It can fund some of the cost of development, but once you factor in the time spent on doing the specification work, finding the right components, doing the schematic, doing the layout, making a BOM, after sales technology support etc., whatever small amount of money I make on selling a few PCBs is just drop in a bucket.
Last edited:
I hear you! I could never make back the time invested to get the product launched and working. it wasn't meant as a business, but you are right that even board sales is not enough to cover the huge amount of time to fully design, develop AND support a product like this.
I worked, off and on, close to 3 years on my version. it ended up being more of a software project, by the time spent in software devel vs hw devel. hw was done in a few months. the rest was figuring out what options, UI style (etc) to include and testing the hell out of it until it was good enough
folks who have not done this full design/ship/support experience have no idea how much work is involved!
I worked, off and on, close to 3 years on my version. it ended up being more of a software project, by the time spent in software devel vs hw devel. hw was done in a few months. the rest was figuring out what options, UI style (etc) to include and testing the hell out of it until it was good enough
folks who have not done this full design/ship/support experience have no idea how much work is involved!
So those PCBs I sent out?
All except one has received them, they have been "under transportation" since April 23.
Sent some PCBs to 2 different persons in the same country and one of them received the PCBs last week. This was by standard mail. The other "under transportation" was a considerably more expensive, recommended/signed letter option and what do I/the person who ordered the PCBs get for that extra money?..........so far 1 week extra shipping time to the same damn country. Argh.
I hope it either shows up soon or the post office can tell me what they have done to the shipment.
What does this have to do with my project? Nothing really, just wanted to let out some steam.
All except one has received them, they have been "under transportation" since April 23.
Sent some PCBs to 2 different persons in the same country and one of them received the PCBs last week. This was by standard mail. The other "under transportation" was a considerably more expensive, recommended/signed letter option and what do I/the person who ordered the PCBs get for that extra money?..........so far 1 week extra shipping time to the same damn country. Argh.
I hope it either shows up soon or the post office can tell me what they have done to the shipment.
What does this have to do with my project? Nothing really, just wanted to let out some steam.
Last edited:
Hello,
I was wondering if there are any news concerning that headphone amp project of yours?
best regards,
hallodeletue
I was wondering if there are any news concerning that headphone amp project of yours?
best regards,
hallodeletue
.....
I am going to use this attenuator for a "high-end" headphone amplifier.
Basically it would be like this :
Input buffer with gain > Attenuator > High current unity gain output buffer.
Input and output buffer will both be placed on the same small 70 mm x 40 mm, 4 Layer PCB.
Maximum output would be 10Vrms into 50 Ohm - 600Ohm loads, 9Vrms into 32 Ohm loads.
Everything built with SMD components, yes this means OP-AMPs are used as well. Like OPA1612, OPA827, OPA140, BUF634.
Total noise with x4(12dB) gain should be lower than -115dBv.
Totally DC coupled, output offset less than 120uV.
It will be dual mono, so I am going to use a 4-Pin XLR as output, this is done to eliminate crosstalk.
But I am not ready to share any more details at this moment.
It is well under way, still in the board layout phase but nearly finished.
It will be balanced input, with dual mono output. Power supply will be dual LT3080 floating regulators mounted on the headphone amplifier PCB, excellent low noise regulators.
LT3080 and BUF634 will be in the TO220-5 package and they should be mounted to the headphone amplifier bottom plate.
PCB will be 4-Layer 145mm x 50mm. Regular lead HASL. Spending money on lead free HASL or ENIG is a waste of money.
Balanced input is for avoiding ground loops, dual mono output is for eliminating crosstalk and it gives lower noise than balanced output.
Noise should be lower than -115dBv if all goes well.
There is no gain in this amplifier design.
The source will be a DAC with TI PCM1794A with a 9Vrms Output in balanced mode, which is why there is no gain in the headphone amplifier. The DAC provides all the voltage needed, the headphone amplifier basically just acts as a very low noise, low distortion, volume controlled current buffer.
Doing it this way, with the source providing all the voltage minimizes total source+headphone amplifier system noise.
It is a little uncommon to do it this way, no commercial systems I am aware of does something similar.
The drawback is that if you want the full 9VRMS output from the headphone amplifier you are forced to use the TI PCM1794A DAC with 9Vrms balanced output or something similar.
But 9Vrms is absolutely an overkill and ridiculous output level for most headphones. Even something like Audeze planar magnetics and Hifiman planar magnetics do not need 9Vrms, it is close to instant hearing damage.
Pairing the headphone amplifier with the more common 3Vrms-6Vrms balanced output DACs on the market is more than enough for most users.
I personally never listen at more than 75-80dB while 9Vrms is enough for 120+ dB with Audeze planar magnetics and most Hifiman planar magnetics.
It is built for my own personal needs, so not everyone would agree with my design philisophy, but it is what it is and for me it makes perfect sense.
It will be balanced input, with dual mono output. Power supply will be dual LT3080 floating regulators mounted on the headphone amplifier PCB, excellent low noise regulators.
LT3080 and BUF634 will be in the TO220-5 package and they should be mounted to the headphone amplifier bottom plate.
PCB will be 4-Layer 145mm x 50mm. Regular lead HASL. Spending money on lead free HASL or ENIG is a waste of money.
Balanced input is for avoiding ground loops, dual mono output is for eliminating crosstalk and it gives lower noise than balanced output.
Noise should be lower than -115dBv if all goes well.
There is no gain in this amplifier design.
The source will be a DAC with TI PCM1794A with a 9Vrms Output in balanced mode, which is why there is no gain in the headphone amplifier. The DAC provides all the voltage needed, the headphone amplifier basically just acts as a very low noise, low distortion, volume controlled current buffer.
Doing it this way, with the source providing all the voltage minimizes total source+headphone amplifier system noise.
It is a little uncommon to do it this way, no commercial systems I am aware of does something similar.
The drawback is that if you want the full 9VRMS output from the headphone amplifier you are forced to use the TI PCM1794A DAC with 9Vrms balanced output or something similar.
But 9Vrms is absolutely an overkill and ridiculous output level for most headphones. Even something like Audeze planar magnetics and Hifiman planar magnetics do not need 9Vrms, it is close to instant hearing damage.
Pairing the headphone amplifier with the more common 3Vrms-6Vrms balanced output DACs on the market is more than enough for most users.
I personally never listen at more than 75-80dB while 9Vrms is enough for 120+ dB with Audeze planar magnetics and most Hifiman planar magnetics.
It is built for my own personal needs, so not everyone would agree with my design philisophy, but it is what it is and for me it makes perfect sense.
Last edited:
Sounds promising, I hope you find the time to document it here!
In the meantime, I started soldering one of my attenuator boards together, and I realised I fuc**ed up the order, I happened to order only 3 instead of 10 1uF caps. However, I found a couple of 100n caps in my parts bin, so I was wondering if I could swap them in instead. As far as I understood, C5-C15 are supply bypass caps, no? For the remaining 3 1UF caps, are there places where they would be beneficial? I would definitely put one across the ADC pins....
If 1uf is mandatory, I will have to place another order, and wait another week.....
Many thanks for your effort, and best regards!
hallodeletue
In the meantime, I started soldering one of my attenuator boards together, and I realised I fuc**ed up the order, I happened to order only 3 instead of 10 1uF caps. However, I found a couple of 100n caps in my parts bin, so I was wondering if I could swap them in instead. As far as I understood, C5-C15 are supply bypass caps, no? For the remaining 3 1UF caps, are there places where they would be beneficial? I would definitely put one across the ADC pins....
If 1uf is mandatory, I will have to place another order, and wait another week.....
Many thanks for your effort, and best regards!
hallodeletue
Last edited:
Sounds promising, I hope you find the time to document it here!
In the meantime, I started soldering one of my attenuator boards together, and I realised I fuc**ed up the order, I happened to order only 3 instead of 10 1uF caps. However, I found a couple of 100n caps in my parts bin, so I was wondering if I could swap them in instead. As far as I understood, C5-C15 are supply bypass caps, no? For the remaining 3 1UF caps, are there places where they would be beneficial? I would definitely put one across the ADC pins....
If 1uf is mandatory, I will have to place another order, and wait another week.....
Many thanks for your effort, and best regards!
hallodeletue
Sorry for the late answer, didn't see your post before now.
C5-C13 are supply bypass caps.
C14-C16 serve other functions and should be kept at their orginal 10nF, 100nF, 150pF values.
The 1uF are not critical.
I actually ended up using 10nF caps instead, because I screwed up when mounting the components and used the wrong caps and I did not feel like desoldering them and putting in the 1uF caps.
It is no problem at all, any size between 10nF-1uF will work just fine.
However, I would still advise to put a 1uF cap across the ADC pins. A lower value might work but I used 1uF and it seems to be working fine.
Wrt. the other project I mentioned, I will document it when I have something to show.
Thanks for the answer! I went to my local FabLab, and they had a full pack of 1uFs lying around. Lucky on that one. I soldered the last missing caps in, and fired the thing up powered by a regulated lab supply. No smoke, no excess current draw, just the onboard voltage regulator was a little warm to the touch. The relays reacted promptly to any turn on the pot, all fine so far....
So i was confident and installed the attenuator in my "The Wire" headphone amp (with changed input resistors to make up for the attenuator impedance), and discovered that the left channel was dead except for maybe the loudest 10%.
I measured the voltage after the regulator, on spot, i measured the attenuator resistors, and there we are: they were all way off spec, the smaller ones for a small percentage, but the higher ones were completely off... so maybe i burned them up during soldering. My old, crampy unregulated soldering iron has problems soldering on massive ground planes.... and maybe higher value thin film resistor are more sensitive to thermal abuse?
anyways, i will order replacements, and see where i can get.
best regards,
hallodeletue
So i was confident and installed the attenuator in my "The Wire" headphone amp (with changed input resistors to make up for the attenuator impedance), and discovered that the left channel was dead except for maybe the loudest 10%.
I measured the voltage after the regulator, on spot, i measured the attenuator resistors, and there we are: they were all way off spec, the smaller ones for a small percentage, but the higher ones were completely off... so maybe i burned them up during soldering. My old, crampy unregulated soldering iron has problems soldering on massive ground planes.... and maybe higher value thin film resistor are more sensitive to thermal abuse?
anyways, i will order replacements, and see where i can get.
best regards,
hallodeletue
Youre welcome.
I glad to hear you got it up and running.
I am a little mystified why your left channel is almost dead, I would not really expect that to happen.
Since the right channel is working(according to your testing) we now know that the control logic itself is working 100% and since it controls 7 dual relays both the left and right channel should be working identically. The only thing they do not have in common is the resistors and one half of each relay.
Interesting. Hard to say what it is without looking at it IRL.
Did you compare the left channel resistor values with the right channel resistor values?
Left and right channels are mirrored on the PCB layout so should be fast and easy to do.
I'm sorry I can not give you a better answer.
Fault finding IRL can be tricky, doing it over the internet just makes even trickier.
I do a fair amount of fault finding on PCBs, professionally, and sometimes you just want to take the easy way out and throw everything in the garbage.
If I had your PCB in front of me I could most likely find the error within 5 minutes. But that doesn't help you much.
Hi Neutrality,
thanks for your effort. I am really sorry for not getting back to you, but life has been pretty busy, and DIY-time was reduced to virtually zero.
Anyways, I troubleshooted as far as I could go, and in the end I found the usual culprits: bad solder joints, a short close to the SN74LVC2G08DCTR (and I managed to rip off one of it's tiny legs... damm, those things are small......). This tiny SMD-stuff is really getting to me. But so far, I managed to confirm that all the relais are opening and closing in a consistent way, so I think the general logic is OK. I ripped out all the attenuation resistors ( and lost a few in the process), and reordered new ones. Upon soldering them in, I wondered where the 10k output resistor that you mention in the thread, and that the marvellous online Logarithmic Attenuator Calculator indicates, was on the board? It's not there, isn't it? So I suppose I have to solder it, say, accross the in/output connector (left out/GND + right out/GND).
I hope that solves the riddle. By the way, do you recommend a logarithmic or a linear taper for the controlling pot?
I have to say, during the time that I spent in understanding your circuit for troubleshooting purpose, I had to admire the mere elegance and beauty of that approach! Many thanks for all the time you put into it, and for sharing it and making it accessible for noobs like myself!
Thanks for your efforts, and best regards,
hallodeletue
thanks for your effort. I am really sorry for not getting back to you, but life has been pretty busy, and DIY-time was reduced to virtually zero.
Anyways, I troubleshooted as far as I could go, and in the end I found the usual culprits: bad solder joints, a short close to the SN74LVC2G08DCTR (and I managed to rip off one of it's tiny legs... damm, those things are small......). This tiny SMD-stuff is really getting to me. But so far, I managed to confirm that all the relais are opening and closing in a consistent way, so I think the general logic is OK. I ripped out all the attenuation resistors ( and lost a few in the process), and reordered new ones. Upon soldering them in, I wondered where the 10k output resistor that you mention in the thread, and that the marvellous online Logarithmic Attenuator Calculator indicates, was on the board? It's not there, isn't it? So I suppose I have to solder it, say, accross the in/output connector (left out/GND + right out/GND).
I hope that solves the riddle. By the way, do you recommend a logarithmic or a linear taper for the controlling pot?
I have to say, during the time that I spent in understanding your circuit for troubleshooting purpose, I had to admire the mere elegance and beauty of that approach! Many thanks for all the time you put into it, and for sharing it and making it accessible for noobs like myself!
Thanks for your efforts, and best regards,
hallodeletue
Real life gets in the way sometimes.
Was unemployed for a year and had LOTS of time for diy, but then I got a temporary position at a comapny in June and worked there until I found permanent employment at another company 3 weeks ago. Suffice to say, my spare time has been limited and have not had much time to do any serious DIY, although I do have a few things in the works.
The 10K resistor should go between output from the attenuator and ground.
But remember, the attenuator has to see a TOTAL resistance of 10K from output to ground, so whatever input resistance you have in the circuit you put after the attenuator affects the termination resistance. If you put a 10K resistor from output to ground and the following circuit has an input resistance of 10K as well, your total termination resistance will only be 5K since 10K//10K is 5K.
I design most of my circuits with a 10K input resistance, which is why there is no place for a 10K termination resistor on the attenuator board itself. If you have 10K input resistance on the circuit that follows the attenuator, you just have to wire everything together and it will work as intended.
As for a volume pot, I would use a linear pot.
SMD stuff can be hard to solder right, especially some of the really small modern component packages. I managed to do it without much issue, using a loupe to check for solder bridges or bad solder joints, but it requires some experience. Now that I once again have permanent employment within the electronics industry, I have access to the proper magnifier equipment needed for serious SMD soldering.
I am happy to hear your like the work I put into this design.
To my knowledge I am the first to have implemented a fully HW based R-2R attenuator with the required timing needed to avoid clicks or pops. In the realm of DIY that is, there might be some companies out there who have done the same thing. I just wonder why no one in the DIY world has come up with something similar before, since once I decided how the overall design was going to be, the rest was really not that hard. Anyone with knowledge of designing digital logic should have been able to do it.
Was unemployed for a year and had LOTS of time for diy, but then I got a temporary position at a comapny in June and worked there until I found permanent employment at another company 3 weeks ago. Suffice to say, my spare time has been limited and have not had much time to do any serious DIY, although I do have a few things in the works.
The 10K resistor should go between output from the attenuator and ground.
But remember, the attenuator has to see a TOTAL resistance of 10K from output to ground, so whatever input resistance you have in the circuit you put after the attenuator affects the termination resistance. If you put a 10K resistor from output to ground and the following circuit has an input resistance of 10K as well, your total termination resistance will only be 5K since 10K//10K is 5K.
I design most of my circuits with a 10K input resistance, which is why there is no place for a 10K termination resistor on the attenuator board itself. If you have 10K input resistance on the circuit that follows the attenuator, you just have to wire everything together and it will work as intended.
As for a volume pot, I would use a linear pot.
SMD stuff can be hard to solder right, especially some of the really small modern component packages. I managed to do it without much issue, using a loupe to check for solder bridges or bad solder joints, but it requires some experience. Now that I once again have permanent employment within the electronics industry, I have access to the proper magnifier equipment needed for serious SMD soldering.
I am happy to hear your like the work I put into this design.
To my knowledge I am the first to have implemented a fully HW based R-2R attenuator with the required timing needed to avoid clicks or pops. In the realm of DIY that is, there might be some companies out there who have done the same thing. I just wonder why no one in the DIY world has come up with something similar before, since once I decided how the overall design was going to be, the rest was really not that hard. Anyone with knowledge of designing digital logic should have been able to do it.
Last edited:
Yeah, real life... ;-)
Glad to hear you're on the payroll again. But then again, if you had more time to create beauties like this one....
Thanks to all your tips, I finished that thingy now. I used the headphone amp's input impedance resistor as termination resistor for the attenuator, and in the end I decided to up the amp's gain to 6dB, and hey presto - what a brilliant combination!
Again, many thanks for all efforts, I really like that attenuator of yours! But excuse me now, I have an appointment with Ludwig van!
Best regards,
hallodeletue
Glad to hear you're on the payroll again. But then again, if you had more time to create beauties like this one....
Thanks to all your tips, I finished that thingy now. I used the headphone amp's input impedance resistor as termination resistor for the attenuator, and in the end I decided to up the amp's gain to 6dB, and hey presto - what a brilliant combination!
Again, many thanks for all efforts, I really like that attenuator of yours! But excuse me now, I have an appointment with Ludwig van!
Best regards,
hallodeletue
Hi Christian,
The +12VDC and GND goes to the power supply but where should the CTRL goes to ? How do you use it?
Is this the correct relay? Seems like the manufacturer part no from element14 does not match with the Mouser (G6K-2F-Y-TR DC12).
Omron Signal Relays
Thanks
The +12VDC and GND goes to the power supply but where should the CTRL goes to ? How do you use it?
Is this the correct relay? Seems like the manufacturer part no from element14 does not match with the Mouser (G6K-2F-Y-TR DC12).
Omron Signal Relays
Thanks
The CTRL should go to +12V. It is a control signal for the digital logic +5V regulator. If you connect it to GND, the +5V regulator goes OFF, if you connect it to +12V, the +5V regulator goes ON. You want it to be ON if you want to use the attenuator, if you turn if OFF the attenuator turns OFF and you get maximum attenuation(-82dB).
That is the correct relay. Farnell just missed the - between 2F and Y.
That is the correct relay. Farnell just missed the - between 2F and Y.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.