HF AC DHT Filament Supplies

[simon]

Whats CMC ?

Common mode choke. As opposed to differential mode.

[pre65]

Whats CMC ?

Common mode choke. As opposed to differential mode.

Ah yes, of course.

I wonder if they do 10A versions ?

[Dave the bass]

Whats CMC ?

I think here its referring to -> Common Mode Choke. http://en.wikipedia.org/wiki/Common_mode_choke

DTB

[pre65]

Thanks Simon / Dave. It must have been that beef/pork/chicken/horse pie I had last night.

I might try them on 833a, there is a tiny bit of hum. They are cheap enough.

http://uk.farnell.com/jsp/search/produc ... 1864499-pr

[jack]

Yes, but the SMPS in those cases were producing DC, not HF AC. I have found though that a couple of stages of CMC and caps will get the noise of most SMPS's down to the mv range.

Why CMC - most SMPS OP noise is typically differential, not common mode.

Have you any measurements, specifically 'scope data on this - its interesting in that it certainly not what I would have expected - LCR/RCRC etc. I would expect to have significant impact, but a CMC?

[simon]

I might try them on 833a, there is a tiny bit of hum. They are cheap enough.

http://uk.farnell.com/jsp/search/produc ... 1864499-pr

As you say cheap enough to try, and it will be interesting to see if they make any difference. Note that those in the link are 85 micro henry inductance so really very little. But it might be enough, if your hum is caused by common mode noise.

FWIW I did use them on my Coleman filament modules. Not because I knew I had a problem, but because they were cheap and I only wanted to build the raw power supplies once so wanged them in anyway. Probably a triumph of eagerness over knowledge!

[Mike H]

Only any use for HF switching noise, won't help with hum.

[Paul Barker]

Yes, but the SMPS in those cases were producing DC, not HF AC. I have found though that a couple of stages of CMC and caps will get the noise of most SMPS's down to the mv range.

Whats CMC ?

The SMPS on my 833a filament is a 12v one, and I use RCRC to get it down to 10V.

I was fortunate mine has sufficient adjustment.

[Nick]

Yes, but the SMPS in those cases were producing DC, not HF AC. I have found though that a couple of stages of CMC and caps will get the noise of most SMPS's down to the mv range.

Why CMC - most SMPS OP noise is typically differential, not common mode.

Have you any measurements, specifically 'scope data on this - its interesting in that it certainly not what I would have expected - LCR/RCRC etc. I would expect to have significant impact, but a CMC?

I did measure when doing this, but I dont have records. When I get a chance I will repeat and post the results.

[steve s]

Just catching as limited Internet in the caneries , the real issue with any power supply is the extra circuitry, if removing a hum pot can make an ac heated dht sound better to me, its clear to me that any cicuitry in that position can make a difference to the sound, better or worse?
What I don't subscribe to is that any ripple or ac on its own, should make make the stage sound worse..
There's more to it than that in my view, seeing what was measured as distortion on my px 25... is not really distortion, it just an extra signal that is being amplified, and shows up as not being on the input, just as any varied music source.
the amp amp and speakers should be able to reproduce that and the music signal without reducing the fidelity of the stage. I'm sure that will be the case when I eventually rebuild the amp. But time will tell.

[Ray P]

I started this thread as I'm planning to build Bruce Rozenblit's 300B OTL amp and I want to minimise noise from the AC filaments that have to be used with this amp - HF supplies seemed to be the way to go.

Anyway, Bruce has come up with his own solution and has designed a filament supply that takes the opposite approach and has come up with a 25Hz supply, below what most systems will reproduce. It'll be interesting how this pans out. No circuit details yet.

Ray

[Paul Barker]

it is probably an excellent idea conscerning audio quality.

It requires a larger core and more turns per volt than 50hz. Core size would be experimental, (size for satisfactory noise and heat on test) then at 25hz compared to 50hz twice the turns per volt required, and so even larger core window (and therefore core) required for the same power. But working for you, once you have a larger core less turns per volt required than for smaller original core. So the two factors affect each other to your advantage. The net result is larger iron more turns but not double of either. it would be evaluated experimentally.

Conversly the Hf transformers are tiny little torroids with just a few turns.

So to make a long winded discussion succinct, near enough twice the size cost and weight to get down to 25hz or use high perm material which is cost prohibitive.


I like the idea though. Would be best to use a pure sign waveform at the 25hz. Might be issues with oscillation at 25hz wearing out speaker mechanicals; though unheard.

I am sure Bruce has a solution in mind.

[Nick]

I am sure Bruce has a solution in mind.

I think the key point is

300B OTL amp

[chris661]

I'd doubt there'd be much output from the amplifier from the 25Hz hum, so I don't think speaker excursion would be a problem.

If there was significant 25Hz signal reaching the speakers, you'd get some very interesting sonic effects, particularly with single-driver systems.

Chris

[Nick]

IMHO, the problem with any form of AC heating is not the hum (25hz, 50hz, 50kHz) its the effect of the intermod products of the hum and signal, so with 25 hz, you get sidebands +- 25Hz instead of 50Hz