Working on the wake-up code to wake up the rfduino from shutdown. So far it's not working.
Also, can someone help with this...
https://1580mixer.shutterfly.com/pictures/59
Along the top edge you can see a trace where I'm bussing the input jack shields. I think I'm going to change jacks to a more secure nut/locking washer type where the shield connects directly to chassis via the jack. Is there any negative to doing this vs. what's shown in the pic above where the shields are bussed?
That1580 mini mixer
Re: That1580/1510 mini mixer
FInally I got an answer about systemoff() on the rfduino. It's a real poweroff (duh, the name does say so), but the docs implied that it was more like running in ultra low power mode. So there was a good reason I could not wake the thing up.
I'm going to just do another revision on the rfduino adapter board and put a physical on off switch on the 3.3v rail. On 3.3v is connected, off 3.3v is floating.
I have a kludged battery supply ready to test everything else - need to find out the source of the remaining noise. One more qfn is wonky, sometimes it works sometimes not so that needs to be reflowed properly.
The battery supply has got me thinking to make a 'real' battery supply for the whole thing. Best idea as of now is 4 9V's on a pcb the size of the top of the tin, 4 9v clips soldered on to it and two regulators, no need to enclose it. two batteries for the +-9, no regulator needed, and two batteries for +18 regulated to +12. It will require me shuffling gear around and using only mic bodies that take 12v but it could be worth it, and of course dynamics and ribbons would be unaffected.
I wanted to do something cooler than the above for the battery supply rather than a free floating pcb sitting on top of the box but I came up blank. No-go ideas were some sort of hinged battery clips that could sit on top of the case or 4 clips sticking/soldered to the outside rim.
I'm going to just do another revision on the rfduino adapter board and put a physical on off switch on the 3.3v rail. On 3.3v is connected, off 3.3v is floating.
I have a kludged battery supply ready to test everything else - need to find out the source of the remaining noise. One more qfn is wonky, sometimes it works sometimes not so that needs to be reflowed properly.
The battery supply has got me thinking to make a 'real' battery supply for the whole thing. Best idea as of now is 4 9V's on a pcb the size of the top of the tin, 4 9v clips soldered on to it and two regulators, no need to enclose it. two batteries for the +-9, no regulator needed, and two batteries for +18 regulated to +12. It will require me shuffling gear around and using only mic bodies that take 12v but it could be worth it, and of course dynamics and ribbons would be unaffected.
I wanted to do something cooler than the above for the battery supply rather than a free floating pcb sitting on top of the box but I came up blank. No-go ideas were some sort of hinged battery clips that could sit on top of the case or 4 clips sticking/soldered to the outside rim.
- mediatechnology
- Posts: 5463
- Joined: Sat Aug 11, 2007 2:34 pm
- Location: Oak Cliff, Texas
- Contact:
Re: That1580/1510 mini mixer
Thanks for the update.
So "Off" really means Off.
I'm still amazed that you got any of the QFNs working after hand-soldering.
So "Off" really means Off.
I'm still amazed that you got any of the QFNs working after hand-soldering.
Re: That1580/1510 mini mixer
I am not familiar with arduino code, i didn't even know that there was such a thing as a RFduino, but I have done some design with battery powered microprocessors.
If you want to maximize battery life you want off to really be off...
I worked up a PS design for my battery powered gadget where the on/off button press, not only turned on an analog power supply (mosfet switch) that held up the PS rail long enough for the micro to boot up and latch the PS on.... A second button press of the same on/off button after it was on would be read by the micro as a turn off instruction so the PS on-latch is released...
Not trivial a couple resistors, transistors, and a series pass mosfet... IIRC I added a second mosfet to block reverse battery voltage so reverse battery voltage is harmlessly ignored.
My design was slightly more complicated because I wanted really long battery life so used 4x AA cells in series for an only 3.3V rail... BUT 4x1.65V in series was too much voltage for the LDO cmos voltage regulator so I made my pass mosfet into a crude pre-regulator. In hindsight I probably could have gotten away with 3x AA batteries, but the 4x AA will run almost forever before bumping into the LDO cut off and then keep running longer..
JR
If you want to maximize battery life you want off to really be off...
I worked up a PS design for my battery powered gadget where the on/off button press, not only turned on an analog power supply (mosfet switch) that held up the PS rail long enough for the micro to boot up and latch the PS on.... A second button press of the same on/off button after it was on would be read by the micro as a turn off instruction so the PS on-latch is released...
Not trivial a couple resistors, transistors, and a series pass mosfet... IIRC I added a second mosfet to block reverse battery voltage so reverse battery voltage is harmlessly ignored.
![Cool 8-)](./images/smilies/icon_cool.gif)
My design was slightly more complicated because I wanted really long battery life so used 4x AA cells in series for an only 3.3V rail... BUT 4x1.65V in series was too much voltage for the LDO cmos voltage regulator so I made my pass mosfet into a crude pre-regulator. In hindsight I probably could have gotten away with 3x AA batteries, but the 4x AA will run almost forever before bumping into the LDO cut off and then keep running longer..
![Laughing :lol:](./images/smilies/icon_lol.gif)
JR
Cancel the "cancel culture", do not support mob hatred.
Re: That1580/1510 mini mixer
It looks slightly more complicated than it is... in a few places I used resistors in series to avoid bringing in another different part number, and quad resistors gave me 4 resistors in one part placement (machine time more expensive than the parts these days). I also used some diode connected transistors and show diodes in the design so I could choose to populate either one... In larger volume it would have made sense to pop the diodes, but for modest production using the transistors as diodes made sense because they were already in use elsewhere and come 5k to a reel.
The top left mosfet is for battery polarity protection (cheap insurance). Voltage regulation is performed by the top NPN transistor in connection with the second mosfet. It is a little bit of a trapeze act. The base of the NPN is driven to 3.3V from the microprocessor power up latch line. This crudely regulated the +UN to a voltage low enough to not release smoke from the low voltage semiconductors.. In hindsight it might be the class D amp that was most sensitive over voltage, not the LDO. The LED drivers were connected directly to the +Batt since they could handle all the voltage, and I didn't want to pull LED current through the pre-regulator.
The on/off momentary switch jump starts the PS to allow the micro power to reset, then the PS is held on by the line connected to NPN base after the microprocessor is up and running. When the on/off switch is pressed again after the micro is up and running the closure to ground pulls down a micro input port through the diode, initiating the shutdown. The diode is so it can float up higher than 3.3V when not pressed.
By reading the actual base voltage of the regulating NPN I can see when it drops out of pre-regulation, and in addition an error flag from the LDO reports when the 3.3V regulator drops out... Roughly 0.8V per AA cell.
![Laughing :lol:](./images/smilies/icon_lol.gif)
JR
Cancel the "cancel culture", do not support mob hatred.
Re: That1580/1510 mini mixer
JR thanks for the input. It's sort of over my paygrade, I'm still in the barely able to get it working stages. btw rfduino is just an arduino (modern pic mcu) plus a bluetooth module.
One step forward two back. It looks like noise problems are solved. I fired it up with both the p/s card and a battery supply. With the rfduino off switch bodged on, output was quiet with no audible clicks or hash, but I will need to test this more rigorously.
Then tons of new bugs appeared, all sorts of intermittencies where the gain would not change, one channel worked properly, and one channel had its gain locked in somewhere around 30db, even with the mcu off. Weird stuff. I'm hoping this is just more bad qfn soldering fallout, so I'll go back to the preamp chips and reflow/touch up a bunch of joints. Weirder still was that the behavior was not the same depending on whether I used the p/s card or the battery supply. I tested the p/s card rails and they are all ok.
Good news is that the gain settings on the 5171 on all channels are stable (watch tomorrow they will go wonky again now that it's in print).
One step forward two back. It looks like noise problems are solved. I fired it up with both the p/s card and a battery supply. With the rfduino off switch bodged on, output was quiet with no audible clicks or hash, but I will need to test this more rigorously.
Then tons of new bugs appeared, all sorts of intermittencies where the gain would not change, one channel worked properly, and one channel had its gain locked in somewhere around 30db, even with the mcu off. Weird stuff. I'm hoping this is just more bad qfn soldering fallout, so I'll go back to the preamp chips and reflow/touch up a bunch of joints. Weirder still was that the behavior was not the same depending on whether I used the p/s card or the battery supply. I tested the p/s card rails and they are all ok.
Good news is that the gain settings on the 5171 on all channels are stable (watch tomorrow they will go wonky again now that it's in print).
Re: That1580/1510 mini mixer
Current problem I need some help with
I've been testing the channel cards' spi gain settings by using the arduino serial monitor. Mostly all channels work, with gain increments being displayed properly. I say mostly because there is some intermittence here and there on 1 or 2 channels, but unplugging the usb programmer causing a reset on all channels solves this.
Now when I start testing for realz using either the ad hoc battery power supply (no phantom) or my switching power supply card only two cards seem to work. The others pass signal (likely at 5.6db), I get an rf connection to the rfduino, but no gain changes actually take place. Also it looks like channel 1 on the mixer aka address 0 is causing some brief vu gain changes on different channels. This leads me to think that when not using the rfduino programmer/usb supply that my chip addresses are not being set properly. Previously I thought the problem was just bad qfn joints on the rg or in/out connection between 5171 and 1580, but now I'm not so sure this is the problem, or it could be only part of the problem.
I followed That's instructions and put an rc on each 5171 reset pin ... 100k R and 1n C which supposedly is a sufficient tau for a proper reset sequence to set chip addresses.
Any ideas on what could be causing this?
I'm wondering if I should replace the reset rc caps or experiment with different values.
I've been testing the channel cards' spi gain settings by using the arduino serial monitor. Mostly all channels work, with gain increments being displayed properly. I say mostly because there is some intermittence here and there on 1 or 2 channels, but unplugging the usb programmer causing a reset on all channels solves this.
Now when I start testing for realz using either the ad hoc battery power supply (no phantom) or my switching power supply card only two cards seem to work. The others pass signal (likely at 5.6db), I get an rf connection to the rfduino, but no gain changes actually take place. Also it looks like channel 1 on the mixer aka address 0 is causing some brief vu gain changes on different channels. This leads me to think that when not using the rfduino programmer/usb supply that my chip addresses are not being set properly. Previously I thought the problem was just bad qfn joints on the rg or in/out connection between 5171 and 1580, but now I'm not so sure this is the problem, or it could be only part of the problem.
I followed That's instructions and put an rc on each 5171 reset pin ... 100k R and 1n C which supposedly is a sufficient tau for a proper reset sequence to set chip addresses.
Any ideas on what could be causing this?
I'm wondering if I should replace the reset rc caps or experiment with different values.
Re: That1580/1510 mini mixer
I've never worked with those parts but have had my share of digital communication problems...
Divide and conquer. Simplify and reduce it to thing you can confirm and then move progressively more complicated.
Sorry for this non-answer but any single line missing can corrupt or thwart data transfer... perhaps the programmer us providing a ground connection that is missing when disconnected? (just a wild guess).
JR
Divide and conquer. Simplify and reduce it to thing you can confirm and then move progressively more complicated.
Sorry for this non-answer but any single line missing can corrupt or thwart data transfer... perhaps the programmer us providing a ground connection that is missing when disconnected? (just a wild guess).
JR
Cancel the "cancel culture", do not support mob hatred.
Re: That1580/1510 mini mixer
No such thing as a non-answer, every little bit helps.
More info and thoughts:
When using the rfduino programmer my CS line measures 1.7V. With my power supply CS is 3.3V. It should theoretically be high/non-active since no data is being sent when probing.
I just remembered today that when I was designing the android<>rfduino code I originally wanted to have rfduino read gain updates after they were written and sned them back to android to act as a confirmation. I conveniently left that code out. Getting that code working will confirm exactly what's going on with this problem and probably just be a good thing to have in place going forward.
I'll put in a .1u cap to replace the 1n in the reset rc just on a lark so see if it works.
More info and thoughts:
When using the rfduino programmer my CS line measures 1.7V. With my power supply CS is 3.3V. It should theoretically be high/non-active since no data is being sent when probing.
I just remembered today that when I was designing the android<>rfduino code I originally wanted to have rfduino read gain updates after they were written and sned them back to android to act as a confirmation. I conveniently left that code out. Getting that code working will confirm exactly what's going on with this problem and probably just be a good thing to have in place going forward.
I'll put in a .1u cap to replace the 1n in the reset rc just on a lark so see if it works.
Re: That1580/1510 mini mixer
Moving along...
I stuck a .1u cap in the reset rc, didn't work. Then tried a 22u and it seems to be working. Things got a lot more stable.
Next I rewrote a lot of the code on the arduino and android so that only actual gain register data was being displayed. Every change on the android gain display has a proper gain read associated. Android app now is only sending increment and decrement commands.
I fixed a few joints, one missing cap, reflowed one 1580 and as of last test all channels were passing audio and responding to gain increments and decrements.
WIll start more in depth audio testing soon. Very eager to finish this, call it a prototype and move onto 2.0.
All those poor qfn chips I reflowed thinking it was an audio signal problem when it was spi all along, well mainly.
Still need to work out some android kinks. There is a lag in the gain textviews displaying properly. Also I made a little status textview. Android coding sux, something as simple as that caused problems in the other gain textviews. One more little thing I'd like to add is a little red or green dot on each channel indicating that the channel is good - something like when the mcu starts up it tries to communicate with each slave, or read the serial number and reports back that the channel is communicating.
I stuck a .1u cap in the reset rc, didn't work. Then tried a 22u and it seems to be working. Things got a lot more stable.
Next I rewrote a lot of the code on the arduino and android so that only actual gain register data was being displayed. Every change on the android gain display has a proper gain read associated. Android app now is only sending increment and decrement commands.
I fixed a few joints, one missing cap, reflowed one 1580 and as of last test all channels were passing audio and responding to gain increments and decrements.
WIll start more in depth audio testing soon. Very eager to finish this, call it a prototype and move onto 2.0.
All those poor qfn chips I reflowed thinking it was an audio signal problem when it was spi all along, well mainly.
Still need to work out some android kinks. There is a lag in the gain textviews displaying properly. Also I made a little status textview. Android coding sux, something as simple as that caused problems in the other gain textviews. One more little thing I'd like to add is a little red or green dot on each channel indicating that the channel is good - something like when the mcu starts up it tries to communicate with each slave, or read the serial number and reports back that the channel is communicating.