Btw. wouldn't a schottky between the ldo and the Tiny fix the issue? Then you're able to program at every voltage you like.
Got the socket in, I've actually got a little programmer jig built for it. Works pretty good:
Once I get the firmware proven, I'm preprogramming the micros for the rest of the boards.
Once I get the firmware proven, I'm preprogramming the micros for the rest of the boards.
Gotcha, forgot that the firmware was still under development, I could see that being a big pain without ICSP. Good luck!
Measuring 12.22 and -11.61V for the two rails. Imbalance doesn't really matter.
I'll look closer later on a scope.
Making a BOM change on the Pro - resistor R25, part of the ripple compensation of the LM25018 circuit. Was 10K, now 75K - I'd say anything in 68K to 82K range is fine.
The low value (despite being OK by TI's calculations) put too much ripple on the FB pin of the LM25018 at higher input voltages, which had the effect of pushing the output voltage up. At 30V, the output voltage was >13V, and potentially endangering the TPA3250 which has 13.x absolute maximum ratings on its 12V supply inputs.
With 75K, the output voltage ranges from 12.2 to 12.4 over a 16-30V input range. Much better.
bk856er, hopefully you've got some 75K's left over from the first Wieners you built.
The low value (despite being OK by TI's calculations) put too much ripple on the FB pin of the LM25018 at higher input voltages, which had the effect of pushing the output voltage up. At 30V, the output voltage was >13V, and potentially endangering the TPA3250 which has 13.x absolute maximum ratings on its 12V supply inputs.
With 75K, the output voltage ranges from 12.2 to 12.4 over a 16-30V input range. Much better.
bk856er, hopefully you've got some 75K's left over from the first Wieners you built.
And bad news so far - I can't program fuses on the AVR at 3.3V. It tries but reads back the old values.
I'll come up with some hack to running the processor at 5V for programming, without damaging anything.
I'll come up with some hack to running the processor at 5V for programming, without damaging anything.
Of course! I learned early to order government style.
BTW, on the single Pro BOM I saved over $5 by increasing the Mouser quantities of many of the parts to 10+, etc. Go figure. Nearly free shipping right there...
And congrats on moving some electrons through the circuit!
BK
Got programming working. The 78L33 doesn't mind its output being driven to 5V. Programming started working around 4.3V.
Have a quick code hack done where if the green LED output is grounded, the code halts with all pins driven low. So I program with VCC driven to 5V and a jumper across the green LED, nothing blows up. Power down the board, remove the jumper and 5V supply, power back up and I'm in business.
Now to start bringing everything up.
Have a quick code hack done where if the green LED output is grounded, the code halts with all pins driven low. So I program with VCC driven to 5V and a jumper across the green LED, nothing blows up. Power down the board, remove the jumper and 5V supply, power back up and I'm in business.
Now to start bringing everything up.
Hi Gmarsh,
Any idea when you will be able to ship my order which I made in Group buy #2 back in Feb 2016 which I never received?
Hope you have not forgotten about it.
Any idea when you will be able to ship my order which I made in Group buy #2 back in Feb 2016 which I never received?
Hope you have not forgotten about it.
apotelyt - you're on the top of the order list, I'm building your order first. Right now my main priority is getting the Pro design validated, once I'm done with that I'll start volume manufacturing of cards.
Current state of things: Wiener Pro software is mostly doing what it should, as far as timers/ADC conversions/supply monitoring/etc. I haven't uncommented the lines to enable the TPA chip yet, that's tonight's plan if I can get to it.
Have a second component value change, R33 (part of power supply sensing divider) is 22.1K in the current BOM, should be 10K. I'll release a new BOM shortly, once I've got everything else proven in the design.
Current state of things: Wiener Pro software is mostly doing what it should, as far as timers/ADC conversions/supply monitoring/etc. I haven't uncommented the lines to enable the TPA chip yet, that's tonight's plan if I can get to it.
Have a second component value change, R33 (part of power supply sensing divider) is 22.1K in the current BOM, should be 10K. I'll release a new BOM shortly, once I've got everything else proven in the design.
Figure I'll write a little blurb on what the microcontroller does.
The uC samples the power supply voltage at a ~50Hz rate, and lowpasses the measurement (1st order IIR, corner frequency ~3Hz) to make an average. If the instantaneous voltage drops 12% below the average, indicating a dropout, the card mutes the TPA until the power supply stabilizes, for a minimum 1 second delay. Also if the voltage falls below 15V (instantaneous or average) the card is muted. This is how pop-free turn-off is done, assuming you just pull power instead of shutting down the amp with the enable switch. The LED flashes amber during a dropout/undervoltage event at a 2Hz rate.
If /CLIP_OTW goes low, indicating clipping or a temperature warning, the LED changes from green to solid amber, for minimum 1/2 second. Amp remains running at this time. If /FAULT goes low on the amp, indicating bad things, a solid red LED is displayed and the amp is muted for minimum 1 second. If all is kosher, the amp displays a solid green LED. If the amp's disabled, the LED is off.
Sounds simple enough but it is a couple hundred lines of code.
The uC samples the power supply voltage at a ~50Hz rate, and lowpasses the measurement (1st order IIR, corner frequency ~3Hz) to make an average. If the instantaneous voltage drops 12% below the average, indicating a dropout, the card mutes the TPA until the power supply stabilizes, for a minimum 1 second delay. Also if the voltage falls below 15V (instantaneous or average) the card is muted. This is how pop-free turn-off is done, assuming you just pull power instead of shutting down the amp with the enable switch. The LED flashes amber during a dropout/undervoltage event at a 2Hz rate.
If /CLIP_OTW goes low, indicating clipping or a temperature warning, the LED changes from green to solid amber, for minimum 1/2 second. Amp remains running at this time. If /FAULT goes low on the amp, indicating bad things, a solid red LED is displayed and the amp is muted for minimum 1 second. If all is kosher, the amp displays a solid green LED. If the amp's disabled, the LED is off.
Sounds simple enough but it is a couple hundred lines of code.
Update!
https://www.youtube.com/watch?v=voOODpYfeOo
Since recording the video I've done short circuit testing (across the outputs, outputs to ground, etc...) and everything's doing exactly what it should so far.
A few things left to do:
- Double check the +-12V power supply over the full power supply range (up to +36V, my bench supply only goes up to +25V)
- push the board into overtemperature warning. Attempting to do that right now, running the board at high volume with a cardboard box over the top for several minutes and nothing yet, might have to move up to styrofoam.
- push the board into clipping.
https://www.youtube.com/watch?v=voOODpYfeOo
Since recording the video I've done short circuit testing (across the outputs, outputs to ground, etc...) and everything's doing exactly what it should so far.
A few things left to do:
- Double check the +-12V power supply over the full power supply range (up to +36V, my bench supply only goes up to +25V)
- push the board into overtemperature warning. Attempting to do that right now, running the board at high volume with a cardboard box over the top for several minutes and nothing yet, might have to move up to styrofoam.
- push the board into clipping.
Mailing out tomorrow: bk856er and apotelyt.
Actually intended to mail them out today, but didn't get a chance to get to the post office after work.
Actually intended to mail them out today, but didn't get a chance to get to the post office after work.
Actually actually mailing out tomorrow: BK, apotelyt. Wanted to build a Pro with a preprogrammed microcontroller just to make sure everything's perfect, before I actually mail out BK's 🙂
Got SMT assembly done on 3 Pros and 1 fully built 4 ohm Pro done. The prototype worked perfectly but this batch of four suffered so many solder bridges on the TSSOPs and a few shorted 0603 parts. Really wishing I invested in a stencilling jig.
Got SMT assembly done on 3 Pros and 1 fully built 4 ohm Pro done. The prototype worked perfectly but this batch of four suffered so many solder bridges on the TSSOPs and a few shorted 0603 parts. Really wishing I invested in a stencilling jig.
No worries - we all benefit from your careful approach.
Tell me about this stenciling jig. I usually use the OSH "jig" for the pcb and tape down the stencil on all edges. But this will be my first time with a SS stencil. Does solderpaste type and stencil thickness play a role in the outcome? My previous record using a kapton stencil was a LQFP144 package, but not without some pucker factor...
BK