Balanced to single-ended converter (XLR to RCA / BAL to SE)

Unfortunately i never had the opportunity to try the usual 1:1 converter but the 1:4. This said i have to say that this converter is working very well, i connected it and disconnected it several times to make sure i wasn't missing something and it definitely is a well made product. Because of my moving i didn't had the chance to try with every amp but i will do so as soon as the moving is done and will post impressions.

 

I recently bought both the Radial J-ISO and Twin-ISO to convert my Wavelight's balanced output to single ended (to feed the T4).

I have been using the J-ISO for about a month and just received my XLRF - RCA cables to use the Twin-ISO. I had tried the Twin-ISO in XLR-XLR mode with my 3F prior and the results were similar to what I'm about to describe.

I get significantly more bass out of the J-ISO than I do the Twin-ISO! As in probably 5 or 6dB more! I would guess in the <60Hz range. Bass extends much lower on the J-ISO as a result (although, it might not be the cleanest? It's hard to compare both since one is so much lower than the other, even with the Utopia...).

The Twin-ISO sound a bit clearer/cleaner overall, but this could be a direct consequence of its lower bass level.

I am at a loss explaining the difference?

The Wavelight has a 50 ohm output impedance and I haven't seen any measurements into a 600 Ohms load, so I can't rule out some weird interaction between it and the Twin-ISO (the J-ISO is 1.3k Ohms).

Anybody noticed something similar?

 

Answering my own question: the problem was due to the Goldpoint SA-2X that sits between the wavelight and the isolator that I suspect increased the source impedance (as seen from the isolator) too much. I guess it affects the Twin-ISO more than the J-ISO (which has > 2x input impedance and 1/3 output impedance).

The (nice) Radial technical rep said that, in any case, the J-ISO is better suited for BAL -> SE application than the Twin-ISO.

 

If I understand correctly you have:
DAC -> 10K stepped attenuator -> XFMR -> Amp

Having the stepped attenuator before the XFMR is not ideal. This is potentially a high impedance sourcing into a low impedance. What you want is:
DAC -> XFMR -> 10K stepped attenuator -> Amp

This would imply using a Goldpoint SA2 rather than the SA2X. Typically you want to organize so low impedance sources feed high impedance inputs.

 

That's correct.

The J-ISO didn't seem to be impacted quite the same. Maybe the extra 700 ohm input impedance made all the difference.

I re-organized the chain to not have the SA2X in front of the isolator, but now I'm wondering if the 5.8Vrms output of my DAC (Wavelight) is too hot for the transformer. The THD seems to be pretty high down low at 10dBV. 5.8Vrms is way above that :/

 

@Clemmaster, this is what I worry about when I use the Convert 2 all the time. At its lowest setting of -14dBFS (or 18 dBU), the DAC outputs 6.15Vrms. I have been using the Rolls Promatch MB15b active converter which is a two-way stereo converter box. Seems to be working well thus far (knock on wood).

 

@Hands, would you use this adapter along with an XLR to RCA cable to connect a pro-audio DAC (with hot outputs) to a consumer amp?

 

Why not just use the built-in pot instead of the extra active stage?

 

Radial J-Iso can handle 20dBu, or 7.75Vrms at 1% THD
Radial Twin Iso can handle 21 dBu, or 8.70Vrms at 1% THD

Using a stepped attenuator in front of the XFMR presents a varying output impedance from the stepped attenuator to the XFMR input. At 6 dB of attenuation a 10K SA2X will present 5K source impedance to either a 600R or 1.3K XFMR input. Not ideal.

Using the J-Iso in front of an SA2 will present 200R output impedance to constant 10K SA2 input impedance. Now the output of the SA2 still varies but it will typically be seeing a constant resistive input for the amp.

No sure why you are having a high distortion issue. Maybe you could create a block diagram of the connections with components listed.

 

Wouldn't that amount to the same as having the Goldpoint before the isolator (increased source impedance)?

I'm not saying I'm hearing distortion, I was just referring to the graphs that show that distortion:

1. Increases with source impedance (graph shows 0 and 600 ohm)
2. Increases with input level (graph shows -10, 0 and +10 dBV)

Seems like - save for controlling the volume in the DAC itself (which has its own sets of problems) - it's a losing battle here...

Can you think of a XLR attenuator that I could fit before the transformer that would not either raise the source impedance too much, or present too low a load for the DAC?

Edit: the Wavelight is 5.8Vrms balanced and 50ohm output impedance.
Edit 2: the whole point is to avoid using the WL's built-in SE outputs, which, for some reason, sound clearly inferior.

 

Here are two graphs that give actual distortion data setup for your specific situation.

Signal generator has been set to 17.48 dBu which is 5.8Vrms for DAC 0 dBFS
Signal chain is Signal Generator -> Twin-ISO -> SA1X -> Analyzer
SA1X has been set for 20 dB attenuation.

With an amp set for 0 dB gain and listening to HD6x0 headphones this would be approximately 97 dBSPL peak for 0 dBFS from the DAC. Assume 10 dB peak to average so average listening level would be 87 dBSPL.

Legend:
Blu: THD+N
Grn: 2nd harmonic distortion
Yel: 3rd harmonic distortion
Red: 4th and higher harmonics + N distortion (crap factor)


The distortion is well below Radial Engineering's specification

Now if the DAC has analog output set for attenuation of 10 dB, adjusting the Signal Generator 10 dB lower with SA1X set for 10 dB attenuation to achieve the same output level as above produces this distortion profile:


Low frequency distortion is clearly lower than the previous gain staging.

You might try it both ways and hear for yourself how the two gain staging methods compare.

 

This is fantastic! Thank you so much

I ended up returning the Twin-ISO and am sticking with the J-ISO.

I will simulate some resistor based attenuator to stick in front of it and see if I can come up with something that can optimize the THD+N overall.

Most of all, I'll just enjoy my setup

 

I need a "bump box" since the Convert has only balanced outputs and my headphone amp has only SE inputs.

 

For the attenuators, I've tried Shure A15AS on my setup. For reasons I can't explain, it seem to suck life out of the music. Everything sounds flatter with attenuators.

Input voltage level does effect the distortion, but you can use it to your advantage. My Gungnir A2 + Cinemag + Magni 3+ sounds so cold that even Final Audio E3000 feels neutral on it. When I switched the Gungnir with Dangerous Music Source, I feel like I'm sitting next to a fireplace after a long winter day. Then I put a JBL Nano Patch before the Cinemag, I verified the Cinemag does sound warmer with higher input levels. To be honest, transformer distortion is pretty unique. It's not the wide band distortion from tubes and transistors. Distortion on the low frequencies makes things sound sweet, while distortion on high frequency makes thing sound harsh. Thus the uneven distortion FR of transformers could serve as condiments in certain cases. I recommend you double check the sound with your ears before adjusting the input level. If it sounds too warm, bring it down for sure. If it sounds neutral to you, then it should be fine to just leave it there.

The load impedance changes the FR slightly with my Cinemag. When I added Nano Patch before my Magni 3+ (impedance lowered from 25k to 5k), there is ever so slightly more bass. I suspect the difference might be more noticeable on other transformers though.

 

I had a quick question: I recently bought the two components recommended for non-DiY folks at the beginning of this thread (Radial Twin-Iso and the cables), and I was wondering if I should be attenuating the signal from my Dacs (Modius and Soekris 2541) before feeding them into my ZMF Pendant through the Radial Twin-Iso? The Soekris 2541 thankfully has a pre-amp to do this with, but should I attenuate the signal through Roon's DSP function with the Modius? And how much should I attenuate the signal by, generally speaking? Thanks in advance for any help, I'm not very good at this techy stuff.

 

Radial Twin ISO is rated for 1% THD at +20 dBu
at 0dBFS dac2541 Bal output is 13.63 dBu
at 0dBFS modius Bal output is 14.17 dBu
You are well under the spec rating.

If you feel your ZMF Pendant volume control range is limited to the very low end, then you might try 3 dBFS attenuation in the digital domain. Many DACs benefit from this as it will counter any filter ringing clipping on transient sounds.

Here is your specific distortion profile when you use modius -> ISO Twin:


If you add a Goldpoint SA1X and use 10 dB attenuation post Twin ISO this is the distortion profile observed:


No significant difference with THD+N being 3rd harmonic dominant.
4+ HD+N (crap factor) is much lower.