Calculating Output Transformer Primary Impedance

[TomCarlos]

I have been reading up on Output Transformers. One of my books has an abbreviated table which shows different tubes at differing voltages, in Push Pull, 2 tubes, 4 tubes, etc and the table gives a range of High and Low Impedance (matching to an Output Transformer). I'm not really sure this table is accurate (comes from "Inside Tube Amps" by Dan Torres).

So... I wanted to take a real world scenario... my Crate Blue Voodoo BV120h. The current options are for 8 or 16 ohm output. I would like to find something that has a 4 ohm output. For a given output section, 4 6L6 tubes in a Push Pull, what is the impedance range and what would be an ideal replacement Output Transformer? (For now, let's exclude the Crate replacement).

I am looking at the specs for a 6L6. I don't think it tells me what I need to know.

There's gotta be a good book or web site that has some diagrams and Impedance calculator? Anyone have anything bookmarked?

By the way, I stumbled across the Hammond 1750R, 100 watts, 2000 ohms CT, 4,8,16 ohm secondary. Looks like it would be a fit?

Thanks, Tom

[TomCarlos]

I just found a Sylvania data sheet - see attached. It looks like a general spec for Class AB using 2 tubes is 6000 to 3800 ohms. And if I understand this correctly, if you are using 4 tubes for Class AB, you take those values and cut them in half - so you end up with a range of 3000 to 1900 ohms (for the primary impedance)? These values are shown for 360V on the plate. But some amps run at a much higher B+. That is why a calculator would be cool !!!!

[guitician]

This old radio site has info... Output Transformer Impedance

[TomCarlos]

Thank you.... I will keep looking. Someone's gotta have a page that has some nice charts and variations of Class Types, Number of Tubes, Varying Voltages, etc.

[Enzo]

Maybe, but keep in mind all these are ballpark figures, none of this is precision stuff. And it all depends upon what the designer is trying to do.

The data sheet you found was for 6L6, which is the old metal tube, then 6L6G, the glass version, and then 6L6GA, a somewhat improved version. But all of those have much lower specs than modern 6L6GC types. The RCA book has similar data. The data is put there for nice polite designs, things that were expected to have long reliable lives, like table radios, and televison sets. Think of the book specs as your mom's car. Now we come along making guitar amps, we abuse the tubes, we like the results. Think of that as you borrowing mom's car and drag racing it against your buddy's mom's car. You will wear out the tires sooner, you will place more stress on the moving parts. Just as Fender routinely ran 6V6 tubes at 100v higher plate voltages than the book.


Someone may have charted it all up, but look at it another way. You have four 6L6 tubes? Fine, look at OTs made for four-tube amps. A Fender 100W something, a Peavey 5150, even a Marshall OT. You will find, wonder of wonders, that they do not all have the same impedance ratios. And yet they are all successful designs for quads of 6L6. My point being this: don;t get hung up on numbers.

[TomCarlos]

Excellent points.... and I think I kinda came to that conclusion. If I have Amp A that used four 6L6 tubes for the output but does not have a 4 ohm secondary, I should be able to use the same OT from Amp B, rated for the same power, also uses four 6L6 tubes, but has the secondary I want. Easy solution!!

Again, thanks Enzo.

[Chuck H]

But the question remains...

How do we determine the priZ that best suits whatever goal we have in mind? So we can consider output vs. distortion. Sometimes even the PARTICULAR distortion. It's all in the tube plots (if only I could get my head around them!?! @#$%) Most often I think the guitar amp designs from the later golden age (think BF Fender and plexi Marshall's) used OT's that offered better power performance since finite distortion figures weren't of much concern for stage performance and the "power wars" had begun. They also chose transformers that could be ordered direct and as cheaply as possible. Quality was a big consideration but they weren't going to pay more for it than they had to. So even with these amps the particulars of the OT primary impedance was likely just roughed in with available parts. This is evidenced all over the guitar amp design world when you observe what transformers were used where and find that part numbers were commonly repurposed for different designs with different parameters otherwise. In other words... It's just like Enzo said! Yes there may be some brainiac way to make finite determinations about OT primary impedances for different purposes specific to output and distortion, but for distorting tubes it would be too long haired and white coat for anything we do. Not to mention all the happy accidents that happen by flying in the face of any stringent parameters.

This is actually a question that comes up a lot. In my amp design travels I've found that I seem to prefer the higher impedances. As in 6.6k for a pair of big bottle type tubes at moderate to high voltage rather than the 4k-ish load often used in the classics. Speaker damping and such, yada, yada. But I always get satisfactory results just leaning that way with whatever I can buy off a shelf. That is, it's just as Enzo said! Close is close enough.

The tonal difference between higher and lower impedances (one set for the best ratio of power to distortion and the other set for max power) is worth experimenting with to find your own preference. These ranges are fairly finite. It's pretty much 4k-ish for a pair of big bottles to get max power and 6.6k-ish to get lower damping and distortion. The actual difference is only a couple of watts. You'll never hear it. But you will "feel" it as a player. So try both impedance ideals if it pleases you. IMHE the difference between two similar sized 4k OT's is tiny. And the difference between two similar sized 6.6k OT's is tiny. The difference between the 4k (per pair) and the 6.6k (per pair) is worth checking out though.

JM2C

P.S. for little bottles (6V6's and EL84's) Just go with the common designs which are mostly 6.6k to 10k per pair with 8k being the most common. Can't go wrong and it always sounds "right".

[olddawg]

And your 4 ohm load will work perfectly fine on your 8 ohm tap anyway....

[Merlinb]

Thank you.... I will keep looking. Someone's gotta have a page that has some nice charts and variations of Class Types, Number of Tubes, Varying Voltages, etc.

The trouble is, the required load depends on the supply voltage, desired output power, desired class of operation, and screen voltage. The screen voltage really throws a spanner in the works because each tube has a different Ia:Vg2 relationship. You would need a four dimensional graph for each tube type. And even then, the graph would only work for a "hifi" perspective. Guitar amps almost never use an 'optimum' load impedance in the hifi sense (or, looking at it another way, guitar amps nearly always run the screen voltage much higher than would normally be considered optimum). You're better off looking at the example data given on the tube data sheet and just eyeballing it (that's what Leo and Marshall did!). Tube guitar amp design is not an exact science.

[nashvillebill]

Go to the 3rd or 4th page of this 1961 Stancor catalog here: http://www.bunkerofdoom.com/xfm/STAN...ancor_1961.pdf where they list various tube types and their recommend OT. Not every tube is listed but it's about the most comprehensive list I've seen. The caveat is that the impedances are (ostensibly) for hi-fi use not guitar amp...so apply the usual disclaimers: YMMV, offer not valid in all 50 states, may cause drowsiness/sleeplessness/death, etc...

[Chuck H]

Not to mention all the complications that start piling up with the introduction and nature of NFB circuits and frequently clipping the piss out of the tubes. Idealizing OT impedance could become an exhaustive exercise WRT guitar amps. Agree about the screens too. Another good reason to make 1k a go to screen grid resistor value rather than the more commonly used 470r. Roughing in the PriZ value is satisfactory and any lack of perfection doesn't come at significant cost in tone or performance so why bother.?.

[TomCarlos]

Lot's of great info in the replies. thanks again!

[guitician]

Here is a clip from a text book on it....

And the "Optimum" load resistance is whatever the tube pair needs based on their curves, which will change with every maker. The thing is that tubes are generally very tolerant of variations in load and "Optimum" isn't all that tight of a spec.

[guitician]

RCA 6L6GC Chart for load vs Dist.

A clip from RCA's Data Sheet.

6000 ohms looks optimum for two tubes. Divide that in half for 4 tubes push-pull. But this is running at the listed screen voltage, which is 100V less than the plate.

[Chuck H]

That's interesting. I think there are parameters unaccounted for though. Like the efficiency of the transformer and how it transfers the power to the load. And the affect of NFB lowering the ideal. I've read many times that the 4k-ish priZ was used my MFG's to get more output watts at the expense of added distortion. According to the chart that shouldn't work. In fact they should have gone higher rather than lower. So there's something to the unaccounted for operating conditions I'll guess. Which is what I always do when picking a load anyway.