Too late!
...it was necessary to split the board, otherwise the connection could not be made, as the trafo is pcb type and pads and traces are on top layer. But this way the trafo can be placed a bit more away from circuits..Yes, on the audio board the relays have their own supply rail and gnd trace, all going back to the psu board.
Regards,
JAY X
Hi!
After the surgery, the noise has reduced a bit, but still have a lot of ripple.
At zero volume 20uv ripple. as the volume increases more and more ripple.
i have attached a wav file of the noise, increasing volume. (no music file, just psu noise). I have edited the file, shortened it a bit to make it smaller to upload. If you can analyse it on a software oscilloscope you will see the spikes...
I wired the primaries in series 2x115v ac and the secondaries in series 2x18v ac. The rectifier capacitors in my psu are JACKON LHK 3x 2200uf 35v 105º
Thank you for your help!
JAY X
After the surgery, the noise has reduced a bit, but still have a lot of ripple.
At zero volume 20uv ripple. as the volume increases more and more ripple.
i have attached a wav file of the noise, increasing volume. (no music file, just psu noise). I have edited the file, shortened it a bit to make it smaller to upload. If you can analyse it on a software oscilloscope you will see the spikes...
I wired the primaries in series 2x115v ac and the secondaries in series 2x18v ac. The rectifier capacitors in my psu are JACKON LHK 3x 2200uf 35v 105º
Thank you for your help!
JAY X
Hi,
I had a look to your sound file.
The FFT spectrum clearly shows a fundamental at 50Hz with a lot of harmonics (100Hz, 150,200,250............), ranging up to 20kHz.
If it comes from the psu, the fundamental would be 100Hz because of the full wave rectification (if your transformer is now connected correctly).
Weird thing is that I can see also spectral lines at 120Hz,180,240.... and so on.
From the point of listening I would assume a problem related to grounding (gnd loop?).
It would be helpful if you post some pictures or sketches of your setup.
I had a look to your sound file.
The FFT spectrum clearly shows a fundamental at 50Hz with a lot of harmonics (100Hz, 150,200,250............), ranging up to 20kHz.
If it comes from the psu, the fundamental would be 100Hz because of the full wave rectification (if your transformer is now connected correctly).
Weird thing is that I can see also spectral lines at 120Hz,180,240.... and so on.
From the point of listening I would assume a problem related to grounding (gnd loop?).
It would be helpful if you post some pictures or sketches of your setup.
Last edited:
Hi Andreas!
Well, the test setup is not the best: because the boards are out of a metal box. so the input jacks are not connected to a chassis. At least i will tie some gnd wires from the jacks gnd pin to the iec filter inlet gnd. All the jacks have a separated gnd area (chassis gnd), on the boards.
But i wanted to test the output of the psu, so i placed a voltage divider this morning from the unregulated rail (+28v) 1k/47k -> so 0,5volts output, and recorded to soundforge. it looks better! but my software oscilloscope has not much resolution!.
Groundloops!.: hmm, the fact is that i have two dual potentiometers whose gnd pin is tied to the gnd plane individually. I will try without the pots, and also with another gnd arrangement for the pots: link all the gnd pins together, and take the gnd wire to the main gnd point in the audio board, where it is placed the input the power rail.
JAY X
Well, the test setup is not the best: because the boards are out of a metal box. so the input jacks are not connected to a chassis. At least i will tie some gnd wires from the jacks gnd pin to the iec filter inlet gnd. All the jacks have a separated gnd area (chassis gnd), on the boards.
But i wanted to test the output of the psu, so i placed a voltage divider this morning from the unregulated rail (+28v) 1k/47k -> so 0,5volts output, and recorded to soundforge. it looks better! but my software oscilloscope has not much resolution!.
Groundloops!.: hmm, the fact is that i have two dual potentiometers whose gnd pin is tied to the gnd plane individually. I will try without the pots, and also with another gnd arrangement for the pots: link all the gnd pins together, and take the gnd wire to the main gnd point in the audio board, where it is placed the input the power rail.
JAY X
Hi!
I have a 26" Samsung LCD tv (year 2006) that i use also as pc monitor on the desk. But it is placed about 1,5m from the boards. -----TV desk--------- and at 90º a keyboard stand with the boards
These boards are a summing mixer project. 34 inputs. 32 in db25 format. 4 xlr outputs, and balanced send return inserts. Many opamps MC33078. Also some line drivers and receivers INA134/DRV134. and two output transformers: lundahl LL2811. Nice project!
I have a 26" Samsung LCD tv (year 2006) that i use also as pc monitor on the desk. But it is placed about 1,5m from the boards. -----TV desk--------- and at 90º a keyboard stand with the boards
These boards are a summing mixer project. 34 inputs. 32 in db25 format. 4 xlr outputs, and balanced send return inserts. Many opamps MC33078. Also some line drivers and receivers INA134/DRV134. and two output transformers: lundahl LL2811. Nice project!
Last edited:
Cool project!These boards are a summing mixer project. 34 inputs. 32 in db25 format. 4 xlr outputs, and balanced send return inserts. Many opamps MC33078. Also some line drivers and receivers INA134/DRV134. and two output transformers: lundahl LL2811. Nice project!
I´m too curious, so I would be very pleased if you post some pics.
Don´t worry if your improvised setup doesn´t look that nice, I will not criticize you therefor.
At first view I would say your psu is quite small for this project, many opamps, relays, line drivers.......
Have you calculated the overall power requirement of the audio boards?
Hmmm, 28V at unregulated power rail. I assume this is without load.
Just measure the voltage at the same position with all boards connected to the psu, then we will see.
Hi!
Yesterday i took some measurements with a picoscope from a friend...
I have some problems:
1. Transformer secondary saturation. the waveform show the peaks crashed (flat topping). This is why i have harmonics, (third harmonic) and so much ripple and noise. Even with load.
2. The circuits represent a load of about 120-150ma per winding. (measured with a multimeter in series). This is maybe a bit light load for the transformer. The transformer is capable to give 278ma per winding. 10va. 18-0-18v sec.
3. possible solutions:
3a. Right now i have two 500ma fuses in the primary iec inlet. maybe i low them to 250ma, to limit the current in the primary..??
3b. Rectifier CRC FILTER. 4400UF -2.2OHMS-2200UF. Maybe replacing the 2.2R for an inductor... say 820uh...??? to have more ripple attenuation.
3c. DC output filtering. Right now i have 220uf in parallel with 100nf.
But maybe i have to increase the cap to 330 or 470uf. or place an RC filter at the dc output, between the psu and audio board.
3d. More filtering at the gain opamp stages: fader amplifiers (Gain=3), and maybe the output line drivers. With an rc filter in the supply tracks: 47uf +560R, this is a 6db filter.
3d. All this problems may be attenuated when i have the metal box made, because the box acts as faraday cage, and so much noise can be diverted to chassis gnd.
Ok, pictures soon.
JAY X
Yesterday i took some measurements with a picoscope from a friend...
I have some problems:
1. Transformer secondary saturation. the waveform show the peaks crashed (flat topping). This is why i have harmonics, (third harmonic) and so much ripple and noise. Even with load.
2. The circuits represent a load of about 120-150ma per winding. (measured with a multimeter in series). This is maybe a bit light load for the transformer. The transformer is capable to give 278ma per winding. 10va. 18-0-18v sec.
3. possible solutions:
3a. Right now i have two 500ma fuses in the primary iec inlet. maybe i low them to 250ma, to limit the current in the primary..??
3b. Rectifier CRC FILTER. 4400UF -2.2OHMS-2200UF. Maybe replacing the 2.2R for an inductor... say 820uh...??? to have more ripple attenuation.
3c. DC output filtering. Right now i have 220uf in parallel with 100nf.
But maybe i have to increase the cap to 330 or 470uf. or place an RC filter at the dc output, between the psu and audio board.
3d. More filtering at the gain opamp stages: fader amplifiers (Gain=3), and maybe the output line drivers. With an rc filter in the supply tracks: 47uf +560R, this is a 6db filter.
3d. All this problems may be attenuated when i have the metal box made, because the box acts as faraday cage, and so much noise can be diverted to chassis gnd.
Ok, pictures soon.
JAY X
No, this is not saturation. Flat sine wave at secondary is just completely normal.
Why? Because of two devices, the rectifier (a non linear device) and the charging cap(s). With only resistive load at the secondary the result would be a very nice pure sine wave.
Measured AC at secondary or DC at regulaters output?
This would not change anything.
Yes, this is the position where it usually should be.
O.K.
The transformer is capable to deliver 278mA AC, this means not that you can get out 278mA DC.
With 150mA DC load at regulator outputs your transformer is slight over it´s limit. The total apparent power is about 11VA.
With rectifier and cap, the current out of the secondaries is not sine-shaped. There are pulses, in your case with peak value of about 0.5A!
PSU calculations are more complex than it seems at first glance. It´s too wide a subject, so I would recomend to read some tutorials about this topic, then you will see things more clearly.
Last edited:
The maximum continuous DC output current from a capacitor input filter is approximately half the maximum continuous AC current rating of the transformer.
i.e. the maximum continuous DC current from a
To allow the transformer to run cool to warm use ~50% of that, i.e. 70mAdc continuous.
i.e. the maximum continuous DC current from a
is ~140mAdc
To allow the transformer to run cool to warm use ~50% of that, i.e. 70mAdc continuous.
Yes Andrew, this is a good rule of thumb.
Jay, you should swap the 10VA transformer to 20VA, but I think for some tests you can still use the 10VA.
Don´t misunderstand these things, the reason for the flattened sine at secondary is not the too small transformer, it will also happen with the 20VA.
You can not avoid current pulses, flattened sine wave or voltage ripple, these things are inherent in the system.
Probably this all will solve your problems.
Jay, you should swap the 10VA transformer to 20VA, but I think for some tests you can still use the 10VA.
Don´t misunderstand these things, the reason for the flattened sine at secondary is not the too small transformer, it will also happen with the 20VA.
You can not avoid current pulses, flattened sine wave or voltage ripple, these things are inherent in the system.
4400µF gives a voltage ripple of about 200mV (with 150mA load), LM317/337 ripple rejection is ca. 60dB. The RC filter gives you a nice improvement of additional 10dB. This is fairly good, imho a inductor is not needed.
You can increase the cap., but I don´t recommend a RC filter.
Not a good idea, quiescent current of 4 mA for one opamp. gives a voltage drop (560R) of more than 2V.
I think you are right, such a complex project will need a careful and clean wiring in a proper enclosure. Use shielded cables for audio, avoid ground loops when you connect pots and plugs at the front/back panel.
Probably this all will solve your problems.
Hi all!
Solved the noise problem!! It was a ground loop!! As i have also a channel imbalnca level problem, (5db), i changed the pots, to see what happens, and i pulled one gnd wire and soldered the two gnd pins together at the pot. No more gnd loop!.
This is part of the problem, still have to track the channel imbalance....
Just it takes time and patience...
Jay x
Solved the noise problem!! It was a ground loop!! As i have also a channel imbalnca level problem, (5db), i changed the pots, to see what happens, and i pulled one gnd wire and soldered the two gnd pins together at the pot. No more gnd loop!.
This is part of the problem, still have to track the channel imbalance....
Just it takes time and patience...
Jay x
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.