I have available +-25volts rails, and want to make an amp of rating 150watts or so per channel, 2 channel.... Is this possible with this low voltage? I was thinking of using the bridge/parallel diagram for the LM3886 ICS.... Not sure how much power this would give... and these parts are EXPENSIVE!!! I would prefer transistor, But don't think I could get this much power from such low rails.... I have no money, and these transformers are cheap... I have no money, but still spend it all.. 🙁
rough 'back-of-the-envelope' calcs would suggest that you would need to be drawing 8.5 amps per rail, per amp !!!
and you would probably need to be driving a 1 or 2 Ohm load to get the power , that would be very hard on the LM3886's even paralleled.
the LM3886 current limiting would restrict what you could drive.
ray
and you would probably need to be driving a 1 or 2 Ohm load to get the power , that would be very hard on the LM3886's even paralleled.
the LM3886 current limiting would restrict what you could drive.
ray
But on the LM3886 Data sheet, it says with a +-25volts I can get 55watts into 4ohms.. bridge 2 into 8ohms, and I should get 110watts?? So, why couldn't get 150 using 4 ICs?? 
BTW, this is the Transforer in Question... I am using 2, one for each rail.. with the secondaries in series..
250VA TOROIDAL TRANSFORMER: 2 x 120V primary, 2 x 9V secondary. No mounting hardware available. The unit weighs 4Kg
BTW, this is the Transforer in Question... I am using 2, one for each rail.. with the secondaries in series..
250VA TOROIDAL TRANSFORMER: 2 x 120V primary, 2 x 9V secondary. No mounting hardware available. The unit weighs 4Kg
An externally hosted image should be here but it was not working when we last tested it.
I cant see you getting more than 100W RMS into an 8ohm load even if you bridge / parallel multiple chips. The mathematics is rather simple. Go study the datasheet and you'll see that you will not get ~200W RMS into anything less than a 4 ohm load. To acheive that, you would need to run a few in parallel and provide good heatsinking to keep away from the protection circuit.
You want to use +/-25V and you plan to use an 8ohm load.
Your going to bridge parallel the chips.
IIRC, these chips are capable of swinging to within 5 Volts of their supply voltage so given the configuration we are using, we can swing 40V. 40V swing into 8ohms gives 200W peak or 100W RMS (average).
Your going to bridge parallel the chips.
IIRC, these chips are capable of swinging to within 5 Volts of their supply voltage so given the configuration we are using, we can swing 40V. 40V swing into 8ohms gives 200W peak or 100W RMS (average).
Hmm... but +-25volts gives 55watts into 4ohms... bridge 2, and you have 110watts into 8ohms??
http://david.lewander.com/projects/amp/amp.htm
http://david.lewander.com/projects/amp/amp.htm
Here is the math approximately:
You have +-25 Volts to play with.
Best case you are going to swing within 3 volts of the rails before the distortion starts to get really bad. That means you are going to swing +-22 Volts, or a maximum excursion in either direction in a bridged configuration of 44 volts
RMS Power = (Voltage/(Sqrt(2))^2
-----------------------------------
Load Resistance
Or in this case (44/1.414)^2/8 = 120 Watts.
If I was being a little more "realistic" with the LM3886, I would say you are not going to swing more than within 5 volts of the rail for good sound which will give you 100 watts.
No matter how many LM3886 you parallel, you are not going to get much closer to the rails. That is a limitation of their internal design and you can't change that.
Your peak current will be somewhere on the order of 50/8 (or higher as your speaker impedance drops lower in all likelyhood). That does not mean your power supply must put out an average of 8 amps, but it does mean sufficient power supply filtering to ensure you can handle the peaks without the power supply drooping.
You are going to use 2-250VA transformers. That does not mean you get 500watts out of them as it is a complex correspondence between VA, how the load behaves, and watts, but I would say they have enough oomph.
Besides, a "good" 100 watts is always better than a crummy 150. A speakers impedance can drop considerably below the nominal, and if your design handles lower impedances well, then you will be surprised how load 100 watts can be.
My primary amplifier puts out only 50 watts into 8 ohms, but it will put 100 into 4 and 200 into 2. My speakers vary in impedance from 3.7 ohms up to about 10 ohms over the audio bandwidth and they have an efficiency of 92db/watt. I love loud music and I never feel like I don't have enough power.
Alvaius
You have +-25 Volts to play with.
Best case you are going to swing within 3 volts of the rails before the distortion starts to get really bad. That means you are going to swing +-22 Volts, or a maximum excursion in either direction in a bridged configuration of 44 volts
RMS Power = (Voltage/(Sqrt(2))^2
-----------------------------------
Load Resistance
Or in this case (44/1.414)^2/8 = 120 Watts.
If I was being a little more "realistic" with the LM3886, I would say you are not going to swing more than within 5 volts of the rail for good sound which will give you 100 watts.
No matter how many LM3886 you parallel, you are not going to get much closer to the rails. That is a limitation of their internal design and you can't change that.
Your peak current will be somewhere on the order of 50/8 (or higher as your speaker impedance drops lower in all likelyhood). That does not mean your power supply must put out an average of 8 amps, but it does mean sufficient power supply filtering to ensure you can handle the peaks without the power supply drooping.
You are going to use 2-250VA transformers. That does not mean you get 500watts out of them as it is a complex correspondence between VA, how the load behaves, and watts, but I would say they have enough oomph.
Besides, a "good" 100 watts is always better than a crummy 150. A speakers impedance can drop considerably below the nominal, and if your design handles lower impedances well, then you will be surprised how load 100 watts can be.
My primary amplifier puts out only 50 watts into 8 ohms, but it will put 100 into 4 and 200 into 2. My speakers vary in impedance from 3.7 ohms up to about 10 ohms over the audio bandwidth and they have an efficiency of 92db/watt. I love loud music and I never feel like I don't have enough power.
Alvaius
My appologies, official specs state that this device can swing to withing 4V of it's rails so in this case, you should be able to get 110W RMS (average) maybe just.
Okay, so, I'll order the transformers then.. 😀 could I use 3??
ANd use all three in series, with the centre tap of the middle trans being the overall centre tap?? 🙂
ANd use all three in series, with the centre tap of the middle trans being the overall centre tap?? 🙂
No you cant do that .... these chips cant take that much voltage into that load. It could be done but you would have to modify the topologies generally adding expense and complication that really is not warrantted.
....2 x 9V secondary....
hmmm.... did you type that in correctly?
how are you getting +-25v rails from 9v secondaries?
did you mean 19v , or are you using more than one transformer?
ah, I see now - just re-read the post, you are using 1 for each rail....
I still think you are pushing it a bit the expect 150W from a bridged pair. You might be ok with at least 2 paralled and then bridged , but you will lose out slightly due to the need for matching resistors, so you wont get 2x the power of one pair...
paralleling the ic's will not improve the voltage swing, but it might enable you to drive a lower impedance load ( normally 8R min in bridge config ) , you might *just* be able to get 150W into 4R , at the cost of 4 devices and some matching....
however, this will still run over the the 4amp current limit of the ic's , since you will be drawing 8.5 amps from each rail, so expect major distortion.
To run clear of the current limiting you would need to parallel 3 devices on each side of the bridge, thats 6 devices / channel !!
ray
ray
hmmm.... did you type that in correctly?
how are you getting +-25v rails from 9v secondaries?
did you mean 19v , or are you using more than one transformer?
ah, I see now - just re-read the post, you are using 1 for each rail....
I still think you are pushing it a bit the expect 150W from a bridged pair. You might be ok with at least 2 paralled and then bridged , but you will lose out slightly due to the need for matching resistors, so you wont get 2x the power of one pair...
paralleling the ic's will not improve the voltage swing, but it might enable you to drive a lower impedance load ( normally 8R min in bridge config ) , you might *just* be able to get 150W into 4R , at the cost of 4 devices and some matching....
however, this will still run over the the 4amp current limit of the ic's , since you will be drawing 8.5 amps from each rail, so expect major distortion.
To run clear of the current limiting you would need to parallel 3 devices on each side of the bridge, thats 6 devices / channel !!
ray
ray
I have decided to just bridge 2... are you saying this won't work?? 🙁 Its bed time now anyway, THANX for all the help.. 🙂 bye bye
Not when they are looking at an effective 4 ohm load ... go check the datasheet. You would need to run to in parallel or you would activate the protection circuit
AudioFreak said:
So, it won't work??? 🙄 The speakers are 8ohms, bridging mean each channel get 4ohms, the data sheet states the chip can do 68watts into 4ohms with +-28volts..... I confused...
Using 1 transformer per rail will work your later idea of using 3 transformers with the centre tap of the middle transformer acting as ground would provide too much voltage. Just stick with 1 transformer per rail and you'll be fine.
Hi all,
Just my 2 cents worth..... Gee whiz, if you are going to series transformers together and get higher voltages, why not find a transistor amp or get an STK I.C.? I suggest putting 2 transformers in series and 2 in parallel with that. Then you have the voltage and current! LM3886 chips aren't all that people rave about. I replace them constantly in cheap guitar amps. I'm NOT impressed......
With respect,
Chris
Just my 2 cents worth..... Gee whiz, if you are going to series transformers together and get higher voltages, why not find a transistor amp or get an STK I.C.? I suggest putting 2 transformers in series and 2 in parallel with that. Then you have the voltage and current! LM3886 chips aren't all that people rave about. I replace them constantly in cheap guitar amps. I'm NOT impressed......
With respect,
Chris
- Status
- Not open for further replies.