An 8mV DC offiset on your OUTPUT is entirely acceptable. If you seem to drift from that value, that's normal too.
I didn't understand your 600mV "on one of the two rails". Are you stating you have a voltage difference of 600mV when you measure each of the two rail voltages to ground, and then compare the values to one another. If this is the case, it's not a problem. Please explain.
Hello!
oh my english is too bad... 🙁
ok, I try to describe: every time I set the voltage through R11/R12 to 590mV and listen to music for an hour or two, the voltage through R11/R12 is getting to >600mV until I turn the pots down a little bit again. R11 is always a little bit higher than R12, but that should be no big problem (about 25mV difference)
the maximum measured voltage was 613mV.
I can say: the warmer it gets, the higher is the voltage through R11/R12. if I then turn the pots down a little bit, everything seems to be ok. An hour or two later, I should again turn the pots down to come under 600mV.
the offset on the output: cold about 25mV for the first minutes, then going to about 12mV , warm between 8 and 3mV... I can turn it to 0mV, but that never lasts long, after a few listening-hours its changing to 3...8mV again.
I hope I could describe it a little bit?
Matthias 🙂
oh my english is too bad... 🙁
ok, I try to describe: every time I set the voltage through R11/R12 to 590mV and listen to music for an hour or two, the voltage through R11/R12 is getting to >600mV until I turn the pots down a little bit again. R11 is always a little bit higher than R12, but that should be no big problem (about 25mV difference)
the maximum measured voltage was 613mV.
I can say: the warmer it gets, the higher is the voltage through R11/R12. if I then turn the pots down a little bit, everything seems to be ok. An hour or two later, I should again turn the pots down to come under 600mV.
the offset on the output: cold about 25mV for the first minutes, then going to about 12mV , warm between 8 and 3mV... I can turn it to 0mV, but that never lasts long, after a few listening-hours its changing to 3...8mV again.
I hope I could describe it a little bit?
Matthias 🙂
Last edited:
Your english is sehr gut.....!
Your F5 is working as expected. Let it have a good warm-up period (1-2 hours) and set the bias across the two resistors to 590 mV. Then tweak one of the pots as you measure your output DC offset, and minimize that offset voltage. Then go back and double-check your two bias voltages across the resistors to confirm they are still "about" 590 mV. The variations you are seeing are typical, and are related to the parameters of the amp's components (MOSFETS, resistors) and their drift as the amp warms up. We see that in all our F5 amps. It sounds like you are doing everything properly, but if you try to "perfect" all three readings (two resistors at 590 mV, and DC offset at <10 mV) you'll end up chasing your tail....!
A tip.... during the warm-up period, keep the case of your amp closed, as you would when you use the amp. If possible, have leads in place to measure the voltages across the two resistors with the case closed. Otherwise, if you have to open your case to check the bias levels, the change in "internal temp" when you open the case can cause your readings to drift.
You are close enough, with your present readings. Relax, and enjoy the amp!
😀
Your F5 is working as expected. Let it have a good warm-up period (1-2 hours) and set the bias across the two resistors to 590 mV. Then tweak one of the pots as you measure your output DC offset, and minimize that offset voltage. Then go back and double-check your two bias voltages across the resistors to confirm they are still "about" 590 mV. The variations you are seeing are typical, and are related to the parameters of the amp's components (MOSFETS, resistors) and their drift as the amp warms up. We see that in all our F5 amps. It sounds like you are doing everything properly, but if you try to "perfect" all three readings (two resistors at 590 mV, and DC offset at <10 mV) you'll end up chasing your tail....!
A tip.... during the warm-up period, keep the case of your amp closed, as you would when you use the amp. If possible, have leads in place to measure the voltages across the two resistors with the case closed. Otherwise, if you have to open your case to check the bias levels, the change in "internal temp" when you open the case can cause your readings to drift.
You are close enough, with your present readings. Relax, and enjoy the amp!
😀
Last edited:
..............and........the other thing to monitor is your heatsink (and MOSFET) temperatures. With good thermal conductivity between your MOSFETs and heatsinks, you should be able to place your hand on your heatsinks for 10 seconds without much discomfort. If your heatsinks are considerably warmer, then you need to perhaps turn down the bias on r11/r12.
@CanAm Man:
thank you very much! 🙂
so everything seems to be ok with my F5.
I will do the last measuring and setting today afternoon after letting it run for 1,5 hours. 😉
thanks again, I think after finishing the case I'm ready to built the next PASS (F5 Turbo maybe)
Greetings from Bavaria,
Matthias 😀
thank you very much! 🙂
so everything seems to be ok with my F5.
I will do the last measuring and setting today afternoon after letting it run for 1,5 hours. 😉
thanks again, I think after finishing the case I'm ready to built the next PASS (F5 Turbo maybe)
Greetings from Bavaria,
Matthias 😀
Just saw your pics on the photo thread. Impressive results....
Do you mind sharing what measuring chips are you using?
Do you mind sharing what measuring chips are you using?
hi dimkasta!
no problem, I used the ACS712 from allegro.
it was sooo hard to find a IC which is able to measure current without using the same GND as the amplifier. (because of the massive problems with the negative voltage then!)
its a Hall Effect-Based Linear Current Sensor IC, which means you don't need to connect GND from the µC to GND from the amplifier.
in the moment I have +-15mV difference between the multimeter measured value and the µC, but its ok for me.
(you can get better values, if you use the circuit showed on page 12 of the datasheet (application 3)).
they are connected in line to R11/R12.
temperature measuring is done by dallas DS18b20 sensors, which I mounted on the topside of the mosfets.
except the levelmeter (still problems with it...) there is no connection between µC and amp. (the µC with all the parts is using another power supply)
in the moment the levelmeter is disconnected, so really no physical connection 😉
I tested the amp before and after the µC-control unit.
no change in sound, so I'm happy with it.
*edit*
found some other photos:
first tests while setting the offset, etc..:
bevor anodizing and µC:
programming, testing, programming, soldering, programming, soldering, testing:
(in the video there is no ACS712 connected, so the values are completely wrong... the light from the buttons can be switched on and off by software, but I think its better when they are always on (except standby))
the levelmeter was badly programmed in the link, too... now it's smooth without flickering all the time.
http://www.youtube.com/watch?v=ZYsMzktt8BM&feature=youtube_gdata_player
😀
and finally- ready to run 🙂
cheers,
matthias
no problem, I used the ACS712 from allegro.
it was sooo hard to find a IC which is able to measure current without using the same GND as the amplifier. (because of the massive problems with the negative voltage then!)
its a Hall Effect-Based Linear Current Sensor IC, which means you don't need to connect GND from the µC to GND from the amplifier.
in the moment I have +-15mV difference between the multimeter measured value and the µC, but its ok for me.
(you can get better values, if you use the circuit showed on page 12 of the datasheet (application 3)).
they are connected in line to R11/R12.
temperature measuring is done by dallas DS18b20 sensors, which I mounted on the topside of the mosfets.
except the levelmeter (still problems with it...) there is no connection between µC and amp. (the µC with all the parts is using another power supply)
in the moment the levelmeter is disconnected, so really no physical connection 😉
I tested the amp before and after the µC-control unit.
no change in sound, so I'm happy with it.
*edit*
found some other photos:
first tests while setting the offset, etc..:
bevor anodizing and µC:
programming, testing, programming, soldering, programming, soldering, testing:
(in the video there is no ACS712 connected, so the values are completely wrong... the light from the buttons can be switched on and off by software, but I think its better when they are always on (except standby))
the levelmeter was badly programmed in the link, too... now it's smooth without flickering all the time.
http://www.youtube.com/watch?v=ZYsMzktt8BM&feature=youtube_gdata_player
😀
and finally- ready to run 🙂
cheers,
matthias
Last edited:
Wow really impressive design. And the video... oh my... I want one 😀
Such an inspiration... 🙂
About the mV readings, are you using the onboard ADC? I think they are 10bit which means 5mV resolution. So if we expect a +-1 unit error, the 15mV make some sense.
Perhaps an external ADC with bigger resolution could give more accurate results, but given the temperature drift, I don t think it is needed anyway.
I was going to buy extra voltmeters for my build, but I like your idea. I will definitely use it with an arduino when it is time to calibrate my own. The code should be trivial, and it can even be programmed to give you instructions on how to adjust the trimmers (in case we want to make it a tool for everyone) 🙂
Oh and about the current metering IC, since you are already measuring the voltage accross the source resistors, you could have gotten away with a simple division V/R using ohm's law to get the current through them. But then you would have to use common ground for teh digital parts... hmmm...
The external IC could be a nice re-validation of the readings, but since you are probably reading its output through the same 10bit ADC, the resolution and error should be similar or worse than those of your mV readings...
Again congrats on an amazing project and execution
Such an inspiration... 🙂
About the mV readings, are you using the onboard ADC? I think they are 10bit which means 5mV resolution. So if we expect a +-1 unit error, the 15mV make some sense.
Perhaps an external ADC with bigger resolution could give more accurate results, but given the temperature drift, I don t think it is needed anyway.
I was going to buy extra voltmeters for my build, but I like your idea. I will definitely use it with an arduino when it is time to calibrate my own. The code should be trivial, and it can even be programmed to give you instructions on how to adjust the trimmers (in case we want to make it a tool for everyone) 🙂
Oh and about the current metering IC, since you are already measuring the voltage accross the source resistors, you could have gotten away with a simple division V/R using ohm's law to get the current through them. But then you would have to use common ground for teh digital parts... hmmm...
The external IC could be a nice re-validation of the readings, but since you are probably reading its output through the same 10bit ADC, the resolution and error should be similar or worse than those of your mV readings...
Again congrats on an amazing project and execution
Last edited:
hi!
yes, I use the onboard DAC with 10bit solution, so I cannot get exactlier.
funny, I want to build the F5 turbo next and therefore I want to go completely analog --> using voltmeters. 😀
your idea is very nice! only adjusting with the µC could be a little hard, because of the bouncing values (+-15mV). every 5 seconds I'm doing 30 measurements, building the average value and display it on the lcd. I'm quite happy with it, but it would be nice if I could get better values. 😉
all the different values can be calculated then. (you know, the ACS712 delivers 185mV/A, so its no problem to calculate the voltage through R11,R12 and the current.)
I hope you make some nice pics of your project!
cheers,
matthias
yes, I use the onboard DAC with 10bit solution, so I cannot get exactlier.
funny, I want to build the F5 turbo next and therefore I want to go completely analog --> using voltmeters. 😀
your idea is very nice! only adjusting with the µC could be a little hard, because of the bouncing values (+-15mV). every 5 seconds I'm doing 30 measurements, building the average value and display it on the lcd. I'm quite happy with it, but it would be nice if I could get better values. 😉
all the different values can be calculated then. (you know, the ACS712 delivers 185mV/A, so its no problem to calculate the voltage through R11,R12 and the current.)
I hope you make some nice pics of your project!
cheers,
matthias
You think that the voltmeter's or the microcontroller's input impedance might affect the voltage measurement?
the hall sensor has the advantage of being off-circuit...
the hall sensor has the advantage of being off-circuit...
my first tries in measuring ended with very crazy sounds in the speakers (veeeery quiet, but present).
I also had big problems with measuring the values. I tried everything: op-amps, crazy circuits, etc.. nothing worked really, except the hall sensor ones. 😉
I'm not sure why the values bounce that hard- maybe its some distortion from somewhere?
next problem is the LCD. when you connect both GND lines, then the display is hanging up everytime you switch the amp on/off. I programmed an initialisation-routine to reset the display in that moment. You can't see it, but its present.
I think I have to make a new video, but in the moment I'm programming a new software with better routines for the switches, VU-meter and standby.
Maybe I add a software-speaker-protection. But I'm not sure how to do that. Maybe when the voltage over R11 or R12 is smaller then 100mV or higher than 1200mV?
(would be a big bonus, you wouldn't need a relais-board then...)
In the moment the amp is going in standby, when it's higher than 1100mV. (only with a audiotest CD it was able to simulate such high voltage 😀 )
cheers,
matthias
I also had big problems with measuring the values. I tried everything: op-amps, crazy circuits, etc.. nothing worked really, except the hall sensor ones. 😉
I'm not sure why the values bounce that hard- maybe its some distortion from somewhere?
next problem is the LCD. when you connect both GND lines, then the display is hanging up everytime you switch the amp on/off. I programmed an initialisation-routine to reset the display in that moment. You can't see it, but its present.
I think I have to make a new video, but in the moment I'm programming a new software with better routines for the switches, VU-meter and standby.
Maybe I add a software-speaker-protection. But I'm not sure how to do that. Maybe when the voltage over R11 or R12 is smaller then 100mV or higher than 1200mV?
(would be a big bonus, you wouldn't need a relais-board then...)
In the moment the amp is going in standby, when it's higher than 1100mV. (only with a audiotest CD it was able to simulate such high voltage 😀 )
cheers,
matthias
Last edited:
Well HAL sensors use electomagnetism. Having big transformers and floating high-current cables near it is a no-no-no-no-no-no.
A nice iron case around it might solve some of your problems.
Would you mind posting a schematic of the hal sensor too?
A nice iron case around it might solve some of your problems.
Would you mind posting a schematic of the hal sensor too?
hi!
ooh you may be right with that! I will try some cases around the sensors, thank you.
I don't have a schematic, because I bought it here:
(output is directly connected to the ADC-pin on the atmega16)
5A range ACS712 module current sensor module | eBay
cheers,
matthias
ooh you may be right with that! I will try some cases around the sensors, thank you.
I don't have a schematic, because I bought it here:
(output is directly connected to the ADC-pin on the atmega16)
5A range ACS712 module current sensor module | eBay
cheers,
matthias
I noticed nothing. (And I tried really a long time to hear differences. My change of RCA cables from preamp to amp had a noticeable change, ACS712 not.)
Maybe its because of high-current on that point in the amp-circuit, and because of no physical connection between the measure-pins and the logic in that IC?
I didn't use the onboard terminal, and soldered the cable directly.
Between R11/R12 and the ACS712 there is a veeery short cable (in 2x2.5mm²) to make sure nothing is getting lost. 😀
cheers,
matthias
Maybe its because of high-current on that point in the amp-circuit, and because of no physical connection between the measure-pins and the logic in that IC?
I didn't use the onboard terminal, and soldered the cable directly.
Between R11/R12 and the ACS712 there is a veeery short cable (in 2x2.5mm²) to make sure nothing is getting lost. 😀
cheers,
matthias
Last edited:
Well it should act like a plain "cable". the only thing that might affect the sound is the quality of the board itself.
And I was already thinking to remove the headers and solder directly to the board.
And I was already thinking to remove the headers and solder directly to the board.
- Status
- Not open for further replies.