I am looking to build an amp of hifi quality that does 300 to 400 watts rms stereo at 8 ohms and can take a 4 ohm load (if possible 2ohm as long as its safe for the amp and doesn't cost an extra million). How much would something like this cost minus the case? It seems all the pieces are fairly cheap, but there are alot of little pieces. This will be my first amp build so I may need some help on finding parts. This forum seems to be full of audio guru so I am sure that I am in good hands. A newbie question: what is the difference between a chip amp and solid state?
Thanks guys
Dan
Thanks guys
Dan
300 to 400 watts??
First a solid state descrete amplifier uses all separate transistors op-amps resistors, capacitors ect...... and and usually much more complicated then a monolithic chip amp.
A chip amp is an all in one deal, take the LM3875 add +volts -volts two resistors and away you go.
I would suggest starting slow a 300 to 400 watt design descrete or a bridged, parellel chip amp are all pretty complex monsters or maybe take a look in the class D forum.
First a solid state descrete amplifier uses all separate transistors op-amps resistors, capacitors ect...... and and usually much more complicated then a monolithic chip amp.
A chip amp is an all in one deal, take the LM3875 add +volts -volts two resistors and away you go.
I would suggest starting slow a 300 to 400 watt design descrete or a bridged, parellel chip amp are all pretty complex monsters or maybe take a look in the class D forum.
Hi,
each chipamp can give about 60W of reliable power.
If you need 300W then you must use at least 5 chipamps.
If you want 8ohm and 4ohm ability then you must double the chipamp count to 10 minimum.
If you omit the 4ohm ability and allow for 8ohm or 2ohm then you can bridge or parallel but you need an even number of chipamps so 6 becomes the minimum required number.
3chipamps in parallel can push upto 180W into 2.7ohms (8/3).
Bridge these so that you have a bridged/parallel arrangement and you get 360W into 5.4ohms or 240W into 8ohms. Since 240W from 6chipamps is just 40W average from each you could increase the supply voltage slightly to get increased power into the 8ohm load.
Parallel all 6chipamps and you get 360W into 1.4ohms and upto 240W into 2ohms. Again the high voltage option is available to extract more power into the 2ohm load
BTW,
bridging and paralleling is very wasteful of smoothing capacitance. It will cost many times more for the capacitance than it will to buy and assemble the chipamps and PCBs.
It would be MUCH simpler to build a discrete amplifier that can be designed for any sensible load and power that you need.
each chipamp can give about 60W of reliable power.
If you need 300W then you must use at least 5 chipamps.
If you want 8ohm and 4ohm ability then you must double the chipamp count to 10 minimum.
If you omit the 4ohm ability and allow for 8ohm or 2ohm then you can bridge or parallel but you need an even number of chipamps so 6 becomes the minimum required number.
3chipamps in parallel can push upto 180W into 2.7ohms (8/3).
Bridge these so that you have a bridged/parallel arrangement and you get 360W into 5.4ohms or 240W into 8ohms. Since 240W from 6chipamps is just 40W average from each you could increase the supply voltage slightly to get increased power into the 8ohm load.
Parallel all 6chipamps and you get 360W into 1.4ohms and upto 240W into 2ohms. Again the high voltage option is available to extract more power into the 2ohm load
BTW,
bridging and paralleling is very wasteful of smoothing capacitance. It will cost many times more for the capacitance than it will to buy and assemble the chipamps and PCBs.
It would be MUCH simpler to build a discrete amplifier that can be designed for any sensible load and power that you need.
There so many people here building high powered amp based on Chipamp, you can get the boards from them. with the chipamp comparable in price or cheaper with output transistors building a high powered chipamp would be cheaper than discrete besides you use less components.
Furthermore using discrete components you need to hunt around for a suitable design and in addition test it extensively before you would be confortable to use it. Chipamp is so easy to make you cannot get it wrong esp using the PCB board.
Smoothing capacitance there is no difference in the total amount of capcitance use in eithe chipamp or discrete transistors for the same amount of power output.
Furthermore using discrete components you need to hunt around for a suitable design and in addition test it extensively before you would be confortable to use it. Chipamp is so easy to make you cannot get it wrong esp using the PCB board.
Smoothing capacitance there is no difference in the total amount of capcitance use in eithe chipamp or discrete transistors for the same amount of power output.
Hi,
I and many others have used two criteria for determining smoothing capacitance built into an amplifier PSU.
1. the RC time constant of output load (Rload) and the smoothing capacitance on each supply rail (C) must be double the high pass RC filter on the input to the amplifier.
2. the capacitance must supply the peak current sent to the load, this is 2 to 3mF/Apk.
A single channel amplifier of 400W into 8ohm is putting 80Vpk and 10Apk into an 8ohm load. This requires +-20mF to meet criteria 1, if input RC=80mS. It also requires +-20mF to +-30mF to meet criteria 2.
It happens in this 400W case that both requirements are similar.
A discrete 400W amplifier needs about +-20mF/channel.
Now consider a bridged arrangement.
To send 400W to an 8ohm load, each half of the bridge must be able to send 200W into 4ohm.
Each half bridge needs +-40mF to meet criteria 1 if the bass response is to match the RC=80mS chosen above. The peak current into the 4ohm load is still 10Apk and so criteria 2 requires +-20mF to +-30mF.
This time the capacitance requirements do not closely match. I suggest that the higher value should be chosen to give equal performance. We have +-40mF/half bridge.
The whole bridged amplfier therefore requires +-80mF/channel to meet that 400W into 8ohms specification. This applies to any bridged arrangement irrespective of whether it is based on discrete or chipamp.
A bridged amplifier requires four times as much smoothing capacitance as a single amplifier. The biggest saving is that the voltage rating of the capacitors is very different. caps above 75Vdc tend to be expensive, but are they four times as expensive?
Do not fall into the trap of accepting lower performance without taking steps to ensure your decisions are fully informed.
There is another capacitance criteria that I do not use:- +-100uF/W, This would result in +-40mF for the 400W amp and +-20mF for each half of the bridged amp. If you decide that this is more appropriate then the same capacitance will be required for either arrangement.
I suggest that amps based on this do not result in the same bass performance.
I wonder if the complaints about poor bass from some chipamp builders are the result of insufficient smoothing capacitance, particularly, when using bridged amps to obtain more power for sub-bass use?
I and many others have used two criteria for determining smoothing capacitance built into an amplifier PSU.
1. the RC time constant of output load (Rload) and the smoothing capacitance on each supply rail (C) must be double the high pass RC filter on the input to the amplifier.
2. the capacitance must supply the peak current sent to the load, this is 2 to 3mF/Apk.
A single channel amplifier of 400W into 8ohm is putting 80Vpk and 10Apk into an 8ohm load. This requires +-20mF to meet criteria 1, if input RC=80mS. It also requires +-20mF to +-30mF to meet criteria 2.
It happens in this 400W case that both requirements are similar.
A discrete 400W amplifier needs about +-20mF/channel.
Now consider a bridged arrangement.
To send 400W to an 8ohm load, each half of the bridge must be able to send 200W into 4ohm.
Each half bridge needs +-40mF to meet criteria 1 if the bass response is to match the RC=80mS chosen above. The peak current into the 4ohm load is still 10Apk and so criteria 2 requires +-20mF to +-30mF.
This time the capacitance requirements do not closely match. I suggest that the higher value should be chosen to give equal performance. We have +-40mF/half bridge.
The whole bridged amplfier therefore requires +-80mF/channel to meet that 400W into 8ohms specification. This applies to any bridged arrangement irrespective of whether it is based on discrete or chipamp.
A bridged amplifier requires four times as much smoothing capacitance as a single amplifier. The biggest saving is that the voltage rating of the capacitors is very different. caps above 75Vdc tend to be expensive, but are they four times as expensive?
Do not fall into the trap of accepting lower performance without taking steps to ensure your decisions are fully informed.
There is another capacitance criteria that I do not use:- +-100uF/W, This would result in +-40mF for the 400W amp and +-20mF for each half of the bridged amp. If you decide that this is more appropriate then the same capacitance will be required for either arrangement.
I suggest that amps based on this do not result in the same bass performance.
I wonder if the complaints about poor bass from some chipamp builders are the result of insufficient smoothing capacitance, particularly, when using bridged amps to obtain more power for sub-bass use?
An amp with this power level is not recommended for your first amp. Can you tell us where you power requirements come from?saabracer23 said:
If you're sure you need it, you should build a small amplifier first, for practice. A power supply capable of powering the larger amp is no small undertaking.
Re: AndrewT
Do you agree with any of my arguments?
I side with the view that: bridged amps NEED more capacitance to match the performance of a single amp doing the same duty.
You choose to not believe that, or are unaware of what affects bass performance.
Another chipamp thread even claimed that a PSU consisting of just +-1mF on the chipamp did not compromise the bass performance. I tried explaining, but left them to dream (=consider mid & treble performance above all other balanced audio considerations) on their own.
more importantly, do you understand my argument?ttan98 said:
Do you agree with any of my arguments?
do you see there can be differences of opinion on this statement?
I side with the view that: bridged amps NEED more capacitance to match the performance of a single amp doing the same duty.
You choose to not believe that, or are unaware of what affects bass performance.
Another chipamp thread even claimed that a PSU consisting of just +-1mF on the chipamp did not compromise the bass performance. I tried explaining, but left them to dream (=consider mid & treble performance above all other balanced audio considerations) on their own.
Hi. My first project was Rod Elliot's 300 watt sub amp which I later upgraded to 500 watts. I ignored all the warnings about not attempting an amp this large as your first project, I didn't want anything smaller. I didn't even use a proper pc board, but stripped all the componets off an old integrated reciever board, sanded off the traces and used that wired p2p.
The final cost of this amp was around $600.00 just for parts ( bought batches of this, of that) I have gone on to use some of these other componets in other projects, therefore this waters down that price a bit.
Working with a power supply that runs 140 volts DC is damgerous, but that point will not be driven home by working with a 40 volt supply, one must be extremely careful when handling electricity.
The final cost of this amp was around $600.00 just for parts ( bought batches of this, of that) I have gone on to use some of these other componets in other projects, therefore this waters down that price a bit.
Working with a power supply that runs 140 volts DC is damgerous, but that point will not be driven home by working with a 40 volt supply, one must be extremely careful when handling electricity.
I have also built chip amps from free samples I got from National Semi. and was impessed by the simplicity of construction and sound quality. There are too many power supply constaints to deal with for these though, so I prefer all discrete.
I am currently working on a hybrid of the two, don't even know for certain if it will work, but will try anyway. Two LM3886 bridged with a pure class B bjt output stage. Should be fun!
I am currently working on a hybrid of the two, don't even know for certain if it will work, but will try anyway. Two LM3886 bridged with a pure class B bjt output stage. Should be fun!
The first time I drove I pulled away and managed to get the car up to 140kph before the screams of the passengers got through the adrenaline... luckily nothing happened...
I see significant chances of you litteraly blowing alot of money if this is your first big project... not to mention safety hazzards.
at least build a small dual chip amp first to see some of the hazzard areas before you do the big ones...
I see significant chances of you litteraly blowing alot of money if this is your first big project... not to mention safety hazzards.
at least build a small dual chip amp first to see some of the hazzard areas before you do the big ones...
I'm working the model now on Multisim, seems to run clean.
saabracer23: Chip amps are dead easy to use, especially if you have a pc board, not too expensive either. I'd buy all parts from one place such as Digi-Key. Biggest expense in any amp is power supply, mainly a good transformer. I've bought a couple on Ebay with good luck( seller is Antek Inc )
Have fun and be safe.
saabracer23: Chip amps are dead easy to use, especially if you have a pc board, not too expensive either. I'd buy all parts from one place such as Digi-Key. Biggest expense in any amp is power supply, mainly a good transformer. I've bought a couple on Ebay with good luck( seller is Antek Inc )
Have fun and be safe.
AndrewT said:
Hi Andrew,
All good stuff! I assume that mF is microfarads? I wouldn't ask but you used uF once and I'd rather not get it wrong while trying to understand all this.
I must admit, my stock standard Chipamp kit lacks bass, but I did wonder if it was because the transformer outputs were only 21.5 volts - my method of impedance measurement of my DIY speakers (to use when deciding the power rail voltage) isn't so accurate, so I'm not sure if I have this right. I followed a fairly rough and ready method from Dickason's book. I have a new transformer for another project of 400VA/25V so I might just try this as it's only a half hour job to change them over.
By the way, do you build audio gear?
Cheers
Stuey
hi,
You can hardly go wrong with LM3886. can be as simple as 5 components per channel and slot into a match box or as complex as your imagination brings...
One helpful link,
http://myweb.tiscali.co.uk/nuukspot/decdun/gaincloneindex.html
Enough ideas and projects to last you a while.
Or do a Google search on gain clone.
Happy reading
You can hardly go wrong with LM3886. can be as simple as 5 components per channel and slot into a match box or as complex as your imagination brings...
One helpful link,
http://myweb.tiscali.co.uk/nuukspot/decdun/gaincloneindex.html
Enough ideas and projects to last you a while.
Or do a Google search on gain clone.
Happy reading
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.