Congratulations Ashade,
you have a very "unusual" and clever way of biasing the output trannies, from datasheet for MJE15030 I see hfe is around 150 at 25 degrees C., maybe quite the same at Vce 55 Volts, with your resistor values of R7 R8 I calculated the bias to be 82,5 mA, looks good, but be sure to use proper heatsinking!
Due to diffrent hfe of MJE 15030 and MJE15031 there will be a "small" DC offset, but don't worry, you have been clever enough to add a capacitor C2 preventing DC offsets into your precious speakers!
Also Q2 and Q3 hfe will probably differ a bit and causinga "slight" DC offset, but thanks to C7 C8 blocking the DC offset caused be the hfe variance you can be safe.
R5 R6 may have too low values, I dont know 5TX795A but if it's equivalent to MPSA43 the hfe will be min 40 at Vce 10 V and Ic 10 mA, which will bias the input stage to around min 3 mA maybe up to 10 mA which is not any very danger biasing current, but maybe you should double R5 R6 value to be on the safe side with Q2 and Q3?
R2 C3 have good combination of the values giving a time constant of 0,2 seconds, thats good news because the gain will fall below 10 Hz making DC on the input to be attenuated a bit, though you have C5 C6 at input but with a much larger timeconstant than R2 C3!
All in all I think it's a "safe" design and suggest you strongly to build it, thats the fastest way of learning and accelerating your skills, and if you have some problems you can always turn back here.
Good luck and tell us about the euphonic sound experiences when you have build your amplifier and listened to it!
Cheers Michael!
you have a very "unusual" and clever way of biasing the output trannies, from datasheet for MJE15030 I see hfe is around 150 at 25 degrees C., maybe quite the same at Vce 55 Volts, with your resistor values of R7 R8 I calculated the bias to be 82,5 mA, looks good, but be sure to use proper heatsinking!
Due to diffrent hfe of MJE 15030 and MJE15031 there will be a "small" DC offset, but don't worry, you have been clever enough to add a capacitor C2 preventing DC offsets into your precious speakers!
Also Q2 and Q3 hfe will probably differ a bit and causinga "slight" DC offset, but thanks to C7 C8 blocking the DC offset caused be the hfe variance you can be safe.
R5 R6 may have too low values, I dont know 5TX795A but if it's equivalent to MPSA43 the hfe will be min 40 at Vce 10 V and Ic 10 mA, which will bias the input stage to around min 3 mA maybe up to 10 mA which is not any very danger biasing current, but maybe you should double R5 R6 value to be on the safe side with Q2 and Q3?
R2 C3 have good combination of the values giving a time constant of 0,2 seconds, thats good news because the gain will fall below 10 Hz making DC on the input to be attenuated a bit, though you have C5 C6 at input but with a much larger timeconstant than R2 C3!
All in all I think it's a "safe" design and suggest you strongly to build it, thats the fastest way of learning and accelerating your skills, and if you have some problems you can always turn back here.
Good luck and tell us about the euphonic sound experiences when you have build your amplifier and listened to it!
Cheers Michael!
I have one thing to start off with...
Why on earth do you want to do this at 110VDC rails????
One off: This is dangerous, second, if anything at all goes wrong, the parts will fry instantly if your power supply can deliver any sort of current.
For something to make a microphone drive a speaker, I would not generally need to use more than five or ten volts swing.
The two stages can easily run on the same power supply rails together, simplifying matters.
If you want a simple biasing scheme for the output that works, just put two diodes from base to base with the anodes up and you'll get a not bad class AB to start with. With this, your resistors could be something around 5 - 10K and will bias nicely.
The input stage is really what I'm wondering about...
For this, I would start with one transistor in your case. It could easily be connected as a common emitter gain stage with a couple of simple calculations. If you want more gain and linearity, you could get a little bit more complex and load the gain stage with a constant current source, but for now you might just do that with a resistor for simplicity.
I'm sure you must know the calculations appropriate for common emitter gain stages...
I hope this might help you into a direction at least.
Why on earth do you want to do this at 110VDC rails????
One off: This is dangerous, second, if anything at all goes wrong, the parts will fry instantly if your power supply can deliver any sort of current.
For something to make a microphone drive a speaker, I would not generally need to use more than five or ten volts swing.
The two stages can easily run on the same power supply rails together, simplifying matters.
If you want a simple biasing scheme for the output that works, just put two diodes from base to base with the anodes up and you'll get a not bad class AB to start with. With this, your resistors could be something around 5 - 10K and will bias nicely.
The input stage is really what I'm wondering about...
For this, I would start with one transistor in your case. It could easily be connected as a common emitter gain stage with a couple of simple calculations. If you want more gain and linearity, you could get a little bit more complex and load the gain stage with a constant current source, but for now you might just do that with a resistor for simplicity.
I'm sure you must know the calculations appropriate for common emitter gain stages...
I hope this might help you into a direction at least.
This would be the best way to kill all of the "outside the box thinking" ashade showed in his design.tlf9999 said:
Ashade, keep on with your approach! Maybe it would be even better if you go in the lab at the university and try to build these designs, and measure them. Use cheap parts, what you have at hand. If the circuit topology is robust enough, it will work without selected expensive parts.....
A simulation before prototyping might be still good to show you the biggest faults in the circuit. But to understand things completely, build it.
Btw, your input stage has some resemblance to the Zen V5, look here:
http://www.passdiy.com/projects/zenv5-1.htm
Since you are using BJT's and no feedback loop, the damping factor will be rather low, due to the limited current gain of only two stages.
And check for PSRR, it could be critical.
Success, Tino
Ultima Thule said:...you have a very "unusual" and clever way of biasing the output trannies, from datasheet for MJE15030 I see hfe is around 150 at 25 degrees C., maybe quite the same at Vce 55 Volts, with your resistor values of R7 R8 I calculated the bias to be 82,5 mA, looks good, but be sure to use proper heatsinking!
Due to diffrent hfe of MJE 15030 and MJE15031 there will be a "small" DC offset, but don't worry, you have been clever enough to add a capacitor C2 preventing DC offsets into your precious speakers!
It is so unusual and clever that it was abandoned sometime in the late 40s for not working very well, considering that production tolerances for BJT hfe are about +-50% or more (and are not easily improved).
Pop quiz: what will the DC offset be on the midpoint of two series connected current sources (i.e. high impedance!) if they are not EXACTLY the same? This is basic stuff...
zinsula said:
FAr be it from me to kill someone's 'outside the box' thinking. The thing with it is, usually people do that once they know how to cover the basic stuff (i.e. have it work AT ALL), and possibly, once they figured out what 'in the box' is and why they don't want it.
At the end of the day, once you have to select from an in-the-box design that works and an out-of-the-box design that does not, what will you do?
I find discussing 'selected expensive parts', damping factor and PSRR of something that will only 'work' inside a simulator (and even that is a question)... well, strange, to put it mildly. In the real world, that design has a fatal flaw that prevents if from working AT ALL. The reson being, a simulator is just that, a simulator: a means of simulating an IMPERFECT model of the real world. You have to know what it does in order for the results to have any meaning.
ashade said:
The temp range is the least of your worries. If you want to learn read carefully the advice given here. For starters, in real world, this WILL NOT WORK because the Q2, Q3 collectors witll either be at +15V or -15V in rest. There will be very little distorted output, unuseable. If you want to be serious try just simulating the first stage in isolation and try to get an undistorted sinewave out of Q2, Q3 of several volts. And remember, if you have to tweak the biasing resistors in the sim to get it right, you KNOW it will not work in real world.
On the lf response, of course it falls from 300Hz. Your output cap is to small. But since you have symmetrical supplies, get rid of that output cap and bias the output at zero volts. That gives you lf response as low as you want. Don't worry if initially you have some output offset, that can be fixed later easily.
But first take the first step, that input stage is no good.
Jan Didden
ilimzn said:
It is so unusual and clever that it was abandoned sometime in the late 40s for not working very well, considering that production tolerances for BJT hfe are about +-50% or more (and are not easily improved).Originally posted by Ultima Thule...
you have a very "unusual" and clever way of biasing the output trannies, from datasheet for MJE15030 I see hfe is around 150 at 25 degrees C., maybe quite the same at Vce 55 Volts, with your resistor values of R7 R8 I calculated the bias to be 82,5 mA, looks good, but be sure to use proper heatsinking!
Due to diffrent hfe of MJE 15030 and MJE15031 there will be a "small" DC offset, but don't worry, you have been clever enough to add a capacitor C2 preventing DC offsets into your precious speakers!
Pop quiz: what will the DC offset be on the midpoint of two series connected current sources (i.e. high impedance!) if they are not EXACTLY the same? This is basic stuff...
Totaly drained out of sense of humor?
janneman said:
Jan,
in reality it may perhaps be much closer to zero than + or - 15 Volts, or don't you think so?
Here's a clue, how much is the base current for Q3 when it's fully on... and how much is the base current for Q2 in the same situation...
Now, you guys should not take away the "opportunity" from the guy entering a learning curve, humor and science is the guiding principle here!
Cheers Michael
EDIT: PS. I agree with Duo!
110 V is a bit too high for a starter, go with + and - 25 Volts to begin with, and rise the voltage afterwards when you find your design stable enough not to burn up!
Ultima Thule said:[snip]Now, you guys should not take away the "opportunity" from the guy entering a learning curve, humor and science is the guiding principle here!
Cheers Michael
[snip]
Absolutely! However, most of us are not professional teachers and it is difficult to find the right "tone" at the very start of the learning curve (no intention to insult). Time, and therefore patience, are always in short supply on this forum it seems (me included).
Especially for newcomers it is very hard to discern the usefull advice from the nonsense; there is no substitute for personal study at some level. I'm sure a 4th year EE is very aware of that
.Jan Didden
Ultima Thule said:
[snip]Jan,
in reality it may perhaps be much closer to zero than + or - 15 Volts, or don't you think so?
Here's a clue, how much is the base current for Q3 when it's fully on... and how much is the base current for Q2 in the same situation...
Michael,
Maybe it is, maybe not. Who knows? Look at the resistor values, they have clearly been tweaked to get it right in the sim. What do you estimate the percentage that there is usefull output swing from the Q2, Q3 collectors with real life transistors? (Hint: it is less than your shoe size. Much less. How's that for humor?
)The significant thing here is that this approach to design is fundamentally wrong. The goal should be to design a circuit that doesn't need any tweaking, that will work with +/- 20% or more resistor variations and +/- 100% transistor gain variations. Then you have a fighting chance that it will work when you built it.
Jan Didden
Originally posted by Ultima Thule
Totaly drained out of sense of humor?
No, not at all, but considering the level of knowledge demonstrated by the thread starter, I would avoid the not-obviously-sarcastic tone as there is the possibility to take the humour literally - which is what I have chosen to do in my reply, say, to demonstrate the effect
BTW the first stage is so-so considering the bias resistors are at least a load for the collectors, so the output has a chance of not being stuck to eitehr rail. The output stage is another matter - this one WILL be stuck to one rail (keep in mind the output is cap coupled, therefore no DC load).
Also, having +-55V rails on a follower stage fed by a stage which cannot under any circumstances produce more than +-15V is not really logical, but in principle, it's not a problem, unless you mind a whole lot of heat and taunting secondary breakdown. Still, that's the EASY part to fix.
ilimzn said:
No, not at all, but considering the level of knowledge demonstrated by the thread starter, I would avoid the not-obviously-sarcastic tone as there is the possibility to take the humour literally - which is what I have chosen to do in my reply, say, to demonstrate the effect
BTW the first stage is so-so considering the bias resistors are at least a load for the collectors, so the output has a chance of not being stuck to eitehr rail. The output stage is another matter - this one WILL be stuck to one rail (keep in mind the output is cap coupled, therefore no DC load).
Also, having +-55V rails on a follower stage fed by a stage which cannot under any circumstances produce more than +-15V is not really logical, but in principle, it's not a problem, unless you mind a whole lot of heat and taunting secondary breakdown. Still, that's the EASY part to fix.
Sarcastic tone... nooo, it's getting sarcastic when one suggest somebody to build something that will for sure not work or burn, but I'm not so sure about it in this particular case, giving the doomsday profhecy as some have already in this post is sarcastic!
About the output stage, well, if one tranny "wins", it will not have any Vbc anymore biasing itself through the base resistor, how can it be stuck to either voltage rail?
About the voltage supply level for the output stage, I admit overlooked it, anything above +-15 V will be a waste, scores for you! ...but maybe we need some headroom due to above mentioned, the voltage offest at output so the output signal won't clipp too early at only negative or positive rail?
Cheers Michael
janneman said:
Michael,
Maybe it is, maybe not. Who knows? Look at the resistor values, they have clearly been tweaked to get it right in the sim. What do you estimate the percentage that there is usefull output swing from the Q2, Q3 collectors with real life transistors? (Hint: it is less than your shoe size. Much less. How's that for humor?
Haha,
thats OK Jan, read my previous post!
Besides I think the guy want lot of AC gain too!
Cheers Michael
Ultima Thule said:About the output stage, well, if one tranny "wins", it will not have any Vbc anymore biasing itself through the base resistor, how can it be stuck to either voltage rail?
Yes, you are right of course. Still, just using 1:2 hfe will give you a similar DC offset, not very good. The solution to this would be to use DC coupling between both stages and reduce two biassing problems to one. Sometimes a little feedback helps
Well, yes, if you put it that way
though that DC offset is not THAT difficult to fix into more sane values.Maybe we should have a contest on best amplifier design with 4 transistors
Ashade,
I strongly suggest you have a look at the amplifier designs from Nelson Pass. He is definitely thinking outside the box and he likes to keep his circuits as simple as possible.
Just follow the Zen series and read his publications about single ended class A and cascodes. You will learn a lot from it, and you can simulate (or build, takes about the same time) the circuits while you read. Later, have a look at the Aleph amplifiers.
The best thing, all this is available for free (no need to buy expensive books right now, although absolutely recommended when you want to go deeper).
Check out http://www.passdiy.com
I strongly suggest you have a look at the amplifier designs from Nelson Pass. He is definitely thinking outside the box and he likes to keep his circuits as simple as possible.
Just follow the Zen series and read his publications about single ended class A and cascodes. You will learn a lot from it, and you can simulate (or build, takes about the same time) the circuits while you read. Later, have a look at the Aleph amplifiers.
The best thing, all this is available for free (no need to buy expensive books right now, although absolutely recommended when you want to go deeper).
Check out http://www.passdiy.com
OliverD said:
You are right, and that is precisely why this is about the worst advice to a newbee. You can't think outside the box, or follow someone who does, unless you know your way inside the box. Suppose you are completely new to lf amp design, do you think Pass' designs make sense? of course not!
Maybe we should stop trying to impress the asker and instead give the best advice FOR HIM.
Jan Didden
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.