I just tried the link and it seems to be working fine with Foxfire.
Good work nwavguy.
Looking at the board layout I was kinda wondering if there would be space for the dip heatsinks. That sounds like a good idea, maybe working them in as an option on the desktop version. Yep, shipping is a killer! I ordered those sinks as just part of a much larger order from Arrow. Good idea about Newark and the case. I wouldn't be surprised if it becomes necessary to order from at least one additional vendor than Mouser occasionally anyway as their stocks come and go.
I was assuming the output ICs were in sockets. I would agree, not good to use those sinks if the chips are directly soldered over PC traces with no socket. The buf634s I'm using the sinks with are socketed. I'll go back and re-read that section of your description.
Another suggestion for the desktop version. Seems like the LM317/LM337s (adjustable) are generally taken to be lower noise with better line rejection than the fixed regulators IF the adjust pin is bypassed with 220uF or so. That would just add two 1/8W resistors for the adjust set and the one capacitor. The 317/337 pricing should be comparable to the fixed regs. The options for the bypass I've seen are around 220uF electo or 22uF tantalum, but I won't use tantalums in anything anymore due to the failure-mode-short issue.
However, it would be *really* interesting to try your design with the 317/337 regulators and measure the amp results with your equipment. This could be yet another one of those urban myths in terms of any measurable difference on amp output. Just because the power supply is a little cleaner it could turn out to be totally irrelevent with good amp PSRR, of course. Just money wasted on extra parts. This particular issue (fixed regs vs. bypassed adjustable regs) has come up enough in various posts on the forum that you would really add to the "body of knowledge" with some test results there on your Dscope.
I'm reading posts bottom to top.
Sounds good on the gate zener/resistor issue. Definitely wouldn't be worth the cost and time of another board spin just for that. I probably will add my own zeners tacked to the traces on the bottom.
I was assuming the output ICs were in sockets. I would agree, not good to use those sinks if the chips are directly soldered over PC traces with no socket. The buf634s I'm using the sinks with are socketed. I'll go back and re-read that section of your description.
Another suggestion for the desktop version. Seems like the LM317/LM337s (adjustable) are generally taken to be lower noise with better line rejection than the fixed regulators IF the adjust pin is bypassed with 220uF or so. That would just add two 1/8W resistors for the adjust set and the one capacitor. The 317/337 pricing should be comparable to the fixed regs. The options for the bypass I've seen are around 220uF electo or 22uF tantalum, but I won't use tantalums in anything anymore due to the failure-mode-short issue.
However, it would be *really* interesting to try your design with the 317/337 regulators and measure the amp results with your equipment. This could be yet another one of those urban myths in terms of any measurable difference on amp output. Just because the power supply is a little cleaner it could turn out to be totally irrelevent with good amp PSRR, of course. Just money wasted on extra parts. This particular issue (fixed regs vs. bypassed adjustable regs) has come up enough in various posts on the forum that you would really add to the "body of knowledge" with some test results there on your Dscope.
I'm reading posts bottom to top.
Sounds good on the gate zener/resistor issue. Definitely wouldn't be worth the cost and time of another board spin just for that. I probably will add my own zeners tacked to the traces on the bottom.
Last edited:
Hey Agdr, thanks very much for your input so far. I agree with sockets everything is fine with the heatsinks as long as they ICs stay in the sockets (as they won't be fully seated due to the thickness of the heatsink clip underneath).
Stephan, in the the 317/337 O2 Details Comments makes a case for the 317/337 being possibly useful at very high gains. So I can see including them in the desktop version as you suggested. If nothing else it mostly fits under my Harmless Excess design philosophy. But it does mean more expensive regulators and 6 more parts so I'd like to try and find some objective real-world justification.
Stephan, in the the 317/337 O2 Details Comments makes a case for the 317/337 being possibly useful at very high gains. So I can see including them in the desktop version as you suggested. If nothing else it mostly fits under my Harmless Excess design philosophy. But it does mean more expensive regulators and 6 more parts so I'd like to try and find some objective real-world justification.
Last edited:
Did you by chance download the PDF a while back? I fixed that a while ago unless I missed another schematic somewhere? I re-arranged some things on the schematic in the name of formatting and somehow managed to flip D7 over.
About 30 seconds before my post above, to check again.
Yea I noticed that many things had been fixed since you unveiled your project. Not being able to see the board routing, I thought I would mention it now. Certainly not a deal breaker.
Digger, I'm confused. There's only one PDF posted on Google Docs that I know of and it hasn't changed since August 4th. Here's the link:
https://docs.google.com/viewer?a=v&pid=explorer&chrome=true&srcid=0B52Awjeyc5zKNjQyOTlhMjAtZTNmOS00ZTNlLWEyMzgtZTQ0YzUwN2ViOTc4&hl=en_US
At least when I view the PDF linked above, D7 is facing the correct way. So I'm not sure what's going on unless it's some sort of browser cache issue and your browser is unaware the file has changed.
If anyone else still has a backwards D7, or spots other problems, please let me know?
https://docs.google.com/viewer?a=v&pid=explorer&chrome=true&srcid=0B52Awjeyc5zKNjQyOTlhMjAtZTNmOS00ZTNlLWEyMzgtZTQ0YzUwN2ViOTc4&hl=en_US
At least when I view the PDF linked above, D7 is facing the correct way. So I'm not sure what's going on unless it's some sort of browser cache issue and your browser is unaware the file has changed.
If anyone else still has a backwards D7, or spots other problems, please let me know?
V1.1 UPDATE - I received the new bare V1.1 PC boards yesterday (8/11), built up a new O2, and I've now completed some of the more critical dScope tests and can say the performance so far is very similar to the previous version at the low gain (3X/10 dB) setting. I haven't tested the high gain setting yet.
I'm pleased the addition of the gain switch, switch to 1/8 watt resistors, and new PC board layout seem to work just fine so far and haven't compromised the performance.
I'll be updating some of the graphs/measurements in the first article and I may post a couple of V1.0 vs V1.1 graphs in probably the second (design) article for anyone who's curious.
I did find a couple minor "mechanical" issues with the PC board. One is due to formed leads on a capacitor (lead spacing) and the other is a component clearance issue. So it's a good thing I didn't release the PC board layout earlier. And the previously mentioned backwards LED would might have also been a significant problem. It's good these 3 issues are easily fixed without needing another prototype board.
If any of you are wanting to order parts from Mouser it's best to wait until I publish a revised documentation package (the PDF and spreadsheet). I also will try to make some improvements to the PDF as some have had some issues with it.
I've added some additional construction tips, tools required, etc. to the O2 Details article in the Construction section that might be especially helpful to novice DIYers.
Finally, I need to regenerate all the documentation along with the PC board files after double checking my work, the latest changes, etc. I'll try to get it all done and published before the end of the weekend if I can.
I'm pleased the addition of the gain switch, switch to 1/8 watt resistors, and new PC board layout seem to work just fine so far and haven't compromised the performance.
I'll be updating some of the graphs/measurements in the first article and I may post a couple of V1.0 vs V1.1 graphs in probably the second (design) article for anyone who's curious.
I did find a couple minor "mechanical" issues with the PC board. One is due to formed leads on a capacitor (lead spacing) and the other is a component clearance issue. So it's a good thing I didn't release the PC board layout earlier. And the previously mentioned backwards LED would might have also been a significant problem. It's good these 3 issues are easily fixed without needing another prototype board.
If any of you are wanting to order parts from Mouser it's best to wait until I publish a revised documentation package (the PDF and spreadsheet). I also will try to make some improvements to the PDF as some have had some issues with it.
I've added some additional construction tips, tools required, etc. to the O2 Details article in the Construction section that might be especially helpful to novice DIYers.
Finally, I need to regenerate all the documentation along with the PC board files after double checking my work, the latest changes, etc. I'll try to get it all done and published before the end of the weekend if I can.
Last edited:
That's not what it says on your rap sheet.
se
It's much easier to sell a product once you have a few fora bans under your belt.
Don't listen to him. I've been banned more times than you and it's not done my sales any good.
se
Then I outta be a millionaire by now.
se
I suppose one profitable prototype a millionaire does not make