Over the past years I've seen that TO-3 transistor packages have fallen out of fashion with "equivalent" and fakes being all that is available. Is my understanding correct? If so, why? What is the current, equivalent package format for similar power semi-conductors?
Metal can have better thermal behavior and is less suitable to DIE contamination, but is more expensive to produce and assembly as you need 4 holes to fix it in heat-sink.
With mass production the tendency is use cheap items and easy to assemble with only one hole like TO3P and TO264 package.
I worked a little period in IBRAPE that produce Philips semiconductors in Brazil in 80's. It stooped operation in Brazil as our costs was high and not competitive with imported parts even in that period (late 80's years)
With mass production the tendency is use cheap items and easy to assemble with only one hole like TO3P and TO264 package.
I worked a little period in IBRAPE that produce Philips semiconductors in Brazil in 80's. It stooped operation in Brazil as our costs was high and not competitive with imported parts even in that period (late 80's years)
Beancounters. Cost too much, nobody uses them, no demand, why manufacture them. Vicious circle. They are still my favorite because all of the equipment I like to work on uses them. Hopefully ON will continue.
Craig
Craig
The semiconductor manufacturers are looking to minimize materials and labor cost in manufacture.
Product manufacturers are looking to minimize component cost, size and assembly complexity.
The TO-3 is a 55 yr old obsolete package design and is rarely used in new designs outside of audiophile targeted amplifiers made in Asia.
Product manufacturers are looking to minimize component cost, size and assembly complexity.
The TO-3 is a 55 yr old obsolete package design and is rarely used in new designs outside of audiophile targeted amplifiers made in Asia.
The "past years" must have been quite a while ago. Mostly only DIY, high power PA and class A specialists have used steel case transistors for audio, since the 1980s. There is no real equivalent to the steel case even now but sadly the largest plastic package, Sanken's MT200, that has itself been around for near 40 years, has just now become obsolete.
Next down in size is TO264 (large, but not common - see On-Semi audio transistors MJL4281/MJL4302.
I think the most popular plastic power transistor size for decades has been the smaller T03P package used by all Japanese manufacturers but followed more recently by T0247 (See NJW0302/NJW0281) and compare TIP3055/2955 for details). Thermal transfer detail comparisons of most styles of plastic power transistors are here:
https://www.fairchildsemi.com/application-notes/AN/AN-4166.pdf
Next down in size is TO264 (large, but not common - see On-Semi audio transistors MJL4281/MJL4302.
I think the most popular plastic power transistor size for decades has been the smaller T03P package used by all Japanese manufacturers but followed more recently by T0247 (See NJW0302/NJW0281) and compare TIP3055/2955 for details). Thermal transfer detail comparisons of most styles of plastic power transistors are here:
https://www.fairchildsemi.com/application-notes/AN/AN-4166.pdf
Reflecting what-all everyone above has already said, I think the best answer is
Its not like the TO–3 baseplate is big-and-thick (thus being an outstanding heat spreader). Its not. Never was. I've seen transistors in TO–3 as high as "200 watt rated"; but that was before the advent of powdered metal loaded silicone oil heat-sink coupling compounds. You needed a fairly big surface area to deliver The Watts to the typically extruded aluminum heat sink baseplate.
The larger size meant that fewer high-power devices could be strapped to the heat sinks per unit length. Good, bad, hardly makes a difference. Today — as one can easily see on any “high end” amplifier — the sand-state devices are crowded shoulder-to-shoulder in longish strips. To the chassis heat sink. Passively cooled too, for the most part. Since fans are so darn noisy.
Fabrication cost cannot be discounted, either. A larger can, with a hermetically sealed cap is WAY more expensive to make than a silicon-bonded-to-copper tab with a heat-and-UV cured epoxy overcoat. Machines at high speed can make the tabs-and-epoxy TO–220 package by the gazillion. Cheaply.
REFLECTING on the “only available are counterfeit parts” comment … I think it is a sad reminder of just how little the mainland (and Taiwanese!) Chinese have in the way of praiseworthy upholding and adhering to published standards. Or, as one wag quipped at a meeting in 1988 when my partner and I were vending memory chips, “Loo there would sell his granmother into slavery if he could save twenty cents per DRAM chip”.
Seriously.
I think people who vend counterfeit crâhp should be prosecuted.
Perhaps even as a “capital offense”.
Certainly so, for repeat offenders.
Just saying,
GoatGuy
(1) unremarkable benefits for
(2) substantially larger size and
(3) higher fab cost
(2) substantially larger size and
(3) higher fab cost
Its not like the TO–3 baseplate is big-and-thick (thus being an outstanding heat spreader). Its not. Never was. I've seen transistors in TO–3 as high as "200 watt rated"; but that was before the advent of powdered metal loaded silicone oil heat-sink coupling compounds. You needed a fairly big surface area to deliver The Watts to the typically extruded aluminum heat sink baseplate.
The larger size meant that fewer high-power devices could be strapped to the heat sinks per unit length. Good, bad, hardly makes a difference. Today — as one can easily see on any “high end” amplifier — the sand-state devices are crowded shoulder-to-shoulder in longish strips. To the chassis heat sink. Passively cooled too, for the most part. Since fans are so darn noisy.
Fabrication cost cannot be discounted, either. A larger can, with a hermetically sealed cap is WAY more expensive to make than a silicon-bonded-to-copper tab with a heat-and-UV cured epoxy overcoat. Machines at high speed can make the tabs-and-epoxy TO–220 package by the gazillion. Cheaply.
REFLECTING on the “only available are counterfeit parts” comment … I think it is a sad reminder of just how little the mainland (and Taiwanese!) Chinese have in the way of praiseworthy upholding and adhering to published standards. Or, as one wag quipped at a meeting in 1988 when my partner and I were vending memory chips, “Loo there would sell his granmother into slavery if he could save twenty cents per DRAM chip”.
Seriously.
I think people who vend counterfeit crâhp should be prosecuted.
Perhaps even as a “capital offense”.
Certainly so, for repeat offenders.
Just saying,
GoatGuy
Already stated, "cans" have become very expensive. They typically require a glass frit to be fired around the leads and welding in a neutral atmosphere to seal the edges, most importantly a one at a time operation.
Modern packages are actually made in large arrays and the metal sprue is cut off with a die stamper to singulate the parts at the end.
Modern packages are actually made in large arrays and the metal sprue is cut off with a die stamper to singulate the parts at the end.
TO-3 are expensive and more difficult to mount.
4 TO220 or 1 TO264 will outperform TO-3.
Limited choice and old process will seal it's doom as no MOSFETS are available in that package. Mosfets are much better for audio outputs then BJT.
Class D is becoming the audio amp perferred build now as it's performance edges up to the best analog amplifiers.
4 TO220 or 1 TO264 will outperform TO-3.
Limited choice and old process will seal it's doom as no MOSFETS are available in that package. Mosfets are much better for audio outputs then BJT.
Class D is becoming the audio amp perferred build now as it's performance edges up to the best analog amplifiers.
TO3 was originally designed in the 50's, and was a real improvement over the options available at the time, but of course, that was more than sixty years ago, and technology and needs have hugely evolved since then.
For really demanding applications, like space and oil-drills, metal cases are still the norm, because sustained operation at 250°C is required, but there are metal variations of the TO220 and other cases which are more practical than the ol' n trusted TO3 (much more expensive though).
As for the performance of class D amplifiers, they might be comparable from the purely commercial, practical POV, but it would be perfectly possible to design a ppb-level THD analog amplifier, if cost and noise considerations are put aside (the THD of any such amplifier would be drowned in noise).
Class D amps have progressed, but are still very far from such a target
For really demanding applications, like space and oil-drills, metal cases are still the norm, because sustained operation at 250°C is required, but there are metal variations of the TO220 and other cases which are more practical than the ol' n trusted TO3 (much more expensive though).
As for the performance of class D amplifiers, they might be comparable from the purely commercial, practical POV, but it would be perfectly possible to design a ppb-level THD analog amplifier, if cost and noise considerations are put aside (the THD of any such amplifier would be drowned in noise).
Class D amps have progressed, but are still very far from such a target
I remember building a Maplin 225WRMS amp in 1980.
What a pig to drill the holes spot on for the 2n3055/mj2955's.
What a pig to drill the holes spot on for the 2n3055/mj2955's.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
I built a lot of amps in the 1970s and early 1980s with 2N3055 and 2N2955 amongst others and always had to take my time drilling those 4 holes. I always drilled the holes for the B and E pins way oversize to make life a bit easier.. No question drilling just one hole is much easier.. LOL
Whoever came up with the TO-3 package never considered usability/serviceability in the design process. Who puts the electrodes that have to pierce through the heat conduction plane? Then requires special fittings so pins don’t short. Specially cut mica spacers, grease, etc. I have done it a few times and hated it each time. Good riddance to TO-3 as far as I am concerned. With TO-264 we have large flat surfaces, heat pipes, and fans that can handle 800w of dissipation. One screw.
Last edited:
But SU’s are complex vacuum mechanical actuators whereas TO-3 just need to cool a small die of silicon and provide 3 electrical contacts. SU’s are a pain though - I replaced the two SU’s with down draft Weber’s on my MG that I had long ago. Lucas Electric gives a whole new meaning to electrical fires and shorts.
I could not believe my eyes yesterday a block from my house was parked a perfect BRG Morgan fully restored.
Quite the opposite. What was originally envisioned was something only a little more difficult to trip out than a vacuum tube. They were originally intended to be mounted on a heat radiator, often on the exterior of a piece of equipment, and almost always easily accessible. Usually socketed and mounted with 2 screws. Change it without desoldering, and often without opening the chassis. Much easier to change out than the TO-36 which had solder lugs on the base and emitter, and one bolt on the wrong side of thae package. Quite an improvement in serviceability.
Then came modern manufacturing processes, and TO3's got soldered to pcbs and embedded in the interior of densely packed equipment that you have to take completely apart to get to the output transistors. Not just one or two per channel anymore, but 16 or 32. In these situations, flatpacks are WAY preferable in all respects.
Honestly .. get real
I love them and would always use in all my projects ...
better then the current crop of smc (surface mount crap) that is flooding the market .. absolutely awful stuff totally banded from my work shop pile of crap ..
good Riddance
does not help in any way for DIY ... crap crap
honestly get rid as soon as possible for diy projects totally **** idea
I love them and would always use in all my projects ...
better then the current crop of smc (surface mount crap) that is flooding the market .. absolutely awful stuff totally banded from my work shop pile of crap ..
good Riddance
does not help in any way for DIY ... crap crap
honestly get rid as soon as possible for diy projects totally **** idea
Last edited:
Jimbo! Listen brother, don't hold back. No need to sugar-coat it!
I used (mostly 'cuz I had to from the few darn decent scrap-recovery parts available in 'my day') TO–3 transistors for almost every power project I had going. The fact that the pins sticking out were very precisely spec'd in diameter meant that you could purchase top-shelf sockets for them that had receptacles with “surround-gold-contacting” internals. Just like the super-expensive military connectors.
The point there was with gold-to-gold contacts — and with sleave-and-pin linear contacts — one could also count on the connectors not having vibration-causing-ohmic-noise in service. Solid as rocks, actually. Bournes made (or 3rd party sourced) beautiful TO–3 sockets.
The only real 'gotcha' was dealing with the fact that the “can is live”, or 99% of the time, the collector (in emitter-base-collector transistor terminology). This meant that you either had to perform dicey mica insulator surgery along with nylon screws to isolate your transistor(s) from the heat sinks … when complimentary pairs were commonly being used, or you were going to see spectactularly expensive uncontrolled fireworks after changing transistors upon power-up. We all did it once or twice. “Once burnt, not repeated” is the motto.
But this seriously didn't in the end work against the utility of the TO–3 package. Being a large bottomed device, even if you had to use mica spacers and wee globs of silicone grease to establish a decent thermal path from transistor to heat-sink, even if so, the fact that once properly screwed down, they weren't going to drift meant that all would be safe until the next time the idiot user(s) hooked up fuzzy-fraying zip cord as speaker wire and accidentally shorted out a channel. Death. By Stupidity. And repairing it, again, consisting of surgery and expensive transistors.
This is why I was so impressed with the great big Marantz amplifiers that had short-circuit protection and multiple TO–3 cans per channel. Like 8 of them. Full complimentary PNP/NPN outputs, all-transistors-isolated, use of heat sink thermal compound and mica spacers, and beautiful Amphenol sockets. In all the years of abuse we gave my Pa's ridiculously expensive — and thus don't touch my stereo, boys!!! integrated amp, hooking it to multiple speakers in parallel, accidentally shorting in numerous times, feeding it WAY too much signal and all the rest, not once did a single 'final' transistor bug out. Worked like an indestructable tank.
But now it is small-tab MOSFETS. The bipolar transistor for reasons often cited, but really rarely appreciated is no longer the workhorse output device for modern amplifiers. Especially in the “Class D” age, with switching frequenices around a megahertz and operating efficiencies exceeding 80%. My friend's brand new, under $200, class D bass guitar amp ("head" only) puts out 350 W continuous, and well over 1,500 W peak-before-clipping. And even under the most ridiculous service the heat sinks barely get warm. Barely.
We've come a long way.
Sorry to see you go, TO–3.
It was a hêll of a journey.
GoatGuy
- Status
- Not open for further replies.