Follower with Gain
I know it is a bad habbit, but I could not resist the temptation again.
I also have not simulated or built. It is the basic circuit without (much) protection, etc. It should have a gain of about 20dB, same output capability as the F4, and no global feedback. I think bandwidth should not be too bad. Back-of-the-paper calculation says over 200kHz.
Please feel free to comment / criticise.
Patrick
I know it is a bad habbit, but I could not resist the temptation again.
I also have not simulated or built. It is the basic circuit without (much) protection, etc. It should have a gain of about 20dB, same output capability as the F4, and no global feedback. I think bandwidth should not be too bad. Back-of-the-paper calculation says over 200kHz.
Please feel free to comment / criticise.
Patrick
Re: Follower with Gain
Nice try! I think an OCL amp should at least have temperature compensation to keep bias and DC offset. Don't you think so?
EUVL said:
Nice try! I think an OCL amp should at least have temperature compensation to keep bias and DC offset. Don't you think so?
Re: Follower with Gain
+42V is a misprint of -42V at minus rail?
Is this procedure of gain estimation right?
2x(20*log(2x((2.2/0.68)x(39/82))))~20dB
Very interesting.
EUVL said:
+42V is a misprint of -42V at minus rail?
Is this procedure of gain estimation right?
2x(20*log(2x((2.2/0.68)x(39/82))))~20dB
Very interesting.
> +42V is a misprint of -42V at minus rail?
Thank you for noting. I thought I should put in some bugs to test attentions (joke). 🙂
Gain is calculated by :
1st stage = 2200/(680+1/Yfs of ZVN3310A at 5mA)
2nd stage = (2x100=39)/(82+1/Yfs of IRF9610 at 100mA)
3rd stage = Rloadxn/(Rloadxn+0.47+1/Yfs of IRFP240 at bias)
where n=no of devices in parallel.
The 100 ohm 5W resistor is shared by both the upper and the lower half, so they get the 2x when calculating gain. One should of course calculate also the lower half and take the average, to be precise. As transconductances do vary from device to device, one should measure AC gain and adjust the value of the 100ohm 5W resistor. My calculations showed that it might want to be between 120 to 135 ohm if you want exactly 20dB gain, at working temperature.
With the values shown, the IRFP240s are biased close to 1A each.
Going further up on voltage for more power is difficult, so in that case better to make it a fully differential amplifier (a la Aleph-X) by converting the first stage MOSFETs to 2 diff pairs. It should then delivery 100W Class A.
I think the circuit is simple enough, even compared to the F4. It will not have the best distortion figures around, but I expect its harmonic spectrum to be mainly second and third. The sex appeal (to me) is that it has no global feedback, something I have yet to try in power amps. (The F4 also has no global feedback.)
As to thermal compensation, I don't think it has been an issue with Aleph type circuits, so I do not expect runaway here as well. The IRFPs has a slight increase in bias with temperature, but they do settle to a steady state value. If you are realy worried, you may also use 2SK1529/2SJ200, which is a bit more stable in that respect.
One last point I want to add. The circuit is nothing new or intelligent. You can find all elements in the first 4 chapters of Hill and Horowitz. But that does not mean that it has to sound bad.
Patrick
Steen, I guess you are building a F4. Fancy having a modular frontend to try this out ? All standard PASS DIY components you have in your drawer. 🙂
Thank you for noting. I thought I should put in some bugs to test attentions (joke). 🙂
Gain is calculated by :
1st stage = 2200/(680+1/Yfs of ZVN3310A at 5mA)
2nd stage = (2x100=39)/(82+1/Yfs of IRF9610 at 100mA)
3rd stage = Rloadxn/(Rloadxn+0.47+1/Yfs of IRFP240 at bias)
where n=no of devices in parallel.
The 100 ohm 5W resistor is shared by both the upper and the lower half, so they get the 2x when calculating gain. One should of course calculate also the lower half and take the average, to be precise. As transconductances do vary from device to device, one should measure AC gain and adjust the value of the 100ohm 5W resistor. My calculations showed that it might want to be between 120 to 135 ohm if you want exactly 20dB gain, at working temperature.
With the values shown, the IRFP240s are biased close to 1A each.
Going further up on voltage for more power is difficult, so in that case better to make it a fully differential amplifier (a la Aleph-X) by converting the first stage MOSFETs to 2 diff pairs. It should then delivery 100W Class A.
I think the circuit is simple enough, even compared to the F4. It will not have the best distortion figures around, but I expect its harmonic spectrum to be mainly second and third. The sex appeal (to me) is that it has no global feedback, something I have yet to try in power amps. (The F4 also has no global feedback.)
As to thermal compensation, I don't think it has been an issue with Aleph type circuits, so I do not expect runaway here as well. The IRFPs has a slight increase in bias with temperature, but they do settle to a steady state value. If you are realy worried, you may also use 2SK1529/2SJ200, which is a bit more stable in that respect.
One last point I want to add. The circuit is nothing new or intelligent. You can find all elements in the first 4 chapters of Hill and Horowitz. But that does not mean that it has to sound bad.
Patrick
Steen, I guess you are building a F4. Fancy having a modular frontend to try this out ? All standard PASS DIY components you have in your drawer. 🙂
EUVL said:
I am still scratching my head on the first 4 chapters of Hill and Horowitz ^^~
Never ending story for me . . .
I do not understand well how the 100 ohms 5W contributes to the gain.
Do you mind if I ask an additional explanation? Thanks.
Ayo . . . babo . . .
The 100 ohms 5W changes the zero potential of the middle point
between the two 39 ohms to above or below zero. I'm sure I
should read the first 4 chapters more #^^#
The 100 ohms 5W changes the zero potential of the middle point
between the two 39 ohms to above or below zero. I'm sure I
should read the first 4 chapters more #^^#
It plays....
I have assembled my f4 clone, and it plays quite nicely, but it doesn't get hot
It seems to be running in class AB with only 3x60mA idle current. I have P1 all bottomed out, but I can only adjust from 2mA to ~60mA.
Is it simply because 4.7k is not enough? Should I increase R9 or have I just placed some components wrongly?
Still, I'm impressed with the details, although it lacks some in the bottom. This will most likely be sorted with some more current and hot transistors. If I don't fix it it will be a nice summer amp!!!! 😀
It's so strange to have a Pass circuit running so cold, I'm getting used to the warm radiation from my modified babbelfish. 😕
I had no problem adjusting the dc-offset.
A picture of the setup:
http://viller.org/audio/2007mar_f4/assembly/P1040594.JPG
Few more pictures of populated boards:
http://viller.org/audio/2007mar_f4/assembly
I have assembled my f4 clone, and it plays quite nicely, but it doesn't get hot
It seems to be running in class AB with only 3x60mA idle current. I have P1 all bottomed out, but I can only adjust from 2mA to ~60mA.
Is it simply because 4.7k is not enough? Should I increase R9 or have I just placed some components wrongly?
Still, I'm impressed with the details, although it lacks some in the bottom. This will most likely be sorted with some more current and hot transistors. If I don't fix it it will be a nice summer amp!!!! 😀
It's so strange to have a Pass circuit running so cold, I'm getting used to the warm radiation from my modified babbelfish. 😕
I had no problem adjusting the dc-offset.
A picture of the setup:
http://viller.org/audio/2007mar_f4/assembly/P1040594.JPG
Few more pictures of populated boards:
http://viller.org/audio/2007mar_f4/assembly
If you measure voltage across TL431 (from D1 to D2), you should be getting something between 7 and 8V. Anything less means you have too little voltage to bias. Then try to change R9 to 5k and P1 to 10k.
> Is it simply because 4.7k is not enough?
I didn't see any 4.7k on the circuit. Which one -- Rxx ??
Patrick
> Is it simply because 4.7k is not enough?
I didn't see any 4.7k on the circuit. Which one -- Rxx ??
Patrick
The voltage diff between gates is 8.03V, but I use transistors with vgs 4.42
So is this logic correct:
Required voltage for 1A through pair: 4.42V+4.42V+2x0.47V = 9.78
I only need approx 0.5A, but still, 8V is way too little. I just need to increase R9 a few kOhm and I should have the beast running.
So is this logic correct:
Required voltage for 1A through pair: 4.42V+4.42V+2x0.47V = 9.78
I only need approx 0.5A, but still, 8V is way too little. I just need to increase R9 a few kOhm and I should have the beast running.
If you need 9.78V (say 10V), you would want to have
R9+T1 = 22k/3
Read datasheet for TL431.
Patrick
R9+T1 = 22k/3
Read datasheet for TL431.
Patrick
EUVL said:
Reading is good. Did you know that you can make basically any circuit with those tl431s, currents source, voltage source, comparator, ohmmeter and phono amp - crazy little thing in my amp.
I'll try lowering it instead then... 😉
There are many many ways to generate the bias voltage.
TL431 is one. If you look at my schematics for Follower with gain a few post earlier, I just use resistors with current sources.
Patrick
TL431 is one. If you look at my schematics for Follower with gain a few post earlier, I just use resistors with current sources.
Patrick
I put in some 5.49k's I had in the drawer and it worked perfectly. Now I can raise the current to almost 2A through each transistor but I have set the voltage across sources resistors to 300mA and then I'll listen some more and see how hot the sink gets.
I must say it sounds great. My preamp is too weak and shitty to show me the full potential of this buffer amp, but I'm one happy builder!!! 😀
Thanks Patrick for helping me with the current and a huge thank you to Papa!!!
I must say it sounds great. My preamp is too weak and shitty to show me the full potential of this buffer amp, but I'm one happy builder!!! 😀
Thanks Patrick for helping me with the current and a huge thank you to Papa!!!
Congrats, cviller. Nice to see a finished and working F4🙂 Really nice job on the boards. You are quite a wizzard. Mine is nearly finished and I am looking forward to try it with different preamps/speakers. Hopefully this weekend.
Steen🙂
Steen🙂
...I am quite surprised! I had my money on your being first Steen! Are you feeling OK?😉
yeah. Congrats, cviller!
mpmarino said:
There is a really good explanation...
Although I have many weak links in my sound-system, I can tell you it sounds great!! Have you started populating your boards?
Yes,
The boards are fully populated with the exception of the pots. There is little time available in the near future so I guess I'll get there someday.
I always insist on having a almost completed chassis before I turn an amp on. I'll let you know how it sounds around ... November maybe!
🙂
The boards are fully populated with the exception of the pots. There is little time available in the near future so I guess I'll get there someday.
I always insist on having a almost completed chassis before I turn an amp on. I'll let you know how it sounds around ... November maybe!
🙂
A few questions
The f4 will be the next amp for me to build. I can follow schematics and get it done but I confess I am no electronics whizz but am interested to learn more.
A few questions please:
Setting bias, what is the voltage range? and should it be positive or can it be negative voltage measured at the outputs with the inputs shorted.
How do you measure amperage output of the amplifier?
Is there a good article on class A amplifier circuits?
The f4 will be the next amp for me to build. I can follow schematics and get it done but I confess I am no electronics whizz but am interested to learn more.
A few questions please:
Setting bias, what is the voltage range? and should it be positive or can it be negative voltage measured at the outputs with the inputs shorted.
How do you measure amperage output of the amplifier?
Is there a good article on class A amplifier circuits?
Unlike mpmarino, I have no high standards about firing up 15 pieces strung together by wire so I guess I'll be done before you (I hope)
My boards are populated and I started drilling the heatsinks, but ran into a problem with lack of any wiggle room, so I'm going to use bars of alu to hold the fets to the heatsink instead of screwing them down individually.
With any luck (and wife permitting), I should be making music by ... say mid-April 😀
so I guess I'll be done before you (I hope)My boards are populated and I started drilling the heatsinks, but ran into a problem with lack of any wiggle room, so I'm going to use bars of alu to hold the fets to the heatsink instead of screwing them down individually.
With any luck (and wife permitting), I should be making music by ... say mid-April 😀