Hiwatt Output Stages, NFB

[bea]

Hi,

in my attempts to understand the Hiwatts in more detail i came across the topic of NFB layout. In the schematics recovered by Mark Huss, the NFB details are drawn in the preamps, and apparently suitable for the 100W power stages.

In theory the NFB should be optimized differently for the other power stages, notably 200W EL34 and 200W KT88. This has been done in the Sound City variants of these output stages, and i wonder how the Hiwatts are designed in this respect.



If, for example, You would simply use the NFB loop of the 100W DR103 in a 200 W stage, the negative feedback (and the open loop gain) would be significantly larger due to the larger voltage at the 16 Ohms tap, and it would be unclear if the phase compensation of the loop would still fit.

(stability in a clone with a different PT might be an issue for itself).

[Gaz]

You are correct, but apparently Hiwatt didn't care. I'm sure the detail was not lost on them (or at least I'd like to believe so). The 200W sounds good with more feedback to keep the lows tight for bass, and the treble less brittle at higher spls. Also, what do you mean by: "..it would be unclear if the phase compensation of the loop would still fit"? I understand that may be something lost in translation...

[bea]

Thanks - You mean that the NFB circuitry is identical in the 100 and 200 W Hiwatts while there are differences in the Sound Cities designed by the same developer (and in some earlier or custom variants as well)? BTW - i am talking mainly of the vintage designs, not of the current amps.

phase compensation might be necessary to ensure stability of the amp. It may be necessary to adjust it in a different power stage (different OT, i.e. different HF resonance, and different open loop gain).

[Gaz]

Thanks - You mean that the NFB circuitry is identical in the 100 and 200 W Hiwatts while there are differences in the Sound Cities designed by the same developer (and in some earlier or custom variants as well)? BTW - i am talking mainly of the vintage designs, not of the current amps.

phase compensation might be necessary to ensure stability of the amp. It may be necessary to adjust it in a different power stage (different OT, i.e. different HF resonance, and different open loop gain).

Yes, I've always just assumed Mark Huss' schematics were corrected, but maybe I took that for granted. I have built the DR201 with the 10K/2.2K divider off the 16 Ohm tap with no stability issues. It's a lot of feedback, but nothing too out of the ordinary. I will have to take a look at the SC schematics though, and I hope Mark will chime in!

[mhuss]

Chime.

The 200s I've seen use the same 10k/470R combination connected to the 16 ohm tap. See, for example, Hiwatt Internals. I've only ever seen one or two 400s, so I can't say for sure there.

Phase compensation is only a worry when the amplifier is only marginally stable to begin with. The standard Hiwatt PI/amp combination is very stable, and so increasing or decreasing the NFB does not have any unexpected effects.

Note that Dave Reeves only did one run of Sound City amps for Arbiter, one of the earliest, and the later designs had the hand of others involved.

[bea]

Chime.

Thanks.


My effort is still about understanding why my Mywatt killed an Eminence 2515 speaker which matches 350 W of thermal load - it worked like a charm during one reheasal and was dead when i switched on the amp again one week later. I do by far not play that loud - actually a power stage of slightly below 100 W would fully suffice my needs (2 EL34 at 750 V and a hiwatt preamp clone...).

So i speculate a bit on oscillations. But Mywatts exist for more than 10 years (mine is one of the oldest) - any such problems would be known. Mhmm.

[J M Fahey]

Thanks.


My effort is still about understanding why my Mywatt

killed an Eminence 2515 speaker which matches 350 W of thermal load -

datasheet claims 300W RMS

it worked like a charm during one reheasal and was dead when i switched on the amp again one week later.

Somebody might have earned some extra ReichsMarks renting the studio while you were away.
Not kidding.

I do by far not play that loud

Hah !!!! famous last words

- actually a power stage of slightly below 100 W would fully suffice my needs (2 EL34 at 750 V and a hiwatt preamp clone...).

Is that what you were using

So i speculate a bit on oscillations. But Mywatts exist for more than 10 years (mine is one of the oldest) - any such problems would be known. Mhmm.

It's nor the head problem.

Neo magnet speakers typically stand way less power, long run, than similar ceramic ones.

It's a problem of physical/thermal mass and dissipation surface, nothing esoteric.

[g1]

I think an autopsy of the speaker would be valuable in determining if this is caused by over powering or oscillation. Perhaps it was just a coincidence and one of the leads to the voice coil gave up that day.

[J M Fahey]

I suggest you read this *excellent* analysis by Rod Elliott about murdered speakers:
Speaker Failure Analysis
Didn't expect anything less from him and so good it should be a sticky everywhere.

[bea]

Somebody might have earned some extra ReichsMarks renting the studio while you were away.
Not kidding.

And the cause of a severe conflict with the band we share our rehearsal room.

Hah !!!! famous last words

No, seriously, i am obtaining the loudness with a 45 W amp (sin, sustained) which is not clipping audibly (but i am aware that it will clip...)

Is that what you were using

No, i was using the Mywatt (Hiwatt 200 clone). But i own a few amps with such power stages. All from the 60s and 70s.

BTW: i am going to move to a 2x115 setup for the large amp, with EV 15L clones (Oberton 15L400). At present, i am recovering from a surgery an may neither carry around the Mywatt nor those heavy speakers. The 13 kg of my old Dynacord (that 80 W / 750 V beast) and the 16.5 kg of my 115 Neo cab are the limit for the next few months.

BTW: thanks for the article. Really interesting. Actually a room resonance by e.g. a badly tuned drum kit might be able to kill a speaker in a vented cab. Never was aware of that possibility.

And maybe i'll open the 2515 between the years and look if i'll find something... (there is still some residual functionality - scratching noises and large DC but unstable resistance)

[J M Fahey]

Problem is thar voice coils by tnhemselves handle very little power.
Rod Elliott shows it: a 40mm one (1.5" , what Jensen 12N uses) reaches 250C and burns with mere 38W (old Jensen paper and nail enamel glue burnt over 100C , boiling water temperature) yet we know speakers standing more than that.
Fact is air around moving VC is turbulent and then some (imagine thousands of tiny hurricanes and tornadoes) , is equivalent to having a strong fan blowing, and carries heat away "as if it were solid silver" (what Fane Engineers measured and their own words).

Where to?
The magnet structure, of course, just hundredths of an inch away.

That's why magnets get hot with high power long use and EV once published 2 power ratings: one in fully sealed cabinets and one (larger) in open back or vented ones.

Enter the tiny lightweight Neo magnet ... now you can imagine the end of the story.

Situation is so bad that designers had to add cast finned heatsinks:

Regular Eminence Delta 15


NASA approved Eminence Deltalite II 2515


sorry about your back, but you will never ever have that problem with your excellent looking Obertons

[bea]

Like with a single EV 15L, the max excursion of the Oberton should be a bit small for a 200 W Amp.

[g1]

Is running only a single pair of power tubes an option for you?

[bea]

Is running only a single pair of power tubes an option for you?

It is something i would like to avoid. One reason: the heater voltage would rise. But anyway: if the cooling of the voice coil should be less problematic with the 15L400 (thinking about it, the heatsink of the 2515 appears indeed comparatively small for a thermal load of 300 W... other manufacturers use larger ones): the limitations of the cone excursion will be audible, and there are sufficient examples of single EV15 L running nicely from amps of that size. Having a stack of 2 cabs allows to use a single cab where it seems sufficient.

Back to the main topic: good to be aware that high frequency oscillations of the amp can be ruled out as a cause of the problem, and again a matter of thinking for a few minutes.

[J M Fahey]

The real confirmation of the cause of death will come after the autopsy.
This which sounds like a cheesy cliche line from NCIS, Bones or The Bone Collector (or plain old Sherlock Holmes), is in fact true.

JBL sends authorized reconing shops (I have a copy somewhere) a poster meant to be hung on the wall, directed more at customers than Techs themselves, showing some 6 or 8 damaged voice coil pictures, showing different destruction scenarios, so Tech can show them to a Customer who claims "oh, I was just running some Classical string quartet music at bedroom levels and it simply stopped working" , point at the corpse and the matching picture and say: "yours matches a voice coil with over 60V DC on it, or 300% overload for more than 15 minutes or 100% power below 20 Hz or ... or ... or ..."
When I find it I'll post it here and someday you'll open your speaker and compare.

Don't know German market costs but you *should* be able to get an Eminence reconing kit for a reasonable price.
Only after you have it, get a box cutter, cut all around the cone and spider, clip tinsel wires and pull full cone/coil assembly out, then post a couple pictures.

But do it only after you have a proper replacement; if unavailable or way too expensive, a local reconer might be able to replace just the voice coil with an aftermarket one, would still need the original cone.
I would