You are currently using just the + output of your DAC through the transformer? If so, you definitely need to use both outputs. You shouldn't need coupling caps either (measure DC across the primary first to make sure there is no DC offset). What size caps are you using?
Last edited:
It's DC, it won't care. The DAC might want more isolation from capacitance though, so maybe a couple of 47Rs on the primaries. This wouldn't account for the lack of bass, unless the DAC is oscillating (?)
Last edited:
No I took out the trannies as the sound completely lacked bass. When I was using them both + and - legs were used.
At present the - legs of the AK4396 are unterminated. The + legs are each going into a 4.7uF poly cap, and then onto the output sockets.
So Bill, you're recommending 500 ohms on each primary, and Spartacus 47R.
Already tried the transformers without any load, and with 10R and horrible tinny sound was the result for both. No tinny sound with the + legs (only) each going into 4.7uF caps but the sound isn't particularly amazing either. Still the bass is back and a normal volume level.
As I've tried 10R I think I'll go for 500R next, then down again to the 47R if no luck.
Thanks,
- John
Already tried the transformers without any load, and with 10R and horrible tinny sound was the result for both. No tinny sound with the + legs (only) each going into 4.7uF caps but the sound isn't particularly amazing either. Still the bass is back and a normal volume level.
As I've tried 10R I think I'll go for 500R next, then down again to the 47R if no luck.
Thanks,
- John
Wires
Is it possible that you do have the wrong wires and it is out of phase? It sure sounds to me like something is wrong. I get an incredibly deep, black , see through sound stage with amazing fine resolution of the reverberant tails with just the + legs of my AK4395 coupled with Dayton foil caps and the - legs left open. Much better than the Benchmark Dac 1. I sold my highly modified Assemblage Dac 2.7 with it's fabled PCM 1704 dacs. They are opaque dinosaurs by comparison.
Is it possible that you do have the wrong wires and it is out of phase? It sure sounds to me like something is wrong. I get an incredibly deep, black , see through sound stage with amazing fine resolution of the reverberant tails with just the + legs of my AK4395 coupled with Dayton foil caps and the - legs left open. Much better than the Benchmark Dac 1. I sold my highly modified Assemblage Dac 2.7 with it's fabled PCM 1704 dacs. They are opaque dinosaurs by comparison.
Nope checked and re-checked that; all in phase.
I guess I just don't like the AK4396 sound then unless something else is wrong. Perplexing!
I guess I just don't like the AK4396 sound then unless something else is wrong. Perplexing!
Just to check .... what did you think of the sound of the DAC in it's standard configuration with opamps?
In standard configuration it used the AK4393 chip, with a - generally agreed on here - way too complicated output stage consisting of el-cheapo several op-amps.
It didn't sound too bad though - you could certainly hear the potential behind it. The sound was definately involving, just a bit rough around the edges. As I said a few posts back, I prefer the AK4393 with stock output stage - at present - to the upgraded AK4396 chip, running straight into 4.7uF cap and to the outputs. It still sounds rather 'take it or leave it' at present - no sparkle or 'snap' to the sound. Not dull exactly, just pretty lifeless.
It didn't sound too bad though - you could certainly hear the potential behind it. The sound was definately involving, just a bit rough around the edges. As I said a few posts back, I prefer the AK4393 with stock output stage - at present - to the upgraded AK4396 chip, running straight into 4.7uF cap and to the outputs. It still sounds rather 'take it or leave it' at present - no sparkle or 'snap' to the sound. Not dull exactly, just pretty lifeless.
Let me make a couple of comments based on my experience with the other AKM DAC (AKM 4394)
1. You need to have a resistance after the legs of the DAC. Otherwise the next device may try to draw power (and current) from the AKM. It was not designed to do this. On the 4394 they recommended 600 Ohm for AC and 1000 Ohm for DC. The AKM 4396 will have a comparable specification. Do not ignore this.
2. The DAC does not do a very good job driving a capacitor without a resistor (even a small one) in front of the cap. This is easily seen on a scope.
3. The transformer you used will have an inductance. The manufacturer may be reluctant to provide the spec. This inductance combined with the load resistance will form a high pass filter. The loss of bass (depending on the values) can be quite audible, and also easily measured.
4. Running this without a low pass (anti-aliasing) filter is not a good idea. This can easily be accomplished with a RC filter, or a transformer may also provide some LP filtering. This is part of the decision making when choosing a transformer. Putting a wide band signal (no anti-aliasing filter) into an amplifier or pre-amp may cause some problems. It may not happen in every case, but it may happen down the road when you swap components. This practice of eliminating the RC filter is freqeuntly recommended around here, but I think it is a bad / risky idea. Incidentally, the output of a DAC does not "drive a cable" very well. If cables are swapped (or whatever) while the unit (or amp etc) is powered on, then bad things could happen to the DAC chip.
5. The issue of loading a transformer secondary with a resistor is usually not done to affect filtering rather it is done to minimizing "ringing". You should talk to the manufacturer about this or at the very least check it out on a scope. Of course the DC blocking capacitor (the cutoof freqeuncy) will be affected by the input impedance of the next device in the circuit
Good luck and please consider some of the warnings being given around here.
1. You need to have a resistance after the legs of the DAC. Otherwise the next device may try to draw power (and current) from the AKM. It was not designed to do this. On the 4394 they recommended 600 Ohm for AC and 1000 Ohm for DC. The AKM 4396 will have a comparable specification. Do not ignore this.
2. The DAC does not do a very good job driving a capacitor without a resistor (even a small one) in front of the cap. This is easily seen on a scope.
3. The transformer you used will have an inductance. The manufacturer may be reluctant to provide the spec. This inductance combined with the load resistance will form a high pass filter. The loss of bass (depending on the values) can be quite audible, and also easily measured.
4. Running this without a low pass (anti-aliasing) filter is not a good idea. This can easily be accomplished with a RC filter, or a transformer may also provide some LP filtering. This is part of the decision making when choosing a transformer. Putting a wide band signal (no anti-aliasing filter) into an amplifier or pre-amp may cause some problems. It may not happen in every case, but it may happen down the road when you swap components. This practice of eliminating the RC filter is freqeuntly recommended around here, but I think it is a bad / risky idea. Incidentally, the output of a DAC does not "drive a cable" very well. If cables are swapped (or whatever) while the unit (or amp etc) is powered on, then bad things could happen to the DAC chip.
5. The issue of loading a transformer secondary with a resistor is usually not done to affect filtering rather it is done to minimizing "ringing". You should talk to the manufacturer about this or at the very least check it out on a scope. Of course the DC blocking capacitor (the cutoof freqeuncy) will be affected by the input impedance of the next device in the circuit
Good luck and please consider some of the warnings being given around here.
I'm sure the Digitecs are output trafos of fine quality, they will not exhibit any resonance peaks, unlike input trafos.
Definitely use an RC filter on the secondaries.
The Digitecs should have sufficient inductance to keep the F3 below 20hz with 500 ohm series resistors.
Saw it, it specifies 1K5. However if there is no offset it won't be supplying any current into the load.
I understand what you are saying, but can you be sure there is nothing else in the chip that this affects? Maybe an email to AKM would be in order.
Have ordered some 510-ohm resistors so will be interesting to see if this makes the Digitecs more palatable to the AK4396.
- John
Well haven't gotten around to re-installing the Digitecs yet, buuuttt I did find out why I was getting that odd slightly 'hollow' sounding effect. Monsieur Dofus here had gotten one set of speakers wired out of phase!!! What a twit - first time in over 20 years of this crazy hobby I've done this... guess there is a first time for everything 😉
So, one problem down...... 🙂
So, one problem down...... 🙂
Phase
I figured something must be wrong. You should have been in sonic bliss with the AK4396 running direct out.
Well it doesn't sound hollow any longer, but far from sonic bliss I'm afraid. It just sounds like any other semi-decent DAC out there. Hoping once I get those Digitec's set up correctly I'll see what the chip is really capable of.
Digitec
I don't know if a transformer with a 20 ohm input can ever be correct for coupling a Vout dac chip. There aren't even many opamps that would be happy driving that load.
20-ohm was the load I was originally using and it resulted in faint, tinny sound.
Bill Fuss has recommended 500-ohms per leg from reading the AK data sheet, and this is what I'm going to try next.
Running Lundahl LL1527 transformers with 8.6ohm resistors on both + & - outputs only. 4K across secondaries, plus 100K amp input impedance. Sounds great (more dynamic than Sowter 3603 used previously, which didn't sound much different with no series resistors on primaries or 270ohm).
My understanding is that the dac output sees an essentially infinite DC resistance, as both output phases have the same positive DC offset. So series resistors add nothing in this regard. This assumes there isn't a centre tap to ground or other route for the DC to take.
Not sure how to calculate/measure impedance at low bass, but assume any good audio transformer happy to work at 20Hz at the required dBu will provide sufficient load to meet the 600ohm minimum requirement (speculating - tell me if this is wrong)
My understanding is that the dac output sees an essentially infinite DC resistance, as both output phases have the same positive DC offset. So series resistors add nothing in this regard. This assumes there isn't a centre tap to ground or other route for the DC to take.
Not sure how to calculate/measure impedance at low bass, but assume any good audio transformer happy to work at 20Hz at the required dBu will provide sufficient load to meet the 600ohm minimum requirement (speculating - tell me if this is wrong)
- Status
- Not open for further replies.