Driver/Interstage Transformers Pt 1: The Driver Load

[_Steve]

Hi! Im researching how to design an amp with an interstage transformer instead of a tube phase inverter. My goal is to experiment with AB2, but also understand more about reactive loaded tubes, and power amp design in general. Im the sort of person that gets more satisfaction understanding the 'why' rather than the 'what'.

The first thing I really dont understand and can't find info for is the load impedance for the driving tube. Several of the schematics i've found have a load resistor(s) either on the primary or secondary such as this:

https://schematicheaven.net/fenderamps/musicmaster_bass_schem.pdf

That makes sense to me, however also i've seen many that have no load resistors at all like this:

https://schematicheaven.net/gibsonamps/ga20rvt.pdf

What's going on here? Is the driving tube using only the reactance of the transformer primary as it's load, or is there something else going on here I don't understand?

My immediate aim is to understand how to determine the transformer characteristics I'd need for a given driver tube type (eg 6V6) and then to be able to draw up some approximate loadlines, set a bias point etc.

Any help appreciated! After this i'll move onto the secondary side, frequency response, minimum grid resistance etc

[shooter]

built 2 amps with IS-PI's
V5's plate load is the transformer


I've fixed a couple GA5's that used IT, they are pretty easy to burn up with too much current, so if a tube goes short, the tranny will soon follow.
i'll try n dig up the notes, I used cap coupled IT, that keeps the tranny safer, but adds an LC component that is interesting when distorted.

[W5FH]

Steve-
    See my recent post here regarding the use of I/S transformer in the highly regarded series of custom built amplifiers by Mr. Bartel (Sugarland, Roseland and Starwood). Of course there is very little technical info on his circuits out there to my knowledge, understandably so as this is his work product of which I greatly respect. I am like you in wanting to know the "why" of things. I have studied the topside views of his chassis, and some underside views. One topside view shows the I/S transformer with Magnetic Components part numbers clearly visible, however I believe it is a propritary part as I cannot find any data on it.
    The reason I say this is that in his technical data he shows a 12AU7 tube as a phase inverter. In the description of amplifier operation the point is made of counteracting blocking distortion by using the I/S. In my feeble mind I envisioned the phase inversion being done by the I/S transformer, driving the output stage tube's grids. But when studying the underside of chassis at 12AU7 phase inverter (tube is identified on chassis drawing so I believe it is correct tube location) you can see what appears as a transformer lead going to each plate of 12AU7. There are 1.6K resistors going from each grid to nearby turret terminals. The cathodes are tied together with a wire lead going to turret. I cannot identify any other portions of this circuitry. It appears that a phase inverter tube is driving the I/S transformer. But this is only conjecture on my part but it sure brings up some interesting things. I have studied the Hammond 124B I/S transformer as it has the special nickel core for wider frequency response. However it is designed to be driven by a single ended stage; there is no center tap on primary.
    It appears you and I are pursuing a common goal here, hope I have helped some.

[tubeswell]

An interstage transformer reflects the AC-load on the secondary winding - back through the transformer - to the driving tube(s) plate(s). So the AC-load at the primary winding depends partly on what the load on the secondary winding is and partly on what the impedance ratio of the transformer is.

[shooter]

fwiw
here's the last IT amp I put together
probably not the exact version that went into the wild, since the tuning phase with 5 guitarist was kinda wild 

[PRR]

> how to determine the transformer characteristics I'd need

Are you planning to wind your own? The selection of ready-made interstage transformers is VERY small.

Do you know the grid resistance in the positive grid zone? And how fast it drops from the zero grid zone?

Also:
* Why does a tube "need" a load? Don't trucks and airplanes fly better without a load?
* There is always a load!! There are no infinite resistances or zero capacitances.

[_Steve]

Thanks for all the answers. I WAS kinda getting a feel from the previous answers that for the driver I would be using the reactance of the primary side as the load (which is what I had kinda assumed) so I was thinking the calculations would be the same as for choke-loaded stages so I was about to research those but..

With PRR's answer now im questioning everything and probably more confused than ever :(

To answer PRRs question directly no I won't be winding my own, but understanding how to calculate what characteristics I need to look for, even if nothing suitable is available, is more important to me than just copying the design and parts from a known build (of which there are few also!).

I do realize i have a LOT to learn, but I think it would be useful to a lot of people to walk through the learning process publicly and document it here?

[_Steve]

Ive attached below the simplest design AFAIK - maybe it will help understanding the problem..

[tubeswell]

Ive attached below the simplest design AFAIK - maybe it will help understanding the problem..

You need to add grid load resistance on the following tubes

[_Steve]

You need to add grid load resistance on the following tubes

I initially thought so too but that doesn't seem to be the case. Lot's of designs without them eg:
https://schematicheaven.net/gibsonamps/ga20rvt.pdf

Steve

[pdf64]

You need to add grid load resistance on the following tubes

I initially thought so too but that doesn't seem to be the case. Lot's of designs without them eg:
https://schematicheaven.net/gibsonamps/ga20rvt.pdf

Steve

What’s your design, eg negative feedback loop?

[tubeswell]

You need to add grid load resistance on the following tubes

I initially thought so too but that doesn't seem to be the case. Lot's of designs without them eg:
https://schematicheaven.net/gibsonamps/ga20rvt.pdf

Steve

The GA20RVT uses cathode biasing and the output tube grids are ground-referenced through the transformer secondary center-tap.
Your schematic had no context.


If you’re copying the GA20RVT,
the 12AU7 driver has a maximum potential voltage swing of about 460v peak-to-peak. Each output tube grid will take 22v peak-to-peak before onset of clipping, and you want more than that for Class B, say 38v. YMMV


If the 12AU7 was actually going to make 460V peak, and if the EL84s were assumed to be comfortably making Class B at that point, then 460:38 = 12:1 VAC or 144:1 impedance ratio. YMMV


If the 12AU7 won’t make 460V peak, then the interstage tranny needs a lower impedance ratio to get the output stage to make Class B. Assuming the 12AU7 only makes 350v peak, then 10:1 vac ratio.


The 12AU7 is rated for 2.75W (even tho this one is dissipating just under 2W). So, belts and braces, the transformer wants to be rated for at least 6V.A

[_Steve]

You need to add grid load resistance on the following tubes

I initially thought so too but that doesn't seem to be the case. Lot's of designs without them eg:
https://schematicheaven.net/gibsonamps/ga20rvt.pdf

Steve

The GA20RVT uses cathode biasing and the output tube grids are ground-referenced through the transformer secondary center-tap.
Your schematic had no context.

Perhaps I misunderstand what you mean by "grid load resistance on the following tubes"? I thought you were referring to grid leaks hence GA20RVT schematic as an example that doesn't use them.

I still feel a long way from understanding the relationship between the driver tube plate and the primary or the transformer given there is zero or infinite (im not sure which) reflected impedance from the secondary.

[_Steve]

Also:
* Why does a tube "need" a load? Don't trucks and airplanes fly better without a load?
* There is always a load!! There are no infinite resistances or zero capacitances.

After sitting on this for a while I think i might be getting what you are saying, but surely the inductance of the TX primary is the critical consideration no?

[PRR]

> surely the inductance of the TX primary is the critical consideration no?

That's important but only in relation to...... what?

If you really mean "big bottle" (expensive experimentation), you want to look at Fender 300. Which also used 6V6 as a driver.

No, you do not need grid-leak resistors, that's one big point of transformers.

EDIT: Fender likes the name "300". You want the one from ~~1970.
https://reverb.com/item/3436526-fender-300ps-bass-amp-rare
https://mhuss.com/PS300/Schematic.pdf

[_Steve]

> surely the inductance of the TX primary is the critical consideration no?

That's important but only in relation to...... what?

If you really mean "big bottle" (expensive experimentation), you want to look at Fender 300. Which also used 6V6 as a driver.

I think you're hinting at frequency response/bass cutoff, but surely it's really the only thing actually providing the voltage swing from the tube's current in this scenario no? Otherwise what is?

TBH Im not really bothered which power tubes right now. Im trying to gain the understanding to be able to design for whatever but im really struggling to understand this primary/driver side first. I know I have a lot to learn!

[shooter]

to understand;
the driver, the IT, the PA tubes, they are ALL interactive and should be "viewed" as a single unit.
much the same with the OT, that circuit is the PA tubes, the OT, and the speaker acting together as a unit


there is no magical differences using a transformer to couple vs resistor-capacitors to couple stages.  There are design considerations, values you don't want to exceed.  If you're well versed in mathematics, you can crank it out using the values from Gibson's design.


If you're well versed in experimentation, build a known working design, verify the important values under active conditions, then think-out what tweaks n changes you want to skew those values closer to your concept.

[_Steve]

If you're well versed in mathematics, you can crank it out using the values from Gibson's design.

Yes I prefer a maths approach and I have tried to find working designs so that I can derive backwards from them - except I can't find any designs with specs of the transformers. I've even found a Classictone replica of the Fender Musicman TX, but its datasheet doesn't even list turns ratio, let alone its inductance :( This is the crux of the question im trying to answer - how do I derive what specs I need for that TX given tube "x"?

[shooter]

just checked the paper on the IT i used, no henrys, it can be wired 1:1 or 1:2 ratio


I did the math to pass the class, now I just steal, measure everything, then do the easy math 

[PRR]

Look for 6F6. This was like 6V6 except it was commonly driven AB2 to get more output.

Read Liliethal, all of it. (I admit I never did.)
http://lilienthalengineering.com/100-amplifiers-chapter-1

[shooter]

Nice find!  was gonna go surfin, but it was 70 and sunny and things needed tending 

[W5FH]

Just for your information regarding interstage transformer coupling -
     I am learning a little more as I study the different designs. I realize that guitar amps are not HiFi amps. The frequency response of an interstage transformer can really suffer on the low frequency end if the core is saturating due to the DC current flow on primary winding. This is a consideration in using a single-ended tube driver stage on primary of interstage transformer. If the core is not "gapped" the transformer core will saturate at a lower DC current flow on primary. However, when gapped the inductance of windings drops. Consider the common reverb driver transformer with a single ended primary, typically 22.5K impedance, designed to be driven with a paralleled sections 12AT7. Usually an 8 ohm secondary for reverb tank. Not many of these transformers are specified as to construction but Edcor offers one that is specified as having a precision gapped core. I realize that in reverb service the lower bass frequencies are not desirable.
     What has really caught my attention is the use of a center tapped primary type interstage transformer. By operating the interstage transformer primary in push-pull you capture the benefits just the same as with a push-pull output transformer ( much less core saturation from DC current flow on primary winding). However, this means that you must do the phase inversion upstream of interstage transformer. Voila - here we have the 12AU7 phase inverter tube driving the interstage transformer primary as in the Bartel amplifiers. For more study on this take a look at the Magnatone MP1 and MP5 amplifier reverb driver circuits. A 12AU7 phase inverter is driving the center tapped primary of a reverb transformer (25K center tapped primary/8 ohm secondary) A 12AU7 is used due to lower plate resistance plus it can serve as a great phase inverter when set up correctly. It has fixed bias in thiese circuits.
      These are very interesting to me. I am going to do some test builds using 12AU7 P/I driving a center tapped interstage transformer to drive push-pull output tubes. So far the Hammond 124 series appears as one of the few choices of suitable center tapped transformers with a high impedance center tapped secondary. I am curious about some of the custom built interstages by Heyboer I have read about in posts, but have not succeeded in getting any info on. I plan to use fixed bias on the output tube grids so that means the center tap of secondary winding of I/S transformer cannot be DC grounded. My plans are to use a good quality film capacitor of several uFD to ground the CT for audio. Need to study a few more designs. Am very intrigued by I/S designs, would like to do reverb drive as described here, in addition to ouput stage drive. Of course, reason for all is to minimize blocking distortion of output stage and try to keep it as clean as possible.