AliExpress AD1865 R2R DAC

Status
Not open for further replies.
Hello all,

What do you think of this DAC?

AD1865 decoding finished board r2r decoding nos mode DAC hifi audio good-in Digital-to-Analog Converter from Consumer Electronics on Aliexpress.com | Alibaba Group

EDIT. The thread almost immediately digressed into one about TDA1387 (from post 8 on) and then later into op amp in the I/V for TDA1387. Sorry. :eek:

And then when my DAC eventually arrived, it was faulty, and was returned. See post 96.

And the saga went on... the seller refused to sign for the faulty goods so he could "negotiate" a partial refund. :whazzat:
 
Last edited:
I thought so too. And it is not the allegedly inferior dip package, and it is current out. The current is feeble compared to other DACs I know but if the I/V is good... The only thing I see that is bad is electrolytic caps on the output. I guess a transformer is a better option?
 
Last edited:
Member
Joined 2008
Paid Member
I have the tda1387 "cousin" of that board, the one Abraxalito mentioned. I haven't spent a tremendous amount of time listening to it yet, but I've spent enough time to feel it's definitely worth $50. I posted a bunch of pics of it here, if you want to get a feel for the workmanship (which looks pretty good IMO). I'm tempted to buy that 1865 board myself...
 
Thank you. The photos are very good! I can see there are two grey electrolytic capacitors in the signal path after the first OP42 op amp. Both boards use this coupling capacitor and as far as I can see, this is the only issue that might degrade the sound. If it is only one component to upgrade, these DACs look like very good value and very easy to use for non technical DIYers.
 
Disabled Account
Joined 2002
Thank you. The photos are very good! I can see there are two grey electrolytic capacitors in the signal path after the first OP42 op amp. Both boards use this coupling capacitor and as far as I can see, this is the only issue that might degrade the sound. If it is only one component to upgrade, these DACs look like very good value and very easy to use for non technical DIYers.

Those 2 caps seem to be a key selling point:

7. It adopts the early Hongbao lime skin electrodeless coupling capacitor which has been discontinued for many years. The charm is very good, delicate and rich in music. There is no general sense of particle coupling in the electroless coupling, and it is very rare to couple the sound.
 
Only if you ignore the elephant in the room - the opamps. IME opamps (most especially when exposed to a lot of RF) are more egregious degraders of sound even than electrolytics.

I don't have a problem with op amps. In fact, I really like them, especially LME49990 etc. The OP42 are jfet I think so they have really good slew and settling, decent bandwidth and their only major downside is noise. Jfets sound nice as IV op amps. However, caps in the signal path can have the awful effect of smothering detail and introducing other issues - I can literally hear a bad cap in the signal path whereas, as a buffer, I cannot hear the LME49990 at all because it outperforms my ears.
 
Last edited:
Noise isn't an issue in an I/V application as the signal level is generally very high (at least 1VRMS) so an opamp can have a very poor noise performance and it'll still be under the hiss on the recording. I agree that JFET opamps are best for the I/V function because of the extreme levels of out-of-band noise encountered at the output of a DAC. The OP42 is only a 10MHz GBW part which I consider to be inadequate, given the bandwidth of the raw signal from the DAC. Something like AD8065 (with more than 10X the GBW) might well do better.
 
........IME opamps (most especially when exposed to a lot of RF) are more egregious degraders of sound....


Interesting. Thanks.


I wonder what you think of the idea of placing caps immediately over the DAC outputs to directly filter out ultrasonics? Should make a simple RC low pass filter (RC Low-pass Filter Design Tool) with R being the output impedance of the DA chip? Seems smart to filter out the ultrasonic hash as soon as possible?



Then in terms of opamp selection I wonder what your opinion is of the popular AD844 IV solution?


"HIGH SPEED DAC BUFFER The AD844 performs very well in applications requiring current-to-voltage conversion."


"20 MHz full power bandwidth, 20 V p-p, RL= 500 Ω"



Is that 10dB gain? If so does GBW= 200?

 
Interesting. Thanks.
I wonder what you think of the idea of placing caps immediately over the DAC outputs to directly filter out ultrasonics?


It is more common to place a cap or cap-resistor in parallel to the I/V resistor so HF is not given any gain, rather than shunt the HF into ground, but I am also interested in the relative merits of these choices.

I noticed this DAC has only the I/V resistor in the op amp feedback loop so I expect I'll add a small smd cap (pps film or cog ceramic) in parallel to limit the gain bandwidth.
 
Last edited:
I wonder what you think of the idea of placing caps immediately over the DAC outputs to directly filter out ultrasonics? Should make a simple RC low pass filter (RC Low-pass Filter Design Tool) with R being the output impedance of the DA chip? Seems smart to filter out the ultrasonic hash as soon as possible?


You might not be asking the right person here, in the sense that my views on this aren't at all mainstream. I have seen people advocating a cap on the output of the DAC (Scott Wurcer in his AD797 applications is one who comes to mind) but I go a few steps further than this. In my lingDAC design I put a 5th order (CLCLC) filter between the DAC and the I/V stage.

So yes I agree totally that it 'seems smart to filter out the ultrasonic hash as soon as possible' but why stop at just one cap, which only gives 6dB/octave roll-off? With a 5th order filter its possible to get of the order of 45dB stop-band rejection above 30kHz but it does require attention to getting the loading at its output correct.

Then in terms of opamp selection I wonder what your opinion is of the popular AD844 IV solution?

AD844 is one of the best choices in my view for I/V conversion. Some like to use it open loop (bypassing the buffer) but another option is to follow it with a classA buffer. The weakness of the chip seems to me to be that its buffer is classAB but as the compensation pin is brought out it ought to be fairly straightforward to substitute your own discrete buffer. I'd try with a single current-source loaded MOSFET myself.

CFB opamps don't follow the usual rules about GBW being a constant as gain is varied so yes, if you use one to give some gain, the GBW will increase, up to a point. 10dB gain though is just X3.1, so not 10X greater.
 
It is more common to place a cap or cap-resistor in parallel to the I/V resistor so HF is not given any gain, rather than shunt the HF into ground, but I am also interested in the relative merits of these choices.

Its quite correct that a cap across the feedback resistor provides HF roll-off. In my experience though, DACs sound better when this cap value is minimized. I tried to find the reason for this in simulation many years ago but failed. I've since come up with a hypothesis - the cap appears to some degree as an output load and so increases the 'bounce' on the opamp's supply rails. No cap at all should be used when the opamp is a CFB type as it will encourage oscillation.

I noticed this DAC has only the I/V resistor in the op amp feedback loop so I expect I'll add a small smd cap (pps film or cog ceramic) in parallel to limit the gain bandwidth.

Could be the designer listened and came to the same conclusion as me :)
 
I think we all hope it's because of extensive listening, measurement, layout revision, etc. but it might simply be to reduce the BOM, get some sales, and leave the choice of op amp open, rather than specify an op amp and optimise for its bandwidth, only to run out of stock of said op amp.... no cap might simply be to allow a CFB option, not because it's best for the OP42.

I always wonder how these low volume folks make money. Are they just selling on their own surplus "toys" or are they in business?
 
Last edited:
Status
Not open for further replies.