Re: minimum gain
Volume controls usually don't actually change the gain of the amplifier.
Make Rf a fixed resistor set for a reasonable gain (maybe 20) and just set up a pot as a voltage divider on the input (often called a "passive preamp") Ipod's signal to one end of the pot, ground on the other, and input of the opamp from the wiper.
You can turn it all the way to 0, no problem.
Full volume will be whatever level the ipod's control is set to.
As for opamps, the 741 is old and slow and probably making your amp sound much worse than it actually is.
I would probably just leave it off, since the Ipod should have enough current capability to properly drive the 3886 directly. Even if it is a bit weak, it'll probably distort less than a 741.
-Nick
melbsp said:
Volume controls usually don't actually change the gain of the amplifier.
Make Rf a fixed resistor set for a reasonable gain (maybe 20) and just set up a pot as a voltage divider on the input (often called a "passive preamp") Ipod's signal to one end of the pot, ground on the other, and input of the opamp from the wiper.
You can turn it all the way to 0, no problem.
Full volume will be whatever level the ipod's control is set to.As for opamps, the 741 is old and slow and probably making your amp sound much worse than it actually is.
I would probably just leave it off, since the Ipod should have enough current capability to properly drive the 3886 directly. Even if it is a bit weak, it'll probably distort less than a 741.
-Nick
Re: Re: thanks Greg
Wow, I think 28Vac is pushing it a tad. The maximum supply differential the chip is rated for is 84V.
With the transformer I'm using at home rated 600VA w/ 2x25VAC secondaries I get 40VDC
Based on that It seems to me that a 28V would put me over the 84V rating.
-Nick
AndrewT said:
Wow, I think 28Vac is pushing it a tad. The maximum supply differential the chip is rated for is 84V.
With the transformer I'm using at home rated 600VA w/ 2x25VAC secondaries I get 40VDC
Based on that It seems to me that a 28V would put me over the 84V rating.
-Nick
Hi,
there's something odd here.
40Vdc implies about 15% regulation, about three times worse than one would expect from a 600VA transformer.
I would expect 28Vac to give 40.8Vdc on open circuit with one rectifier and about 40Vdc with twin rectifiers.
Once the amp is coupled and drawing quiescent current, both these voltages will drop a bit.
When the mains is above nominal supply voltage these voltages will rise, but hopefully not too much and not for too long.
When you measure the transformer AC voltage and DC voltages what is the corresponding mains voltage?
Re: Re: Re: Re: thanks Greg
Yeah, I think my transformers are a bit higher voltage than spec. Still, that puts it right at the maximum voltage rating. If the line is overvoltage at all it'll be beyond the spec of the chip.
I would probably suggest ~20V secondaries assuming you're not aiming for maximum power. I ran my old amp from 18V into some 8ohm speakers and it had plenty of power.
At 20V it'll be a little happier with difficult speakers.
If you're aiming for max power, go for it, but remember that you're stuck with 8 ohm loads.
-Nick
MJL21193 said:
Yeah, I think my transformers are a bit higher voltage than spec. Still, that puts it right at the maximum voltage rating. If the line is overvoltage at all it'll be beyond the spec of the chip.
I would probably suggest ~20V secondaries assuming you're not aiming for maximum power. I ran my old amp from 18V into some 8ohm speakers and it had plenty of power.
At 20V it'll be a little happier with difficult speakers.
If you're aiming for max power, go for it, but remember that you're stuck with 8 ohm loads.
-Nick
input attenuator
If you want to use only the knob on the unit to control loudness, you could simply gain the amp for the maximum and use a potentiometer or step attenuator for the input signal and leave the gain alone.
This way you could achieve the total range of loudness with the smaller heatsink.
If you want to use only the knob on the unit to control loudness, you could simply gain the amp for the maximum and use a potentiometer or step attenuator for the input signal and leave the gain alone.
This way you could achieve the total range of loudness with the smaller heatsink.
if that's 27VAC x 2 then it's a bit high, and it'll run pretty warm (and you'll have to stick to 8 ohm speakers). You'll have to give it a test run to see. I ran some 3875s on 25VAC secondaries, and it did get quite warm, though not enough to be a problem on the chassis I was using. Currently I'm running mine on a pair of 14.4V transformers, and it doesn't even get warm enough to feel it.
It that's 27VDC per rail, it should be no problem.
It that's 27VDC per rail, it should be no problem.
Don't forget that the chip's tab (on non-insulated versions) is at -V supply potential, not ground. You need to electrically isolate the chip from the chassis (or any heatsink) using appropriate measures, like silicone or mica insulating washers and plastic mounting screws or bushings.
heat sink size!
found on datasheet a 'Power Dissipation vs Output Power' curve for 8ohm load and it references +/-40V power supply!
I will limit mine to 30W for speaker...
going to look at some heat sinks now and double check the thermal resistance numbers.
http://www.ti.com/lit/ds/symlink/lm3886.pdf
Figure 36, then Figure 34...
found on datasheet a 'Power Dissipation vs Output Power' curve for 8ohm load and it references +/-40V power supply!
I will limit mine to 30W for speaker...
going to look at some heat sinks now and double check the thermal resistance numbers.
http://www.ti.com/lit/ds/symlink/lm3886.pdf
Figure 36, then Figure 34...
Okay, I ran my transformer, rectifier and 10,000uF caps with three different loads!
The output voltage unloaded is about +/-40, but it runs from 30 to 34 loaded.
Should give me enough overhead for high peak to average signals with 15.5V rms average?!
Chart says thermal resistance of around 3.
Found a nice, small heat sink with thermal resistance of 2.1 and is only about 2.5" square. Wakefield p/n 517-95AB.
guess I'm running out of excuses...need to finish up ChipAmp pcb stuffing and run it a little. going to use 8ohm resistors at first?
The output voltage unloaded is about +/-40, but it runs from 30 to 34 loaded.
Should give me enough overhead for high peak to average signals with 15.5V rms average?!
Chart says thermal resistance of around 3.
Found a nice, small heat sink with thermal resistance of 2.1 and is only about 2.5" square. Wakefield p/n 517-95AB.
guess I'm running out of excuses...need to finish up ChipAmp pcb stuffing and run it a little. going to use 8ohm resistors at first?
The maximum voltage for the 3886 is 84V (±42Vdc).
What is the mains voltage range for your supply?
What was the mains voltage when you measured 40Vdc?
It would take an increase of just 5% on your mains voltage to exceed the 84V absolute maximum for your chipamp.
I recommend that you double the National guide value from their design chart Figure 34.
If @ ±40V & 8r0 test load with Ta=25°C, then National show 1.9C/W
I would suggest you use 0.95C/W per channel, for 8ohms @ ±40V with Ta<=30°C
From Fig34 I see 3C/W for 69V @ 25°C & 8r0 test load. Is that what your read?
Last edited:
Thanks, AndrewT! So good to have knowledgeable feedback!
First, AC Mains: just read VAC @ 124.1 with output +/- 39.4 VDC Unloaded, +/- 34.7 VDC at 0.54 ADC. so the output drops quickly with load.
I don't know how much higher the mains would get...this is already 10% high, i thought!
Also, the datasheet says the Max Supply 'No Signal' is 94 VDC... 'Signal' is 84. I feel like i'm splitting hairs with my poor amplifier's life!
I agonized at the transformer choice...this Triad line goes from 48CT down to 36CT. Maybe i need to go find an in between? there are some 22-0-22 VAC out there, but not many. what do you see for successful LM3886 projects? I do want maximum peak for my 30W speaker, clean signal, which is 15.5VACrms. i thought 2:1 is minimum. is LM3886 bad choice for this?
This is a 'plywood prototype' at this point...maybe i'll try it anyway...Would the $6 amplifier chip fail or would that cause the $50 speaker to fail? Sure wouldn't want the amplifier output stage to short and put that DC power supply across my speaker...my $50 speaker protected my 10cent fuse?
thx,
tom
First, AC Mains: just read VAC @ 124.1 with output +/- 39.4 VDC Unloaded, +/- 34.7 VDC at 0.54 ADC. so the output drops quickly with load.
I don't know how much higher the mains would get...this is already 10% high, i thought!
Also, the datasheet says the Max Supply 'No Signal' is 94 VDC... 'Signal' is 84. I feel like i'm splitting hairs with my poor amplifier's life!
I agonized at the transformer choice...this Triad line goes from 48CT down to 36CT. Maybe i need to go find an in between? there are some 22-0-22 VAC out there, but not many. what do you see for successful LM3886 projects? I do want maximum peak for my 30W speaker, clean signal, which is 15.5VACrms. i thought 2:1 is minimum. is LM3886 bad choice for this?
This is a 'plywood prototype' at this point...maybe i'll try it anyway...Would the $6 amplifier chip fail or would that cause the $50 speaker to fail? Sure wouldn't want the amplifier output stage to short and put that DC power supply across my speaker...my $50 speaker protected my 10cent fuse?
thx,
tom
AndrewT thermal resistance!
I ran three different power supply load points and carefully looked at the charts. i will never have full power at +/-40...the transformer sags too much.
2x20ohm load yielded +/-30VDC; that would be audio power of 26Watts. Chip dissipation of 26Watts, thermal requirement at 40degC ambient 3.2degC/Watt. so my heatsink at 2.1 i thought should be okay...?
Of course, now i'm doubting my chip choice!!! HELP!!!
i don't have a huge investment in the LM3886, and will gladly keep the two ChipAmp PCBs for other projects...now i'm not even convinced about my 48VCT transformer...boy i love design!
But i am determined to play guitar through these two ten inch 30W Celestion speakers. Celestion said that higher peak voltages are not a problem as long as the voice coil heating stays below the 15.5Vrms value...
This still seems almost 'right-sized', please advise!
thx,
I ran three different power supply load points and carefully looked at the charts. i will never have full power at +/-40...the transformer sags too much.
2x20ohm load yielded +/-30VDC; that would be audio power of 26Watts. Chip dissipation of 26Watts, thermal requirement at 40degC ambient 3.2degC/Watt. so my heatsink at 2.1 i thought should be okay...?
Of course, now i'm doubting my chip choice!!! HELP!!!
i don't have a huge investment in the LM3886, and will gladly keep the two ChipAmp PCBs for other projects...now i'm not even convinced about my 48VCT transformer...boy i love design!
But i am determined to play guitar through these two ten inch 30W Celestion speakers. Celestion said that higher peak voltages are not a problem as long as the voice coil heating stays below the 15.5Vrms value...
This still seems almost 'right-sized', please advise!
thx,
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.