Indeed so, I figured this out very quickly.
So surely there is an advantage to consider, pulling it to ground and have a "real" ground rather than faked one - and makes potential direct coupled circuits possible. At least more easy to do.
Is there something like a "virtual earth" that is Zero Ohm? Any signal there would indicate not. So is a "virtual earth" really "near earth"? If there is a signal there, then that signal (as a thought experiment) could be amplified.
So why not explore that... use very Low Z passive - is that not also "near earth" but without some of the negatives.
Looking at the parameters required for "voltage" drive and applying it to "current" drive, this in rounded off numbers is 10:1 or better. Preferably better, of course.
This is what I am doing and trust me, it works. Sounds very good indeed and lots agree.
Cheers, Joe
.
Yes, the dac has mosfet devices on the output
probably cascoded, and probably non-complementary type
because of the single supply of avcc +3.3V
for this pins I put Vref style of power (Jung variation with ad797, one for each channel)
So it is possible to tap the right value from avcc to exact match offset on the riv-s...
other way Joe explained
third way is to extract surrent to external I-to-I device and then convert.
.
the interesting q can be of the issue of amplification and nature of the amplification
1. Riv one level not amplified, 1:1 intersage - out
2. low riv - amplification active - interstage 1:1 optionally - out
3. low Riv - no gain - interstage step UP, passive amplifications - out
probably cascoded, and probably non-complementary type
because of the single supply of avcc +3.3V
for this pins I put Vref style of power (Jung variation with ad797, one for each channel)
So it is possible to tap the right value from avcc to exact match offset on the riv-s...
other way Joe explained
third way is to extract surrent to external I-to-I device and then convert.
.
the interesting q can be of the issue of amplification and nature of the amplification
1. Riv one level not amplified, 1:1 intersage - out
2. low riv - amplification active - interstage 1:1 optionally - out
3. low Riv - no gain - interstage step UP, passive amplifications - out
output impedanse or internal resistance of the generator, of each current output
is about 800ohm litlle bit less, so 4 outputs merged is about 200ohm litlle bit less.
that is for the 9008 chip but it is the same for the 9018.
it is not lets say referenced to ground, but rather this is internal resistance of the generator. so if we have current source conected to ground or some potential, that is R
connected ot the ither side.
is about 800ohm litlle bit less, so 4 outputs merged is about 200ohm litlle bit less.
that is for the 9008 chip but it is the same for the 9018.
it is not lets say referenced to ground, but rather this is internal resistance of the generator. so if we have current source conected to ground or some potential, that is R
connected ot the ither side.
LL1674 have 2 secondary coils. So You can run it in parallel and reduce the transfer ratio,
and voltage if You think that is too much.
unfortunatly in the datas some parameters from importance is ommited
like Lp inductance of primary
inductance of secondary, coupling factor
and capacitances...
and voltage if You think that is too much.
unfortunatly in the datas some parameters from importance is ommited
like Lp inductance of primary
inductance of secondary, coupling factor
and capacitances...
Last edited:
Several. For one thing, it gets rid of the 1.65V offset and means that it is easier to get rid of capacitors in the signal path. The offset voltage becomes and offset current of 2.1mA per phase. The other is the much lower Z. If you don't ground the midpoint, the termination impedance becomes the Z of the other phase, so you are no longer using the DAC in current mode.
BTW, you can certainly try both, but I suspect I know which you are going to like after listening. Sabre DAC sounds better in current mode, a more fully fleshed out balance and not lightweight as it sounds when in voltage mode.
Cheers, Joe
I am not very familiar with balanced outputs but your connections don't look correct. Aren't you supposed to connect the secondaries in series and reference one pin on the XLR output connector to ground?
I am sure somebody here with more experience with balanced outputs will chip in.
In fact, the centre of the secondary need not be grounded to chassis or even to Pin 1. In fact, it may well be advantageous NOT to ground it at all.
What then determines whether it is balanced is the receiving end. If the receiving end grounds the bottom (or one side) of the secondary, it then becomes unbalanced. On the other hand, if the input on the receiving end is truly balanced, then signal will appear on both Pin 2 and 3. But the signal that gets processed is the difference between Pin 2 and 3, and not with any reference to ground. The ground reference is not the reference to the signal and that is a good thing. Hence best not to ground the secondary at all.
The receiving ends sees the difference between Pin 2 and 3, that is all. Hence, perhaps the better description is that it is differential rather than balanced, since signals at Pin 2 and Pin 3 may not be exactly balanced or exactly the same amplitude anyway.
Is that clear as mud?
There are other advantages to a gloating secondary, galvanic isolation. This is an advantage peculiar to transformers as they can be used as isolators.
Cheers, Joe
.
In the case of a balanced interconnect, the XLR pin #1 is not part of the audio circuit ground. The XLR pin #1 is only a shield and should be connected to the chassis very near the connector. Occasionally at the receive end of the cable, the shield is not connected at all or connected through a capacitor. But this modification should be done inside the cable connector not inside the chassis.
I'm ready to give this a shot, but can anyone tell me if there is any significant difference, aside from the DCR of the windings, between a "600:600" transformer and a "10k:10k" transformer in either the voltage out or the current out configurations suggested in this thread? I'm looking at the Lundahl 1527 for the former or the Jensen 11p1 for the latter. TIA - Pat
Last edited:
LL1527 is Lundah's general purpose 1:1 transformer.
I am sure others will give more complex explanations, but when I see a transformer with reasonably low DCR and reasonably low insertion loss, the main thing I look at is the ratio of the transformer.
But there are other factors to consider, the max level of signal and the current loads and the magnetics to support it, frequency response under those conditions and so on. But with 1:1 transformers and 43R DCR per winding, for our use here, a well designed transformer should do the job. Try it and see if you like what that one does for you.
Yes, I would try LL1527, if I had one on hand to try. The specs are quite good, 0./1% distortion at +6dBu (1.5V), <1% at +16dBu (5V).
Cheers, Joe
Last edited:
Thanks, Joe, got a quote from Kevin at K&K. Just need to find someone to do the assembly and SMD soldering for me on my Acko isolator/reclocker, and I'll be good to go. (Digikey sent me some envelopes with what they claimed to be 0603 components (or perhaps caraway seeds) but I'm 63 and my eyes are not that good anymore, and my hands are so shaky that the only way to stop them is to drink to excess, and that makes my eyesight just that much worse.) ;-)
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.