| PicancoNet |
| Anyone has a simple Class B Amplifier ? |
|
|
| Leolabs |
| What for?????:confused: :confused: :confused: |
|
|
| Duo |
I do have some very simple class B ideas.
If you want extreme simplicity, I suggest an opamp and diode biased output stage. This can be accomplished with a minimal amount of components and operate with fair audio quality at low distortion if done properly. |
|
|
| PicancoNet |
I want a simple audio amplifier with discret components.
The input stage can be a op-amp. |
|
|
| amplifierguru |
Here you are then.
Still leaves room for embellishment. RtoR. Good for sub. |
|
|
| AndrewT |
Hi Amplifierguru,
Looks like a CFP using FETs
Should the Vgs +Rs be fed from the Vce of the driver?
or from a resistor in the collector supply of the driver?
How does the VAS determine it's voltage gain? There is no feedback resistor, just a 330p cap. Or is it 4k7/100r? Seems a bit large @ times 47. |
|
|
| Miles Prower |
Anyone has a simple Class B Amplifier ?
Here, they don't come any simpler than this. Here is part of the audio strip for a longwave receiver I designed and built awhile ago. Pay attention to the final that is connected to the active CW filter.
Works just great for this application. Of course, as for audio quality: :bawling:
Then, again, high fidelity is neither needed nor desireable in this application.
Edit: Forgot a decimal point. The emitter resistors are 1.0/1.0W, not 10 ohms. |
|
|
| Duo |
Lol, the PNP transistor in the push-pull output stage on that CW radio amplifier is in backwards.
By changing that amplifier to work with negative global feedback, it would be not bad. |
|
|
| Duo |
Here's a design of great simplicity that I've used in a lot of projects when I needed a quick amplifier.
If you like, I can show you one with even more simplicity that uses three transistors, however, it's nowhere near as nice as the one I've shown here.
(NOTE) The input cap is 2.2µF. I guess if you want a better low roloff you could put 22µF lol. |
|
|
| Miles Prower |
Lol, the PNP transistor in the push-pull output stage on that CW radio amplifier is in backwards.
Yeah, I have to redraw that someday. Anyway, that is not how I built it. It's two complementaries.
By changing that amplifier to work with negative global feedback, it would be not bad.
Wouldn't do much good for what it's used for. The total gain of the entire audio strip is 120db(v). Since this is a TRF design, that's where all the gain comes from. So to add global feedback, and maintain the gain, the open loop gain would have to be even greater, and for a useage where fidelity isn't necessary. Almost all longwave work is CW, and so all it has to do is go: "beep, beep". The DBM also will cause the same sort of distortion on AM that you'd get with SSB. So I designed for stability, which I've got (exception: open input, then you get a 2.5KHz oscillation) and as little tendency for microphonics as possible. It does what it was meant to do.
Edit: Here's a corrected schematic. |
|
|
| mod_evil |
Hi friends, he necessite a simple amplifier because in rio de janeiro, we arent found a diverse components.
Portuguese version;
Olá amigos, ele necessita de um simples amplificador, porque aqui no rio de janeiro, nós não achamos diversos componentes" |
|
|
| amplifierguru |
Hi Andrew T,
I don't follow you're comments. Rest assured everything is in the right place and it will deliver serious power at low THD. the FETS are driven from the low output Z complementary EF and the the predrivers all in a local capacitative local loop to rolloff gain to unity at VHF. The main (global) FBloop is 22K//39p=5K1 to the 1K
at the chip input. All BJT's are BC546/556. DC coupled. |
|
|
| Duo |
Miles Prower: I understand what you're saying. I just mentioned global feedback as in the last amplifier stage there because it would improve the performance for audio listening.
I am an amateur radio operator and I know you don't need/want good fidelity for listening on CW. ;) |
|
|
| Miles Prower |
OK, Duo, I see what you're saying. I thought you meant global feedback around the entire audio strip. As for the PA and driver, usually it would be best to take the feedback for the op-amp from the output. I simply didn't do it. Regardless, you're not going to get outstanding performance with simple, 741-type op-amps anyway. :whazzat:
For whatever the original poster wanted to do, I would endorse that mod. Still, he wanted "simple", and I gave that to him. |
|
|
| Duo |
| lol, with a 741 you'll need all the help you can get :D |
|
|
| AndrewT |
Hi Amplifierguru,
If the output is EF then the fets are transposed.
If its complementary follower pair then the gate should be on the collector of the driver. |
|
|
| amplifierguru |
Hi Andrew T,
It is indeed CFP with gain rolling to unity at VHF and an interposed complementary EF driver! Elegant simplicity. |
|
|
| AndrewT |
Hi amplifierguru,
CFP ok.
Why is the gate on the emitter of the driver?
The driver has a Vce of about 56v-V(100r/2) or about 50v
The FETs will have this 50v across their gate/source connections. |
|
|
| amplifierguru |
Hi Andrew T,
It is early morning in Scotland? Splash some water and look again. It is CFP (with gain) with interposed complementary EF driver (pair BC546/56 as dissipation typ 50mW). The top MOSFET is Pch , bottom Nch.
Cool. Yes it is. many +++ es. Can you see it? |
|
|
| AndrewT |
Hi,
I will have to look again as you suggest. There is something I cannot understand here. |
|
|
| PicancoNet |
| quote: | Originally posted by mod_evil
Hi friends, he necessite a simple amplifier because in rio de janeiro, we arent found a diverse components.
Portuguese version;
Olá amigos, ele necessita de um simples amplificador, porque aqui no rio de janeiro, nós não achamos diversos componentes" |
Thanks for the post. I have the book: "Designing Power Amplifier and Hi Quality Amplifier"
It's grate, but i want a more amplifier base. |
|
|
| Mr Evil |
If you want simple, then they don't come any simpler than this one. One op-amp, two MOSFETs and two resistors; just five components for a complete amp (best to use at least two 100nF supply bypass capacitors too, although it is often possible to get away without them depending on exact component choices and layout). I often throw this one together on a breadboard for testing purposes.
Use lateral MOSFETs for lower Vgs. Yes, crossover distortion will stil be high with no biasing at all, but it's surprising just how low it can be with a good fast op-amp. Certainly good enough for non-critical applications and it can be all but unnoticable at high output powers. For high output power use an op-amp like OPA445, for up to +/-45V supply, or thereabouts.
Minor increases in complexity can yield improvements. For instance, a single resistor from the op-amp output to the MOSFET sources will partially compensate for crossover distortion by allowing the op-amp to supply the first few mA of current to the load. A high output current op-amp helps with that. Alternatively, for slightly more complexity, a few diodes and a couple of resistors can bias the output stage into class AB. |
|
|
| PicancoNet |
| quote: | Originally posted by Mr Evil
If you want simple, then they don't come any simpler than this one. One op-amp, two MOSFETs and two resistors; just five components for a complete amp (best to use at least two 100nF supply bypass capacitors too, although it is often possible to get away without them depending on exact component choices and layout). I often throw this one together on a breadboard for testing purposes.
Use lateral MOSFETs for lower Vgs. Yes, crossover distortion will stil be high with no biasing at all, but it's surprising just how low it can be with a good fast op-amp. Certainly good enough for non-critical applications and it can be all but unnoticable at high output powers. For high output power use an op-amp like OPA445, for up to +/-45V supply, or thereabouts.
Minor increases in complexity can yield improvements. For instance, a single resistor from the op-amp output to the MOSFET sources will partially compensate for crossover distortion by allowing the op-amp to supply the first few mA of current to the load. A high output current op-amp helps with that. Alternatively, for slightly more complexity, a few diodes and a couple of resistors can bias the output stage into class AB. |
So good and simple this amplifier. To decrease the crossover distortion, can i put a 1.2V bias in gate of MOSFETs ?
For the MOSFET can i use the IRF640 ?
I am using the Circuit Maker program to test circuits. is it a good program ? |
|
|
| Mr Evil |
| quote: | Originally posted by PicancoNet
So good and simple this amplifier. To decrease the crossover distortion, can i put a 1.2V bias in gate of MOSFETs ? |
1.2V is about right for lateral MOSFETs. You'll want to set it so that quiescent current anything up to about 100mA. Less gives more efficiency, more gives lower distortion.
| quote: | Originally posted by PicancoNet
For the MOSFET can i use the IRF640 ? |
You could (plus its complement for the P-channel). Being a vertical MOSFET you will need a lot higher bias voltage, plus you need to ensure thermal stability via source resistors and thermally coupled Vbe multiplier. |
|
|
| PicancoNet |
| quote: | Originally posted by Mr Evil
1.2V is about right for lateral MOSFETs. You'll want to set it so that quiescent current anything up to about 100mA. Less gives more efficiency, more gives lower distortion.
You could (plus its complement for the P-channel). Being a vertical MOSFET you will need a lot higher bias voltage, plus you need to ensure thermal stability via source resistors and thermally coupled Vbe multiplier. |
I made this circuit. Is it realy can give 50W at 2V@4ohms in input ? |
|
|
| AndrewT |
Hi,
you just blew up your 741 putting 80v across it.
you have to limit the Vrail to the same voltage as your opamp requirements.
OR use a voltage amplifier between the opamp and the output stage.
That one by Amplifier guru does that, even though I still can't understand it. |
|
|
| PicancoNet |
| quote: | Originally posted by AndrewT
Hi,
you just blew up your 741 putting 80v across it.
you have to limit the Vrail to the same voltage as your opamp requirements.
OR use a voltage amplifier between the opamp and the output stage.
That one by Amplifier guru does that, even though I still can't understand it. |
I put the 741 only to reference. It is a hypotetic Op Amp |
|
|
| Upupa Epops |
| Why are you all still fascinated by some simply ***** ? Wake up guys, now is the time, when useable transistors cost only 5 cents ! :D Simply Zen, simply GC, simply ... Beauty is in complicated connections, only with them you can make right amp. ;) |
|
|
| edl |
What about this?
With a 2*12V, cheap and small psu trafo.
With cheap darlingtons (BDW93/94C for example)
In class AB or class B, or even in class A |
|
|
| Mr Evil |
| quote: | Originally posted by PicancoNet
I made this circuit. Is it realy can give 50W at 2V@4ohms in input ? |
I expect that would work ok.
| quote: | Originally posted by Upupa Epops
Why are you all still fascinated by some simply ***** ? Wake up guys, now is the time, when useable transistors cost only 5 cents ! :D Simply Zen, simply GC, simply ... Beauty is in complicated connections, only with them you can make right amp. ;) |
Simple is beautiful too :) Ok, so most of these amps aren't going to give the greatest quality, but they are useful. Like I said, I find the one I posted is useful when I need a quick test amp. Simple amps are also the best place to start learning. |
|
|
| Dozuki |
Simple CAN also sound good.
see:
http://www.diyaudio.com/forums/show...&threadid=33912
Not much more complicated, very adaptable (I've used it with IRF devices). I also made a version using 2 green LEDs in place of the Vbe multiplier. No Global Feedback, only local. I would add Gate stopper resistors tho. Especially with lateral Mosfets. If the Output seems to be too hot for what its doing, then its oscillating and the gate stoppers will stop this. SK1058/SJ162 seem to need higer values for the stoppers. 220 should be fine for IRF Devices and 470 for SK1529/SJ200 seems OK. The Cap between the Opamp and the Output is probably not necessary, just a little protection.
-D. |
|
|
| Dozuki |
Here's a bare bones Basic AB circuit. R2 sets the bias for the amp and should be set so that 1/2 the supply voltage (~12V) is seen between the + side of C2 and Ground.
No Global Feedback and 2V should send it to clipping. Will post Distortion Spectrum Later.
1V in will swing about 13Vpp. I figgure that is about 20W into 8Ohms with a 1K source impeadance.
C1's polarity is reversed. sorry. |
|
|
| Dozuki |
I did some real measurements on the above amp:
1000 Hz sine input at .66v (1v out of source unloaded)
Unloaded output of amp : 12.7V
Output with 8 Ohm speaker: 10.5V
13 Watts output
Distortion Spectrum in db with 8 Ohm speaker load: |
|
|
| AndrewT |
Hi,
your calculator needs new batteries.
13Vpp is 2.6W into 8R
10.5Vpk is 6.9W into 8R.
Is that a typo 2sj163?
can this amp really swing to within 1.5v of the rails with that drive circuit? |
|
|
| amplifierguru |
Good morning Andrew T,
You have figured out my offering yet? It's audio cryptography - the seed. |
|
|
| AndrewT |
Hi Amplifier Guru,
I planted some currents into the circuit & calculated the standing voltages. It seems to indicate that the right level of voltage appears across the CFP 2nd stage.
But all the CFPs I've seen before had a collector resistor on first stage & used the volts drop across the R to generate the driving voltage for the second stage.
Is yours connected to the emitter because of the voltage amplification? or is that a separate issue?
I still think previous circuits with voltage amplification had a local gain loop ireturning to the emitter but your's does not. |
|
|
| PicancoNet |
| quote: | Originally posted by Dozuki
I did some real measurements on the above amp:
1000 Hz sine input at .66v (1v out of source unloaded)
Unloaded output of amp : 12.7V
Output with 8 Ohm speaker: 10.5V
13 Watts output
Distortion Spectrum in db with 8 Ohm speaker load: |
Hello,
What program did u use to made this measurements ? |
|
|
| richie00boy |
| Quick guess on Amplifierguru's circuit. Is the voltage gain achieved by the first transistor stage, then the second transistor stage is to buffer the the gate drive requirements. The MOSFETs are connected apparantely upside down to instigate an inverting action so that the overall feedback is negative, as the first transistor stage causes an inversion. Also to facilitate rail-to-rail swing. |
|
|
| Dozuki |
| quote: | your calculator needs new batteries.
13Vpp is 2.6W into 8R
10.5Vpk is 6.9W into 8R.
Is that a typo 2sj163?
can this amp really swing to within 1.5v of the rails with that drive circuit?
|
2.6W and 6.9W RMS, yes (although 13 Vpp would be 10.6W RMS). I guess I should have clearified that I was measuring Peak Voltages and Peak Power, not RMS.
I=E/R and W=I*E in which case
10.5e/8r=1.3125i, 1.3125i*10.5e=13.78 Watts Peak to Peak
Yes, that's a typo, should be 2sj162
And since I was measuring Peak to Peak, the amp stays within 5 volts of the positive rail and 5 volts of ground with about .66v input.
| quote: | | What program did u use to made this measurements ? |
An Oscilloscope that shows voltages
and Signal Scope for the distortion specs.
With the circuit built on a breadboard.
Edited Schematic, and input cap is correct: |
|
|
| AndrewT |
Hi,
W=I^2*R = V^2/R = I*V
If V is rms then W is rms equivalent.
If V is peak then W is peak = 2*Wrms when dealing with sine waves.
You use the version of the formula that you have values for. You don't need to do a current calculation in between if you don't have a value.
If the +Vrail is 24Vdc and you can swing to within 5V of either rail when loaded then the pk to pk swing is 24-(2*5)=14 and Vpk=7v.
Close to your 12.5Vpp quoted but not near 10.5Vpk=21Vpp.
10.5Vpp =5.25Vpk Wpk=Vpk^2/R =5.25*5.25/8 = 3.4Wpk = 1.7Wrms.
If you don't agree on any of this come back.
But please define your voltages, currents and power as either rms, pk or pp if there is risk of confusion to the reader. |
|
|
| Dozuki |
Like I said, I should have clearified that I measure from peak to peak. And that's what i said in the second post:
| quote: | And since I was measuring Peak to Peak, the amp stays within 5 volts of the positive rail and 5 volts of ground with about .66v input.
|
As for the swing, with the bias set at 12V at the sources of the Fets, the swing is from 6.75 v to 17.25 v. This is within 5v of positive and 5v of ground (IE, within: 5v to 19v).
I have to admit that it's post like this that make me refrain from posting at all. I was simply demonstrating a 'Simple AB amp', with my own interpretation of it. :xeye: |
|
|
| AndrewT |
Hi,
please keep posting useful info especially to beginners asking for guidance.
I reluctantly had to correct the error otherwise that newbie and any others reading the thread would be totally confused.
If you had avoided the watts conversion the whole problem would have gone away and the relevance of your post would have been just as good. |
|
|
|
