Custom OT Impedance Question

[dynaleadamps]

I recently had an OT made at Heyboer. SE 40 watt, switchable 2k5, 4k and 6k pri/0, 4, 8, 16 sec and it was *supposed* to be for 2xoctal tubes in parallel. They didn't realize that I am using 2 power tubes, so my question is what happens to the primary impedance when using 2 parallel tubes versus 1 tube?

[MWJB]

Your speaker taps become 2, 4 & 8ohms.

[Steve Conner]

I guess I don't really understand the question here. Are you saying that you ordered a 2k5/4k/6k to 4/8/16 OT, and they made you a 5k/8k/12k one?

Or are you saying that you really wanted one that was 1k25/2k/3k, but you forgot to divide the impedance per-tube by the number of tubes when ordering?

A good OT should have at least one "step" of leeway either way. Using it higher than its design impedance, you lose bass. Using it lower, you lose treble.

[R.G.]

Forgive me, but I'm struck by the non-sequitur nature of someone selling $2400 amplifiers, complete with press releases and advertising talking about "to engineer an amplifier from the “ground up” " and not understanding what the effect of paralleled output tubes in SE does, not understanding that paralleling tubes does nothing to the primary impedance of a transformer, nor that transformers have ratios, not impedances, nor that the SE nature of the output makes this much, much stickier than just changing the impedance ratio of the transformer.

But I should not be surprised.

[Steve Conner]

Oh yeah, that's a good point. If you run a SE OPT at twice the DC primary current it was designed for, it's not going to be a happy trans-bunny.

[dynaleadamps]

Wow. Such Holier-Than-Thou venom in your post. The 3 designs I sell are solid, tested designs ... This is an experiment, something we 'lower class' folk do for fun. We build. We share information as we learn it. We try to foster an environment where questions and answers are welcome.

It must be swesome being a super-genius EE, MBA whatever.

Why so superior? No, wait don't answer it REALLY doesn't matter.

[Steve Conner]

RG has been sharing and fostering for a good 10 years with his geofex.com site. Maybe he just had an off-day today.

I think he was also a transformer designer in a previous life.

[dynaleadamps]

I see. I apologize for being dick-ish, I was just caught off guard. I am used to AX84 where there are very few negative comments in regards to sometimes obvious questions.

Maybe point me in the direction of some reading and learning rather than putting down my love for building and designing?

Thanks, Gents!

[dynaleadamps]

Back to the original issue, The OT was designed for 40 watt, 290ma with the above pri and sec impedances, but for a single octal output tube.

So when using two octal output tubes in parallel, how does that change the way the pri and sec impedances react?

is it as simple as MWJB states about the sec impedances?

[Steve Conner]

Well, since I'm in a good mood and have nothing better to do, I'll tell you what RG would have said if he had been bothered. (Well, I think. We're both EEs, but I'm not a mind reader.)

When you're ordering a SE transformer from a transformer maker, you tell them the power, the bandwidth, the primary and secondary impedances, and the DC current rating of the primary. This last thing is important, because if you exceed it the core will saturate and it'll sound crappy.

I'm guessing that you didn't specify a DC current rating to Heyboer, or you'd have spotted the problem before ordering. So there was another layer of decision making internal to them, to get the full transformer spec. How did they think you were going to get 40 watts out of one tube in SE?! Maybe call them and ask them.

So, if you want to use your transformer with two tubes in parallel, the DC current rating of the primary should be greater than or equal to the combined idle currents of your two tubes. That's the most important thing.

[Steve Conner]

Oh, I just read your simulpost above. 290mA is plenty, I hope the transformer came with a free forklift to move it around.

The transformer doesn't know or care how many tubes are hooked up to its primary. It just transforms a 4, 8, or 16 ohm load to 2.5, 4 or 6k, according to which taps you hook to.

If 2.5, 4 or 6k are inappropriate loads for your parallel combination of tubes, then you need to do some creative mismatching as MWJB suggested. If you hook two tubes in parallel to a 6k tap, then each tube is seeing a 12k load. Mismatch the speaker load one step as he suggested, and that 6k tap becomes 3k, so each tube is seeing a 6k load.

But this isn't the transformer maker's problem. It's yours as the amp designer. You asked for a transformer with those impedances: if you didn't realise that connecting two tubes in parallel needs the load impedance to be halved, and so accidentally ordered a transformer with twice the primary impedance you really needed, that was too bad.

[MWJB]

Yes.

[loudthud]

Holly Smokes !

A quick calculation says 40W at 2.5K will be 316V(RMS) and 126mA(RMS). So B+ need to be 540V or so and a pair of KT88s will be red plating. I heard a rumor there is a KT120 with a 60W plate rating or maybe you'll want to plan for more octal sockets and a fan. Maybe a 211 or 811? Oh, those don't take octal sockets.

[Steve Conner]

Champ 1000 Watt Tube Amp that'll do it

[MWJB]

Loudthud, why would the KT88 redplate at 540v?