I am in the process of planning a gainclone type amplifier around the LM3886. It will be a 4.1/5.1 system and i would probably be bi-amping each speaker (tweeter and mid).
so the question is really about hte power supply. Rather than having 5/6 power supplies and the same number of transformers, could i use one larger transformer of the necessary VA rating to supply multiple rectifier/regulation boards to power each chanel?
I am somewhat confused on the issue of "crosstalk" i understand that it can occur when using a single power supply, but was that referring to a PS with only one reguulation circuit?
Ive searched and seen people saying to use transformers with dual secondaries IE 24-0-0-24, but i dont really understand why this would help......
If what i am trying to do isnt possible, is there an easy way to achieve this with minimal cost?
THanks, Brian
so the question is really about hte power supply. Rather than having 5/6 power supplies and the same number of transformers, could i use one larger transformer of the necessary VA rating to supply multiple rectifier/regulation boards to power each chanel?
I am somewhat confused on the issue of "crosstalk" i understand that it can occur when using a single power supply, but was that referring to a PS with only one reguulation circuit?
Ive searched and seen people saying to use transformers with dual secondaries IE 24-0-0-24, but i dont really understand why this would help......
If what i am trying to do isnt possible, is there an easy way to achieve this with minimal cost?
THanks, Brian
Hi,
dual secondary allow one to use dual rectifier or single rectifier.
centre tapped only allows one to use single rectifier.
Some chipamp PCBs are designed for dual sec + dual rectifier, but these can usually be modified to use centre tapped.
A single transformer can supply any number of amplifiers.
1.)
the cheap and nasty way is single bridge and single smoothing pair and run +ve & -ve lines to each amplifier.
2.)
better is single rectifier +smoothing bank dedicated to EACH amplifier. couple each rectifier in parallel to it's neighbours from the same secondary tappings.
3.)
one pair of secomdary windings feeding a single rectifier + smoothing to EACH amplifier. This achieves some of the separation of the true monoblock. Multiple secondaries are not common.
4.)
full monoblock with transformer and rectifier and smoothing to EACH amplifier.
A short while ago I gave advice on VA rating for a multi-channel.
I think it ran along the line of transformer VA=1.5times largest amp power rating + sum of all the other amplifier power ratings, IF they all run at the SAME voltage. eg. 5.1 with 5channels @ 50W into 8r and a parallel pair for sub-bass @ 100W into 4r would work quite well with 400VA and needs 7banks of +-10mF. The extra bank is due to the parallel pair needing smoothing for each half of the sub-bass pair. That's 14off 10mF 50V caps for single amplified speakers. Bi-amped will need 14off 10mF 50V for the bass/mid duty and 10off 6,800uF 50V for treble duty (or maybe quite a bit smaller, some advocate +-1mF at the chipamp ONLY).
dual secondary allow one to use dual rectifier or single rectifier.
centre tapped only allows one to use single rectifier.
Some chipamp PCBs are designed for dual sec + dual rectifier, but these can usually be modified to use centre tapped.
A single transformer can supply any number of amplifiers.
1.)
the cheap and nasty way is single bridge and single smoothing pair and run +ve & -ve lines to each amplifier.
2.)
better is single rectifier +smoothing bank dedicated to EACH amplifier. couple each rectifier in parallel to it's neighbours from the same secondary tappings.
3.)
one pair of secomdary windings feeding a single rectifier + smoothing to EACH amplifier. This achieves some of the separation of the true monoblock. Multiple secondaries are not common.
4.)
full monoblock with transformer and rectifier and smoothing to EACH amplifier.
A short while ago I gave advice on VA rating for a multi-channel.
I think it ran along the line of transformer VA=1.5times largest amp power rating + sum of all the other amplifier power ratings, IF they all run at the SAME voltage. eg. 5.1 with 5channels @ 50W into 8r and a parallel pair for sub-bass @ 100W into 4r would work quite well with 400VA and needs 7banks of +-10mF. The extra bank is due to the parallel pair needing smoothing for each half of the sub-bass pair. That's 14off 10mF 50V caps for single amplified speakers. Bi-amped will need 14off 10mF 50V for the bass/mid duty and 10off 6,800uF 50V for treble duty (or maybe quite a bit smaller, some advocate +-1mF at the chipamp ONLY).
ok, i think i get what your saying.
So i could use a single rectifier board of a sufficient current rating for all my amplifier circuits, then run a regulation (smothing) board for each amplifier i want to run....? I would have 4 satelite speakers (each with 2 amplifiers one for tweeter one for mid), so i would probably use the same regulation board to feed those two. So 4 power reg. boards for the satelite speakers, and 8 amps. Whew... this is gonna be a monster
.
As for the sub, i would take the front L&R chanels and amplify both. They would probably each have their own dedicated amp and regulation board. So the total would be 10 banks of regulation and amplifiers.....
BTW, what is the difference between #2 and #3 that you described?
By doing it this way would i get any crosstalk between channels?
and is there any way to reduce the chances of getting it?
400-500 VA was about what i figured, thanks for confirming.
So i could use a single rectifier board of a sufficient current rating for all my amplifier circuits, then run a regulation (smothing) board for each amplifier i want to run....? I would have 4 satelite speakers (each with 2 amplifiers one for tweeter one for mid), so i would probably use the same regulation board to feed those two. So 4 power reg. boards for the satelite speakers, and 8 amps. Whew... this is gonna be a monster
. As for the sub, i would take the front L&R chanels and amplify both. They would probably each have their own dedicated amp and regulation board. So the total would be 10 banks of regulation and amplifiers.....
BTW, what is the difference between #2 and #3 that you described?
By doing it this way would i get any crosstalk between channels?
and is there any way to reduce the chances of getting it?
400-500 VA was about what i figured, thanks for confirming.
Hi,
hold on a moment,
one rectifier feeding parallel banks of smoothing caps is electrically just the same as one giant bank of smoothing caps. All one would achieve is moving them about a bit. It would be the same as option 1. The cheap and nasty way.
The difference going to option 2. is that each rectifier isolates each amplifier from the remainder. Assume for a moment that one or two channels suddenly play VERY loud. the voltage on the supply caps will fall rapidly as the transformer tries in vain to recharge them at the frequency mains provides (100Hz or 120Hz recharge pulses). the other amps with rectifiers and smoothing on low power will run from the charge in the smoothing banks and their voltage will fall more slowly and NOT feel any recharging pulses coming in because their stored voltage is above the pulled down voltage due to the very high demands from the greedy channels.
This has two effects.
a.) the lower power channels do not get fed dirty pulses on their supply lines and so the amp has an easy job trying to give a clean output.
b.) the gain of the lower power amps has not been affected by the low supply voltage. An ideal amp should have a gain that is completely independant of Vrail voltage but all suffer to some extent. This can cause the image to move and other distortion artefacts.
Eliminating both effects allows the quieter channels to play cleaner while the loud channels get on with playing loud. It is this same mechanism that allows true monoblocks to sound so stable and clean.
option 2. and option 3. differ in that separate secondaries will interact with each other less than common secondaries.
The regulation of a transformer is related to the primary losses, the secomdary losses and the iron losses. In a common secondary arrangement, all channels will suffer a reduced voltage due to all losses at any one moment. In a separate secondary transformer each rectifier will see the common primary and iron losses but their own secondary losses will be due to the current flow in their own circuit. i.e. less crosstalk.
All dual and multichannel amplifiers will suffer crosstalk. The audibility will depend on how closely the channels are coupled. Even a dual monoblock will suffer some cross talk due to load on the mains supplying the independant primaries. Whether this is audible/measureable will be debateable. I wonder if this is why some commentators put down the practice of sharing a cable with other ancilliaries fed from a multi outlet extension lead?
hold on a moment,
one rectifier feeding parallel banks of smoothing caps is electrically just the same as one giant bank of smoothing caps. All one would achieve is moving them about a bit. It would be the same as option 1. The cheap and nasty way.
The difference going to option 2. is that each rectifier isolates each amplifier from the remainder. Assume for a moment that one or two channels suddenly play VERY loud. the voltage on the supply caps will fall rapidly as the transformer tries in vain to recharge them at the frequency mains provides (100Hz or 120Hz recharge pulses). the other amps with rectifiers and smoothing on low power will run from the charge in the smoothing banks and their voltage will fall more slowly and NOT feel any recharging pulses coming in because their stored voltage is above the pulled down voltage due to the very high demands from the greedy channels.
This has two effects.
a.) the lower power channels do not get fed dirty pulses on their supply lines and so the amp has an easy job trying to give a clean output.
b.) the gain of the lower power amps has not been affected by the low supply voltage. An ideal amp should have a gain that is completely independant of Vrail voltage but all suffer to some extent. This can cause the image to move and other distortion artefacts.
Eliminating both effects allows the quieter channels to play cleaner while the loud channels get on with playing loud. It is this same mechanism that allows true monoblocks to sound so stable and clean.
option 2. and option 3. differ in that separate secondaries will interact with each other less than common secondaries.
The regulation of a transformer is related to the primary losses, the secomdary losses and the iron losses. In a common secondary arrangement, all channels will suffer a reduced voltage due to all losses at any one moment. In a separate secondary transformer each rectifier will see the common primary and iron losses but their own secondary losses will be due to the current flow in their own circuit. i.e. less crosstalk.
All dual and multichannel amplifiers will suffer crosstalk. The audibility will depend on how closely the channels are coupled. Even a dual monoblock will suffer some cross talk due to load on the mains supplying the independant primaries. Whether this is audible/measureable will be debateable. I wonder if this is why some commentators put down the practice of sharing a cable with other ancilliaries fed from a multi outlet extension lead?
ok, so what your saying makes sense....
to modify my previous post, i will have a bridge and regulation for each satelite speaker......
This should, if im understanding correctly, allow canels that are requiring alot of power, to draw from the transformer, while not bleeding off power from the others. As for "decoupling" the amplifiers, would using better quality diodes help eliminate that?
I planned on using schottky diodes for everything. Hoping that their speed, and low voltage drop would be a plus.
I realize that crosstalk is a dynamic thing and totally subjective but i want to do as much to eliminate it as i can.
THanks andrew
to modify my previous post, i will have a bridge and regulation for each satelite speaker......
This should, if im understanding correctly, allow canels that are requiring alot of power, to draw from the transformer, while not bleeding off power from the others. As for "decoupling" the amplifiers, would using better quality diodes help eliminate that?
I planned on using schottky diodes for everything. Hoping that their speed, and low voltage drop would be a plus.
I realize that crosstalk is a dynamic thing and totally subjective but i want to do as much to eliminate it as i can.
THanks andrew
understandable, that i suppose is a minor consideration at this point.....
Im trying to decide on a transformer to use. Id rather use a toroid but im pretty much open to anything cheap.....
I looked at this thread http://www.diyaudio.com/forums/showthread.php?s=&threadid=85346&perpage=10&highlight=&pagenumber=2
and the chart for supply voltage vs. output power for 4 6 and 8 ohm loads. the output power for a 4 ohm load peaks at around 27 volts with 55 watts while at the same voltage, the 8 ohm load would only get around 40 watts. (easily enough for the size room im trying to fill)
I would be working with speakers of varying ratings. Tweeter around 5 ohms, mid around 8 ohms and my sub is also 8 ohms.
now, since the tweeters generally wouldnt need to be driven as hard as the mids and sub, would i be able to use a higher voltage transformer and regulate, with a preamp, the signal to the tweeter amps? I was thinking about a 20-25 v +/- transformer for the supply. Again, the 600 VA range. HOwever, if i went with a higher supply 30-40v, 30 is the absolute top end of the graph for 4 ohm loads, which would probably kill the amps for the tweeters, however the output power for the 8 ohm loads would be aroudn 60 watts, which is great for bridging 2 into my 150 watt sub..... Many tradeoffs, maybe going with two transformers, one for the tweeters, and one for the sub/mids.
Im trying to decide on a transformer to use. Id rather use a toroid but im pretty much open to anything cheap.....
I looked at this thread http://www.diyaudio.com/forums/showthread.php?s=&threadid=85346&perpage=10&highlight=&pagenumber=2
and the chart for supply voltage vs. output power for 4 6 and 8 ohm loads. the output power for a 4 ohm load peaks at around 27 volts with 55 watts while at the same voltage, the 8 ohm load would only get around 40 watts. (easily enough for the size room im trying to fill)
I would be working with speakers of varying ratings. Tweeter around 5 ohms, mid around 8 ohms and my sub is also 8 ohms.
now, since the tweeters generally wouldnt need to be driven as hard as the mids and sub, would i be able to use a higher voltage transformer and regulate, with a preamp, the signal to the tweeter amps? I was thinking about a 20-25 v +/- transformer for the supply. Again, the 600 VA range. HOwever, if i went with a higher supply 30-40v, 30 is the absolute top end of the graph for 4 ohm loads, which would probably kill the amps for the tweeters, however the output power for the 8 ohm loads would be aroudn 60 watts, which is great for bridging 2 into my 150 watt sub..... Many tradeoffs, maybe going with two transformers, one for the tweeters, and one for the sub/mids.
Hi,
if designing a speaker system from scratch you must know the sensitivity as well as the reactance of each driver.
If the sensitivity of all drivers were the same then the drive voltage to each would be the same for the same volume out.
Many treble drivers are much more sensitive than their partners in a system and the result is they need less voltage to drive them to the same volume.
In a passive crossover this can sometimes be allowed for by adding a series resistor to the treble driver to reduce the treble output to match the other drivers.
If your treble driver is more sensitive than the other drivers then your power into small impedance problem goes away. All you need is at least a 2ohm resistor to reduce the sensitivity by about 2db and the driver reactances nearly match.
if designing a speaker system from scratch you must know the sensitivity as well as the reactance of each driver.
If the sensitivity of all drivers were the same then the drive voltage to each would be the same for the same volume out.
Many treble drivers are much more sensitive than their partners in a system and the result is they need less voltage to drive them to the same volume.
In a passive crossover this can sometimes be allowed for by adding a series resistor to the treble driver to reduce the treble output to match the other drivers.
If your treble driver is more sensitive than the other drivers then your power into small impedance problem goes away. All you need is at least a 2ohm resistor to reduce the sensitivity by about 2db and the driver reactances nearly match.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.