EL84 Parallel SE transformer

[dunit]

Hello;
Hoping to get some knowledgeable suggestions about choosing an output transformer for a future amp build I'm working on.
I've done a couple of SE EL84's and loved the results.
I've tried using 5K primary OT's rated at 5W, 10W and 15W. The 3W offered some interesting results, but I was concerned whether or not
my house insurance would cover the house fire that seemed looming. :-) That said, the 5K primary was definitely the way to go, and I like the
10W the best.
My next project is to implement parallel, single-ended EL84's. Again, I plan to test with various wattage output transformers to dial in the
tonal characteristics they add. My question is; (I'm not sure if this is right) - it seems to me that the two output tubes in parallel would prefer
to be matched with a 2.5K primary. Sound right?
The math seems to work, but I can't find any technical writing to clearly state this is so, or is wrong.
Anyone have a better answer than mine?
Thanks.

[Dave H]

My question is; (I'm not sure if this is right) - it seems to me that the two output tubes in parallel would prefer
to be matched with a 2.5K primary. Sound right?

Yes, that's correct.

[dunit]

Yes, that's correct.

Thank you for the quick reply and confirmation that math is still my friend. :-)
Much appreciated.
Have fun!

[loudthud]

A lot of people seem to think 5K is the ideal impedance for a single ended output transformer primary. Primary impedance should be based on B+ Voltage, allowable Plate dissipation and peak Plate current. Look around online and you can find equations to help with the math.

[mozz]

Instead of the math, you can put a 4 ohm, 8 ohm, 16 ohm load on the output and see which produces the most power. But it may not be correct by the math method but it will show you what the tube likes.

[Helmholtz]

My question is; (I'm not sure if this is right) - it seems to me that the two output tubes in parallel would prefer
to be matched with a 2.5K primary. Sound right?

Yes.
Imagine each tube having its own 5k plate load.
By the parallel wiring also the plate loads are in parallel, resulting in a shared load of 2.5k.

The ~5k plate load per tube with SE class A is confirmed by the EL84 datasheet with B+ = 250V, Rk = 135R.

[dunit]

Yes.
Imagine each tube having its own 5k plate load.
By the parallel wiring also the plate loads are in parallel, resulting in a shared load of 2.5k.

The ~5k plate load per tube with SE class A is confirmed by the EL84 datasheet with B+ = 250V, Rk = 135R.

Yes; and, thanks.
Without going into the full-on calculations, the simple math suggests sharing the load would reduce the load needing to be balanced at the transformer.
I came here just hoping to get some feedback/suggestions if I'm on the right track or not.
As was mentioned in a different reply, what works out best is relative to the voltage being applied. All my amps, so far, have been
SE, single tube, EL84 seeing around 280-300V through the primary (depending on each particular design).
When I get this one started, I'll play around with how much voltage sounds best.

[Helmholtz]

Without going into the full-on calculations, the simple math suggests sharing the load would reduce the load needing to be balanced at the transformer.

Don't understand what you mean with "the load needing to be balanced at the transformer".

[stokes]

I built something similar about 10 yrs ago. I used a 5k SE tranny with a 6V6 and El 84 tube. Basically a Tweed Cham pre into that. I used either a 4 or 16 ohm speaker, dont recall, I know I tried all 3 impedances 4,8 &16ohm. I gave it to my sons girlfriend and she's still using it today.

[dunit]

Don't understand what you mean with "the load needing to be balanced at the transformer".

Maybe matching would have been a better word. But I don't like repetition if I don't have to. :-)

[dunit]

I built something similar about 10 yrs ago. I used a 5k SE tranny with a 6V6 and El 84 tube. Basically a Tweed Cham pre into that. I used either a 4 or 16 ohm speaker, dont recall, I know I tried all 3 impedances 4,8 &16ohm. I gave it to my sons girlfriend and she's still using it today.

Sounds cool.
Hammond makes a transformer which shows on the data sheet that you can adjust the "matching" of the tubes and primary (or approximate that) by simply applying the speaker to different taps on the secondary side.
This makes complete sense as the balance of the load translates through the OT each end affecting the other.
It seems to me that the difference with this one might be that it was designed to handle what for another transformer a mismatch of impedences which some OT's really don't like.

[dunit]

A lot of people seem to think 5K is the ideal impedance for a single ended output transformer primary. Primary impedance should be based on B+ Voltage, allowable Plate dissipation and peak Plate current. Look around online and you can find equations to help with the math.

Yes. Perfect theoretical matching gives the "best" results. BUT; we're talking about guitar amps. Everything we're doing is, in some way, going against the
rules so the resultant sound is NOT perfect. :-)
My first experience with this notion was a long time ago. Didn't really know too much about this stuff yet. I did know that output tubes were supposed to be matched.
I didn't really know what that meant, and relied on the guy behind the counter to make it happen. Until I had just one of the two tubes fail and couldn't afford a new set.
So I bought the single and hoped for the best. The amp suddenly sounded much more interesting - I liked it. As it didn't blow up, I kept it that way until the other tube
failed (cheap set). This time I just yanked it out with only one tube in use. I actually liked this sound even better!! :-D It was raw, nasty, noisy - but it worked for me.
Soon after the amp (Reverberocket II) died. I didn't know what to do, but that's when I decided I was going to learn what's going on in there. It's been a treasure trove
of fun and adventure ever since.

[g1]

Hammond makes a transformer which shows on the data sheet that you can adjust the "matching" of the tubes and primary (or approximate that) by simply applying the speaker to different taps on the secondary side.
This makes complete sense as the balance of the load translates through the OT each end affecting the other.

This is true for all output transformers, Hammond just includes it in their documentation in this case. It is what mozz was referring to in post #5.

The output transformer just provides a ratio of impedances from primary to secondary. If you change the loading on the secondary, it reflects back a changed impedance to the primary.
So, for example, if an OT is rated at 5K primary with 8 ohm secondary tap, putting a 4 ohm load on it will make the primary 2.5K instead.
Or in the Hammond example, maintaining the same speaker impedance but putting it on different taps will increase or decrease the primary impedance accordingly (by the impedance ratio).
This is also why when removing a pair of tubes to drop a 100W amp down to 50W, ideally we should change the impedance switch to half of what the speaker impedance is, which doubles the primary impedance to match the new power tube scenario (making impedances matched again).

[mozz]

Must squeeze out every available watt. Sometimes a 12at7 will get you more than 12ax7 in the phase splitter. But I test at 1khz, I should try 400hz. The vintage 16 Carvin chassis I just picked up here uses the same output transformer in the 16w, 33w 50w models. It's pretty big for a 16w. They cut off the extra leads. I'm measuring 4.3k for 8 ohms so in the 2 tube model you are putting 8ohm on the 4ohm tap. I don't know if you can drive a pair el84s at 4.3k with 350v plate.

[Helmholtz]

This is true for all output transformers, Hammond just includes it in their documentation in this case. It is what mozz was referring to in post #5.

True, but there are drawbacks.
The primary winding of an OT is optimized for some nominal impedance.
Using it with lower impedance increases the primary current and thus the DCR losses.
DCR losses increase with the current squared, so the power loss can be significant.

Using the OT with higher than nominal impedance sacrifices bass response.