You often needed a couple amps of base current, and drive the base emitter into avalanche during turn off (said so in the data sheets). Transformer drive of Borg - resistance is futile you must comply. If not, the vce would rise (out of saturation) as the collector current ramp did naturally from the inductance and that spelled h.e.a.t. D.e.a.t.h.
Designing the circuit to operate safely and reliably probably was easier with tubes.
Designing the circuit to operate safely and reliably probably was easier with tubes.
Hi Jim,
All due respect, but not one decent amplifier was ever marketed using a 2N3055 - ever. The transistor was never well suited for audio use for one. Parts designed for that purpose were vastly better in every regard.
The 2N3055 design had involvement from Lambda, the linear power supply manufacturer, and somewhere I have copies of some original documents that related to the early specs. It was designed not to oscillate when mounted on a heat sink away from the control circuit. That's why the low fT.
Yes, an amplifier can be designed and built using these parts. No, they have not be manufactured for ages so anything new is a copy. Real 2N3055 have a collector - emitter breakdown voltage of 60 VDC, so that is a hard limit. Collector - base breakdowns were listed at 100 VDC, but running these at a full supply voltage of 60 VDC or higher puts you beyond the rated maximums and into the safety margin built in. It would be completely irresponsible to operate in the region. So, 60 VDC total means that if you are using a bipolar supply common in most designs, your hard limit is 30 VDC at you are running at absolute maximum.
I wouldn't call an NAD product a good amplifier. That is a stretch, and if you looked at build quality, they were pure junk. We are talking extremely low budget design. Measured performance of these back my statements up, and anyone with eyeballs can clearly see these were built as cheaply as possible.
Jim, I designed and built various amplifiers back then, and was professionally servicing them as well. Still do these things. Back then there were better transistors designed for audio service, but they cost more. That's fine, you either used the right parts, or got by using parts not designed for the purpose. However, you would never get the performance from a 2N3055 like say a 2N5631.
Getting back to horizontal output transistors. They are, as stated, always NPN and typically include a flyback diode. They may be good for 5 amperes typically with 1,500 volt breakdown voltages. This combination means they will have low beta (8 is common) and low fT because of how the dimensions of the die must be. So while you can make them function in audio applications, they are not the best choice. I have used them for high voltage regulator pass transistors, but the low beta is always a concern as the driver will also dissipate a lot of power.
-Chris
All due respect, but not one decent amplifier was ever marketed using a 2N3055 - ever. The transistor was never well suited for audio use for one. Parts designed for that purpose were vastly better in every regard.
The 2N3055 design had involvement from Lambda, the linear power supply manufacturer, and somewhere I have copies of some original documents that related to the early specs. It was designed not to oscillate when mounted on a heat sink away from the control circuit. That's why the low fT.
Yes, an amplifier can be designed and built using these parts. No, they have not be manufactured for ages so anything new is a copy. Real 2N3055 have a collector - emitter breakdown voltage of 60 VDC, so that is a hard limit. Collector - base breakdowns were listed at 100 VDC, but running these at a full supply voltage of 60 VDC or higher puts you beyond the rated maximums and into the safety margin built in. It would be completely irresponsible to operate in the region. So, 60 VDC total means that if you are using a bipolar supply common in most designs, your hard limit is 30 VDC at you are running at absolute maximum.
I wouldn't call an NAD product a good amplifier. That is a stretch, and if you looked at build quality, they were pure junk. We are talking extremely low budget design. Measured performance of these back my statements up, and anyone with eyeballs can clearly see these were built as cheaply as possible.
Jim, I designed and built various amplifiers back then, and was professionally servicing them as well. Still do these things. Back then there were better transistors designed for audio service, but they cost more. That's fine, you either used the right parts, or got by using parts not designed for the purpose. However, you would never get the performance from a 2N3055 like say a 2N5631.
Getting back to horizontal output transistors. They are, as stated, always NPN and typically include a flyback diode. They may be good for 5 amperes typically with 1,500 volt breakdown voltages. This combination means they will have low beta (8 is common) and low fT because of how the dimensions of the die must be. So while you can make them function in audio applications, they are not the best choice. I have used them for high voltage regulator pass transistors, but the low beta is always a concern as the driver will also dissipate a lot of power.
-Chris
If you are buying new transistors for audio amps, explain the need to use a fixed speed switching transistor in a variable speed audio amplifier, as it handles a much larger frequency range than a SMPS transistor.
How will the sound reproduction be affected?
Is there a very big price advantage?
Or are you just curious?
How will the sound reproduction be affected?
Is there a very big price advantage?
Or are you just curious?
No need today, that’s for sure. Buy MJ21193’s and be done with it. No old switching transistor, no 3055.
Even class D wouldn’t use one of these old klunlers because they are too slow and require too much POWER to drive at several hundred kHz.
But what to do with old HOTs laying around, or the ones that can be found on the surplus market cheap? That’s likely what all the hubbub is about. Not the best choice, even if you can get them cheap. No gain, and worse is the SOA in the critical 20 to 50 volt VCE range. I may actually buy older audio types, even the 80 volt types on the surplus market but I leave the HOTs alone. No need to waste even a buck apiece. Horizontal output tubes on the other hand, are worth something. Especially if another iron curtain goes up. Saw it coming years ago and started buying.
I did use them as drivers for output banks (which were operated above rated VCEO) when no other high voltage devices were available to me. But that’s ancient history.
What I meant was that the switching speed on SMPS and similar is in the same range, does not fluctuate so much, most are PWM controlled by chips like the TL494 family...and the device is handling one frequency at a time, which will vary with load, as required.
The frequency range would start about 3 kHz, I think, and extend to 30 kHz, maybe more...to hundreds of kHz, as wg_ski says...
In audio, there will be bass and treble notes together, a polyphonic frequency range is being amplified.
At the same point in time the transistor is handling frequencies in the 20 Hz to 20 kHz range, so its construction may be different.
And those frequencies will not have the same amplitude...
For that reason, I think a HOT is not suitable as an analog audio amplifier.
To save a few bucks, why compromise?
And make a new design?
Too much work...
The frequency range would start about 3 kHz, I think, and extend to 30 kHz, maybe more...to hundreds of kHz, as wg_ski says...
In audio, there will be bass and treble notes together, a polyphonic frequency range is being amplified.
At the same point in time the transistor is handling frequencies in the 20 Hz to 20 kHz range, so its construction may be different.
And those frequencies will not have the same amplitude...
For that reason, I think a HOT is not suitable as an analog audio amplifier.
To save a few bucks, why compromise?
And make a new design?
Too much work...
Last edited:
absolutely.Just looking in the old T D Towers Transistor Selector. BU208 has a gain of 2
@ 4.5 amp. The 208A pushes that to 2.5. Power dissipation 12 watts and 12.5 watts respectively, the first at 25C and the second on a heatsink at 25C ambient. How big is the heatsink though.Fast switch off is so critical to all these applications including SMPS. Many SMPS used capacitor coupling to the base of the switching transistor together with a diode clamp. The cap was often around 4.7uF and a typical failure item. If you looked at the drive waveform the base was sluggish in being pulled back down due to a rise in e.s.r. of the cap.
Hi Karl,
Yes, some spec'd out at 2.5 in the book. Abysmal performance for any design apart from HOT use where the voltage rating was the most important thing.
I agree with wg_ski. Just use the correct parts and be done with it. I can't understand wasting time with sorting out fakes, used parts and below spec parts - coupled with shipping costs. It just doesn't make any sense to play with question marks. Unless a part comes from authorised distributors, it is a question mark.
Yes, some spec'd out at 2.5 in the book. Abysmal performance for any design apart from HOT use where the voltage rating was the most important thing.
I agree with wg_ski. Just use the correct parts and be done with it. I can't understand wasting time with sorting out fakes, used parts and below spec parts - coupled with shipping costs. It just doesn't make any sense to play with question marks. Unless a part comes from authorised distributors, it is a question mark.
Make that $1 or $1.50 FOB tops and I´m ordering 1000 right now, just for starters.
I´ve delivered over 14000 MI amplifiers so far, some 8000/9000 of them using 2N3055, or 2N3773 for the higher power ones
Most are in use today, oldest I have data of is a 200W PA power amp, still in use every weekend at an Evangelical Church.
Last reliable 2N3055H I used was a batch of about 1000 USHA/UR , Indian made, impressively rugged.
Later I learnt that USHA was an Indian Army custom supplier.
Just one Month ago I found a US surplus supplier who claimed to have some 9800 of them so I ordered 1000 , my price of course.
Maybe they didn´t like it, maybe they were bluffing and actually have no stock, they answered "they are out of stock for now, ask again later"
Not a fanatic, "golden memories" or passionate type guy , I like them simply because they have worked before, very reliably.
If you have a reliable source
Again: these are Guitar amps, MUCH abused, roughness beats anything else.
EDIT:
back in the thread : WAY back in the day, think mid 70´s, I tried to use Horizontal Output Transistors since they were the only ones easily available with HIGH Vce , I was happy with 120Vce, 150V tops, and thought I had reinvented the wheel.
Sadly, big failure, transistors exploded with speaker loads (although they just held with resistive ones)
Reading and reinterpreting datasheets plus actual testing showed me they did stand voltage, but they had ABYSMAL second breakdown characteristics.
Still "usable", sort of, but needed a lot of parallel transistors each side, and cost stopped being an advantage.
Only commercial MI amp I know used them was a WEM/Watkins amp (of Pink Floyd fame) who used 4 of them to get 100W into 16 ohm,
For some unknown reason, Brits LOVE 16 ohm speakers.
I didn´t like bridged outputs way back then, so instead I used series transistors in a totem pole configuration , similar to what Ampzilla did, but mine was much simpler.
Last edited:
1) Ya Momma!!!!!
Guitar amps have very high quality, in their own field, don´t compare pears to oranges.
A tractor is terrible in a Formula 1 track, a Ferrari or Mc Laren is the worst in the World pulling a plough in a muddy field, to each his own.
A dead amplifier has the worst quality of them all, guess which will die first driving quite inductive speakers clipping full blast for 8 hours in a row, as in a Club night or busy rehearsal room.
As a side note, Yamaha amplifiers are a good horse in the wrong race, some of them use circuits exact same as some of same vintage Hi Fi, a couple including differential Fets inside, etc.
Main problem being they invariably have very high NFB, so distortion is almost zero when clean ... and walking barefoot over glass shards when they clip, that´s why only Jazz players favour them.
2) I measure all transistors I use from day 1, any real 2N3055 stands 70V (surprise surprise, it´s even in the datasheet) up to 2008 most routinely stood > 80/85V, my own supply rails were classic +/-42 V, they lasted for years, many still do.
A few selected ones stood up to 100V , go figure.
Typically used in transformer driven amps, which like it or not, when pushing one base forward, apply negative voltage to the opposing base.
That rating, called Vcex is even higher.
Not sure you even heard about it (you apparently only know Vceo, that´s why you insist on "60V" all the time) and clearly never measured them or you would agree with my findings.
Here, straight from:
FWIW I don´t like transformer driven amps so I designed two circuits to provide negative bias to power transistor when turned OFF, go figure.
Both work well.