I assume you mean this one?
It shows 80w @ +/- 40v 😱
Has anyone ever got this much power from a single 3886?
So for 60w into 8 ohm I need approx 33/34v at the chips pins.
Is there anything else to consider?
yes,
1.) cooling and
2.) PSU sag under full load and
3.) PSU overvoltage when mains is at 254Vac and the load is light or the fuses have blown to one channel.
1.) cooling and
2.) PSU sag under full load and
3.) PSU overvoltage when mains is at 254Vac and the load is light or the fuses have blown to one channel.
Thanks.
I think I'll leave the CA as-is other than maybe installing a bigger trafo in the future.
Mike.
I think I'll leave the CA as-is other than maybe installing a bigger trafo in the future.
Mike.
Not sure if this is relevant to the LM3886, but I use my LM3875's at 42v per rail (30+30v tranny)
So thats about 45v dc at the chip then?
I'm definitely interested in raising the voltage to obtain more power (my TDL's are hungry beasts!) but not at the expense of sound quality due to increased distortion.
I will have to get my head around all those graphs! 😕
Mike. 🙂
I'm definitely interested in raising the voltage to obtain more power (my TDL's are hungry beasts!) but not at the expense of sound quality due to increased distortion.
I will have to get my head around all those graphs! 😕
Mike. 🙂
Ok.
The CA 340A SE service manual states the output from the trafo is 24V. based on this we have:
24v AC x 1.414 = 33.94v DC.
According to the graph above I have approx 60w per channel into 8 ohms. However the service manual also states a claimed output of 45w/pc with 8 ohm speakers. Could this difference be due to the power supply not meeting the requirements necessary to produce the 60w we see on the graph? If so whats needed to get the 60w figure? (Trafo too small in VA perhaps)?
Mike.
The CA 340A SE service manual states the output from the trafo is 24V. based on this we have:
24v AC x 1.414 = 33.94v DC.
According to the graph above I have approx 60w per channel into 8 ohms. However the service manual also states a claimed output of 45w/pc with 8 ohm speakers. Could this difference be due to the power supply not meeting the requirements necessary to produce the 60w we see on the graph? If so whats needed to get the 60w figure? (Trafo too small in VA perhaps)?
Mike.
Hi,
measure the voltage across the chipamp supply pins when the input is shorted. Check the mains voltage at the same time.
Now add an 8r0 load to one channel and drive the specified 45W into that 8r0. Measure the chip supply voltage and the mains voltage.
Now add an 8r0 load to the other channel and drive the specified 45+45W into dual 8r0. Measure the chip voltages and the mains voltage.
What is the PSU sag when delivering 45W? and 90W?
You can also measure an approximation of the PSU ripple voltage using a suitable DMM set to 200mVac. Ripple peak to peak is approximately 3times the AC(~rms) reading from the DMM. It will vary between Quiescent and 45W and 90W.
measure the voltage across the chipamp supply pins when the input is shorted. Check the mains voltage at the same time.
Now add an 8r0 load to one channel and drive the specified 45W into that 8r0. Measure the chip supply voltage and the mains voltage.
Now add an 8r0 load to the other channel and drive the specified 45+45W into dual 8r0. Measure the chip voltages and the mains voltage.
What is the PSU sag when delivering 45W? and 90W?
You can also measure an approximation of the PSU ripple voltage using a suitable DMM set to 200mVac. Ripple peak to peak is approximately 3times the AC(~rms) reading from the DMM. It will vary between Quiescent and 45W and 90W.
Morning Andrew.
When doing this do I just turn the volume to max in order to get the full power while taking my readings?
Mike.
When doing this do I just turn the volume to max in order to get the full power while taking my readings?
Mike.
no, calculate the AC voltage required to produce 45W into 8r0.
apply a sinewave to the input. Increase the volume until the output voltage matches the calculated Vac.
Switch off or attenuate (-20dB) the input signal before the amplifier overheats. Leave it to cool. We are not testing the chip or the protections inside the chip, so we don't want to run it warm nor hot.
Now DMM the PSU, first 200Vdc and switch on the signal. Take your Vdc reading and immediately switch down the signal. Let the amp cool. This is important, the load and the amp must be cool at the start of every measurement. Run the 45W output for as short a time as you can manage to get a stable voltage reading on your DMM probably less than 2 seconds.
Now add the 8r0 Load.
repeat the Vdc check with the load connected. and switch off the sgnal to allow to cool.
Now turn your DMM to 200Vac and take the AC reading across the PSU.
If it's less than 2Vac switch the DMM to 20Vac. Take the ripple reading again. If it's less than 0.2Vac then turn the DMM to 2Vac and take a third reading. If it's less than 20mVac, then set the DMM to 200mVac and take the final ripple reading at quiescent state it should be <50mV. It might be a lot less.
Now send 45W to a single 8r0 and note the ripple reading. It will have increased maybe ten fold.
You're ready to test the other channel. Repeat the whole single channel procedure for the 2nd channel. The numbers should be almost identical.
Now apply both 8r0 loads and feed both channels with that sinewave signal.
Noting the PSU DC and AC readings.
Allow plenty of time to cool between each measurement. Small heatsinks and power testing increase temperatures very rapidly.
Come back with a table of numbers.
apply a sinewave to the input. Increase the volume until the output voltage matches the calculated Vac.
Switch off or attenuate (-20dB) the input signal before the amplifier overheats. Leave it to cool. We are not testing the chip or the protections inside the chip, so we don't want to run it warm nor hot.
Now DMM the PSU, first 200Vdc and switch on the signal. Take your Vdc reading and immediately switch down the signal. Let the amp cool. This is important, the load and the amp must be cool at the start of every measurement. Run the 45W output for as short a time as you can manage to get a stable voltage reading on your DMM probably less than 2 seconds.
Now add the 8r0 Load.
repeat the Vdc check with the load connected. and switch off the sgnal to allow to cool.
Now turn your DMM to 200Vac and take the AC reading across the PSU.
If it's less than 2Vac switch the DMM to 20Vac. Take the ripple reading again. If it's less than 0.2Vac then turn the DMM to 2Vac and take a third reading. If it's less than 20mVac, then set the DMM to 200mVac and take the final ripple reading at quiescent state it should be <50mV. It might be a lot less.
Now send 45W to a single 8r0 and note the ripple reading. It will have increased maybe ten fold.
You're ready to test the other channel. Repeat the whole single channel procedure for the 2nd channel. The numbers should be almost identical.
Now apply both 8r0 loads and feed both channels with that sinewave signal.
Noting the PSU DC and AC readings.
Allow plenty of time to cool between each measurement. Small heatsinks and power testing increase temperatures very rapidly.
Come back with a table of numbers.
The actual difference in volume between 50W and 60W is surely going to be tiny... nothing I would spend much time worrying about
Guys,
I was just curious if I could get an extra 15 watts out of the amp. I dont know how much difference that would make, I was just wondering. I know theres not going to be much difference in volume, thats not what Im interested in. I was assuming an extra 15w would help with control of the woofer and improve bass a bit? Andrew's responses are amazing though and hugely usefull. At some point I will get myself clued up on 'scopes and when I start building amps from scratch Andrews posts will be a usefull reference. In the meantime I think I'll not get too carried away with the CA in this respect.
I would like to know what causes PSU sag and what the cure is though? would a higher VA rated trafo help?
Mike. 🙂
I was just curious if I could get an extra 15 watts out of the amp. I dont know how much difference that would make, I was just wondering. I know theres not going to be much difference in volume, thats not what Im interested in. I was assuming an extra 15w would help with control of the woofer and improve bass a bit? Andrew's responses are amazing though and hugely usefull. At some point I will get myself clued up on 'scopes and when I start building amps from scratch Andrews posts will be a usefull reference. In the meantime I think I'll not get too carried away with the CA in this respect.
I would like to know what causes PSU sag and what the cure is though? would a higher VA rated trafo help?
Mike. 🙂
power ratings are usually stated in reference to clipping distortion, be it 0.1%, 10% or what else.
A rating without distortion reference doesn't tell you very much, as an amplifier can have several power ratings then.
regards
A rating without distortion reference doesn't tell you very much, as an amplifier can have several power ratings then.
regards
mikesnowdon said:
No Mike, like everything else it's just old age 😀
I built a bridged LM3886 with the potential for 100 Watts per side. It sounds good, but not as good as a single 3875 IMO.
He does not want to test the power output of the chipamps. He wants to know if the PSU is reasonably sized as made or if he is wise to spend money on upgrading the transformer and/or smoothing capacitance.Juergen Knoop said:
It's the PSU he is measuring.
measuring is fine, nothing against that! 🙂
But if Mike wants an explanation, why the amp is stated with 45W instead of maybe 50W, he might want to know about the usual practice regarding such specs.
regards
But if Mike wants an explanation, why the amp is stated with 45W instead of maybe 50W, he might want to know about the usual practice regarding such specs.
regards
Lets take into consideration what I have changed in the PSU, Caps and Diodes. I swapped out the stock caps (10'000uf) for equivalent value Mundorf M-Caps. And the diodes were changed from 1N5402 to MUR860. Sound wise I notice slightly more bass extension, slightly more defined and natural bass with clearer mids and highs. Its not a striking difference but its there nonetheless. I haven't done any critical listening as I still have the minor buuz to fix. What differences have these changes made in technical terms and will a trafo capable of delivering more amps help the chips produce power more easily?
My new input selector ic arrived today. When I get time in the next few days get it in. Lets hope that cures my buzzing! 🙂
Any tips on soldering SMT microchips?
Any tips on soldering SMT microchips?
- Status
- Not open for further replies.