Hey , did anyone seen hv deflection transistors with diode and base-emitter resistor inside , about 30 ohms . I have few such ,but don't think they are usable because of resistor . Typically they had few volts saturation voltage ,that was ok in tv or monitors with 130v b+. They needed high current to drive them with minimal loss , so transformer was used . 1 ampere base current, in comparison to mosfet its crap! About fixed frequency - such switching transistors were designed for different speeds , i remember bu2525 or 2527 , which was 64 khz and many amperes . Parts made for special purpose have limited use for different usage . If anyone still wants to use them for audio , can try , why not ? But expect not best results ...
That was one of many. Most of their house numbers were for specific customers and not generally available. 40636 was probably for Peavey. They had a handful that were generally available (but still, only from the factory) like the 90 volt 40411. They had a whole matched transistor set that went with it.
Motorola had all manner of house numbered parts as well, but they also had a full line of 60 and 80 volt epi-base pairs at various current ratings in TO-3 as well. Any one of those could be selected for 100 volts, so a house number might not have been a true epi- 2N3055 (might have been better to start with). Didn’t see as many of them after their 140 volt process became common. Then they did the same thing with those - selecting them for 200 volts.
Hi Jim,
Yamaha MI amplifiers are some of the best. You are supposed to use whatever distortion effect box with them. They are also deadly quiet when on. They are some of the highest quality amplifiers on the market. They are also very tough, I had very few failures with them over the years doing service for Yamaha. When a guitar amp has it's own built in distortion it colours everything. If it is a classic like a Marshall (don't care for them) or a Fender (love twins!) then it is expected. Note those are tube amplifiers. Transistor amps distort differently and not musically. So, when you have a transistor amp design you are best to let a distortion circuit voice the sound. Low gain, non-linear solid state amps never sound good compared to a tube amp. That is why Yamaha and a few other great solid state guitar amps are quiet with very low distortion of their own. Obviously you design differently, and that's fine. But the clean solid state amps work best in my area. Messy solid state is like Traynor, don't like those at all.
I am more than familiar with the higher breakdown ratings with base shorted to emitter or driven negative. However it is very clear that because you are getting away with it, you are pushing those parts beyond the specs. You're not the first to do this, but it is hardly good design practice. That and you've got to be careful not to allow the emitter - base junction go into reverse breakdown. That isn't very good for the transistor but normally doesn't cause immediate catastrophic failure. Before you assume the limits of someone's knowledge and experience you should maybe know something about them first. I actually understand how semiconductors work on an atomic level, as many others do. That gives me insight as to why these maximum ratings are important and should be respected. By the way, the ratings you are relying on normally apply when the part is used as a switching transistor. Not typical in linear applications where the outputs and drivers (all the transistors except current protection) are supposed to remain biased on to some degree. What that means is that you do not use emitter - base shorted or reverse biased specs, they do not apply. In your application they might, but I will then assume your amplifier is a solid class "B" design and may certainly sound like it.
The higher actual breakdown values you measure is exactly what is expected. That's called the safety margin and you are not supposed to operate in that region. Your designs have zero safety margin in case you get a high supply voltage, and I'm not sure if you protected against voltage kick-back from transformers or the load. Whatever, you decided to design this way and your failure rate is low enough for you to accept. You also decided that you like the distortion characteristics of your design. That's also fine, but it does not mean Yamaha and others are doing it wrong at all. I will say that Yamaha amplifiers are designed correctly with safety margins respected, using parts designed for the application.
What I have an issue with is your recommendation to others that they should ignore good design practice and "wing it" as you have. Also, these are audio amplifiers and not guitar amplifiers, so your criteria for an amplifier doesn't apply - as you noted above.
The musical instrument industry is one where equipment designed for musicians is typically built as cheaply as you can get away with. Quality is almost completely unimportant as long as it makes noise. Everyone is cutting costs so deeply with marginal designs that the result is pretty much where you are. It is an industry wide thing. The better musicians and recording studios use better equipment, either the "real thing" or quality amplifiers and good effects boxes. Someone trying to survive and coming up will use equipment such as the type you sell. There is a market for that, but those designs are nothing to aspire to. You have commercial success, so I applaud you for that (having run a business for 16 years myself - sold it).
-Chris
Yamaha MI amplifiers are some of the best. You are supposed to use whatever distortion effect box with them. They are also deadly quiet when on. They are some of the highest quality amplifiers on the market. They are also very tough, I had very few failures with them over the years doing service for Yamaha. When a guitar amp has it's own built in distortion it colours everything. If it is a classic like a Marshall (don't care for them) or a Fender (love twins!) then it is expected. Note those are tube amplifiers. Transistor amps distort differently and not musically. So, when you have a transistor amp design you are best to let a distortion circuit voice the sound. Low gain, non-linear solid state amps never sound good compared to a tube amp. That is why Yamaha and a few other great solid state guitar amps are quiet with very low distortion of their own. Obviously you design differently, and that's fine. But the clean solid state amps work best in my area. Messy solid state is like Traynor, don't like those at all.
I am more than familiar with the higher breakdown ratings with base shorted to emitter or driven negative. However it is very clear that because you are getting away with it, you are pushing those parts beyond the specs. You're not the first to do this, but it is hardly good design practice. That and you've got to be careful not to allow the emitter - base junction go into reverse breakdown. That isn't very good for the transistor but normally doesn't cause immediate catastrophic failure. Before you assume the limits of someone's knowledge and experience you should maybe know something about them first. I actually understand how semiconductors work on an atomic level, as many others do. That gives me insight as to why these maximum ratings are important and should be respected. By the way, the ratings you are relying on normally apply when the part is used as a switching transistor. Not typical in linear applications where the outputs and drivers (all the transistors except current protection) are supposed to remain biased on to some degree. What that means is that you do not use emitter - base shorted or reverse biased specs, they do not apply. In your application they might, but I will then assume your amplifier is a solid class "B" design and may certainly sound like it.
The higher actual breakdown values you measure is exactly what is expected. That's called the safety margin and you are not supposed to operate in that region. Your designs have zero safety margin in case you get a high supply voltage, and I'm not sure if you protected against voltage kick-back from transformers or the load. Whatever, you decided to design this way and your failure rate is low enough for you to accept. You also decided that you like the distortion characteristics of your design. That's also fine, but it does not mean Yamaha and others are doing it wrong at all. I will say that Yamaha amplifiers are designed correctly with safety margins respected, using parts designed for the application.
What I have an issue with is your recommendation to others that they should ignore good design practice and "wing it" as you have. Also, these are audio amplifiers and not guitar amplifiers, so your criteria for an amplifier doesn't apply - as you noted above.
The musical instrument industry is one where equipment designed for musicians is typically built as cheaply as you can get away with. Quality is almost completely unimportant as long as it makes noise. Everyone is cutting costs so deeply with marginal designs that the result is pretty much where you are. It is an industry wide thing. The better musicians and recording studios use better equipment, either the "real thing" or quality amplifiers and good effects boxes. Someone trying to survive and coming up will use equipment such as the type you sell. There is a market for that, but those designs are nothing to aspire to. You have commercial success, so I applaud you for that (having run a business for 16 years myself - sold it).
-Chris
Hi wg_ski,
Yes, service people absolutely hated the house number parts system and you could never get the true specs - selected or not. Selecting for higher breakdown voltages was very common. Back then everything was a secret and we resorted to looking at the circuit and using engineering principles to select a commercially available part. Your industry involvement gives you greater insight into that practice. All the semiconductor manufacturers had a house numbering system used on request by equipment manufacturers. The details were always between the semiconductor company and the manufacturer and was confidential. It was never in the best interests for the end user of that equipment.
Typically the manufacturer would run out of, or cease buying the special part so you could not buy the part that was used. That created "orphans" that were officially unserviceable. Thank goodness for higher education in electronics!
-Chris
Yes, service people absolutely hated the house number parts system and you could never get the true specs - selected or not. Selecting for higher breakdown voltages was very common. Back then everything was a secret and we resorted to looking at the circuit and using engineering principles to select a commercially available part. Your industry involvement gives you greater insight into that practice. All the semiconductor manufacturers had a house numbering system used on request by equipment manufacturers. The details were always between the semiconductor company and the manufacturer and was confidential. It was never in the best interests for the end user of that equipment.
Typically the manufacturer would run out of, or cease buying the special part so you could not buy the part that was used. That created "orphans" that were officially unserviceable. Thank goodness for higher education in electronics!
-Chris
I’ve been lurking for some time on the forum. I think this is a cool idea, just to hack some HOT transistors into an audio amp. I’m guessing it has to be in quasi output. Does anybody has a schematic to try as a starting point?
I absolutely loved Elvee’s threads about using voltage regulators in the output stage: https://www.diyaudio.com/community/threads/just-for-fun-the-regulator-chip-amplifier.175457/
and a coil deflection amplifier https://www.diyaudio.com/community/threads/yet-another-funny-chip-amp.280777/
So,I’m saying this is just for fun in the DIY spirit!
I absolutely loved Elvee’s threads about using voltage regulators in the output stage: https://www.diyaudio.com/community/threads/just-for-fun-the-regulator-chip-amplifier.175457/
and a coil deflection amplifier https://www.diyaudio.com/community/threads/yet-another-funny-chip-amp.280777/
So,I’m saying this is just for fun in the DIY spirit!
Hi wg_ski,
Yes, absolutely.
Guess what I did instead of digging up old output transistors? The MJ2119x family is even better, including the plastic packages.
The issues you might run into with these newer transistors comes from stability. They were faster than earlier transistors and recompensating the amplifier (or at least checking stability) became a necessary step. I can tell you from experience that most people didn't even check this.
Yes, absolutely.
Guess what I did instead of digging up old output transistors? The MJ2119x family is even better, including the plastic packages.
The issues you might run into with these newer transistors comes from stability. They were faster than earlier transistors and recompensating the amplifier (or at least checking stability) became a necessary step. I can tell you from experience that most people didn't even check this.
I Also have a bunch of Original Motorola 014-614 Aluminum TO-3's, Does this also a variant of 2n3055? can't find exact datasheet for it.
My experience is that replacing outputs with a somewhat faster one never caused problems if the driver was already sufficiently fast. Using slower outputs could cause oscillation because they would eat into phase margin - non-dominant poles not far enough out. An output transistor can be too fast or too slow for the drivers - but if that doesn’t happen you’re usually ok.
There are many. Here is a single rail design from the late 70's. The JLH69 also comes to mind as a super simple design but I don't know if you could get enough base current to bias the outputs into Class A. You would have to try... which is what you are after doing
in the end you always need one PNP somewhere. Unless you do the JLH or a transformer driven circlo/totem pole. I don’t recommend trying the JLH with HOTs - probably not enough power handling at 20 volts to run class A. AB, probably - but when you get above 40 V, 2N3055’s are actually better.
You probably need to increase the idle current in the VAS stage to make up for the low beta, and strongly advised to use modern sustained-beta Jap driver transistors. You know, the ones that are rapidly disappearing from the market in thru-hole versions. Older vintage 2N types may have too low a beta at half an amp to fully drive HOTs - unless you’re using a triple. And you’ll smoke a 2N3904/6 or MPSA06/56 pair from the current demand.
Last time I used HOTs for audio was in 1980. I used them as the driver in a triple, with a 2N3439/5415 predriver stage, and running three 2N5631’s that I selected for 170 volt operation. Crime of necessity back then when I barely had five bucks to my name and had to salvage everything or simply do without. The other “crime” was the 170 volts, and you can guess where that came from. But it is the only way to build 200 watts per channel for “nothing”, and HOTs we’re often free from discarded TVs. You literally found them on the side of the road, and I’d go and pull all the transistors. Video outputs were perfect for VAS transistors. They’d never go bad, and you get three from a color set.
You probably need to increase the idle current in the VAS stage to make up for the low beta, and strongly advised to use modern sustained-beta Jap driver transistors. You know, the ones that are rapidly disappearing from the market in thru-hole versions. Older vintage 2N types may have too low a beta at half an amp to fully drive HOTs - unless you’re using a triple. And you’ll smoke a 2N3904/6 or MPSA06/56 pair from the current demand.
Last time I used HOTs for audio was in 1980. I used them as the driver in a triple, with a 2N3439/5415 predriver stage, and running three 2N5631’s that I selected for 170 volt operation. Crime of necessity back then when I barely had five bucks to my name and had to salvage everything or simply do without. The other “crime” was the 170 volts, and you can guess where that came from. But it is the only way to build 200 watts per channel for “nothing”, and HOTs we’re often free from discarded TVs. You literally found them on the side of the road, and I’d go and pull all the transistors. Video outputs were perfect for VAS transistors. They’d never go bad, and you get three from a color set.
Well spotted, yes I remember that getting mentioned back in the magazines errors and corrections. I better correct it on the diagram I have because I just don't notice now
I think emitter resistors would be a worthwhile addition because I remember it was very critical on bias setting and could verge toward thermal runaway. I always remember it sounding really good though.
It was the 'Europa' from Practical Wireless:
https://worldradiohistory.com/UK/Practical/Wireless/70s/PW-1978-03.pdf
And as if by magic:
Last edited:
Emitter resistor doing few opposite things . 1. Local feedback,which makes output stage more linear and more stable ,less depending on temperature and etc . 2. Decreases total gain gain before global feedback ,that may increase distortion a little . 3. Reduces output power ,clipping point comes earlier ,so maybe only this effect can be heard at max output or near clipping .
Mooly, I'm surprised that amp sounded good. With a high VAS resistor (12k+2,.2K) and a 1nF compensation cap it would probably exhibit a poor slew rate.
And yes it certainly should have had emitter resistors to stabilise the quiescent current - or a very large heatsink (even then warming up could have increased the current significantly over a period of 30 mins or so.) If you had that amp did you build it and wre there any followup letters complaining about thermal runaway?
And yes it certainly should have had emitter resistors to stabilise the quiescent current - or a very large heatsink (even then warming up could have increased the current significantly over a period of 30 mins or so.) If you had that amp did you build it and wre there any followup letters complaining about thermal runaway?
As for HOT as outputs, I never tried as a Class AB for any significant power because the published data sheets showed very low second breakdown voltages. And low gain as others have pointed out. Though i recall once using a valve output transformer with one running at low current once. Worked, but not much power (class A).