ashok said:"..............reversed printing................"
This is great ! I made four boards and spent the last couple of days getting it done. This morning while trying to put in some parts I realised that the pattern was reversed !
I certainly will not be making the board again . So I thought for a while and then decided that I will use it as it is and invert the power transistors and all the other active devices !
Will put up pictures after it is done ! And I thought that only I make such mistakes !
Oh man, sorry to hear about that. Please everyone take care when making this board. Use the layout image with components shown to guide you in the orientation. Remember this image shows the tracks looking "through" the board from the component side. Whatever technique you use please double check by flipping the board over and comparing how it should look.
Cheers
Q
Over-current protection idea.
.
Given this thread is called "Power amp under development", here is the next idea; Over-current protection.
This is just an idea for comment by the clever people who have help thus far.
The idea is to make a tiny board (under 1 inch square) that can be retrofitted to completed modules if desired. I decided on an opto-coupler approach to make the circuit universal.
One application is as shown; connected across the Vbe multiplier and shunting drive from the next stage during over-current conditions. Another application could be to use the optocoupler transistor to trigger some other safety system.
Anyway please comment and add what I missed.
Cheers
Q
.
Given this thread is called "Power amp under development", here is the next idea; Over-current protection.
This is just an idea for comment by the clever people who have help thus far.
The idea is to make a tiny board (under 1 inch square) that can be retrofitted to completed modules if desired. I decided on an opto-coupler approach to make the circuit universal.
One application is as shown; connected across the Vbe multiplier and shunting drive from the next stage during over-current conditions. Another application could be to use the optocoupler transistor to trigger some other safety system.
Anyway please comment and add what I missed.
Cheers
Q
Attachments
Hi Quasi,
I think the idea is very fine, but I have some concerns that even at lowish currents, say 3A, which would cause 1V41 drop across the 0.47R emitter resistor, low currents would flow into the optocoupler threatening to turn on the transistor shunt prematurely. One way to prevent this would be to simply place a diode in series with the optoLED, so that triggering would not be possible until it was forward biased, at around 5.5A.
I feel the principle is very good, however, quite elegant. There was a special chip released by Linear Technologies years ago if I recall which served this purpose; it was a low voltage differential Schmitt trigger, quite expensive, but reputedly effective. It's possible other chip makers have something cheaper which now does an even better job.
Protection is a difficult issue. Most current designs have sonic impact, particularly where they act directly on the VAS drive.
Cheers,
Hugh
I think the idea is very fine, but I have some concerns that even at lowish currents, say 3A, which would cause 1V41 drop across the 0.47R emitter resistor, low currents would flow into the optocoupler threatening to turn on the transistor shunt prematurely. One way to prevent this would be to simply place a diode in series with the optoLED, so that triggering would not be possible until it was forward biased, at around 5.5A.
I feel the principle is very good, however, quite elegant. There was a special chip released by Linear Technologies years ago if I recall which served this purpose; it was a low voltage differential Schmitt trigger, quite expensive, but reputedly effective. It's possible other chip makers have something cheaper which now does an even better job.
Protection is a difficult issue. Most current designs have sonic impact, particularly where they act directly on the VAS drive.
Cheers,
Hugh
current limiting
Hi,
this is monitoring current.
Could you modify the sensor side to make it a V-I limiter.
It will be more useful, allowing larger currents to pass when Vds is low and reducing the acceptable current when Vds is high.
If you retain the proposed current limit, what value do you select for trip current?
The maximum dissipation that the heatsink can cope with?
or the maximum that the cold FETs can pass?
or the maximum that the hot FETs can pass?
or the maximum short circuit current that can pass when Vds=Vrail?
or many others all of which have a downfall when V is excluded from the protection locus.
Hi,
this is monitoring current.
Could you modify the sensor side to make it a V-I limiter.
It will be more useful, allowing larger currents to pass when Vds is low and reducing the acceptable current when Vds is high.
If you retain the proposed current limit, what value do you select for trip current?
The maximum dissipation that the heatsink can cope with?
or the maximum that the cold FETs can pass?
or the maximum that the hot FETs can pass?
or the maximum short circuit current that can pass when Vds=Vrail?
or many others all of which have a downfall when V is excluded from the protection locus.
Re: current limiting
I guess it could be set for any of the above. My initial idea was simply to protect against an output short, where the current would be limited (albeit at a high level) until the fuses blew. I.e; if the peak current into 4 ohms was 15 amps (5 amps per FET in the case of the Nmos350) then the sensor could be adjusted to activate on 20 amps. This should blow the fuses (5 amp) quick enough to save the FETs. It is something I have not tried yet though.
The other alternative is to connect the transistor output to a logic circuit (or SCR etc), where if activated would completely shut the amp down, requiring a power down reset.
Cheers
Q
AndrewT said:
I guess it could be set for any of the above. My initial idea was simply to protect against an output short, where the current would be limited (albeit at a high level) until the fuses blew. I.e; if the peak current into 4 ohms was 15 amps (5 amps per FET in the case of the Nmos350) then the sensor could be adjusted to activate on 20 amps. This should blow the fuses (5 amp) quick enough to save the FETs. It is something I have not tried yet though.
The other alternative is to connect the transistor output to a logic circuit (or SCR etc), where if activated would completely shut the amp down, requiring a power down reset.
Cheers
Q
Working on it!!
Here there's a pic of Nmos350 board i'm making. There are a lot of components missing yet i'm gonna get them next week and hopefully finish it.
PS: Before anyone notices
T8 goes underboard, i know, it's just placed there but not welded, i'll weld it when i put the FETS.
Cheers!
Here there's a pic of Nmos350 board i'm making. There are a lot of components missing yet i'm gonna get them next week and hopefully finish it.
PS: Before anyone notices
T8 goes underboard, i know, it's just placed there but not welded, i'll weld it when i put the FETS.Cheers!
Attachments
Re: Re: current limiting
let's take your 20A trigger current.
each FET will take 10A (2pair) and generate about 700W per FET and 2800W into the sink.
The FET junction is going to heat very rapidly and the safe current is falling as Tj & Tc is rising. So the rail fuses are required to blow quickly to prevent overheating damage.
If your planned maximum output current is 15A then F8A fuses should be fitted.
Now I see a problem:
the output current can never exceed 19.9Apk no matter how short the transient. This may sound as needless limiting on transient signals that are well within the safe pass ability of the output pairs.
FETs can benefit from V-I limiting just as BJTs can.
Hi,quasi said:
let's take your 20A trigger current.
each FET will take 10A (2pair) and generate about 700W per FET and 2800W into the sink.
The FET junction is going to heat very rapidly and the safe current is falling as Tj & Tc is rising. So the rail fuses are required to blow quickly to prevent overheating damage.
If your planned maximum output current is 15A then F8A fuses should be fitted.
Now I see a problem:
the output current can never exceed 19.9Apk no matter how short the transient. This may sound as needless limiting on transient signals that are well within the safe pass ability of the output pairs.
FETs can benefit from V-I limiting just as BJTs can.
Nice work Hernanstafe and it's good tho see the odd second hand component in there, just like mine. Let us know how you get on.
Hi AndrewT,
The peak current of 15 amps (into 4 ohm load) occurs only for an instant, the RMS current being 0.7 time less than that, and the average current with music (what fuses really care about) is much less. So running 5 amp fuses with 65-70 volt rails works and I've never changed my fuses yet.
The peak current into 4 ohms will never exceed 15 amps because this is the hard limit set by the voltage rails. A detection headroom of 5 amps means that the circuit will only activate under short circuit conditions. In this case the potential current will be dependant on too many variables but could have exceeded 100 amps were it not limited to 20. Across 3 pairs of FETs this amounts to a theoretical 470 watts per FET, although in real terms the power supply would have collapsed so it would be considerably less. This can be improved by adding more FETs, as in the Nmos500.
So the question is, will the FETs survive? With 5 amp fuses they should. Of course the detection point can be set lower (by changing the voltage divider) say at 16 amps, but this runs the risk of turning on the detector early resulting in distortion. Is this a bad thing? Well it could be depending on your point of view.
The above is based on a typical setup for an Nmos350 with 70 volt rails and allows for some power supply sag and FET drop.
Cheers
Q
Hi AndrewT,
The peak current of 15 amps (into 4 ohm load) occurs only for an instant, the RMS current being 0.7 time less than that, and the average current with music (what fuses really care about) is much less. So running 5 amp fuses with 65-70 volt rails works and I've never changed my fuses yet.
The peak current into 4 ohms will never exceed 15 amps because this is the hard limit set by the voltage rails. A detection headroom of 5 amps means that the circuit will only activate under short circuit conditions. In this case the potential current will be dependant on too many variables but could have exceeded 100 amps were it not limited to 20. Across 3 pairs of FETs this amounts to a theoretical 470 watts per FET, although in real terms the power supply would have collapsed so it would be considerably less. This can be improved by adding more FETs, as in the Nmos500.
So the question is, will the FETs survive? With 5 amp fuses they should. Of course the detection point can be set lower (by changing the voltage divider) say at 16 amps, but this runs the risk of turning on the detector early resulting in distortion. Is this a bad thing? Well it could be depending on your point of view.
The above is based on a typical setup for an Nmos350 with 70 volt rails and allows for some power supply sag and FET drop.
Cheers
Q
Hey quasi! i had some old caps and resistors that i tested and were in great condition so i used them instead of buying new ones! the electrolityc caps and 2 of the ceramic are used, but like i said they're just fine so why not using them Everything else is new (sadly, i didn't have any of the expensive transistors )
Tomorrow is a national holiday so shops will be closed, so on tuesday i'm going to buy the components i still don't have.
PS: I have a doubt about the "Isulated wire link underboard". Why is it underboard? any special reason?
Everything else is new (sadly, i didn't have any of the expensive transistors )Tomorrow is a national holiday so shops will be closed, so on tuesday i'm going to buy the components i still don't have.
PS: I have a doubt about the "Isulated wire link underboard". Why is it underboard? any special reason?
hernanstafe said:
The insulated wire link is used to connect the main PCB ground to the front end. It is under the board to avoid a lot of components alont the way. The wire does not have to be very thick as it only carries milliamps but it must be insulated. It must be connected only at the points shown in the layout to ensure low distortion.
Cheers
Q
oh i see...but is it my imagination or there are no pinouts for that cable? cause i don't see a whole to put the wire!!. One more thing, the caps C11 and C13 (330uf 200v) have to be of that value? because i have some 1000uF and 470 uF 200v caps and i'd like to know if i can use those instead without affecting performance!
Cheers
Cheers
Re: What ..Isulated wire link underboard..?
Send me an email, and I'll send you the latest PDF file showing this. By the way the insulated wire link only applies to the Nmos350 and the Nmos500. It does not apply to the Nmos200.
Cheers
Q
ashok said:Insulated wire link below the board ...........where ?
I don't recollect seeing one in a picture of any board . So what points does this connect ?
Send me an email, and I'll send you the latest PDF file showing this. By the way the insulated wire link only applies to the Nmos350 and the Nmos500. It does not apply to the Nmos200.
Cheers
Q
Quasi:
I'm continuing building the amp and since i'm close to finishing it, i took a look at the setting up instructions you made. I have a doubt which makes me feel a bit ignorant but i need to know in order to set the amp up.
There's a part where you say i have to measure the offset voltage...my doubt is, is there a way to do this with a multimeter? because i know how to do it with an oscilloscope but i don't have one at home, so i'd like to know if i can measure it with some other instrument. If not, i'll have to get an oscilloscope from a friend or at college.
I'm sorry for the question, i'm sure it's quite silly *blushes*.
Thanks again!
I'm continuing building the amp and since i'm close to finishing it, i took a look at the setting up instructions you made. I have a doubt which makes me feel a bit ignorant but i need to know in order to set the amp up.
There's a part where you say i have to measure the offset voltage...my doubt is, is there a way to do this with a multimeter? because i know how to do it with an oscilloscope but i don't have one at home, so i'd like to know if i can measure it with some other instrument. If not, i'll have to get an oscilloscope from a friend or at college.
I'm sorry for the question, i'm sure it's quite silly *blushes*.
Thanks again!
Hi Hernanstafe,
Connect a multimeter from Ground to the output track before the output relay. This is easily done using a component lead connected to the output track.
Using your digital meter's 200mV scale and without anything connected to the input measure the voltage. It should be near zero. Use VR1 to adjust to as close as zero as you can. Try to get under 10mV.
Cheers
Q
Connect a multimeter from Ground to the output track before the output relay. This is easily done using a component lead connected to the output track.
Using your digital meter's 200mV scale and without anything connected to the input measure the voltage. It should be near zero. Use VR1 to adjust to as close as zero as you can. Try to get under 10mV.
Cheers
Q
Hi,
to measure the output offset, you must disconnect the load and fit a shorting plug to the input. Use a voltmeter set intially to a high value DC volts and switch to lower values as you confirm that the voltmeter will not be overloaded.
Then set the offset after the amp has reached it's final warm temperature.
You will probably find that the final setting results in a cold offset, but this usually reduces rapidly at first switch on and then more slowly as the devices warm up.
to measure the output offset, you must disconnect the load and fit a shorting plug to the input. Use a voltmeter set intially to a high value DC volts and switch to lower values as you confirm that the voltmeter will not be overloaded.
Then set the offset after the amp has reached it's final warm temperature.
You will probably find that the final setting results in a cold offset, but this usually reduces rapidly at first switch on and then more slowly as the devices warm up.