Okay...
1. I would have thought that separate transformers would be more expensive. If they're identical, you could probably parallel their output, but if you're using a number of smaller caps anyway I would just keep things separate.
2. If you parallel transformer outputs, if you wire them wrong you'll blow fuses like crazy. Like wiring batteries in parallel, if you don't hook + to +, they'll burn each other out.
You are planning for fuses, right?
3. Again, it's probably cheaper to buy bigger caps than a number of small ones. I'm assuming you're building the dual-supply version + and - power supplies. If you're building a single-supply version you have the speaker output capacitor to consider as well.
And to repeat: if you want to build a lower-power amplifier to save money on the transformer, you should reduce the voltage of the transformer as well as its VA rating.
1. I would have thought that separate transformers would be more expensive. If they're identical, you could probably parallel their output, but if you're using a number of smaller caps anyway I would just keep things separate.
2. If you parallel transformer outputs, if you wire them wrong you'll blow fuses like crazy. Like wiring batteries in parallel, if you don't hook + to +, they'll burn each other out.
You are planning for fuses, right?
3. Again, it's probably cheaper to buy bigger caps than a number of small ones. I'm assuming you're building the dual-supply version + and - power supplies. If you're building a single-supply version you have the speaker output capacitor to consider as well.
And to repeat: if you want to build a lower-power amplifier to save money on the transformer, you should reduce the voltage of the transformer as well as its VA rating.
arrgh
Power supplies are as clear as mud.
anyway.. 50VA ones are only rated to about 1A I think
I have lots of those around from those power bricks. the 100VA one they're selling is 68 bux, while the 50VA one is only 13 or so...
I don't understand that bit about lower power = lower voltage.
if I'm running 2 channels from 150W, whats the diff with running a channel on 50W?
the rating for the transformers should still remain 18-0-18 right? or should I use a 15V or lower trans?
I'm just planning to split the mains line to run 2 transformers in parallel, which will then connect to each channel seperately.
Power supplies are as clear as mud.
anyway.. 50VA ones are only rated to about 1A I think
I have lots of those around from those power bricks. the 100VA one they're selling is 68 bux, while the 50VA one is only 13 or so...
I don't understand that bit about lower power = lower voltage.
if I'm running 2 channels from 150W, whats the diff with running a channel on 50W?
the rating for the transformers should still remain 18-0-18 right? or should I use a 15V or lower trans?
I'm just planning to split the mains line to run 2 transformers in parallel, which will then connect to each channel seperately.
Look at the graph on p. 3 of the LM1875 datasheet of "Power Output vs. Supply voltage". This shows what voltage you need for a certain power level.
But more power requires both more voltage and more current. Add the power output to the power dissipation (another graph on the same page), and you have the total power the power supply must provide.
DC voltage from a full-wave bridge is about 1.4 times the AC voltage from the transformer. VA rating for the transformer should be at least double the DC power you need to supply. This plus the graphs should give you enough information.
By all means if you have the transformers or they are way cheaper then use 2 of them. I'd suggest keeping one per channel, each using a separate bridge rectifier and caps.
If the DC voltage is higher than what you need for a given power output, the rest is dissipated in the IC. If you use a +/- 25 VDC supply (from your 18-0-18 transformer), even with no power to the speaker, the IC will be dissipating about 8 watts (see the graph). Never run these ICs without a heat sink attached! It looks like you can get about 8-9 watts per channel; each IC will then be dissipating about 17 watts. 25 watts total, 50 VA transformer.
Notice that if you ran from a +/- 20 volt supply and produce 10 watts per channel, the power dissipation per IC will only be 12 watts. 22 watts total, 44 VA (14-0-14 volt) transformer. You're actually drawing less power from the supply to produce more power into the speaker. Weird, eh?
I hope that made sense; email me if it didn't.
But more power requires both more voltage and more current. Add the power output to the power dissipation (another graph on the same page), and you have the total power the power supply must provide.
DC voltage from a full-wave bridge is about 1.4 times the AC voltage from the transformer. VA rating for the transformer should be at least double the DC power you need to supply. This plus the graphs should give you enough information.
By all means if you have the transformers or they are way cheaper then use 2 of them. I'd suggest keeping one per channel, each using a separate bridge rectifier and caps.
If the DC voltage is higher than what you need for a given power output, the rest is dissipated in the IC. If you use a +/- 25 VDC supply (from your 18-0-18 transformer), even with no power to the speaker, the IC will be dissipating about 8 watts (see the graph). Never run these ICs without a heat sink attached! It looks like you can get about 8-9 watts per channel; each IC will then be dissipating about 17 watts. 25 watts total, 50 VA transformer.
Notice that if you ran from a +/- 20 volt supply and produce 10 watts per channel, the power dissipation per IC will only be 12 watts. 22 watts total, 44 VA (14-0-14 volt) transformer. You're actually drawing less power from the supply to produce more power into the speaker. Weird, eh?
I hope that made sense; email me if it didn't.
Looking at those graphs
it seems that the THD is lowest at slightly more than 10W or so and 10 watts of output power means I need 15V & 60mA into that thing, I need to dissipate 7-13W of heat. Is that all correct?
If I want my 20W then I'll have to get a 23V transformer and 65mA.
Now I'm getting somewhere.. I can probably take out the 15V transformer in my CPU powersupply, and use their bridge rectifier. Caps should also be aplenty inside there.
Alternatively, would you reccomend I use one of those power bricks? I have one thats 15V, and I could just run the DC to a bunch of caps in parallel for smoothing. It'd also provide for easier wiring.
Question. The input on the ESP plans say 25W, with 1A fuses.
and I see no way I can actually reach 1A without blowing the thing. If I use the power bricks method, does it mean that if I buy one of those adjustable bricks, I can just up the voltage for more power?
: ^) thanks alot,
I'll be done with the soldering tonight, leaving the testing for the weekend. Any ideas for boxing up the project? its on a pcb thats 12cm by 8 cm... I'm thinking of mounting it on a wooden cutting board lying around.. easy to drag, and very stable. + the transformers look cool.. with that copper gleam
it seems that the THD is lowest at slightly more than 10W or so and 10 watts of output power means I need 15V & 60mA into that thing, I need to dissipate 7-13W of heat. Is that all correct?
If I want my 20W then I'll have to get a 23V transformer and 65mA.
Now I'm getting somewhere.. I can probably take out the 15V transformer in my CPU powersupply, and use their bridge rectifier. Caps should also be aplenty inside there.
Alternatively, would you reccomend I use one of those power bricks? I have one thats 15V, and I could just run the DC to a bunch of caps in parallel for smoothing. It'd also provide for easier wiring.
Question. The input on the ESP plans say 25W, with 1A fuses.
and I see no way I can actually reach 1A without blowing the thing. If I use the power bricks method, does it mean that if I buy one of those adjustable bricks, I can just up the voltage for more power?
: ^) thanks alot,
I'll be done with the soldering tonight, leaving the testing for the weekend. Any ideas for boxing up the project? its on a pcb thats 12cm by 8 cm... I'm thinking of mounting it on a wooden cutting board lying around.. easy to drag, and very stable. + the transformers look cool.. with that copper gleam
To get 10 watts output, with a +/- 15 V power supply, the IC will be dissipating about 6 watts (with an 8-ohm speaker). The power supply will have to provide 16 watts at 30 volts, or 0.53 Amps, per channel.
When I say +/- 15 volt supply, this means a total of 30 volts. If you have a 15 volt "brick" (this is a DC power supply?), you'll need two of them in series to provide the +/- 15 V.
You can run the IC from a single power supply (a different circuit diagram applies), but 15 V is getting pretty low for this IC. National has other parts that will be happier at this lower DC voltage (LM383 comes to mind).
When I say +/- 15 volt supply, this means a total of 30 volts. If you have a 15 volt "brick" (this is a DC power supply?), you'll need two of them in series to provide the +/- 15 V.
You can run the IC from a single power supply (a different circuit diagram applies), but 15 V is getting pretty low for this IC. National has other parts that will be happier at this lower DC voltage (LM383 comes to mind).
Sounds good. If you use one transformer per channel, I think you'll be able to get about 10 watts per channel.
Fuses: 1 amp is probably okay, but you may end up having to increase it depending on how loud you play your music.
Enclosure: whatever turns your crank. It's you that's looking at it. The cutting board idea sounds cool. The SAF (spousal acceptance factor) in my house would inhibit having a bunch of parts on display. Should be lots of ideas in the thread archives at this site.
Fuses: 1 amp is probably okay, but you may end up having to increase it depending on how loud you play your music.
Enclosure: whatever turns your crank. It's you that's looking at it. The cutting board idea sounds cool. The SAF (spousal acceptance factor) in my house would inhibit having a bunch of parts on display. Should be lots of ideas in the thread archives at this site.
For the ppl following this project, DON"T save money by buying a smaller trans. Its not worth the extra cost.
I made that mistake.. sighz.. now stuck with a 18-0-18 1.1A
think I'll use it for centre channel amplification or something
That's cos the bridge rectifiers and caps can easily go to 13+ dollars for this project. and that's for only one channel. If you wanna do 2 channels then you have to do 13+ Again + the cost of the trans.. for me its 10
so I'll spend 13+ extra
I think I'll get another trans 18-0-18 3A and wire the two in parallel for 4A total... But more work and more expense
so don't do something stupid like this.
I also bought a 10cm by 5cm heatsink,with fins 4 cm tall
hope it'll work well. PaulB, should I lap the LM1875? or should I just leave it be? I have no heatsink compound and 7.50 for a tube seems too much since i just need a small dab.
2nd, Heatshrink tubing.. required for the transformer connections? or should I just use electricians tape or not bother at all.
3)Mistakes. I didn't read the schematics closely enough, so my 100nF and 220nF caps are electrolytic instead of plastic
should I go buy the plastic ones or should I just use the electrolytic.. If I do, which direction should I connect the electrolytics?
4)I'm using bare speaker wire. Should I buy those expensive nakamichi binding posts or should i just use those springy cheapo things? the Red and black thing that traps the wires
Thanks
I made that mistake.. sighz.. now stuck with a 18-0-18 1.1A
think I'll use it for centre channel amplification or something
That's cos the bridge rectifiers and caps can easily go to 13+ dollars for this project. and that's for only one channel. If you wanna do 2 channels then you have to do 13+ Again + the cost of the trans.. for me its 10
so I'll spend 13+ extra
I think I'll get another trans 18-0-18 3A and wire the two in parallel for 4A total... But more work and more expense
so don't do something stupid like this.
I also bought a 10cm by 5cm heatsink,with fins 4 cm tall
hope it'll work well. PaulB, should I lap the LM1875? or should I just leave it be? I have no heatsink compound and 7.50 for a tube seems too much since i just need a small dab.
2nd, Heatshrink tubing.. required for the transformer connections? or should I just use electricians tape or not bother at all.
3)Mistakes. I didn't read the schematics closely enough, so my 100nF and 220nF caps are electrolytic instead of plastic
should I go buy the plastic ones or should I just use the electrolytic.. If I do, which direction should I connect the electrolytics?
4)I'm using bare speaker wire. Should I buy those expensive nakamichi binding posts or should i just use those springy cheapo things? the Red and black thing that traps the wires
Thanks
Ah, the joy. Don't get discouraged, this is your first one, right?
"should I lap the LM1875?" - don't understand. It NEEDS a heatsink. The metal tab on the IC is connected to one of the pins, so if you don't insulate it from the heatsink, the heatsink will need to be insulated from everything else (including the box - this is a pain). You normally use a mica or other kind of pad to insulate the part from the heatsink; this interferes with heat transfer so you use the compound to fill in the air gaps and help transfer. Radio Shack used to sell a small tube size.
2. electrical tape is okay, heat shrink is expensive and just looks better.
3. I don't think the values of your caps are correct - 100 nF and 220nF are unlikely to be electrolytic. You probably have 100 uF and 220 uF if they are metal and have a + or - sign marked on them. So you'll need to buy the correct ones.
4. I'd stay cheap on the speaker posts, this is a "cheapo" amp, remember.
What did you decide to build it on? You mentioned a PCB in a previous post, did you buy one?
"should I lap the LM1875?" - don't understand. It NEEDS a heatsink. The metal tab on the IC is connected to one of the pins, so if you don't insulate it from the heatsink, the heatsink will need to be insulated from everything else (including the box - this is a pain). You normally use a mica or other kind of pad to insulate the part from the heatsink; this interferes with heat transfer so you use the compound to fill in the air gaps and help transfer. Radio Shack used to sell a small tube size.
2. electrical tape is okay, heat shrink is expensive and just looks better.
3. I don't think the values of your caps are correct - 100 nF and 220nF are unlikely to be electrolytic. You probably have 100 uF and 220 uF if they are metal and have a + or - sign marked on them. So you'll need to buy the correct ones.
4. I'd stay cheap on the speaker posts, this is a "cheapo" amp, remember.
What did you decide to build it on? You mentioned a PCB in a previous post, did you buy one?
Oh I just used perfboard, and joined the leads together with solder... saved me lotsa trouble.
Regarding the heatsinking stuff... lapping means using sandpaper and rubbing it until its flat. I think its a computer term, used to mean sanding till you expose the copper of the heat spreader.
Heatsinks are a pain. I can't get them to stand on the perfboard properly. as in... they're elevated off the PCB and I'm afraid of the stresses on the IC. They're not soldered in yet. as for electrical insulation, I'm thinking of buying those thermal interface pads. but I've heard that they have aluminum powder as a heat conductor, so they're out too. How much do mica sheets cost anyway? They also mean that I need heatsink paste right? : ^(
2 seperate large heatsinks would be much cheaper. Which is probably the path I'll take
Anyone knows why they actually connected the pins to the metal block? isn't that counterintuitive?
Regarding the heatsinking stuff... lapping means using sandpaper and rubbing it until its flat. I think its a computer term, used to mean sanding till you expose the copper of the heat spreader.
Heatsinks are a pain. I can't get them to stand on the perfboard properly. as in... they're elevated off the PCB and I'm afraid of the stresses on the IC. They're not soldered in yet. as for electrical insulation, I'm thinking of buying those thermal interface pads. but I've heard that they have aluminum powder as a heat conductor, so they're out too. How much do mica sheets cost anyway? They also mean that I need heatsink paste right? : ^(
2 seperate large heatsinks would be much cheaper. Which is probably the path I'll take
Anyone knows why they actually connected the pins to the metal block? isn't that counterintuitive?
Okay, no, don't "lap" the heatsinks if they're already finished, unless they're bare metal. The usual black anodizing is okay as-is.
Yes, you'll have to figure out a way to stabilize the heatsink. The ESP PCB mounts the IC right on the edge of the board so the heatsink is not on the PCB.
Hey, if the heatsink is right on the board and you can ensure the heatsink is insulated from everything else, I'd consider skipping the pads. You should still use the compound, but you may be able to get away without it. Try it without - if the IC gets too hot to put your finger on it, that's too hot.
You'll probably end up taking it on and off the heatsink a few times anyway, and the compound gets messy.
Yes, you'll have to figure out a way to stabilize the heatsink. The ESP PCB mounts the IC right on the edge of the board so the heatsink is not on the PCB.
Hey, if the heatsink is right on the board and you can ensure the heatsink is insulated from everything else, I'd consider skipping the pads. You should still use the compound, but you may be able to get away without it. Try it without - if the IC gets too hot to put your finger on it, that's too hot.
You'll probably end up taking it on and off the heatsink a few times anyway, and the compound gets messy.
Gonna get another 50VA 18-0-18 for a total of 18x(4.1A) = 73.8VA And thus 1.48A at 50V, 0.75A per channel, putting me comfortably at the 25V supplies.
Considering getting another 10 dollars worth of caps cos I'm running so close to the limit.
Question. I'm worried about connecting the outputs of the two transformers to one bridge rectifier. Will there be like phase problems or anythign making the bridge rectifier run hotter than rated? or should I just buy another rectifier and run both directly to the resistor arrays in parallel?
Mounting the stuff is still a problem. Think I'll mount the T0-220 the wrong way round, leaving it up in the air where I can solder the leads to wires. Think I'll also hang the 0.1uF cap in the air to prevent oscillation. I'm gonna get Heatsink paste and the mica plates since there's no decent solution to the thing. Using Elliot's heatsink calculator, I'm at 1.28 for that heatsink which is pretty close to the 1C/W I need for disspating 32W of heat.
[Edited by downhere on 10-14-2001 at 07:51 AM]
Considering getting another 10 dollars worth of caps cos I'm running so close to the limit.
Question. I'm worried about connecting the outputs of the two transformers to one bridge rectifier. Will there be like phase problems or anythign making the bridge rectifier run hotter than rated? or should I just buy another rectifier and run both directly to the resistor arrays in parallel?
Mounting the stuff is still a problem. Think I'll mount the T0-220 the wrong way round, leaving it up in the air where I can solder the leads to wires. Think I'll also hang the 0.1uF cap in the air to prevent oscillation. I'm gonna get Heatsink paste and the mica plates since there's no decent solution to the thing. Using Elliot's heatsink calculator, I'm at 1.28 for that heatsink which is pretty close to the 1C/W I need for disspating 32W of heat.
[Edited by downhere on 10-14-2001 at 07:51 AM]
If the transformers are identical, you should be able to parallel the secondaries (if the identical leads from each transformer are connected). But it sounds like you want to use two different transformers? The 18-0-18 @ 1.1A (about 40 VA) and a 50 VA unit? It may be difficult to tell how to connect the secondaries together; they have to be "in phase". Also, they may not share the load very well; you may have to add small series resistors to each.
I would use two bridge rectifiers, and combine the outputs after them. It's easy to make sure they're wired correctly: + to +, - to -.
"or should I just buy another rectifier and run both directly to the resistor arrays in parallel?" you mean capacitor arrays, right?
It sounds like you've learned a lot doing this.
I would use two bridge rectifiers, and combine the outputs after them. It's easy to make sure they're wired correctly: + to +, - to -.
"or should I just buy another rectifier and run both directly to the resistor arrays in parallel?" you mean capacitor arrays, right?
It sounds like you've learned a lot doing this.
Yup haha, I've always been a digital kind of guy. Used to do robotics and make cute little LEGO robots. Still not too used to analogue circuits and AC power. VA is measured by max voltage x current right? but if I take power from say +18 to 0 instead of +18 -18 the VA should still be the same right? As in, I'll get half the AC waveform, and half the voltage, and half the current.
I think I'll screw the one of the extra heatsink holes into the side of the chopping block as support for the heatsink and make a clear acrylic box around the whole chopping board, leaving two rectangular holes for the heatsinks. I'm gonna buy the parts and leave the assembly for december or so. Using this forum as a kind of diary of what to buy and how to assemble. I can kinda imagine it now, two transformer blocks, one large one small, a PCB with 20 1000uF caps and another PCB with the circuit and 2 heatsinks. and inputs and outputs + LED + switch on the front side.
Thanks alot PaulB for the help over these few days : ^) Will post pics when I'm finally done.
Next project is the 63W single chip amp, when I get money from my job in December.
I think I'll screw the one of the extra heatsink holes into the side of the chopping block as support for the heatsink and make a clear acrylic box around the whole chopping board, leaving two rectangular holes for the heatsinks. I'm gonna buy the parts and leave the assembly for december or so. Using this forum as a kind of diary of what to buy and how to assemble. I can kinda imagine it now, two transformer blocks, one large one small, a PCB with 20 1000uF caps and another PCB with the circuit and 2 heatsinks. and inputs and outputs + LED + switch on the front side.
Thanks alot PaulB for the help over these few days : ^) Will post pics when I'm finally done.
Next project is the 63W single chip amp, when I get money from my job in December.
The total VA will be the same, but if you're only using half of the secondary winding (18 - 0), you'll get half the voltage but the full current. Therefore you can only use half the VA. Look at using a full-wave CT instead of a bridge rectifier if you only want to use half the voltage, then you can use the full VA.
You're welcome. I've enjoyed hearing about your progress. Looking forward to pictures.
I'm about to replace my LM1875-based amp with Rod Elliott's Project 3A 60W discrete amp. It's an addictive hobby, and this is a great site to get help and inspiration. Cheers to Jason for putting it together and keeping it running!
You're welcome. I've enjoyed hearing about your progress. Looking forward to pictures.
I'm about to replace my LM1875-based amp with Rod Elliott's Project 3A 60W discrete amp. It's an addictive hobby, and this is a great site to get help and inspiration. Cheers to Jason for putting it together and keeping it running!
My LM1875 amp
My amp is based on an article from Popular Electronics many years ago. It uses LM1875s, each in the feedback loop of an op-amp (AD712), two per channel in a bridged configuration.
The power supply uses a 288 VA transformer, 18-0-18 V, with 12,500 uF per channel.
I did the circuit boards myself, and used a rackmount chassis from a place called Mark V Electronics (not sure if they're still around).
I also built a "standard" LM1875 amp into a desktop speaker enclosure. It uses a 16 VAC "wall-wart" transformer to drive a 40-ohm speaker.
I'll try to post pics when I borrow a digital camera.
My amp is based on an article from Popular Electronics many years ago. It uses LM1875s, each in the feedback loop of an op-amp (AD712), two per channel in a bridged configuration.
The power supply uses a 288 VA transformer, 18-0-18 V, with 12,500 uF per channel.
I did the circuit boards myself, and used a rackmount chassis from a place called Mark V Electronics (not sure if they're still around).
I also built a "standard" LM1875 amp into a desktop speaker enclosure. It uses a 16 VAC "wall-wart" transformer to drive a 40-ohm speaker.
I'll try to post pics when I borrow a digital camera.
Sounds interesting
I'll be waiting for pics.
now, anyone has any ideas on an amp case that cannot be drilled?
The tools I have are a lousy handsaw, and no drill. Means no holes watsoever.
I'm thinking of mounting the PCB by slotting it into rubber feet with slits in them.(the taller 1 inch type ones)
the transformer I'll duct tape to a board. the output and input jacks I'll duct tape too
question. I have no chassis. Where do I connect the leads that say chassis ground to?
I'll be waiting for pics.
now, anyone has any ideas on an amp case that cannot be drilled?
The tools I have are a lousy handsaw, and no drill. Means no holes watsoever.
I'm thinking of mounting the PCB by slotting it into rubber feet with slits in them.(the taller 1 inch type ones)
the transformer I'll duct tape to a board. the output and input jacks I'll duct tape too
question. I have no chassis. Where do I connect the leads that say chassis ground to?
Chassis ground
The chassis ground connection doesn't apply if you don't have a metal case. But for safety reasons (in case of an accidental short to the power line) it should connect to all exposed metal parts. That way you blow a fuse or circuit breaker instead of electrocuting yourself.
With a metal case, it also provides shielding from radio interference, etc.
The chassis ground connection doesn't apply if you don't have a metal case. But for safety reasons (in case of an accidental short to the power line) it should connect to all exposed metal parts. That way you blow a fuse or circuit breaker instead of electrocuting yourself.
With a metal case, it also provides shielding from radio interference, etc.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.