...
Another week later, and now they are up to $22.50!!!
Maybe this thread is monitored by chip sellers, we should all chat in secret code from now on!!
exactlyThey monitor sales/interest and many people bought this chip in a short time, simple law of demand and offer ... this thread helped
Hi,
Could I use the follow bridge rectifier with a 2200uf cap before this board?:
DF005-G Comchip Technology | Mouser Espana
Thanks
Could I use the follow bridge rectifier with a 2200uf cap before this board?:
DF005-G Comchip Technology | Mouser Espana
Thanks
yes it is possible, like this:
transformator -- diode bridge + filtration -- voltage stabilizer -- DAC
... or this bridge rectifier is more robust: 625-PB4008-E3/45
the output DC voltage after filtration is:
AC voltage from transformer multiply by 1.414, if AC voltage from transformer is 2x 12Vac than after filtration it will be 2x 17Vdc ... this goes into voltage stabilizer which produces the clean voltages for DAC (+-12Vdc, +-5Vdc)
transformator -- diode bridge + filtration -- voltage stabilizer -- DAC
... or this bridge rectifier is more robust: 625-PB4008-E3/45
the output DC voltage after filtration is:
AC voltage from transformer multiply by 1.414, if AC voltage from transformer is 2x 12Vac than after filtration it will be 2x 17Vdc ... this goes into voltage stabilizer which produces the clean voltages for DAC (+-12Vdc, +-5Vdc)
Last edited:
That's of course the best way to feed it. The AD1862 is more versatile for that task than the PCM63P cause the digital side is not limited to 5.5/6.0 V.
Best choice of course LiFePo cells. they are 3.3V at their max but show very fastly 3.2 V into load, at least the genuine A123 brand I have. To be a noticeable improvment the LiFePo cells must be directly connected on the load without regs in between. the length of wires must be the shortest possible and the 2 wires wounded : ask a little 3D upon the board diy and creativity.
So two serie LiFePo4 give 6.6 V max for the digital side : perfect !
And 4 in serie, ait I do the math... suspense.... : 13.2 V.
Now it's possible it will ask for a short decoupling on the power pins of the ad1862 : 0.01 uF ; 0.1 uF ; or 1 uF ??? : jusst avoid X7R caps - my personal subjective biased choice my ears thank- Try instead 2.5 mm pitch cap : mkt,mkp or smd NPO or paired SMD PPS 5% precision or less if exist - for this last: my personal biased etc- ....- @ JosephK: oh the japonese AK9099 guys did that also -
Only a scope measurment could tell if deoupling is needed here or a trade off choice made by listening if no scope imho according how it distors best for our ears. Anyway the real improvment here should come from a I2S front end alsopower fed by the cells and a ground breaker between the data source -playback- and the I2S input.
The temperature climbs with my diy fever and about the opa choice, I'm more and more attracted by adapting the opa 861(that means no headphones if no further buffer) or a Jfet 1556 from Ti or 8610... maybe the opa210... but without being tech and re read some post the feedback oaps sucks for I/V cause linearity... but the opa 861 that is the ideal oap, the rest are mostly good because also their buffer quality in that one oap design we tal about in the thread. However that's theory and ears don't care : the 1556 from Ti -double channel that has not really same spec in single channel, but has a non audio cunterpart in the TI catalog iir- which is not bipolar as the opa861 but Jfet and transimpedance vs transconductance shown subjective good critics in I/V purpose... the numbers are so low perhaps the ears are not ensible anymore...
the LM6176 seems ideal caue very fast and ideal to load headphones- but my understanding after some readings is in I/V its current imput sensisivity copes bad ith a low current output dac cause it's creating instability that seems less terrible with a discrete diamond stage as the Painkiller DAC made for the AD1862 on a discrete design published by P.Rodgic. Btw because the TDA1541A +/- outputt curent there is in the AYAs design of Pedja Rogic a discrete JFET as CCS before the opa861, and my very basic understanding that is not technical is also it is contributing to the gound sound of the opa861 in his design
BiFet could be cool... wait, they are opa627 AD797 (can be used with no external feedback?)....
Time to make the BOM at Mouser, I waited answers of the FifoPi tweak thread in the PC forums 2 days ago about a BOM cause I have few time to read it, but had no answers - = RTFM - I only purchase some supercaps too and some more PPS caps from Rubycon that comes in several uF values
Vishay Roe for the mkp - personal choice - and 2.5 mm ptch MKT from Wima.
Sorry to be a little late for the populating- but summer, sars2 - that's not an oap- ; distanciation- sometimes my laptop is also scarying me... dunno if the cat is chatting at night lol all that are slowling me.
that's cool to have a diy pcb where we can diy around it, thanks Miro1360 again for that.
Best choice of course LiFePo cells. they are 3.3V at their max but show very fastly 3.2 V into load, at least the genuine A123 brand I have. To be a noticeable improvment the LiFePo cells must be directly connected on the load without regs in between. the length of wires must be the shortest possible and the 2 wires wounded : ask a little 3D upon the board diy and creativity
.So two serie LiFePo4 give 6.6 V max for the digital side : perfect !
And 4 in serie, ait I do the math... suspense.... : 13.2 V.
Now it's possible it will ask for a short decoupling on the power pins of the ad1862 : 0.01 uF ; 0.1 uF ; or 1 uF ??? : jusst avoid X7R caps - my personal subjective biased choice my ears thank- Try instead 2.5 mm pitch cap : mkt,mkp or smd NPO or paired SMD PPS 5% precision or less if exist - for this last: my personal biased etc- ....- @ JosephK: oh the japonese AK9099 guys did that also
- Only a scope measurment could tell if deoupling is needed here or a trade off choice made by listening if no scope imho according how it distors best for our ears. Anyway the real improvment here should come from a I2S front end alsopower fed by the cells and a ground breaker between the data source -playback- and the I2S input.
The temperature climbs with my diy fever and about the opa choice, I'm more and more attracted by adapting the opa 861(that means no headphones if no further buffer) or a Jfet 1556 from Ti or 8610... maybe the opa210... but without being tech and re read some post the feedback oaps sucks for I/V cause linearity... but the opa 861 that is the ideal oap, the rest are mostly good because also their buffer quality in that one oap design we tal about in the thread. However that's theory and ears don't care : the 1556 from Ti -double channel that has not really same spec in single channel, but has a non audio cunterpart in the TI catalog iir- which is not bipolar as the opa861 but Jfet and transimpedance vs transconductance shown subjective good critics in I/V purpose... the numbers are so low perhaps the ears are not ensible anymore...
the LM6176 seems ideal caue very fast and ideal to load headphones- but my understanding after some readings is in I/V its current imput sensisivity copes bad ith a low current output dac cause it's creating instability that seems less terrible with a discrete diamond stage as the Painkiller DAC made for the AD1862 on a discrete design published by P.Rodgic. Btw because the TDA1541A +/- outputt curent there is in the AYAs design of Pedja Rogic a discrete JFET as CCS before the opa861, and my very basic understanding that is not technical is also it is contributing to the gound sound of the opa861 in his design
BiFet could be cool... wait, they are opa627
AD797 (can be used with no external feedback?).... Time to make the BOM at Mouser, I waited answers of the FifoPi tweak thread in the PC forums 2 days ago about a BOM cause I have few time to read it, but had no answers - = RTFM
- I only purchase some supercaps too and some more PPS caps from Rubycon that comes in several uF values Vishay Roe for the mkp - personal choice - and 2.5 mm ptch MKT from Wima.
Sorry to be a little late for the populating- but summer, sars2 - that's not an oap- ; distanciation- sometimes my laptop is also scarying me... dunno if the cat is chatting at night
lol all that are slowling me.that's cool to have a diy pcb where we can diy around it, thanks Miro1360 again for that.
LM6172 opamp has very good input stage for the I/V conversion created from BJT transistors. BJT has a significant advantage in I/V, because it's characteristics is linear while FET's have non-linear characteristics. This is perhaps the reason why jFETs sound softer on mids and highs, and lows are even deeper, less natural a kind of dynamic equalizer or effect, some people like it
a kind of dynamic equalizer or effect, some people like it
Yes in theory. But here from what I read he would create an instability due to the impedance iirc. And Jfet in the case of the 1556 are so good that the correct theory exposed is not true anymore. I try to found the link whee I read that. The LM will be better in buffer position. Read some posts from Joko Homo, Pedja Rogic, Elso Kwak... no doubt one should avoid feedback when it comes to I/v and linearity for the few I understood and to make it short. Hence opa660, opa861, AD844...
Of course I cut the sound wave in 4... but as I have many dacs, it's about to push your design the best I can with having some fun and the illusion I can tweak it a little.
Now all these question of input stage, chips beeing so good is less important ihmo, it's certainly also about to find a trade off that cope with the rest of the devices... LM4562, opa2604 have certainly things to say; opa2604, only 100 uV slew rate but its mellow sound can help in some designs, etc... why not the ad80665/66 as well ? too much things to test... if we accept not going for a single channel oap and play with creativity & adaptators as per the how to wrote some posts above !
Or one chip for the tw channels an AD13700 dip 16 chip ?
Of course I cut the sound wave in 4... but as I have many dacs, it's about to push your design the best I can with having some fun and the illusion I can tweak it a little
.Now all these question of input stage, chips beeing so good is less important ihmo, it's certainly also about to find a trade off that cope with the rest of the devices... LM4562, opa2604 have certainly things to say; opa2604, only 100 uV slew rate but its mellow sound can help in some designs, etc... why not the ad80665/66 as well ? too much things to test... if we accept not going for a single channel oap and play with creativity & adaptators as per the how to wrote some posts above !
Or one chip for the tw channels an AD13700 dip 16 chip ?
Last edited:
@surfparadise
I am unable to share the original files
It is free design, so you can easily follow/replicate all traces in Kicad. Alternatively you can try importing the gerbers to your design if Kicad has such functionality. Or try to merge your gerbers with my gerbers with some merge tool, if something exists.
I still recommend doing it modularly with short wire connections.
@diyiggy
open loop design is something we need to finish and test
I am unable to share the original files
It is free design, so you can easily follow/replicate all traces in Kicad. Alternatively you can try importing the gerbers to your design if Kicad has such functionality. Or try to merge your gerbers with my gerbers with some merge tool, if something exists.
I still recommend doing it modularly with short wire connections.
@diyiggy
open loop design is something we need to finish and test
this story smells aops that are about to finish their life as piercing jewells...
.. ...
@surfparadins : pleasure is in the tweaking and power supply experiment is your friend, no needs here to redesign and no real advantages if not designing a front end and another outputt stage and the grounding for all of that... which drive you to the whole diy circuit again from scratch : beginning from the ground, add layers like lasagna, then see what happen
take the verroboard way or the modular way (ebay kits with ldo reg chips, etc). The square board is little enough to use the wire road to experient and it is the fun to improve it little by little... imho, ymmv though
.. ...@surfparadins : pleasure is in the tweaking and power supply experiment is your friend, no needs here to redesign and no real advantages if not designing a front end and another outputt stage and the grounding for all of that... which drive you to the whole diy circuit again from scratch : beginning from the ground, add layers like lasagna, then see what happen
take the verroboard way or the modular way (ebay kits with ldo reg chips, etc). The square board is little enough to use the wire road to experient and it is the fun to improve it little by little... imho, ymmv though
“.....beginning from the ground, add layers like lasagna, then see what happen”
Now we are takin’ my favorite language FOOD!! Yummy
Fellas,
I’ve been fascinated with the OPA861 I/V conversion-buffer that Diyiggy has first hand experience with. I tried a simple circuit using the OPA860, but didn’t like the results. Whether it was my non-ideal layout possibly causing oscillation or it was just not a good match, I don’t know.
I decided to try and implement the Aya2 opa861 I/V output section with Miro’s Dac board. I have a small pcb on the way that will be populated with SMD components and utilize the original opamp socket as the interface for power, ground, Iin and Vout. I hope it works as planned...... stay tuned.
”Now we are takin’ my favorite language FOOD!! Yummy
Fellas,
I’ve been fascinated with the OPA861 I/V conversion-buffer that Diyiggy has first hand experience with. I tried a simple circuit using the OPA860, but didn’t like the results. Whether it was my non-ideal layout possibly causing oscillation or it was just not a good match, I don’t know.
I decided to try and implement the Aya2 opa861 I/V output section with Miro’s Dac board. I have a small pcb on the way that will be populated with SMD components and utilize the original opamp socket as the interface for power, ground, Iin and Vout. I hope it works as planned...... stay tuned.