AKM4393 transformer output LL1539

[smallbutfine]

Hi guys,
I took a look on an AKM4393 based DAC i have around here and would like to use a transformer on the output.
Since I have some LL1539 around, I now wonder if I can hook up the transformer directly to the converter chip's outputs (passive design)?
From the datasheets, the AKM outputs max. +-2.55 Vpp and has an output current of 3.5mA, load resistence 600Ohms.
The LL1539 +31 dBU balanced driver transformer could be used in a 1:1 instead of 2:1+1 configuration.
(Going unbalanced to my unbalanced amp for monitoring.)
Is there a serious flaw here? would I be forced to have an active buffer?

Kind regards and thanks in advance

Martin

 

[playboss]

http://www.jensen-transformers.com/as/as093.pdf

http://www.jensen-transformers.com/datashts/11ssp8ma.pdf

http://www.jensen-transformers.com/prices.html

 

[smallbutfine]

Thanks, playboss...
actually the jensen notes brought me to that idea of passive outs...
Nice design with dc cancellation trimming (I saw how to do this elsewhere) and lowpass filters....
but they are based on a a specific jensen splitter transformer - specs are different.
I also saw brian sowters dac output transformers, but actually I like the 1539 quite alot in opamp based outputs from other projects, so I wonder if the AKM chip would be able to drive it directly. Also, I have them just in front of me...and the dac price might not justify the expenses for another transformer pair...

The LL1539 are actually quite big transformers (for lundahls) - I feared the poor AKM could be too weak to drive them...but I am not sure from the specs, so I wanted to ask if someone could bring me to the right track of the theory behind this.
(I.e. advantages ansd disadvantages of going passive in a DAC, possible impacts on linearity of the dac, maybe the math behind when a buffer should/must be used...)

Kind regards,
Martin

PS: I just saw the price list - no, the dac definitely would NOT justify 400 additional bucks for transformers.........

 

[smallbutfine]

Or did I misunderstood you and you wanted to say that the Jensen specs are close enough to the 1539 for the task?
???

 

[playboss]

no , we just had this talk about these on local forums , i think 8legs (opamp) is good , anmd that DA chip definately wont justify the 400 bucks as it has switched MOS capacitor chains inside (and opamp?)  hehehe.

 

[audiox]

Some converter types may require low-pass filtering higher than first order. But there are types that have analog filtering included in the chip too.

Output drive capability of the chip can also be limited.

 

[playboss]

these V-out cirrus / akm have HF crud between 400-500 khz.

Input is more important if you ask me , then comes local clock and that sorta stuff.

 

[Tubemooley]

Your idea may or may not work for you. If you read the datasheet closely, the Analog Characteristics on page 5 are obtained when tested using a load greater than 600 ohms. The converter manufacturers are very stingy with the technical details on their devices. They give minimal spec's on the DAC output capabilities. And if you ask them any specific questions, they probably won't tell you. They expect you to test and figure out your own design without their help. I am a HUGE fan of using transformers on the output of my DAC's. I used Jensen JE-11P-1 input transformers on the output of the DAC's in that ART DIO box. Do you know that one? It's this tiny little box with a tube on the ADC inputs. They are hard to find now. We still use one with the Jensen trannie's on the DAC's. It is really nice. But I can't remember exactly which AKM codec is inside that box. I can look around and find my files on that project. I guess what I'm suggesting here is that you are going to have to try your idea and see what it sounds like and hope that you don't fry the DAC. I would try to have a spare DAC IC on-hand if possible. Does this help at all? DW.

 

[smallbutfine]

Hi guys,
first thanks for all your comments, they are really helpful.
I will test the akm chip with direct out via caps first, then do another test with the lundahls and just try and compare.
Also I think I will try an improved  and more simple opamp output stage (like playboss suggested) with probably OPA1632 as line driver. I think I have some samples of these around from the time they came out that I did not use until now. And then compare again....

In case I fry the akm chip (I hope I do not), I might replace the AKM4393 with an AKM4395- might be a good improvement as well. Specs look much better on paper.

I found a nice small simple but clean clock replacement circuit that might fit the bill, but will do a better 5V psu replacement first to improve the converters performance.

Also, the recommendation of fixing/improving the input first is very welcome, playboss.
I found a few things maybe could be improved on the digital in, but have to check again how to do this.

Fortunately I got my hands on the manufacturers schem of the box in the meantime, this helps alot. 8)
(This very well known huge chinese/german company is known to be very picky about this! I think it is easy to guess who I am talking about now, isn't it?----) ;D

Hm, so at least I have some interesting options now to fiddle around.

Unfortunately I discovered my monitoring has a phase related problem in the amplifier/speaker combination. I do not fully understand the problem now, but it seems the staging and stereo image is seriously crippled since I replaced Van den Huul/Array amplifier with an Adcom. Interestingly on other speakers the Adcom performs extremely well (Impedance mismatch kind of thing?). I have to sort that out before I am able to do serious listening tests with the DAC.
Meanwhile I might check the dac performances with the 3 different output stages by doing a short DAC-ADC loop and measure with RMAA (word clocking from the ADC, I might use the Mytek to have a good enough adc quality for serious comparisons).

I will come back to this thread as soon as I have some practical findings.
I appreciate all feedback you gave me so far.

Thanks alot and kind regards,
Martin

 

[gridcurrent]

"I now wonder if I can hook up the transformer directly to the converter chip's outputs (passive design)?"

Yes, been there, done that.  Direct differential connection performs flawlessly.  Best implemented (IMHO) with a step-down transformer, 4:1 ratio.  A 1:1 (600 to 600) is not ideal for a variety of reasons.  A transformer offers a measure of HF filtering, of course passive, without any dreaded cheap op amps.  Coupling capacitors are not required.  You will need a great line amp after the step-down transformer for line level (+4 dBu) operation.  I can assure you that a implementing a quality vintage transformer directly to the converter's outputs yields stunning sonic results.  Note:  I said vintage transformer.