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

Yes for me it was that exact one. But for all I know it was a dud that was returned a few times already!! I can say nothing but positive about the Radial though. A serious piece of kit.

 

I am looking at these Chinese specials and this store in Ali catch my attention, it seems to specialize in audio transformers. Will this work for BAL > SE conversion?

https://www.aliexpress.com/item/4001065465438.html

 

Aliexpress is kind of a roll of the dice.

Radial is a known and relatively trusted entity (they also bought Jensen from what I understand).

 

It seems to be able to do SE to balanced and vice versa, but it's got an awfully high DCR, and fairly low inductance for a core that big. From the same shop, this seems like better value to me:

https://www.aliexpress.com/item/4001068431205.html

The cores are permalloy / 80% nickel, which is really the better choice for an input transformer. Looks like the one you linked is GOSS and may be gapped as well, as it mentions it can pass 30mA?

I've seen some amorphous transformers on Ali, too, but I'd also save my pennies for a known quantity.

 

Thanks @peef @Armaegis ... I just started reading up on transformers so my knowledge is still limited.

There is this one too from ampsandsound but it seems awfully expensive.

https://ampsandsound.com/products/xlr-to-rca-converters

 

Got to admit I'm a bit tempted to purchase one of the ART units from Amazon and see what it sounds like and return it and buy the Jenson if it adds hum. But wondering if I (or someone more talented than I) should be measuring it and/or tearing it down as well as listening to it...

https://www.amazon.com/ART-DTI-Transformer-Isolator-Eliminator/dp/B0009GUOQA/

 

I just want to express thanks for this thread. I finally got the parts and carved some time out to create a converter with some Cinemags. It was simple following the diagram that @purr1n posted (except I grounded the white wire to the xlr tab—also thanks @batriq for the pix on his post). The part that took the most time was actually cutting the aluminum DIY case with the stepped bit jamming. The soldering of the wires to the XLR and SE was fairly straightforward.

Does anyone know if it makes sense cobble RCA attenuators with some resistors to tame the heat? Would that degrade the sound?

 

It should not get physically hot if used to convert low level signals

 

As a rule of thumb, transformers offer the lowest distortion when driven from a low impedance and a low signal level. Adding an attenuator after the converter will keep the source impedance low, but the transformer will see the source's full output voltage. Conversely, putting it before bumps up driving impedance, but reduces the signal across the transformer primary.

It's probably worth experimenting, but in either case a resistor attenuator brings a small penalty. If you're sure that you need the attenuation, swapping out the transformer for one with a 1:4 or 1:2 stepdown usually offers the best fidelity.

 

Apologies for the misunderstanding, it’s not physically hot, just a lot of juice to the single ended side. The amp now gets loud at around 8 o’ clock on the pot. I was hoping to get more room turning the volume knob.

 

Thanks…I reckon I’ll have to try to find out when I find the time.

Anyway, I thought I read somewhere that the 1:4 also affects the sound a bit. I guess converters, as with any audio equipment, are compromises.

 

I was interested in the suitablity of the Cinemag CMOQ-3 as an input transformer or balanced to SE interface between a DAC and amp. It's a quadrifilar transformer on a core larger than most of the ones discussed here, and is available on steel, nickel, or pinstripe cores.

I got Cinemag to send me a few transformers with the cores unassembled a little while back, and finally got around to analyzing a few different core stacks. The hypothesis is that the pinstripes or lower inductance core configurations might do better with high signal levels. I tested these at 4VRMS, which is more or less what the Yggdrasil spits out of its balanced outputs.

Nickel full interleave: each EI lamination is flipped, resulting in the smallest effective gap and highest inductance.
Nickel 2x17: EIs are alternated in pairs. There are a total of 17 pairs and two end Es with no Is holding it all together. Some info about this from Intact, but in short the BH loop gets slightly linearized.
Nickel Sandwich 2x17: Same stacking as above, but 8 of the pairs are nickel, and the remaining of the stack is made up of steel lams above and below. This should limit interaction between the nickel and steel, which would cause the core to prematurely saturate.
Steel Sandwich 2x17: The opposite of the above, steel goes in the middle. Electra-Print did their pinstripe core for me in this manner (with full interleave) when I got one years ago.
Cinemag Pinstripe: Full interleave alternating between steel and nickel (i.e. all nickel Es facing the same direction, and making up half the stack).




This is a very limited test, I think the takeaway is that CMOQ3H could be viable as an input transformer if driven from a sufficiently low impedance. I was surprised that the two sandwiches offered basically identical performance, and that they were a bit worse than the "pinstripe" config. I was also surprised that the 2x17 stack didn't really offer better level handling or lower distortion than the full interleave.

Edit: For comparison, here's the full nickel stack CMOQ vs the CMLI-15/15B.

 

I'm going to purchase one of these 1:1 boxes for my Gungnir A1 and I'm wondering if there is any general preference towards Radial TWIN ISO or the Jensen PI-2XR (PO-2XR)? Quite an investment and I only understand the basics, so I would love the input of someone more experienced. Would use short cable runs of no longer than 3 feet each way if that's important.

 

My own experience with the Gungnir A1 and the Radial twin ISO is that the sound was noticeably colored. So much so that I returned it. I've never heard the Jensen.

 

Curious question: instead of spending ~300$ on one of these devices (or making one using transformers for ~200), why not put that money into a preamp like the Freya (S) that would allow you to cross between balanced and unbalanced ins/outs?

 

There were some who went that route in this thread, IIRC. Freya S is more money than building one of these, though, I suppose. They do all seem to be colored to some degree, just different kinds of color, from what I've gathered.

After reading an ASR thread where the Amir cult guys (including solderdude, FWIW) were poo pooing the need for these and just saying to use a XLR --> RCA cable, I'd be interested to hear folks' views on those. I've read bob's post on the matter, and if I understood him correctly, the primary risk with those cables is ground noise/hum at minimum and blowing something up at ("magic smoke") worst... right? I honestly don't think I parsed him very well there.

 

I agree that the Freya is a lot more by itself, I was thinking about the following situation:

Sell saga (additional tubes, if any, not considered): ~200
Buy new Freya S: 599
Net spent: 399

Price of jensen/radial device: ~300

My prices may be a bit off, i didnt do a thorough search of the best prices but I assume the general idea is understood

 

As mentioned above, everything will add some coloring. I would go with the Freya for the added functionality (fully featured preamp with remote).

I expect the Freya to add a little more "flavor" than the basic transformers mentioned in this thread since the parts count is much higher. But I haven't done an A/B so what do I know.

 

Going with a Freya would probably be the most easy and flexible way. I built my Cinemag box because I was curious and I wanted something discrete. The Freya would be taking real estate and another socket…but having a tube option would be cool.

The Cinemags do impart a color to the sound, like Fujifilm Velvia vs Reala (or photo filters on your phone camera). I find it very pleasing and not distracting. I’m very pleased with them out of the Gungnir Multibit, not that the SE on that is bad but there is just a little more out of the XLR, and the C’mags let me tap into them.

 

The XLR->RCA cable often ties the negative leg to ground (ie: pin3 is tied to pin1). What that means for the source is that the negative leg is a short circuit, which means any kind of voltage pulls a lot of current which could potentially burn out the output device. Even if it doesn't, that much current imbalance is not ideal and could cause unwanted witchery on the positive leg.

Can you just snip the negative leg off so it's not tied to ground? Well yes, that eliminates your worst case. Well, one of the worst cases. The other one is if the balanced signal is not balanced around ground, in which case hooray you're now feeding some potential DC into your amp. There's less critical things like the whole common mode rejection blah blah or ground gremlins, but whatever.