Biasing amplifier transistors

Biasing the pair of power transistors (NPN + PNP) in the output amplifier of my function generator (TR-0458/B). The previous owner probably changed the transistors, but they also needed to be re-biased, as their base to emitter voltage drop was higher. They were under-biased, causing signal distortion. The same can apply to bipolar discrete transistor audio amplifiers.
Please support my channel on Patreon:
/ diodegonewild
Instagram:
/ savage_danyk

[@zyghom]

apart from making the electronics very nice and easy to listen and learn, you have also HUGE KNOWLEDGE - very rare today ;-)

[@BritishEngineer]

The truth about huge knowledge in this context, is qualification levels. I’d say he has at least a bachelors in engineering, possibly a masters. In other contexts your statement is more correct though..

[@zyghom]

@@BritishEngineer I know engineers who finished uni about electronics and they passed exams (somehow) but today they remember nothing. I don't care what is HIS education level by means of school - I was talking about his knowledge (hint: knowledge not necessarily means good school finished). That is my opinion of course.

[@TazkoHoZabit]

@@BritishEngineer @zyghom ... guys, the most fascinating thing is, that DiodeGoneWild (Dan or danyk) has no bachelor or master in electrotechnic engineering, he has also no standard school education in electronics - he graduated high school (gymnasium) and then studied economics and management at the faculty, ... everything he knows about electronics he learned by himself ☺

[@jasonvoorhees3282]

@@TazkoHoZabitTrue, I have read a comment made by him about that. He is an actual genius.

[@robertatpierpontbeach]

Excellent description of the transistor biasing process; even your cat's comment only lamented about the number of the little discrete devices😂

[@TheEmbeddedHobbyist]

Nice, i do love dc coupled amplifiers. the fact that the fault affects all part of the amp and its impossible to break it into parts without affecting the dc conditions. Having the circuit diagram is a great help.

[@thisisyourcaptainspeaking2259]

I judge the quality of design based on capacitor count.

[@TheEmbeddedHobbyist]

@@thisisyourcaptainspeaking2259 but is 1 cap a higher quality circuit or 10 caps better. As there are engineers out there who just load a shotgun with caps and give the PCB both barrels. 🙂

[@thisisyourcaptainspeaking2259]

@@TheEmbeddedHobbyist Have a closer look, you'll notice things like better performance metrics, construction, materials, durability, failsafes, surge protection, on and on. Generally speaking, every electrolytic capacitor that exists in a device represents a time-bomb waiting to take the circuit down. Careful engineering and many times, extra cost can eliminate the use of "crutch" capacitors, etc.

[@TheEmbeddedHobbyist]

@@thisisyourcaptainspeaking2259 My point exactly! but it does make it harder to fault find as each stage relies on the one before it and the one after it. break the loop and the DC conditions go out the window. all components should only be there if needed. When working out MTBF's the less parts the better.

[@thisisyourcaptainspeaking2259]

@@TheEmbeddedHobbyist Yes they're gonna try cutting costs every way they can, for instance bean counters won't allow using a 50v capacitor where a 6.3v capacitor can be used to smooth a 5v potential. Same reason we never find a fiberglass laminated pcb anymore, where a recycled popsicle stick will do, Remember the hey-day of video cassette recorders? The better VCR machines used to have fiberglass pcb's with about half the number of electrolytics, the low quality machine mechanisms were stamped sheet metal, wonky plastic idlers and wheels while the better machines employed cast aluminum tables and noticably better quality mechanisms. As far as ac and dc coupling go, quality lab equipment are designed and constructed to a higher standard that far outperforms by using dc coupling as opposed to ac coupling wherever possible. Largely depends on the intended design parameters but I generally correlate value based on capacitor count as a necessary evil, with high counts being indicative of low performance characteristics. Yes, I loath high capacitor count circuits, lol. True, it can be an easier task of troubleshooting, just replace all the caps in a section and it's back in business, also it's easier to decouple the various stages without upsetting dc biases, during troubleshooting. I've replaced way more leaky capacitors than any other type of component, usually with a higher voltage rating electrolytic if possible. There can be occasional drawbacks to be aware of when using this strategy but in general it works fine, in my experience.

[@rolfts5762]

Good educational video. Thanks for sharing your quality knowledge Dan. (..yes if have time, as you said it would be interesting to see if any optimizing could be done on leading-circuitry. The triangle-conversion etc)

[@Alchemetica]

Some more education towards me understanding biasing. Thank you,

[@jim9930]

Danny, Look up Texas Instruments AN-263 Sine generation techniques application note. For 'diode breakpoint' sine approximation, figure 12 is similar to the design in your function generator. SNOA665 1999 revised 2013 You should be able to reduce the distortion by an order of magnitude (

[@stevebahr7782]

You're videos are amazing. Thank you!

[@TheTemporalAnomaly]

Very good explanation of this effect 15:50 Your videos and the material on your website make a great online reference source.

[@erikziak1249]

Just a couple days ago, my good old Onkyo TX-300 has blown its STK4893 amplifier and sent -44,2V DC to the left speaker. While the tweeter and mid-range were protected against DC by series capacitors in the passive cross-over, the woofer got completely "sucked in" and was damaged. I already replaced it, as I had a spare, it just needed its dust cap replaced by the one from the blown woofer and a paint job to look nice. I was quite surprised that the relay connecting the speakers to the output of the amplifier works, even when there is a huge DC voltage. Maybe it checks only for positive DC? I do not know. The easy way to fix the TX-300 would be to replace the STK4893, but I am not sure if there is not something different blown and even if not, the STK4893 is not being produced for a long time, only cheap Chinese copies of questionable quality (parameters). The best way to repair would be to install discrete components, but that is way beyond my knowledge. I studied mechanical engineering, not electronics. And even my studies are two decades in the past. I mostly forgot everything. Nobody wants to fix the old Onkyo TX-300 anymore, as it is too much work and it is generally cheaper and better to buy something new. And that is true. But I want my old amp for nostalgic reasons. It is like fixing an old car for occasional driving. The money invested is not worth the driving quality nor safety or efficiency. Using vintage things - and not just cars or amplifiers - is not about rationality.

[@2pist]

I get it. I'm the old guy driving old cars in your example.

[@MrGuano11]

Wow this goes over my head. Is there a book where discrete audio amplifiers are explained for dummys?

[@WolfgangMahringer]

If you are interested in transistor circuit design, I'd recommend Horowitz/Hill's "Art of Elelectronics". It's a lot of pretty detailed stuff to read, but worth having it around.

[@piconano]

Those 2N3055 transistors are hard to come by. Bought some from 3 different Chinese sellers. All fake. Opened one up to compare to a real die. the die was 1/2 the size of the original!

[@mrnmrn1]

These 3055s in the video are Hungarian from Mikroelektronikai Vállalat. They have a *huge* die, with massive connection to the leads, not even bonding wires, rather bonding rails. And a massive copper heat speader underneath the die. It is following the original RCA design from the '60s. Tungsram also made the same version. Checzoslovakian TESLA KD503 was an alternative for it, it is in a full copper TO-3 package, similarly big die. Looks the same as Motorola's aluminium TO-3 package, it's just copper (the cap is aluminium though).

[@FelinoAnticomunista]

Maybe you can got some real from utsource

[@SandWire]

@@mrnmrn1 Original and cheap Tesla KD503 is also not easy to get these days. There is cheap modern alternative TIP35C :)

[@mjouwbuis]

@@SandWire which might be forged as well... There's also some current models that can be got directly from ON. I think MJ150xx something.

[@LozzoAmiga]

Best bit at 2:30 psspsspss

[@liam3284]

Ah, analog electronics, feels at home

[@TESLA_KF507]

Myslím, že ten průběh by se mohl dát ještě trochu vylepšit, přestože to asi nikdy nebude jako z například RC generátoru. Hnát tento generátor do dokonalosti v harmonickém průběhu však asi nemá smysl.

[@nsfeliz7825]

nope nope i would buy an ic tamplifier. no drama. like tqhe tda xxx series of ic amps.

[@ProdigalPorcupine]

Will they have flat response to 1MHz?

[@TheDefpom]

@10:30 it probably had germanium transistors in it originally rather than silicon

[@DiodeGoneWild]

I also thought this, but for most germanium ones the bias would be too high. Maybe just a different type of silicon ones... Can't tell for sure, because almost all transistor types are different than in the schematic, despite everything else including the resistor values in the schematic is the same.

[@mrnmrn1]

@@DiodeGoneWild Nope, the originals are 2N2219A - 2N2905A pair. The BC313 installed here is quite different. The output should be replaced to a proper complementary pair, preferably to a selected pair. The 2219 / 2905 has 300 / 200 MHz transition frequency, the BC313 is 50MHz.

[@wsc112782]

Great analysis! For the triangle to sine converter, it would be interesting to compare it to a JFET sine shaper, they're popular in the synth community. Try the J201 JFET if you can find them. Supposed to have amazing performance as a sine shaper.

[@cezarcatalin1406]

There’s also a “self-biasing” configuration that works by measuring the current passing through both transistors (the sum current minus the differential current divided by two) aka the minimum between source and sink currents. The measurement is, of course, purely analog and uses makeshift differential amplifiers and shunt resistors but the effect is that the power transistors are dynamically auto-biased to a fixed current and it’s pretty much independent of temperature so no thermal runaway is possible since if the common current increases, the self-biasing circuit applies negative feedback and tries to restore the common current to the previously set value given by a constant current source/sink/diode.

[@boomer9900]

My first repair on this output topology was a stereo receiver. One of the output transistor shorted and the thing got so hot that it melted the record player turntable rubber pad. It looked more like a potato chip. On this repair I learned about bias voltage and diode compensation on the heatsinks. The burned components were replaced and I used this receiver as a dance floor amplifier for my high school graduation ceremonies. I balanced the bias by feeling the output transistor heat level, given I had only rudimentary test equipment.

[@SaltyPuglord]

For anyone who's curious: The bias transistor is often called a "Vbe Multiplier" or "rubber diode". There's a Wikipedia page on it, but I can't give the URL here since YT silently deletes any comment containing a URL! :P Google search the above terms, you'll find it. Here's a quote: "It is commonly used in the biasing of push-pull output stages of amplifiers, where one benefit is thermal compensation: The temperature-dependent variations in the multiplier's VBE, approximately -2.2 mV/°C, can be made to match variations occurring in the VBE of the power transistors by mounting to the same heat sink."

[@ohmware2020]

i idolize you, you explain nicely, your very knowledgeable in electronics

[@DiodeGoneWild]

Thanks, but no person should be ever idolized ;)

[@thisisyourcaptainspeaking2259]

The negative feedback signal into the 1st stage is smoothed by the +15v supply, this is low pass feedback.

[@JohnnyX50]

Dekuji :D Hope the thanks help you a little :) Really appreciate your hard work :)

[@constantinjitariu1439]

Hi there is there any way i can contact you i want to ask for help with a strange amplifier circut and i dont have anyone to help me please help if you see the coment nice video btw🙂

[@MrEdwinHubble]

Wow you got some nice resistors and other components from Tesla. Where did you manage to get your hands on such good stuff?

[@celsoneves2368]

Bueno

[@tajtrlik1111]

Veľmi pekne vysvetlená teória o koncovom tranzistorovom stupni, najmä to o tom predpätí (bias) tranzistorov a ako to musí bez predpätia ten OZ v spätnej väzbe dorovnávať. Určite uvítam viac aj takýchto teoretickejších videí a budem sa tešiť aj na video o náprave toho sínusového (harmonického) priebehu v tvarovači, ak teda toto video natočíš.

[@Gengh13]

I would love to see a video adjusting/improving that sinewave.

[@amersa89]

me after 16min. 😵🥴😵🥴

[@lunedemai2]

That's a good one. Thank you!

[@antibrevity]

Missing screws are bizarre to me, because I find it strange that someone would go to the trouble of diagnosing and replacing components, yet do so with shoddy workmanship. A lazy person would throw the unit away or attempt to pawn it to an unsuspecting buyer, so who are these people that are willing to spend hours diagnosing and repairing a circuit, but not spend the extra 5 minutes it would take to do a good job? I can only assume that they are getting paid to repair the unit and thus need to get it done as fast as possible. My culture uses the word "professional" to mean high quality or highly skilled, but it's the professionals that I trust the least.

[@SeanBZA]

Also might have had it on a shelf for a month, waiting for the spares, and the screws were not in a bag in the unit, so got lost during that time, used elsewhere, or the cat got them.

[@saruthapa2732]

lovely cat❤❤

[@Speeder84XL]

Very interesting! I have never built an amplifier like that. For signals, I mostly use operational amplifiers (but this gives a good insight how those works - since they also consists of similar circuits integrated in an IC). For higher power signals (from a few watts and up) I have tried building a small inverter type one (class D) that was just a few watts - which worked quite well (just 0.3 % THD - but that's close to maximum output level and goes down at lower levels, since it doesn't suffer from crossover distortion such as type A/B linear amplifiers do) and had ok efficiency of about 75-80% at high output signal levels (which also has room for improvement with better gate drives and lower circuit "idle" power consumption). I have no videos of that - but earlier this summer I made one of a "singing arc", that works very similar - the difference is that instead of modulate both "positive" and "negative" pulses equally, the positive pulses will increase while the negative decrease - and the opposite. Silence will be close to 50% (actually 50% minus the "dead time" always needed for the transistor bridge), with both positive and negative pulses having the same length. An LC filter is then put on the output to block the high frequency AC output, while letting the slower "offset" voltage changes caused by the duty cycle modulation, thru - resulting in a AC voltage that looks like the audio signal between bridge terminals. For that reason, the pulse frequency should be at least 10 times higher than the highest parts of the audio output - to give room for the filter to filter out most of the high frequency "carrier wave", while not affect the audio significantly - so minimum 200 kHz to play back audio up to 20 kHz (but that also means, an amplifier that's only built to drive bass and midrange speakers can have lower pulse frequency and use a bigger inductor and/or capacitor in the LC filter - which will increase efficiency right where most of the power is usually needed for a sound system) I will build a big one as a later project when I need an amplifier to drive some speakers and put up a video (maybe at the end of this year).

[@edgeeffect]

It's always really useful when I'm half way through learning about something and then you make a video about it. :)

[@SS-mj2mq]

Nice ty for sharing ❤ I really don't know why I'm telling you this but I have built a NPN and PNP complimentary differential pair it is very complicated the voltages have to be precise I used fixed voltages to set the voltage for each side of the collector of the NPN and PNP using both power rails +VCC and - VEE if the voltage's are not precise on each side of the collectors it will not work at all.. Now I'm going to ask you a question what was the purpose behind me making this differential pair? I'll give you a clue the PNP transistor is used for feedback and the NPN transistor is used for the input.... By the way this is a patent design by me. I should make this clear 1 PNP transistor and 1 NPN transistor only.... And you won't find the answer in any handbook or online.... You have to use your brain..❤Anyone can answer

[@d.k.9406]

Danke! Thanks! Good work. Extra thank you for all the extras Ir image Synchronized shematics and real part placement And showing on osci And i liiiike your drawings! Always helpful.. not showing only straight the one fault and repair it.. as always ty for showing all sorts of fault causes or mistakes which can be done from people who dont know the right way, how to fix the problem the right way I'm highly interested in the detailed explaination How the sinewave is generated.

[@atmel9077]

In an electronics course I took at the university, they say that for a proper class-AB biasing, the voltage drop on the resistors should be around uT=26 mV at 25°C

[@jozsiolah1435]

Same kind of circuit is inside the IC of transformer power supplies, variable power supplies for household devices. They are a little bit more expensive than regular transformer supplies.

[@Pirelli.]

Another great vid dear DGW! Thanks! ☺

[@LawpickingLocksmith]

Very good engineering with very pretty bonus cat!

[@2pist]

I understand constant current now! Why doesn't everybody teach like you do? Thank you for sharing.

[@andymouse]

Nice ! would enjoy watching you fix the ugly sine !.....cheers.

[@TheEmbeddedHobbyist]

that comment was a bit cheesy Mr squeak. 🙂

[@andymouse]

@@TheEmbeddedHobbyist LOL ! Squeak !

[@Alexelectricalengineering]

Even more transistors, lol. Thumbs up 😎👍

[@ahurtadon]

Pero el condensador ese grande , no está montado al revés?

[@LousyPainter]

Very well explained! Thanks.

[@waynegram8907]

DIODEGONEWILD< Why do transistors need to be biased? capacitors, resistors, inductors don't need to be biased. Transistors need to be biased because of the Emitter, base and Collector junction regions?

[@lumifaza]

capacitors, resistors and inductors are passive components and are mostly linear while transistors are semiconductors and are not

[@waynegram8907]

@@lumifaza so you have to bias transistors and fets to make them Linear?

[@lumifaza]

@@waynegram8907 not in all applications, but in case of audio/signal amplifiers yes, otherwise you'll get cross-over distortion which @diodegonewild shows at 13:59

[@noneofyerbeeswax8194]

13:59

[@SojournerDidimus]

Thanks for the nice videos!

[@adrian_sp6def]

I have same oscilloscope❤

[@lmwlmw4468]

Great video.

[@Swenser]

Thanks for explaining, but im too stupid to understand no matter how hard i try. My question is why not use 1 transistor. Small signal at base that is amplified by 1 transistor into a speaker. Done.

[@agasobanuyekezachannel464]

I'm your stupid mate ! But I think it's not stupid what I see is that we need explanation from starting till getting this level

[@stevebollinger3463]

On the schematic in that big grid of resistors on the right why are there dots and lines in the resistor symbols? What does this signify?

[@DiodeGoneWild]

Not sure, but some regions in the old days were using lines inside the resistor symbol to indicate its power rating. This probably never got universally standardized. A perpendicular line indicates a 1W resistor and 2 such lines would indicate a 2W resistor, a V (roman 5) a 5W resistor, a line along means 0.5W, a diagonal line indicates a 0.25W resistor, 2 diagonal lines a 0.125W resistor. I really don't know what the dot is. Maybe an alternative symbol for a 0.5W resistor or some indication of a higher accuracy version.

[@sharedknowledge6640]

You have the circuit theory mostly correct but you don’t understand the levels of distortion that matter in an audio amplifier. You can’t measure them with an oscilloscope let alone your crappy 8 bit Chinese Owon scope. You need a measurement dynamic range in excess of 90db of signal versus distortion components. You need some thing like a $20000 dollar Audio Precision audio analyzer to properly set bias for minimum distortion versus amplifier efficiency. This video is very misleading and displays your own lack of knowledge about audio amplifier performance. Stick to subjects you actually understand. I just unsubscribed.

[@DiodeGoneWild]

... and solid gold cables :) sorry, this is not an audiophile channel ;)

[@2pist]

Most of us "subscribers" wouldn't or couldn't spend so much on equipment so this video is perfect for the real world. Lock the door on your way out please. 😅

[@mjouwbuis]

I don't think we will be missing you much. You seem to only understand small part of the theory yourself.