DAC ouput using Transformer

[Pano]

Hey John, glad to see so much discussion here!

Bill is right on. It's a simple circuit and has a lot of advantages. In best practice you should not tie the low point of the secondary to DAC ground, but I have found that sometimes it's the only way to kill the hum. It seems to depend on the power supply of the amp, more than anything. Same amp -different supplies gives different results.
Be a little careful with your center tap connection! measure first. You may have 2.5V on the primary side CT.

You are correct about the Tripath connection. As the input is biased at 2.5V, one end of the transfo needs to be tied to 2.5V, not ground. Otherwise DC would flow thru the winding. The 2.5V end becomes a virtual ground, so to speak.

Keep going, you're on the right path. You'll get there.

It's nice to see so many people getting hip to the transformer output DAC. I love 'em. I sold one for awhile called the "MagiDAC" and have a new one coming out next month thru Virtue Audio. So you can tell I like them.

[jkeny]

Originally Posted by rossl

I agree with Bill. The bottom diagram by panomaniac (post 13) should work without hum.

I thought what Bill was saying, in relation to this scheme, that DC should not pass through the trafo?

If you have hum them either something isn't hooked up right or you have a power supply fault. Or a bad transformer.

I say that without knowing what the amp input schematic looks like. Do you have a link to it?

Here's the Lepai Tripath TA2020 schematic. I've since bypassed the whole op-amp input tone control section but haven't hooked up the trafo without grounds - I'll try that. Ohh, so many options now, I'm losing track of them.

isolating the grounds is the main benefit you will see in using this transformer approach. You shouldn't need to connect grounds between the DAC and amp.

Yep, I like the theory just can't seem to achieve it in practise, (yet! ) I guess that's why I started this thread - transformers are well understood in (and well represented) in the tube section but not a lot about them in this configuration.

P.S. I'm glad to see no pro/anti transformer wars have started in this thread - let's keep it clean!

[rossl]

Originally Posted by jkeny

I thought what Bill was saying, in relation to this scheme, that DC should not pass through the trafo?

Yes, that's what he is saying. And I'm saying that you seem to have a path for energy at the mains frequency getting across the transformer. 50/60 Hz hum. Are the power supply grounds on either side referenced to the mains low wire? floating? earth ground? mains hot?

Originally Posted by jkeny

Here's the Lepai Tripath TA2020 schematic.

You should be able to use the voltage divider as the 2.5V reference for the transformer secondary low wire. It would have been nicer if it was a lower impedance reference, but it should still work.

[jkeny]

Originally Posted by panomaniac

Hey John, glad to see so much discussion here!

I agree, nice to see it escaping from the realm of the Tube section Although I must get Shoog's expertise over here as I know he played around with this very thing

Bill is right on. It's a simple circuit and has a lot of advantages. In best practice you should not tie the low point of the secondary to DAC ground, but I have found that sometimes it's the only way to kill the hum. It seems to depend on the power supply of the amp, more than anything. Same amp -different supplies gives different results.

Ah, how I'm powering the Tripath is with a 12V SMPS with no ground pin - hmm, more experimenting

Be a little careful with your center tap connection! measure first. You may have 2.5V on the primary side CT.

Thanks, I think there is 1.4V bias on the DAC differential V outs - ouch, as you say - I'll check. But thinking about it, I wonder if there is a similar 1.4V bias pin I could connect CT to ala the Tripath input stage?

You are correct about the Tripath connection. As the input is biased at 2.5V, one end of the transfo needs to be tied to 2.5V, not ground. Otherwise DC would flow thru the winding. The 2.5V end becomes a virtual ground, so to speak.

Yes, that's my understanding - I take it you've tried this & it works? It should also work for the DCA output above?

Keep going, you're on the right path. You'll get there.

It's nice to see so many people getting hip to the transformer output DAC. I love 'em. I sold one for awhile called the "MagiDAC" and have a new one coming out next month thru Virtue Audio. So you can tell I like them.

Cool, thanks for the postings & encouragement! I will keep going, if I don't blow up something first

[jkeny]

Originally Posted by rossl

Yes, that's what he is saying. And I'm saying that you seem to have a path for energy at the mains frequency getting across the transformer. 50/60 Hz hum. Are the power supply grounds on either side referenced to the mains low wire? floating? earth ground? mains hot?

It may be the SMPS power supply (with no ground) I'm using?

You should be able to use the voltage divider as the 2.5V reference for the transformer secondary low wire. It would have been nicer if it was a lower impedance reference, but it should still work.

There is a biascap pin 14 which outputs the 2.5V (currently running to ground through a cap) - this may be a lower impedance source & where this trick usually takes it's 2.5V reference from.

[rossl]

Originally Posted by jkeny

It may be the SMPS power supply (with no ground) I'm using?

If the SMPS is transformer isolated, you should be able to connect the amp ground to earth ground, either directly or through a 10 ohm 1/2 watt resistor.

Then if your DAC power supply is also transformer isolated, connect DAC ground to earth the same way.

That should kill the hum.

[Bill Fuss]

Originally Posted by rossl

If the SMPS is transformer isolated, you should be able to connect the amp ground to earth ground, either directly or through a 10 ohm 1/2 watt resistor.

Then if your DAC power supply is also transformer isolated, connect DAC ground to earth the same way.

That should kill the hum.

That could work brilliantly, but dont ground any part of the secondary wiring at the DAC chassis or you will create a ground loop.

[jkeny]

Originally Posted by rossl

If the SMPS is transformer isolated, you should be able to connect the amp ground to earth ground, either directly or through a 10 ohm 1/2 watt resistor.

I'm not sure, but I presume the SMPS is isolated as you say (I can't open it up without breaking into it)

Then if your DAC power supply is also transformer isolated, connect DAC ground to earth the same way.

That should kill the hum.

This could be the tricky one as it's a USB DAC which I'm running from a Dell laptop. I can't run this on batteries as my laptop battery is busted & I'm not paying for another exorbitant laptop battery. But I think the Dell laptop power supply is not connected to earth - I'll check it out?

Ohh, more experiments - I better start doing something & report results.

Lets see, first off, I'll try this last one - so earth the AMP ground, earth the DAC ground connect the trafo as in Panomaniacs top diagram. This seems to be the best bet.

[Bill Fuss]

Originally Posted by jkeny

I'm not sure, but I presume the SMPS is isolated as you say (I can't open it up without breaking into it)
This could be the tricky one as it's a USB DAC which I'm running from a Dell laptop. I can't run this on batteries as my laptop battery is busted & I'm not paying for another exorbitant laptop battery. But I think the Dell laptop power supply is not connected to earth - I'll check it out?

Ohh, more experiments - I better start doing something & report results.

Lets see, first off, I'll try this last one - so earth the AMP ground, earth the DAC ground connect the trafo as in Panomaniacs top diagram. This seems to be the best bet.

There is a big difference between earth and signal ground. If your amp chassis is earth grounded do not ground anything on the secondaries inside the DAC chassis, let the amp take care of that.

[BudP]

Just a note to clarify the "load" issue with reference to transformers. The inductance is what provides the load to the driver and driven stage. As an example, with Dave's fine product, with 130 henries of inductance out into the permeability range, this unite will provide a 15k load all it needs to have to provide - 3db @ 20 Hz. To obtain -0.5 db would require 2.76 times this inductance. So, these are 5k to 5k transformers at -0.5 db @ 20 Hz.

The AC drive signal can be ignored so long as it remains at the typical DAC level. To handle DC voltage is a non issue, but as little as 5ma dc current will cause core saturation at low frequencies and they will roll off.

Bud