Good day gentlemen....
I already have this things, 220vac trafo with 100v single winding secondary (isolated) at 1000VA used as isolation trafo, 30pcs 1200/250v caps as filters in parallel, 35A br rect., 2-10,000uf at output coupling and several N-Ch power mosfets as outputs. Now all i need is a good single rail amp working at around 140vdc.
To the designers and members out there can you help me out and give me a working schematics please!....I just want to turn this things to usable amp rather than sitting on storage as junks..
I already have this things, 220vac trafo with 100v single winding secondary (isolated) at 1000VA used as isolation trafo, 30pcs 1200/250v caps as filters in parallel, 35A br rect., 2-10,000uf at output coupling and several N-Ch power mosfets as outputs. Now all i need is a good single rail amp working at around 140vdc.
To the designers and members out there can you help me out and give me a working schematics please!....I just want to turn this things to usable amp rather than sitting on storage as junks..
you need to add DC blocking at both the input and the output to allow the quiescent voltage in the amplifier to sit @~70V above Zero Volts.
The input cap is easily solved a 100V 1uF feeding Rin~100k will let through all the audio signal you want/need.
The output cap is different. It must be able to pass speaker currents and block voltages <=70V (+mains tolerance)
A 10mF 100V cap will allow >=20Hz signals to flow to 8ohm speaker with virtually no attenuation, and progressive roll of from 20Hz down to 2Hz.
If this 10mF is connected to a 4ohms speaker then the filtering effect of that DC blocking cap moves up an octave to passing >=40Hz and F-3dB~4Hz.
The input cap is easily solved a 100V 1uF feeding Rin~100k will let through all the audio signal you want/need.
The output cap is different. It must be able to pass speaker currents and block voltages <=70V (+mains tolerance)
A 10mF 100V cap will allow >=20Hz signals to flow to 8ohm speaker with virtually no attenuation, and progressive roll of from 20Hz down to 2Hz.
If this 10mF is connected to a 4ohms speaker then the filtering effect of that DC blocking cap moves up an octave to passing >=40Hz and F-3dB~4Hz.
BUT, these amps still can be made to sound very, very good. And they have intrinsic speaker protection, no small consideration.
Hugh
Hugh
Hi andrew thanks for the info, at what capacitance do i need to compensate even at 4 ohms load.
to CBS 240.
Yes it can be by having it at full bridge it double the rails at +-140v and 1 thing more is the filter caps must be very large.
Hi aksa
...thanks for dropping in.. you're an expert here can you recommend a good design schematics? I am not after a low THD as long as it will work and stable, anyway this will be use for a disco or videoke purposes.
to CBS 240.
Yes it can be by having it at full bridge it double the rails at +-140v and 1 thing more is the filter caps must be very large.
Hi aksa
...thanks for dropping in.. you're an expert here can you recommend a good design schematics? I am not after a low THD as long as it will work and stable, anyway this will be use for a disco or videoke purposes.
each time you halve the load impedance, you double the coupling/blocking capacitor value to maintain the same RC.
Jun,
Any stable, good quality two rail design, such as Quasi's design on this forum, can very simply be adapted to single rail operation by adding an output coupling cap and creating a bias voltage around 70V by using a simple voltage divider. You can even use the same pcb with minimal additional components. Take a look at the Quasi thread and ask around; lots of people there have built this amp with outstanding results and could help you.
AndrewT's recommendation for 10,000uF coupler rated to aboutg 100VW is fine, don't worry too much about low Z loads, that will be fine.
Hope this helps,
Hugh
Any stable, good quality two rail design, such as Quasi's design on this forum, can very simply be adapted to single rail operation by adding an output coupling cap and creating a bias voltage around 70V by using a simple voltage divider. You can even use the same pcb with minimal additional components. Take a look at the Quasi thread and ask around; lots of people there have built this amp with outstanding results and could help you.
AndrewT's recommendation for 10,000uF coupler rated to aboutg 100VW is fine, don't worry too much about low Z loads, that will be fine.
Hope this helps,
Hugh
Hi andrew and AKSA... Thanks bro for your kind advise....
Actually i had in mind the Quasi design suitable since it is quasi complementary using Nmos. I had already ask Quasi on how to modify the schematics to work on single rail.
In my knowledge it is not posible to just put a coupling cap on the input and output since there is a differential input. Am I correct?
Actually i had in mind the Quasi design suitable since it is quasi complementary using Nmos. I had already ask Quasi on how to modify the schematics to work on single rail.
In my knowledge it is not posible to just put a coupling cap on the input and output since there is a differential input. Am I correct?
No, Jun, wrong.
You can easily drive such an amp through coupling caps at input and at output.
No problem at all, just have to be careful of switch on spikes, nothing more.
Hugh
You can easily drive such an amp through coupling caps at input and at output.
No problem at all, just have to be careful of switch on spikes, nothing more.
Hugh
The power on/off thump will be huge, Hundreds of watts.
You will need at least a 47ohm 20w resistor and a relay to switch the
speakers off / bleed resistor on.
5- 10 second on delay instant off.
Leaving the resistor on will just waste output power and raise the low cutoff frequency.
You will need at least a 47ohm 20w resistor and a relay to switch the
speakers off / bleed resistor on.
5- 10 second on delay instant off.
Leaving the resistor on will just waste output power and raise the low cutoff frequency.
ST,
Only if you wired it up without careful thought. What is wrong with a protective relay, and a 1K resistor from coupling cap cold end to ground? Then, when after five seconds all is stabilised, you switch in the relay and all is well. With a 1K resistor, the peak charge current would only be 70mA, but in practice much, much less if you set up a slow bias voltage rise at the input, which would carry the output with it, of course?
Single rail amps were common years ago. The technology is all sorted out; read the history, look at the various designs around, such as the Armstrong.
Rather than saying it can't be done, investigate how it WAS done years ago, and don't be so discouraging to Jun.
Only if you wired it up without careful thought. What is wrong with a protective relay, and a 1K resistor from coupling cap cold end to ground? Then, when after five seconds all is stabilised, you switch in the relay and all is well. With a 1K resistor, the peak charge current would only be 70mA, but in practice much, much less if you set up a slow bias voltage rise at the input, which would carry the output with it, of course?
Single rail amps were common years ago. The technology is all sorted out; read the history, look at the various designs around, such as the Armstrong.
Rather than saying it can't be done, investigate how it WAS done years ago, and don't be so discouraging to Jun.
going to single supply is just about equivalent to moving one of the dual polarity smoothing caps from the PSU to the amplifier output to feed the speaker. It costs no more and should perform just as well if designed properly. All we do is move the location of one cap.
If you think about it this way then you see the correlation between smoothing capacitance and DC blocking capacitance and how both need to suit the load impedance.
Move the DC blocking cap back to the PSU. Does that require the cap value to be changed? Should it be necessary to change the value?
If you think about it this way then you see the correlation between smoothing capacitance and DC blocking capacitance and how both need to suit the load impedance.
Move the DC blocking cap back to the PSU. Does that require the cap value to be changed? Should it be necessary to change the value?
Last edited:
Hello
If I remember correctly, for a single rail amp you need to bias the input transistor base to have 1/2 of the DC rail voltage at the output node, before the output capacitor, so you will have a simetrical clipping.
And using the Quasi in a single rail without mods may raise the output bias so the value of the VBE multiplier bias trim-pot can't re-adjust it correctly.
So it may need some little mods anyways.
Bye
Gaetan
If I remember correctly, for a single rail amp you need to bias the input transistor base to have 1/2 of the DC rail voltage at the output node, before the output capacitor, so you will have a simetrical clipping.
And using the Quasi in a single rail without mods may raise the output bias so the value of the VBE multiplier bias trim-pot can't re-adjust it correctly.
So it may need some little mods anyways.
Bye
Gaetan
Thanks gentlemen for your informative views...
I have studied the NMOS350 from Vas to output no problem for me, but setting half voltage reference for differential pair is my problem i'm a little bit confused and also grounding the negative half.
If will be gald if someone can do it from me, to set the half point of differential pair.
Adding a delay using relay on after the output cap, will this help the speaker s from avoiding turn on thump?
I have studied the NMOS350 from Vas to output no problem for me, but setting half voltage reference for differential pair is my problem i'm a little bit confused and also grounding the negative half.
If will be gald if someone can do it from me, to set the half point of differential pair.
Adding a delay using relay on after the output cap, will this help the speaker s from avoiding turn on thump?
Jun,
My suggestion is to ask Quasi, just to be sure it's done right.
A simple two resistor divider, each of say 47K, from +140V pos rail to ground, will give you a half voltage, around 70V. You would then need a 220uF 80VW cap at the mid-point to ground, to give smoothing, and a slow (roughly 10 second) build up from ground, and then a 22K R1, the input stage bias resistor, from this mid-point to the input transistor base. This will give a nice, smooth, slow rise +70V supply for bias purposes. Then you need a speaker protection relay, which fires at around 60V at the bias point, plenty of circuits on this forum. Above it, the relay connects output to speaker, below it, it disconnects speaker from output.
The output coupler should have a 220R 5W resistor from output side to ground to keep that output referenced and permit the coupler to charge up at switch on when relay is disconnected.
Very simple, very effective. Input signal cap must be rated to 100V, of course.
Cheers,
Hugh
My suggestion is to ask Quasi, just to be sure it's done right.
A simple two resistor divider, each of say 47K, from +140V pos rail to ground, will give you a half voltage, around 70V. You would then need a 220uF 80VW cap at the mid-point to ground, to give smoothing, and a slow (roughly 10 second) build up from ground, and then a 22K R1, the input stage bias resistor, from this mid-point to the input transistor base. This will give a nice, smooth, slow rise +70V supply for bias purposes. Then you need a speaker protection relay, which fires at around 60V at the bias point, plenty of circuits on this forum. Above it, the relay connects output to speaker, below it, it disconnects speaker from output.
The output coupler should have a 220R 5W resistor from output side to ground to keep that output referenced and permit the coupler to charge up at switch on when relay is disconnected.
Very simple, very effective. Input signal cap must be rated to 100V, of course.
Cheers,
Hugh
Hi Hugh...
Thanks for the pointers. I had already post a message for quasi at Power amp under dev thread the aother day but still no answer but i am still waiting for his reply since he is the designer of such a good ckt.
Anyway, I have a draft of what you have suggested. kindly check if there are errors.
Jun
Thanks for the pointers. I had already post a message for quasi at Power amp under dev thread the aother day but still no answer but i am still waiting for his reply since he is the designer of such a good ckt.
Anyway, I have a draft of what you have suggested. kindly check if there are errors.
Jun
Hi Jun,
Move R3 from base to ground to base to the midpoint of the two 47K bias resistors and you have it. That will do it nicely.
Oh, and reduce C4 from 39pF to 33pF.
That's it!!
Hugh
Move R3 from base to ground to base to the midpoint of the two 47K bias resistors and you have it. That will do it nicely.
Oh, and reduce C4 from 39pF to 33pF.
That's it!!
Hugh
I think there should be a ~100k (new R3) in the lead from +IN base to the bias voltage.
Why change R21 to 22k? 1k is sufficient.
Why change R21 to 22k? 1k is sufficient.
Last edited:
Hi
The output cap or DC blocks value, historically are in the range of 2,000 uF for Hi-Fi 8 ohm speakers. They can be even smaller if the amp is driving a specific box design and incorporated in it's modeling or low frequency sound shaping.
The output cap or DC blocks value, historically are in the range of 2,000 uF for Hi-Fi 8 ohm speakers. They can be even smaller if the amp is driving a specific box design and incorporated in it's modeling or low frequency sound shaping.
that will attenuate some of the bass signal and will introduce a phase change extending into the upper bass frequencies.
We want better than "historically" where little below 100Hz was recorded or reproducible.
- Status
- Not open for further replies.