thommy, did you get this sorted? Sounds very much like one of your op-amps has border-line stability, perhaps decoupling is not optimum on that board. Which op-amps are you using? May be worth trying something slower.
The avcc has dual LT1963-eq reg's one for left and one for right. Should I add a dual TPA trident?
The crystal is a JYEC, should I replace it with an Epson and adapter pcb?
XG-1000CA 106.2500M-CBL3
The crystal is a JYEC, should I replace it with an Epson and adapter pcb?
XG-1000CA 106.2500M-CBL3
Last edited:
The LTs are decent regs, but there are certainly better. You could go with the Tridents, ro better still, the Paul Hynes regs. DEXAs are supposed to be good too, although I've not tried them.
I don't know how good the JYEC is, but, like Avcc, make sure you power it with the best regs you can find.
I don't know how good the JYEC is, but, like Avcc, make sure you power it with the best regs you can find.
I figured the 5 and the 7 couldn't be too far apart, since the rest of the boards are pretty similar. The mkp I found were surplus ero, not the roe I had mentioned before.
Another big step up was the addition of some X type film caps on the incoming ac, wired between the hot and the neutral. By getting rid of some of the noise up front, less amplified noise. This benefit may have been partly due to an industrial environment, but I suspect most would benefit from it, was inexpensive.
Rubber feet on the case, and shielding the transformer can help lower the noise floor a bit too.
Everything else I did had a less effective cost/benefit.
Another big step up was the addition of some X type film caps on the incoming ac, wired between the hot and the neutral. By getting rid of some of the noise up front, less amplified noise. This benefit may have been partly due to an industrial environment, but I suspect most would benefit from it, was inexpensive.
Rubber feet on the case, and shielding the transformer can help lower the noise floor a bit too.
Everything else I did had a less effective cost/benefit.
Update: I put a trident into the 4layer dac, removed every other cap on the Avcc and powered it with a salas reg. It now sounds better than it ever has but is still short on the top end and further on the low end, in comparison with the dac7.
I have the es9018 put into current mode with an 11k resistor. Perhaps I should load it further - use the opamp outputs.
I have the es9018 put into current mode with an 11k resistor. Perhaps I should load it further - use the opamp outputs.
Do you mean putting an 11k resistor from DAC output to ground? That won't put it into current mode, you'll need a much smaller resistor, but then you'll need to get some voltage gain from somewhere else.
Try using the oipamps (good ones) and feed the transformer from them.
I'm sorry I totally forgot how I hooked up the transformers.
This is similar to what I've implemented.
http://www.diyaudio.com/forums/quanghao-audio-design/246579-super-v-i-converted-ll1684.html
I am struggling to figure out what would be the best direction to go from here. By using op-amps I am doubting that the sound will become any clearer, in theory it should only amplify the signal. Perhaps the clock is an issue, but how much can the clock possibly improve the low end?
This is similar to what I've implemented.
http://www.diyaudio.com/forums/quanghao-audio-design/246579-super-v-i-converted-ll1684.html
I am struggling to figure out what would be the best direction to go from here. By using op-amps I am doubting that the sound will become any clearer, in theory it should only amplify the signal. Perhaps the clock is an issue, but how much can the clock possibly improve the low end?
Using opamps will do two positive things:
- The DAC chip will see a low impedance, making it work in current mode resulting in lower distortion.
- The transformer will be driven by a low impedance, which they like. You'll get lower distortion from the transformer, particularly in the bass region.
Imporving the clock's power supply will give definately give you better sound.
- The DAC chip will see a low impedance, making it work in current mode resulting in lower distortion.
- The transformer will be driven by a low impedance, which they like. You'll get lower distortion from the transformer, particularly in the bass region.
Imporving the clock's power supply will give definately give you better sound.
Today I've removed the transformer output, threw in four LME79410 for I/V, a pair of OPA827's for balanced to RCA conversion and I must say things have improved. The ESS9018 does not have the output that the WM8741 and requires an I/V circuit to sound its best. The Low end is more complete and the top end is clearer. Next I'd like to try replacing the LME49710's with my favorite LME49990. Maybe six will do the trick.
It's now very close to the DAC7. The ES9018 top end is a little more relaxed, almost missing the sparkle and the bottom end is slightly over emphasized. But the best part is how it presents the sound stage with more air and less focus towards the speakers.
It's now very close to the DAC7. The ES9018 top end is a little more relaxed, almost missing the sparkle and the bottom end is slightly over emphasized. But the best part is how it presents the sound stage with more air and less focus towards the speakers.
I now have six LME49990's to listen to. They sound the best in all locations, comparing them with OPA827's, LME49710HA's and stock.
My current setup is with just four LME49990's in the I/V and summed through a 1:1 transformer and am very pleased. I'm in no hurry to modify anything more, I don't even want to try the DAC7 because it's working so good now.
The best part of this journey was experiencing the signatures of each chip, and the ES9018 has become a real winner, once you replace the AVCC power supply. I'd like to try tweaking the lt1963 regs to supply the trident 2.1 with 5.25volts, but seeming I have a salas reg why bother.
I bought an Epson SAW TCXO 100mhz but need an adapter. Does anyone have an idea on how I can adapt it from SMD?
My current setup is with just four LME49990's in the I/V and summed through a 1:1 transformer and am very pleased. I'm in no hurry to modify anything more, I don't even want to try the DAC7 because it's working so good now.
The best part of this journey was experiencing the signatures of each chip, and the ES9018 has become a real winner, once you replace the AVCC power supply. I'd like to try tweaking the lt1963 regs to supply the trident 2.1 with 5.25volts, but seeming I have a salas reg why bother.
I bought an Epson SAW TCXO 100mhz but need an adapter. Does anyone have an idea on how I can adapt it from SMD?
Last edited:
hello
I bought on ebay version 4 layers, connecting it to my Blueray player feels perfectly and without any problem; before starting any update I tried to connect it to some CD-Rom using its exit SPDF and inserting the filter described
S/PDIF output - HwB
(with and without transformer, with and without logic gates, also by varying the ratios of the resistors)
in all configurations have the same problem, an annoying dropout every 10 seconds, after several days of testing have not been able to solve, has anyone had a similar problem ??
(I apologize for translation errors)
I bought on ebay version 4 layers, connecting it to my Blueray player feels perfectly and without any problem; before starting any update I tried to connect it to some CD-Rom using its exit SPDF and inserting the filter described
S/PDIF output - HwB
(with and without transformer, with and without logic gates, also by varying the ratios of the resistors)
in all configurations have the same problem, an annoying dropout every 10 seconds, after several days of testing have not been able to solve, has anyone had a similar problem ??
(I apologize for translation errors)
The clock is probably causing the issue. I had static with 352.8Khz sample rate and the Epson SAW TCXO 100mhz solved the issue. ESS9018 is fussy with TCXO.
HI damuffin, I already have a DAC7 and I have replace the OPA2604 with 3x OPA 1642. It improve the sound quite a bit.
I am looking for a second DAC and was wondering if I should stick with the DAC7 or try the EC9018 DAC. Minishow sell on Ebay, different version of this DAC witch differs by the component quality and single or dual supply (analog / digital). I am not shure if the price difference between the basic to the higher end version is worth.
What is your opinion? Thanks
Hi!
I'm considering buying this DAC from ebay: Finished DAC Decoder ES9018 32bit 192kHz Hi End DAC Optical Coaxial | eBay
Is it worth the money? According to this thread at head-fi.org the board is poorly designed and problably delivered with fake parts... Any sugestions?
I'm considering buying this DAC from ebay: Finished DAC Decoder ES9018 32bit 192kHz Hi End DAC Optical Coaxial | eBay
Is it worth the money? According to this thread at head-fi.org the board is poorly designed and problably delivered with fake parts... Any sugestions?
Hello.
I have a 4 layer board DAC with TCXO 100 Mhz 0.1 ppm and 6x AD797.
The sound is for me, the best DAC I have, and the best DAC I listened. The 6xAD797 aren’t original. But they sound well.
And I buy 2x metal LME49710 to Weillian. They are original.
I put original 4xOPA604 and 2xLME49710.
It´s the difference with false AD797.
I have a 4 layer board DAC with TCXO 100 Mhz 0.1 ppm and 6x AD797.
The sound is for me, the best DAC I have, and the best DAC I listened. The 6xAD797 aren’t original. But they sound well.
And I buy 2x metal LME49710 to Weillian. They are original.
I put original 4xOPA604 and 2xLME49710.
It´s the difference with false AD797.
Attachments
The DAC have a problem. After a half hour power ON, the sound is still good but has a lot of noise in the measure.
You can see in the last images the measure with 4xOPA604 and 2xLME49710 the DAC have 1 hour ON. And see the problem in the J-Test Jitter. The noise is a multiple of 100 Hz. You can see nine spectral noises between 19 and 20 kHz. It is a problem of power supply. The problem is greater in L channel than R channel. So it's not a Jitter, TCXO or TCXO power supply, or analog 15 V common L and R power supply.
You can see in the last images the measure with 4xOPA604 and 2xLME49710 the DAC have 1 hour ON. And see the problem in the J-Test Jitter. The noise is a multiple of 100 Hz. You can see nine spectral noises between 19 and 20 kHz. It is a problem of power supply. The problem is greater in L channel than R channel. So it's not a Jitter, TCXO or TCXO power supply, or analog 15 V common L and R power supply.
Attachments
Your problem is most like likely related to the Cheap TCXO. The Stock TCXO gave me some issues as well. Try lowering your sample rate.
The 9018 AVCC L&R supplied to DAC is terrible. You should replace it's power source as recommended in earlier posts.
The 9018 AVCC L&R supplied to DAC is terrible. You should replace it's power source as recommended in earlier posts.
I measured the spectrum of all power supply. The TCXO and TCXO power suply are good. The TCXO is not the problem.
The problem is the regulator TL1963 3,3 v. AVCC-R is worse than the rest when heated. The three regulators get very hot. But this in the middle of the other 2 dissipates heat worse.
I put a heatsink with thermal paste over the TL1963 and the problem is solved for a hour. But just warming up heatsink also.
The problem is the regulator TL1963 3,3 v. AVCC-R is worse than the rest when heated. The three regulators get very hot. But this in the middle of the other 2 dissipates heat worse.
I put a heatsink with thermal paste over the TL1963 and the problem is solved for a hour. But just warming up heatsink also.
Attachments
The voltage on LT1963 is 14 V. at the input. I put first a 7805 to reduce it to 5 V. at the two LT1963 L and R input and the problem stop.
- Red only 2xLT1963 (14 V.->3,3 V.) 4xOPA604 2xLME49710.
- Gray the new measures 7805 + 2xLT1993. (14 V. -> 5 V. -> 3,3 V.) Several hours of operation. 6xOPA604.
Now is the same measures in channel L and R. TL1963 do not get hot.
- Red only 2xLT1963 (14 V.->3,3 V.) 4xOPA604 2xLME49710.
- Gray the new measures 7805 + 2xLT1993. (14 V. -> 5 V. -> 3,3 V.) Several hours of operation. 6xOPA604.
Now is the same measures in channel L and R. TL1963 do not get hot.
