Gates Radio 7X tube preamp

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Yeah, I had a long conversation via phone, wanted quotes on the various tube outputs, and they wanted to know what I was going to make so they could design something.  Wouldn't address stock pricing.  Very strange. 
 
There's also the possibility of using Lundahl iron.  The LL1578 or LL1578XL will do for microphone inputs or the LL1922 (1:8) for 600 ohm line level sources.  The Lundahl output transformer perfect for the output stage job is the LL1680/5mA using the 15K:600ohm connection.
 
We need a 20 mA transformer here, or at least a 10 mA for a modified version.

I realize the K&K Audio BC-2B repro preamp used this Lundahl.  Lundahl needs to remove the claim from their LL1680 data sheet that the LS-27 was used in the RCA BC-2B; it wasn't, though one would work.  The RCA transformer is nothing like an LS-27; ratio is different, winding resistances are different.
 
How does one calculate the current the output transformer needs to handle? How is it in relation to the tube bias current?

Also, that LL1680 has three separate current ratings, which one am I supposed to be looking at?

Recommended DC current (or Primary DC current for 0.9 Tesla): 5mA
Core Saturating DC current: 10mA
Max standing DC current through any primary section: 50 mA

And further, if the rating we should be looking at is "Primary DC current for 0.9 Tesla", then Lundahl transformer we want is the LL1671/30mA (SE Line Output 4:1). I'm under the impression you can also order LL1671/15mA SE types for improved frequency response and perhaps inductance.

Prices are not as exorbitant as I would have thought. about €90 for a standard model. But then €230 for "amorphous core" version, whatever that means.
 
Nice find with that GXSE5-600-15K especially. But 40~18K Hz., <1dBu is not great. 18K I could live with but that bass cut might turn out to be a problem.

The 25W model is strange though, since it's not a 4-16ohm out for guitar amps or hifi speakers. What 600-ohm line level gear needs that kind of headroom? Some kind of obsolete speaker format?

It has better frequency response (20~20K Hz., <1dBu) so if you have a small power tube like maybe ECC99/6N6P it should make a decent line driver with quite extreme headroom. But that's a redesign already.
 
On the tube side of this topic:

The price of 6SN7 has got a bit silly, has anyone got any thoughts on using 6N7 for this app?

The cathodes are tied internally but that's no problem if the tubes doubled up.

best
DaveP
 
DaveP said:
The price of 6SN7 has got a bit silly, has anyone got any thoughts on using 6N7 for this app?

I was going to suggest trying a 6CG7 but by my calculations that 6SN7 is dissipating 5W per triode.

Cheers

Ian
 
DaveP said:
The price of 6SN7 has got a bit silly, has anyone got any thoughts on using 6N7 for this app?

I hear Electro Harmonix and especially Tungsol makes very good 6SN7. Sovtek is also quite cheap.

http://uraltone.com/kauppa/advanced_search_result.php?XTCsid=dqrt89jicfj41t234pna6u6947&keywords=6SN7&x=0&y=0

Then there's 6N8S on ebay. When you browse those Ukrainian surplus stores, don't forget to check out their PIO caps *nudge* *nudge* *wink* *wink*
 
Kingston said:
How does one calculate the current the output transformer needs to handle? How is it in relation to the tube bias current?
When you know the tube runs with 5mA quiescent, you want the transformer to be specified for at least that, and the saturation current to be about twice. The output stage runs in class A so the peak current is about twice idle.
Also, that LL1680 has three separate current ratings, which one am I supposed to be looking at?

Recommended DC current (or Primary DC current for 0.9 Tesla): 5mA
That's the normal operating area.
Core Saturating DC current: 10mA
You want the peak current the stage is capable of delivering to be less than that.
Max standing DC current through any primary section: 50 mA
Not a normal condition. More than that and you fry the windings.
Prices are not as exorbitant as I would have thought. about €90 for a standard model. But then €230 for "amorphous core" version, whatever that means.
Amorphous metal is supposed to have superior magnetic properties. Some are evident, like lower losses, which have made them attractive for high-power transformers (I'm talking MVA's), hence have made the price drop consequently, making them somewhat less expensive than hi-Ni cores. Amorphous cores being superior for audio applications has not been demonstrated.
 
That Edcor looks interesting.  Hadn't seen it.  1dB down at 18K is just fine; a lot of older gear doesn't do that well.  That's the first of the Edcor types I've seen that did respectable bottom.

I have the benefit of knowing 20mA, as the same output stage and transformer is used on another Gates piece which has a current meter jack, and a published reading.  

6N7 is used in similar designs older than this.  Worth a try.   I like the 6F6 and 6K6 too; in many ways the 6F6 fits a certain niche better than any other small power tube, downside being that it's hard to source enough from any one place to make decent pairs for PP operation.  
 
Kingston said:
abbey road d enfer said:
Core Saturating DC current: 10mA
You want the peak current the stage is capable of delivering to be less than that.

RE: distortion

I wonder if these air-gapped SE output transformers have a softer clipping/saturation point than non-gapped transformers.
They have a softer curve, not as steep as non-gapped, but for a given voltage, distortion comes earlier, because the inductance is lower.
 
has anyone got any thoughts on using 6N7 for this app?


I've used them before and have a vintage piece that uses one in SE operation.  It appears the issue of concern is the higher source Z of N7 vs SN7.  N7 Data sheet lists plate Z for // plate operation as ~11K.

But 40~18K Hz., <1dBu is not great

30Hz, 40Hz corners.  Question is will you notice it to a great degree?  I think for apps where you must be flat to some point below this then something SS might be better.  For tube circuits the CF transformerless output ala WE141, or NYDaves One Bottle sans OT offers the different bottom characteristic without worry of the OT messing things up.

The 5687 looks like it might be a contender for getting lower source Z without getting into big tube current levels.  2K @ 23ma. It costs more heater power but not that big a deal.

http://www.mif.pg.gda.pl/homepages/frank/sheets/127/5/5687.pdf

 
I should have written that the LL1680 is a good candidate for the output transformer and it's available gapped for a wider range of currents than just 5mA.  Whether the lineage is related or not to the LS-27 or BC-2B output is irrelevant in this context, although you are welcome to take it up directly with Per Lundahl if you think that it's important.

On the point of amorphous iron, I would daresay, on experience with several designs comparing transformers with identical coils and either amorphous iron or conventional grain-oriented steel cores, that there is a repeatable sonic difference between the two that many would deem significant (and have with their dollars).  However, in a single ended circuit, such as that starting this thread, the difference is pretty minor to my ears.  In push-pull applications, the difference is much more striking.
 
about SE output transformers in general,

how do they affect the tube curve in comparison to a simple plate resistor? Since the signal is formed directly across the transformer primary, will this have any plate choke like current sourcing effects making the plate curve more linear?

Is there any advantage in using SE output transformer other than not needing a plate resistor and a coupling cap?

when using an SE output transformer I still need to bias the tube optimally using datasheet plate curves. But what is the "plate resistor"?
 
One major advantage is that you can have a lot more voltage swing compared to the resistor
loaded stage. Even beyond the supply voltage.

Not sure what the "current sourcing effects" are that you mention.

The plate load is whatever is connected to the secondary reflected to the primary by the
impedance ratio of the transformer.

Here and here are nice descriptions how to design a SE amplifier.
 
Thanks for those links.

volker said:
Not sure what the "current sourcing effects" are that you mention.

I mean that a place choke DC resistance plots as a near vertical line on the plate curve as it can be made a near ideal current source on audio frequency range, replacing plate resistor. I was wondering if SE transformer would do the same since it also replaces plate resistor.

Reading those links now, it's apparent these things are not related.

I suppose parafeed design combines the advantages of a plate choke and SE output transformer. Or does it?
 
Kingston said:
about SE output transformers in general,

how do they affect the tube curve in comparison to a simple plate resistor? Since the signal is formed directly across the transformer primary, will this have any plate choke like current sourcing effects making the plate curve more linear?

Is there any advantage in using SE output transformer other than not needing a plate resistor and a coupling cap?

when using an SE output transformer I still need to bias the tube optimally using datasheet plate curves. But what is the "plate resistor"?
The advantages of a transformer are that gain is higher since the xfmr impedance is very high, supposed to be several times the actual load, and output swing is not limited to B+ because of the energy magnetically stored in the core.
The problem with a directly connected xfmr is that it has to be designed for handling a certain amount of DC.
Parafeed with plate resistor has no advantages; it just provides impedance adaptation and galvanic isolation.
Parafeed with plate choke has all the advantages; it is often preferred because of the questionable notion that the problems are addressed separately: the choke takes care of DC handling and energy accumulation, the xfmr takes care of the "noble" signal. In fact, since it is in parallels, the choke must be designed with the same constraints as the xfmr - minimised parasitic capacitance, saturation headroom. All in all, it is a much more costly option than a good SE xfmr, with no definite advantage IMO. RF applications are a different matter.
 
This thread is great!

I get SE now. The "magnetically enhanced" output swing range gives it a great advantage. Tube can be set to a much wider linear area on the plate curves compared to a "dumb" resistor on an equal B+. And no parasitics from a coupling cap either.

Obviously I now have to experiment with this. Perhaps I will design a guitar amp with some cheapo SE transformer.
 
Back
Top