Crosstalk in 2503/2623 outputransformers/builds? (Anybody measure?)

[JW]

I see a lot of builds where 2503 or 2623 API style output transformers are stacked right next to each other, like in API style summers, even on Capi's ACA board. I guess this is okay? Crosstalk is not an issue?
I'm finishing off an 8 channel unit that sort of requires this, although I might be able to get away with a bunch of 90 degree angles, but then this would place the transformers closer to the power supply.

 

[ruffrecords]

I seem to rememeber quite some time ago someone had a crosstalk problem with Carnhill output transformers right next to each other but I have not come across it with API ones.

In Neve consoles this was generally not a problem because there was usually two sheets of steel between them but I don't know anything about the mechanics of API consoles.

Cheers

Ian

 

[Whoops]

I seem to rememeber quite some time ago someone had a crosstalk problem with Carnhill output transformers right next to each other but I have not come across it with API ones.

It was probably me Ian.
When output transformers are side by side I find that you should rotate one of them 90degrees, that also a quite common procedure in Tube guitar amplifiers.

At the time I made 2 videos showing the crosstalk between Neve output transformers at different positions.
Music is on channel's 2 input, and we are just listening the ouput of channel 1 (nothin connected to CH1 input).
Changing channel 1 output transformer positioning relative to CH2 output transformer changes greatly the amount of crosstalk. remember in the videos the music that you're listening is just the Crosstalk





I didn't do the same test with API transformers, but you should do it as I'm pretty sure that if the transformers are side by side directed in the same way that the crosstalk will be high, they can be side by side but rotate one of them 90 degrees.
Try it out,do the same test as I did

 

[Whoops]

This is what happens with most transformers, and I believe the same will apply to the API output transformer

 

[JW]

Ah, that was you Whoops. Yeah. I remember that one, and it prompted me to build my 4 ch. Neve unit doing exactly that (90 degree angles) But that's the big Carnhill units. With this 312 clone unit It will place the 8th Xformer quite a bit closer to the power transformer if I situate them at 90 degree angles.

In an API console, I assume these transformers are in the chassis bay cavity behind the busing rails or something? Anybody have some pics? I'm curious how close/shielded the are.

 

[Whoops]

But that's the big Carnhill units. With this 312 clone unit It will place the 8th Xformer quite a bit closer to the power transformer if I situate them at 90 degree angles.

As I said in the last post, I don't see any reason for not happening also with the smaller API output transformers...
You should do the same test that I did, with your API type transformers and then decide if the crosstalk difference is negligible or not. Try it out

 

[CJ]

there is a copper foil wrap on the API original output transformers,

if your transformers have this then cross talk might be lessened over the neve transformers which do not have the foil. experiments are nice,

 

[Whoops]

there is a copper foil wrap on the API original output transformers,

Hi CJ, does the 2623 have copper foil?
or it’s just the 2503 that has it?

thanks

 

[FIX]

2623s where never that close to one another. 2503s had the copper for the coil, but the steel case helped but it was for damage prevention.

No one had issues with it in those days, as tape hiss and crosstalk well exceeded anything that came elsewhere, though I never had any complaints nor testing issues with them. Plus, I don't think anyone cared...

 

[Whoops]

No one had issues with it in those days

Hi Paul, it will for sure depend on how close are the transformers to each other. How close are we talking about?
Are the transformers almost touching each other?

being it 1/4 inch close or being 1inch makes a big difference. The crosstalk at 1inch distance might be neglectable...
So maybe output transformers in API consoles and gear had enough distance from eachother to have a neglectable crosstalk figure,
the space inside and APi console is not comparable to a DIY project that you try to fit 8 output transformers inside a small rack case were space is limited.

Plus, I don't think anyone cared...

That I'll disagree with you, I'm pretty sure the engineers at API cared when they designed the consoles, and I'm sure that the spacing between output transformers inside the consoles was decided with crosstalk in consideration.
I don't how close are they inside the consoles, if anyone has photos it would be great to see.

But what I'm sure is that for a DIY project fitting 8 output transformers inside a small case I would test the positioning for sure

 

[Whoops]

this can also be a solution

 

[Walrus]

If you take the right hand side of this PCB and duplicate it on the left, that is very similar to how an API ACA output board is.
https://www.capi-gear.com/catalog/images/gallery/PCBs/2-ACA-Bo/Rev-B-stuffed.jpg

 

[anders]

This is how Trident solved Swedish Radio´s crosstalks demands when implementing 52 x Jensen 123-S output transformers in the +30 batch of Series 80 recording/broadcast consoles in the early 1980ies. Very tight clearancees underneath, but I guess this arrangement met the strict Nordic broadcast specs.

 

[Whoops]

This is how Trident solved Swedish Radio´s crosstalks demands when implementing 52 x Jensen 123-S output transformers

Thanks for sharing,
The way they did it it’s the best way possible when the transformers have to be really close to eachother.
It’s a mix of the 2 diagrams I posted in this thread, they’re rotated 90 degrees and the part closer to the neighbor transformer it’s the Lams and not the bobbin.

They knew well what they were doing for sure

 

[FIX]

Hi Paul, it will for sure depend on how close are the transformers to each other. How close are we talking about?
Are the transformers almost touching each other?

being it 1/4 inch close or being 1inch makes a big difference. The crosstalk at 1inch distance might be neglectable...
So maybe output transformers in API consoles and gear had enough distance from eachother to have a neglectable crosstalk figure,
the space inside and APi console is not comparable to a DIY project that you try to fit 8 output transformers inside a small rack case were space is limited.



That I'll disagree with you, I'm pretty sure the engineers at API cared when they designed the consoles, and I'm sure that the spacing between output transformers inside the consoles was decided with crosstalk in consideration.
I don't how close are they inside the consoles, if anyone has photos it would be great to see.

But what I'm sure is that for a DIY project fitting 8 output transformers inside a small case I would test the positioning for sure

Well, Owning API (a 27 year span) and employing Saul and both of us being "the engineers", I can assure you that it wasn't something we spent a lot of time on, if any. The smaller transformers in the modules had about .5" between them, the ACA cards were like Jeffs, but all in a line the same way, and the larger transformers were in a steel chassis. To get crosstalk, Saul told me once "If you put a tone in at level, you can measure it, but no one ever does that so we never worried about it" No one single engineer I knew (and still know) agonized about it, from 1976-today. Not one single question.

Crosstalk between one transformer, driven by a low-Z source and another with a low-Z source (Basically shorted to ground) could only crosstalk on a 1:1 ratio, attenuated by the source impedance.

I think you have to keep in mind that it may be measurable, but I doubt that any records wasn't released because someone said "Listen to the crosstalk between the snare and the HH". Tape had worse track bleed, records were worse than that, and AM/FM had their own issues, yet it's still some of the highest selling music even today.

Sure, when CDs came out, you could do a stereo bass, but who did?

So sometimes you can find problems that seem like problems, but you have to weight the fact that it may not effect the end result at all.

James Brown did one of his biggest hits with no mic on the piano because they were out of mic pres. They put it in the middle of the room and it bled into everything. For the solo, James said "I'll just move my mic over to it" and that was the record.

Don't over agonize.

 

[CJ]

Nice! ^

those API outputs were low turns which means less likely to pick up stuff,
and if one end of the secondary is grounded, you now have a shield wire circling the primary wires due to the quad-filar design.

 

[Brian Roth]

It's a different universe in a broadcast system. It is totally common to have (say) a 1 kHz tone appearing on random remote lines....into a channel or out a bus...preparing for the next segment of the programming.Then any sort of crosstalk becomes a problem.

Bri

 

[Brian Roth]

I'm thinking back to maybe 1970-ish. I was in High Screwel and making a whopping $1.65/hour at a small commercial radio station where I worked on the weekends. Sunday mornings I was running plastic spools of 1/4" tape on the Maggie machines, playing 15 or 30 minute sermons, paid for by Reverend Doctor Sendmeyourmoney and his various friends.

At 11:00 AM, the local Church Of the Gooey Death had bought the next hour. We had what I now know as a "dry broadcast loop" from the church to the station. Essentially it was a hard copper path through the Bell Tel exchanges. Every Sunday at around 10:30 AM, I'd get a call on the "hot line" from a Bell Tech who was at the church and was sending tones to me. I'd put that remote feed into "Audition" on the CCA air desk and read back levels to that Bell guy while he tweaked EQ at that far end. I wonder if that was a Pultec there????

Anyway, the test tones were slightly audible on air as Rev. Youboughtmeanewcadillac was playing from tape. Everything in that facility was balanced 600 Ohm with transformers....but a rats nest of wiring.

In broadcast, crosstalk IS an issue. MANY years later I designed/built several streaming TV news facilities and once again became aware of crosstalk. Remote feed from Vyvx via fiber or satellite? They throw up color bars and a 1K tone when idling.

Bri

 

[abbey road d enfer]

View attachment 88447

 

[Matt Syson]

I was 'brought up' on Broadcast desks so crosstalk between almost any non related paths was an 'issue'.
So between adjacent channels of the desk, faderand switch cutoff etc, all had to be designed to prevent 'leakage' between paths. 'Better than 50dB between left and right of stereo channelss was the minimum requirement at 20kHz (IIRC) so 70dB at 1kHz. Of course since stereo material was either grams or tapes at that time even that seemed a bit of a moot point.
Desk operators had to feel confident that if you had one channel with someone saying things quietly, would not be 'interupted' by an adjacent channel with a 'presenter' shouting obscenities causing all of the mic chain up to the final mute/fader running at clipping.
Then we had the 7 second 'profanity delay' units. That caused a bit of a headache for 'split broadcasting'. One show broadcast to two different areas which carried different advertisements.
I miss the old days!