading tonight.
Thanks
compare the graphs in the datasheets rather than the "headline" claim for fT at one frequency.
You will find that they are all within a band from ~30MHz to ~70MHz, over at least some value of Ic.
You will also find that "typical" hides a lot of variation that most Manufacturers do not divulge.
When a manufacturer invents a new technology and others find that it has useful performance advantage, the others have options.
Copy it and risk prosecution, get a license to make "approved" copies, come up with an alternative, but very slightly different technology that achieves similar performance.
Most reputable adopt the last as the cheapest way to compete for market share.
Sometimes an alternative turns out to be better, then the chase starts all over again.
Last edited:
I believe ring emitter transistors were a Fujitsu only dealie, and the transistors in question are from Toshiba, so no RET.
Here's a small blurb about it:
Hi,
Likely the "unique" point of the Fujitsu's is the claimed
added ballast resistors, which are not strictly necessary
due to the instrinsic bulk Re of each multiple emitter.
rgds, sreten.
Actually many manufactirers tried the 'multiemitter' or 'multitransistor' approach. Toshibas mentioned in the thread are derivatives of a slightly older tripple diffused technology which add one more layer in emitter doping to produce an effect similar to ballast resistors in multi-designs. As they say, there are many ways to skin a cat (don't blame me, that's a n english proverb, I'd never do it 😛 ), and manufacturers tried different ones to avoid patent infringement claims.
In some cases Vbe-Ib graphs are plotted in the datasheets, and they are a giveaway of some sort of emitter ballast technique, because they converge into a linear rise rather than an exponential one, suggesting built-in emitter resistors.
Companies that I know tried these technologies:
Sanken (first and foremost, and are at it today), Fujitsu (pushed out of the market by Sanken, also had some reliability problems), NEC (pulled out of the audio power semi market altogether at some point, and were heavily involved, even making VFETs), and also Sanyo. Their technology was half-way between tripple diffused and LAPT, a very popular output pair as used for years by Rotel in their amps.
In some cases Vbe-Ib graphs are plotted in the datasheets, and they are a giveaway of some sort of emitter ballast technique, because they converge into a linear rise rather than an exponential one, suggesting built-in emitter resistors.
Companies that I know tried these technologies:
Sanken (first and foremost, and are at it today), Fujitsu (pushed out of the market by Sanken, also had some reliability problems), NEC (pulled out of the audio power semi market altogether at some point, and were heavily involved, even making VFETs), and also Sanyo. Their technology was half-way between tripple diffused and LAPT, a very popular output pair as used for years by Rotel in their amps.
So ok - everyone has their own unique way of making multiple emitter ballasted transistors to keep from stepping on each others' patents. But does anybody know how MANY emitters are on the A1302?
Every time I break open a real A1302 (or anything close) it destroys the die so I can't tell ya....
Every time I break open a real A1302 (or anything close) it destroys the die so I can't tell ya....
Well, apparently nobody knows, i tryed hard in the net and didnt find anything...maybe the factories keep this things secret...