Quick update:
I got a new drixo to set as well the 6.1 frequency. two weeks italian poste is a mess (to germany)...but build in a day...burning in...
Finished the PP reg and will now observe its performance vs the linear reg version...battery version in final stage as well...next.
De-soldered the I2S connectors of the fifo and exchanged the chinese 50ct connector/cable I2S to a teflon solid core 5cm/2" connection from the i2soverusb card to the fifo...maybe its just imagination, but seems to be better resolution with more body/substance...i highly recommend this little mod.
Prepared the sdtrans384ßßßfully connected already with battery supply, but need a bid more time to really start comparison...got other private committments, a bit time contrained...
I got a new drixo to set as well the 6.1 frequency. two weeks italian poste is a mess (to germany)...but build in a day...burning in...
Finished the PP reg and will now observe its performance vs the linear reg version...battery version in final stage as well...next.
De-soldered the I2S connectors of the fifo and exchanged the chinese 50ct connector/cable I2S to a teflon solid core 5cm/2" connection from the i2soverusb card to the fifo...maybe its just imagination, but seems to be better resolution with more body/substance...i highly recommend this little mod.
Prepared the sdtrans384ßßßfully connected already with battery supply, but need a bid more time to really start comparison...got other private committments, a bit time contrained...
Ok...the PP is still burning in...but I am a bit shocked to be honest.
I compared the burning in PP with the serial reg of Andrea tonight. All on Drixo.
The change in sound of the whole system is at least at the same magnitude of exchanging the fifo dac against the soekris (maybe a bit exegurated but the tendency is correct):
The burned in serial reg plays extremly transparent, the trebles are just vibrating through the room...effortless...like a very good ribbon delivers...(got 12 of Fountek Fp-2 in my line source)...BUT the substance in the mids is thin. A bit fleshless.Wood does not sound like wood...but the trebles are...just awesome.
The PP...gives you the substance...earth...wood....BUT reduced the resolution in the trebles considerably...really a BIG difference....OK..its just playing now for 48hours...maybe this will be cured...and anyhow, I plan to get rid of the AC part on both regs and give them the same LT4320/Nichicon Superthrough Raw supply...
BUT
the most important learning for me:
The Power Supply of the clocks determines /dominates the sound of the whole system in a way I never expected.
You could say...no news...BUT the difference between the crystek 957 and the Drixo was not as big as the difference now in the PSU of the same Drixo...
Which is impressive.
I am glad to have Andrea's battery supply next to be finished...but lets gice the PP a chance to burn in as well...and the Salas reg is on the list still as well...and if nothing completely shines I will try the Ucpure solution of Ian...but thats another 500 Euro ticket...just for an experiment ?!?
I compared the burning in PP with the serial reg of Andrea tonight. All on Drixo.
The change in sound of the whole system is at least at the same magnitude of exchanging the fifo dac against the soekris (maybe a bit exegurated but the tendency is correct):
The burned in serial reg plays extremly transparent, the trebles are just vibrating through the room...effortless...like a very good ribbon delivers...(got 12 of Fountek Fp-2 in my line source)...BUT the substance in the mids is thin. A bit fleshless.Wood does not sound like wood...but the trebles are...just awesome.
The PP...gives you the substance...earth...wood....BUT reduced the resolution in the trebles considerably...really a BIG difference....OK..its just playing now for 48hours...maybe this will be cured...and anyhow, I plan to get rid of the AC part on both regs and give them the same LT4320/Nichicon Superthrough Raw supply...
BUT
the most important learning for me:
The Power Supply of the clocks determines /dominates the sound of the whole system in a way I never expected.
You could say...no news...BUT the difference between the crystek 957 and the Drixo was not as big as the difference now in the PSU of the same Drixo...
Which is impressive.
I am glad to have Andrea's battery supply next to be finished...but lets gice the PP a chance to burn in as well...and the Salas reg is on the list still as well...and if nothing completely shines I will try the Ucpure solution of Ian...but thats another 500 Euro ticket...just for an experiment ?!?
Blitz
Interested in your observations. If I understand correctly, you are reporting significant change in sound using the TWRPS-pp regulator on the drisco clock vs a serial reg. Could you describe the rest of the power supply please?
I had come to the conclusion that the power supply on the Drisco was not as critical as it is in other areas of the system. So I built a very good linear supply with low noise reg, but really did not do much experimentation. So, I am interested to follow your adventure.
Interested in your observations. If I understand correctly, you are reporting significant change in sound using the TWRPS-pp regulator on the drisco clock vs a serial reg. Could you describe the rest of the power supply please?
I had come to the conclusion that the power supply on the Drisco was not as critical as it is in other areas of the system. So I built a very good linear supply with low noise reg, but really did not do much experimentation. So, I am interested to follow your adventure.
Try loading a serial reg with a resistor to ground. If the sound improves with some particular value of resistance then at a minimum it tells you the reg was not optimal as it was. Suggest to use a good quality resistor such as Dale RN65 or something more or less like that. End-cap distortion could introduce some excess noise into the load if not a high quality resistor.
I am with you...I did nit expect a big difference either...I have the recommended two chamber ei xformer on the serial reg and a small (0.8A) two chamber xformer on the PP...rest on the fifo, dac is ultrabib. Which i used as well for the Soekris.
On the DHT front I experimented a lot with Rod colemans regs and learned that each version of rods regs have a different sound signature and that the raw supply is as important and audible....through the reg...but there we are on the filament/cathode of the Triode...so fully in the analog signal,path....now i learn the same sensitivity in PSU for clocks...OK, so lets have some fun with this...
OK, found the schematic of the PP reg...did a quick fix:
Got rid of the 3.3 ohm resistor and integrated a 0.17Ohm/20mH amorphous choke instead...pulled the ceramic snubber caps (have experimented with this and quasimodo for a while, but never convinced me...sounds always kind of damped)...and voila...we have a different piece of equipment already. Rest ripple is now 4mV vs 30mV (from the raw supply) before which is a nice extra, but in my experience a low Rdc Choke always sounds much better than a Resistor...and lowers the impedance of the raw supply as well...and yes, In theory this all should not matter as we have a reg next...but that is just theory .....It does matter. Now we have a lot more air AND substance as well...you CAN have both.
...but this is only an intermediate step...
If I get it right, Andrea has as well some regs on the Drixo, no ? So I wonder if a complete passive solution is not worth to try...I mean 4mV...I can get this dow to less than 1mV with some bigger caps...and with a little cap multiplier even further....plus the benefit of a noise floor down by 70db...and a cap multiplier is nearly not audible compared to a reg, its much more open sounding in my experience (with tube filaments).
Got rid of the 3.3 ohm resistor and integrated a 0.17Ohm/20mH amorphous choke instead...pulled the ceramic snubber caps (have experimented with this and quasimodo for a while, but never convinced me...sounds always kind of damped)...and voila...we have a different piece of equipment already. Rest ripple is now 4mV vs 30mV (from the raw supply) before which is a nice extra, but in my experience a low Rdc Choke always sounds much better than a Resistor...and lowers the impedance of the raw supply as well...and yes, In theory this all should not matter as we have a reg next...but that is just theory .....It does matter. Now we have a lot more air AND substance as well...you CAN have both.
...but this is only an intermediate step...
If I get it right, Andrea has as well some regs on the Drixo, no ? So I wonder if a complete passive solution is not worth to try...I mean 4mV...I can get this dow to less than 1mV with some bigger caps...and with a little cap multiplier even further....plus the benefit of a noise floor down by 70db...and a cap multiplier is nearly not audible compared to a reg, its much more open sounding in my experience (with tube filaments).
Last edited:
I am using 15V...on all the regs I tried so far on the clocks...
I desoldered the white cheap connector for I2S from the board as I hate unnecessary material changes or surfaces of connectors etc...I soldered a solid core ptfe twisted pair wire directly from i2sover usb card to the pcb...
Later I will as well desolder the power connectors from the Fifo and the dac and solder the wire straight into the pcb...it does not make sense to think about ultra low esr solutions and than add many mohm through these kind of connectors...i will as well solder the lithium cells straight onto the pcbs as i did with ian battery supply as well, so no connection losses.
Currently I am listening and the plus in resolution using this little choke trick is stunning...its like room depth and blackness of backround has improved a lot...I will change the inductance from 20mH to 80mH later (build some switches in, so easy) and see if the extra in filtering is a benefit or the extra in Rdc is a malus.
I desoldered the white cheap connector for I2S from the board as I hate unnecessary material changes or surfaces of connectors etc...I soldered a solid core ptfe twisted pair wire directly from i2sover usb card to the pcb...
Later I will as well desolder the power connectors from the Fifo and the dac and solder the wire straight into the pcb...it does not make sense to think about ultra low esr solutions and than add many mohm through these kind of connectors...i will as well solder the lithium cells straight onto the pcbs as i did with ian battery supply as well, so no connection losses.
Currently I am listening and the plus in resolution using this little choke trick is stunning...its like room depth and blackness of backround has improved a lot...I will change the inductance from 20mH to 80mH later (build some switches in, so easy) and see if the extra in filtering is a benefit or the extra in Rdc is a malus.
Hi .. No, I don't think so. The U1 is a voltage reference and Q3 & Q4 are transistors, if I am not mistaken.
If it were me, IMHO I would try first with batteries directly connected to the DRIXO(s). E.g. a lead acid 12V or 4 LiFePO4s in series. That would give the lowest noise, lowest impedance connection to the DRIXOs. Just my two cents ...
Cheers,
Jesper
The Ultrabib in 15V is still in the waiting queue...I work slowly through the list as I want each to have 3-4 weeks full burnin time before coming to conclusions...
Jesper, I like to enjoy learning the differences, thats why I left the battery supply as the last step as I expect it to be maybe the best...and is the most building work...I want to see first how well I can do with normal psu/regs or even passive solutions...
Jesper, I like to enjoy learning the differences, thats why I left the battery supply as the last step as I expect it to be maybe the best...and is the most building work...I want to see first how well I can do with normal psu/regs or even passive solutions...
... Well, then ... might I then suggest that you make the last trial be batteries soldered together and directly to the DRIXO e.g. with just copper wires .. ? Just a thought Hi again,
To my understanding the Ultracapacitors can be very low impedance but I have not yet found any individual cell that is capable of working at more than ~3-3.3 volts. Thus to achieve the voltages needed for the DRIXO one needs to series connect at least four cells and then I guess the difference between such a setup and a battery (e.g. 12V/7AH SLA) will only be a few milliohms. I doubt this in itself will be audible - likely the impedance curve vs frequency for the individual full solution (e.g. including inductances in wires) may then be what leads to differences - if those are found.
Also, I did some listening tests (series connection with headphones) on some A123 LiFePO4 cells and a 50F Maxwell ultracapacitor and found that the sound difference was quite marked. The Ultracapacitor was relatively dull sounding & with a grayish tonal quality whereas the LiFePO4 cells sounded dynamic & just with a tad brightness and sonority in the lower upper frequencies.
Just FYI should it be of interest.
Cheers, Jesper
To my understanding the Ultracapacitors can be very low impedance but I have not yet found any individual cell that is capable of working at more than ~3-3.3 volts. Thus to achieve the voltages needed for the DRIXO one needs to series connect at least four cells and then I guess the difference between such a setup and a battery (e.g. 12V/7AH SLA) will only be a few milliohms. I doubt this in itself will be audible - likely the impedance curve vs frequency for the individual full solution (e.g. including inductances in wires) may then be what leads to differences - if those are found.
Also, I did some listening tests (series connection with headphones) on some A123 LiFePO4 cells and a 50F Maxwell ultracapacitor and found that the sound difference was quite marked. The Ultracapacitor was relatively dull sounding & with a grayish tonal quality whereas the LiFePO4 cells sounded dynamic & just with a tad brightness and sonority in the lower upper frequencies.
Just FYI should it be of interest.
Cheers, Jesper
Yes, this is very interesting...did you try as well LiFePo4 cells with a cap ? Some people reported an improvement ?
My favorites so far on the low voltage front are the Nichicon Super-through...compared those with many highly regarded other caps, but these are very neutral and transparent sounding while maintaining a nice tone...(in DHT-filament application). ON the HV front clearly CDE 947D, so big polyprops with around 2mOhm.
My favorites so far on the low voltage front are the Nichicon Super-through...compared those with many highly regarded other caps, but these are very neutral and transparent sounding while maintaining a nice tone...(in DHT-filament application). ON the HV front clearly CDE 947D, so big polyprops with around 2mOhm.
@Blitz:
). As to great capacitors I have not yet found the "ideal" solution but Nichicon KZ, KG (not the super through but the "B" version) currently are on my short list.
Have you BTW read into the "best electrolytic capacitors" thread? Some good suggestions there ...
Cheers, Jesper
No, I have not tried this. Would not be surprised if it made a difference though since both Ultracapacitors & LiFePO4s have an increasing impedance curve going up in frequency. My guess would be that in relation to the Drixo some additional 22uF - like the ones used in the DRIXO - and some larger capacitances would improve sound (my guess that is
). As to great capacitors I have not yet found the "ideal" solution but Nichicon KZ, KG (not the super through but the "B" version) currently are on my short list.Have you BTW read into the "best electrolytic capacitors" thread? Some good suggestions there ...
Cheers, Jesper
Hi @Blitz, I appreciate that you've already responded to this but I didn't understand your response. Can you help me to identify the following parts please:
- The PCB with the u.fl headers. Is that from IanCanada?
- The name of the white sockets that Andrea has used for the I2S connections and their corresponding plugs. If I can't find your u.fl header PCB I am going to buy some and crimp u.fl leads to them from here: https://rfsupplier.com/customcable.html
White I2S connectors look a lot like the following type: https://www.amazon.com/gp/product/B01MCZE2HM/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1
Also might mention IME tin connectors can lose contact pressure and develop corrosion over time. When that happens clock jitter produced by such connections can become a problem. IME gold plated or gold flashed connectors can prevent the corrosion problem from occurring.
One other thing maybe worth mentioning is that I have seen stray coupling between USB boards and dac boards (and or possibly FIFO boards) through radiated emissions. Experiments showed that putting the USB board underneath the dac board with the USB board ground plane facing the dac board ground plane resulted in a minimum of radiated stray coupling and best SQ. More generally, for best dac performance its best not to take anything for granted. Finding and fixing problems sometimes resembles peeling away layers of an onion -- as one problem is fixed an formerly masked problem becomes audible, so troubshooting can then begin for that new problem. Stray coupling mechanisms can be one of those things that seemed not to be a problem before, but later may become detectable.
Also might mention IME tin connectors can lose contact pressure and develop corrosion over time. When that happens clock jitter produced by such connections can become a problem. IME gold plated or gold flashed connectors can prevent the corrosion problem from occurring.
One other thing maybe worth mentioning is that I have seen stray coupling between USB boards and dac boards (and or possibly FIFO boards) through radiated emissions. Experiments showed that putting the USB board underneath the dac board with the USB board ground plane facing the dac board ground plane resulted in a minimum of radiated stray coupling and best SQ. More generally, for best dac performance its best not to take anything for granted. Finding and fixing problems sometimes resembles peeling away layers of an onion -- as one problem is fixed an formerly masked problem becomes audible, so troubshooting can then begin for that new problem. Stray coupling mechanisms can be one of those things that seemed not to be a problem before, but later may become detectable.
Last edited: