• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Might speaker filters be in the wrong place?

I am reading up on a lot of threads about OPT and big wattage amps. It seems that bandwidth is the problem, as is size.

Broskie has written about parallelling OPT's for BIG-as-s amps.

Now somewhere down the audio line, the music is split into differnt chunks to suit the different drivers (for multiway speakers).

Apart from the normal +amping and speaker crossovers, couldn't the crossover be before the OPT(s) and have different OPT's for every driver, being able to wind transformers to suit the specific bandwidth. Like amorphous high frey, c-core mids and big toroids for low freq...

It would be even possible to add multiple power stages in an amp. Like a DHT for mids and highs and hefty penthodes for bass...

Just a mindfart...
 
I was assuming that Bi-amping and active crossovers were known paths to go down.

I want, for practical reasons, one amp. If i'd go to a transformer winder and say "I want -3dB @ 8 Hz" he'd say it would ruin my highs. Now what if i could add a separate transformer for the highs.

As far as i can see, the OPT is the biggest limiter of bandwidth in the audio chain, apart from the speaker itself.

Bas
 
Hi Beamnet

Sorry, I did not really understand what you meant with your original post. My suggestion on both Broskie's article was based upon your last sentence

It would be even possible to add multiple power stages in an amp. Like a DHT for mids and highs and hefty penthodes for bass...

About crossovers between output tubes and output transformers: I have never heard about it being applied (I don't understand much of higher order passive filters at all, too). My thoughts (for what it is worth): Broskie shows that, SPL wise, a large fullrange amplifier can be substituted for two (three) small power amplifiers preceeded by an active crossover. So, as you are already pondering the use (and cost) of multiple output transformers, I would imagine that the additional costs of building multiple small power amplifiers will not be more than the one big amplifier: An example: a big amplifier built up of 3 pairs of EL84 paralleled can be split in 3 small amplifiers, still using the same power supply. You would have to add some extra driver tubes and triodes for the crossover functions, but hey, those russian valves are cheap, or otherwise I know a source of GE5670 for 1$ each. And then you can adjust Xover frequencies, baffle step corrections, with small and accurate (and cheap) capacitors and resistors, skipping larger inductors and big capacitors that would invariably be needed when mounting a passive crossover (and probably also a crossover between the output tubes and the output transformers).

Erik
 
Sorry eric, but the things you write are really OT and confusing. As I stated, bi-amping is very common and popular.

I was talking about the technicalities of OPT's. They are hard to make large bandwidth ant therefore are expensive. I was just wondering why the crossover is behind the OPT and not before. Just food for thought.

I never spoke of cost. I talked about practicality. Active x-over and multiple amps is the best way to go. But , just hypothetically if you whish, if I wanted just one integrated amp. I think 30 or even 40 hz is unacceptably high as a -3dB point, yet most tube builders have accepted thet that is inherent for tubes.
 
I have read about this concept somewhere before. I think that there was a thread here before concerning the concept of placing the crossovers before the OPT's, or using the inductance of the OPT as part of the filter. It was a while ago, maybe a year. I tinkered around for a while but I could never get it to work. The frequencies near the crossover point get all messed up due to phase shifts.

You just planted another idea in my head though. I never thought of running two or more output stages off of the same driver, each tailored for its own frequency band. More experimenting is needed.
 

Brett

Member
2002-01-07 6:02 pm
I see some issues with it.
- filters will need to be symmetrical
- a hefty driver will be needed
- all the output tubes will need to be the same; eg a 300B typically needs >60V for full output whereas an EL84 needs about 10V
- what happens when the grids go positive and current drawn

Better to biamp fully IMO. Select your designs and OPT's based upon the frequency range required and optimise. Seems to me like less compromise, and as you're 2/3 of the way to a whole second amp, do the job right.
 
Brett said:
I see some issues with it.
- filters will need to be symmetrical
- a hefty driver will be needed
- all the output tubes will need to be the same; eg a 300B typically needs >60V for full output whereas an EL84 needs about 10V
- what happens when the grids go positive and current drawn

Better to biamp fully IMO. Select your designs and OPT's based upon the frequency range required and optimise. Seems to me like less compromise, and as you're 2/3 of the way to a whole second amp, do the job right.

I tend to agree that going all the way is the more practical approach, but hey, if we were practical we'd not use many tubes in the first place.

I would also think that filtering between output tube and OPT would be dicey. However, a single input and driver stage and multiple outputs is certainly doable.

To the issues cited:

Why would the filters would have to be symmetrical?

Yes hefty driver, or MOSFET follower(s).

Yes, easiest if output tubes are the same, but just need similar grid swing and plate voltage requirements - doable. Maybe single to multiple parallel tubes. As for grid current, either stay out of A2 or refer to previous comment.

Not surprised that George has wheels turning.

Sheldon
 

Brett

Member
2002-01-07 6:02 pm
Sheldon said:
Why would the filters would have to be symmetrical?
I assumed a PP O/P stage
Sheldon said:
Yes hefty driver, or MOSFET follower(s).
No MOSFET followers here thanks.
Sheldon said:
Yes, easiest if output tubes are the same, but just need similar grid swing and plate voltage requirements - doable. Maybe single to multiple parallel tubes. As for grid current, either stay out of A2 or refer to previous comment.
Unless you have way more power than you need, it's not uncommon to find grid current on peaks in output stages.