Hello everyone,
In my experience, Gustard has more creativity on its sound tuning than those two SOTA DAC Chinese manufacturer Topping and SMSL.
I got some information from China audiophile forum couple of years ago that Gustard plan to design a R2R DAC and announced R26 on 2022.
The front panel is made with aluminum combine a black acrylic window, the ID design is quite different from previously DAC and I like the new style.
The volume/function key changed back as A18.
Base plate, the new aluminum feet are nice look and stable. The sandblasting finish on new chassis is also much better than previously brushed finish.
There are 4 blocks in side, toroidal transformer in the left, voltage regulator and clock generator are in the middle.
Net bridge board stands in the center and rest DAC circuit is in the right.
Gustard did transformer isolation block with aluminum plate couple of years ago.
It used two customized transformer 50VA 15V*2 for both digital/analog voltage regulators.
Compared with X22, transformer changed from 18V to 15V which may reduce heat from regulators in next stage.
The power supply, the first stage of analog power used TIP122/TIP127 to drop DC20.2V to 17.2V
I believe the first stage is only for voltage drop with 5pcs diode and BJT to reduce heat on LM317.
The digital power source came from 2pcs SCT2432 and step down from DC 19V to 7.2V.
I am not sure there is any benefit to use step down converter.
The clock generator used Gustard K2 ultra low noise sync module with 201.1M Femto clock.
It looks complex design but unfortunately the 12K jitter measurement from l7audiolab is not good.
it's pity on this luxury design.
Net bridge used SigmaStar SSD202D, and the SanDisk 32G TF card stored OS as an embedded Raspberry Pi.
Net bridge is different design between my version and l7audiolab, and BT solution changed from 8675 to QCC5125.
The FPGA of digital processer is covered by sticker, not sure why to cover FPGA.
The analog overview, R2R board put on base board. And the FPGA on R2R board is also covered by sticker.
The architecture of R2R board looks similar as Soekris DAM1021, including those level shifter ICs and Vref made by OPA.
But the ladder resistors seem normal SMD resistors not the precision resistor as dam1021, not HiFi for me.
LDO SGM2209/SGM2211 used to regulate from ±12.6V to ±6.2V and used OPA1612 to ±4V.
The reference voltage of positive voltage came from LT3042 which is really good, but I am not sure how to do for negative voltage.
LPF used TI OPA1678 which is the same as DAM1021 ver.7.
Regarding to Soekris Soren that OPA1678 sounds a little bit slower than OPA1656 with lower power consumption and cheaper.
R26 add 0ohm resistor on ladder looks weird for me which may make tolerance higher since they didn’t use precision resistors.
The discrete SDM to DSD circuit. Unfortunately, user feedbacks of DSD output are not good as PCM from another forum.
DSD LPF used OPA1602 which is an ultra-low noise OPA.
Analog output side, there is 6 sets of LM317/337 as power supply to offer R2R board/ DSD circuit / buffer circuit.
The output buffer circuit is in the right side with discrete components and OPA1642 as DC servo.
All 6 sets of LM317/337 are regulated to ±12.6V
The major capacitors used Nichicon Gold tune 6800uF 35V as usual, 4pcs for analog and 2pcs for digital power source.
In first stage of analog power output caps used Nichicon KW 470uF 25V with WIMA film caps.
The bypass caps of analog output used Nichicon KW 100uF 50V.
Since R26 sounds like the original DAM1021, take a pic with my DAC with DAM1021 module call " The Racket DAC".
Let me know if you have any questions on this teardown report, thanks.
In my experience, Gustard has more creativity on its sound tuning than those two SOTA DAC Chinese manufacturer Topping and SMSL.
I got some information from China audiophile forum couple of years ago that Gustard plan to design a R2R DAC and announced R26 on 2022.
The front panel is made with aluminum combine a black acrylic window, the ID design is quite different from previously DAC and I like the new style.
The volume/function key changed back as A18.
Base plate, the new aluminum feet are nice look and stable. The sandblasting finish on new chassis is also much better than previously brushed finish.
There are 4 blocks in side, toroidal transformer in the left, voltage regulator and clock generator are in the middle.
Net bridge board stands in the center and rest DAC circuit is in the right.
Gustard did transformer isolation block with aluminum plate couple of years ago.
It used two customized transformer 50VA 15V*2 for both digital/analog voltage regulators.
Compared with X22, transformer changed from 18V to 15V which may reduce heat from regulators in next stage.
The power supply, the first stage of analog power used TIP122/TIP127 to drop DC20.2V to 17.2V
I believe the first stage is only for voltage drop with 5pcs diode and BJT to reduce heat on LM317.
The digital power source came from 2pcs SCT2432 and step down from DC 19V to 7.2V.
I am not sure there is any benefit to use step down converter.
The clock generator used Gustard K2 ultra low noise sync module with 201.1M Femto clock.
It looks complex design but unfortunately the 12K jitter measurement from l7audiolab is not good.
it's pity on this luxury design.
Net bridge used SigmaStar SSD202D, and the SanDisk 32G TF card stored OS as an embedded Raspberry Pi.
Net bridge is different design between my version and l7audiolab, and BT solution changed from 8675 to QCC5125.
The FPGA of digital processer is covered by sticker, not sure why to cover FPGA.
The analog overview, R2R board put on base board. And the FPGA on R2R board is also covered by sticker.
The architecture of R2R board looks similar as Soekris DAM1021, including those level shifter ICs and Vref made by OPA.
But the ladder resistors seem normal SMD resistors not the precision resistor as dam1021, not HiFi for me.
LDO SGM2209/SGM2211 used to regulate from ±12.6V to ±6.2V and used OPA1612 to ±4V.
The reference voltage of positive voltage came from LT3042 which is really good, but I am not sure how to do for negative voltage.
LPF used TI OPA1678 which is the same as DAM1021 ver.7.
Regarding to Soekris Soren that OPA1678 sounds a little bit slower than OPA1656 with lower power consumption and cheaper.
R26 add 0ohm resistor on ladder looks weird for me which may make tolerance higher since they didn’t use precision resistors.
The discrete SDM to DSD circuit. Unfortunately, user feedbacks of DSD output are not good as PCM from another forum.
DSD LPF used OPA1602 which is an ultra-low noise OPA.
Analog output side, there is 6 sets of LM317/337 as power supply to offer R2R board/ DSD circuit / buffer circuit.
The output buffer circuit is in the right side with discrete components and OPA1642 as DC servo.
All 6 sets of LM317/337 are regulated to ±12.6V
The major capacitors used Nichicon Gold tune 6800uF 35V as usual, 4pcs for analog and 2pcs for digital power source.
In first stage of analog power output caps used Nichicon KW 470uF 25V with WIMA film caps.
The bypass caps of analog output used Nichicon KW 100uF 50V.
Since R26 sounds like the original DAM1021, take a pic with my DAC with DAM1021 module call " The Racket DAC".
Let me know if you have any questions on this teardown report, thanks.
Last edited:
