Yea or Nea on input bandwitdh filtering?

[jeffdavison]

Considering using the Van Alstine input filter mod. I read both good and bad.
Van Alstine's "white paper" seems very reasonable and well thought out and the mod seem appropriate to do the keep the trannies and what not from going beyond their limits and messing with the sound quality.

on the other hand I've read and one source told me that the filter muddies up things sonically. If top quality components are used..i.e. Auricaps/Musicaps and the like and Vishay/ Caddock resistors are used (only one cap and one resistor in the signal path... should that help? (not that I want to get into the "sound of a resistor conversation here)..just want to use good componets that don't "add" color.

Simple little filter that can be taken out easily and made easily. Probably going to answer myself and say use it and if "I" dont like it, it's easily removed...

Just want to get others inputs and experience with the filter/mod suggested by Van Alstine.

Happy Holidays y'all

Jeff Davison

[Shannon Parks]

Hi Jeff,

Could you post a link?

[WA4SWJ]

Shannon,

The article is here:

http://www.avahifi.com/root/audio_basic ... ab1982.pdf

- begining on page 14 of the pdf or so. It's simply a bandwidth limiting RC filter on the input of the amplifier. The theory is that some of the extreme low frequencies can overload the power supply in the ST-70 and cause distortion. The proposed solution is to limit the bandwidth of the amplifier.

Jeff - in my opinion, any filter, regardless of the component quality, introduces phase shift in the signal path and that phase shift will vary with frequency. Can you hear that? Maybe - maybe not. As for me, I would not even think of doing that mod. The new ST-70 I built now uses a solid state power supply with much more capacitance and a larger Triode Electronics power transformer. The amp sounds wonderful and shows absolutely no sign of needing bandwidth limiting.

The best thing you could do to improve your amp is to put one of Shannon's driver boards in and use NOS 12BH7's as the phase inverters. I also have an old, original ST-70 I did that to and it sounds wonderful as well. I have no interest in Shannon's products other than being a happy customer.

Check out Shannon's published curves for his design.

All the best, good luck and Happy Holidays and Happy New Year to all my friends at diytube!!!!

Regards,

[dcriner]

I read the article. The theory is that the ST70 output xfmrs saturate when trying to handle low frequencies, "clear down to DC." But I didn't see any test results that actually confirms that they actually do saturate while playing music. If the xfmrs were to saturate, the current would go way up and I think you'd hear distortion pretty clearly. During the moment while the xfmr's core is saturated, I would expect that all frequencies would be highly distorted, not just the low parts.

(Don't worry about DC - that should be cancelled out in a push-pull amp, correct?)

If you start hearing distortion accompanied by low-frequency passages, couldn't that be corrected with a little adjustment of the pre-amp's volume and/or tone control? But, in any case, has anyone actually noticed distortion accompanied with low-freq passages?

Testing should be easy: input a low-freq signal with an audio signal generator, with the speakers connected. Observe the ouput voltage waveform with a 'scope. If the xfmr is saturating, the waveform will appear distorted.

The original ST70 specs published by Dynaco claimed a freq response of 20Hz-20kHz, +/- 1dB of 35W per channel, less than 1% distortion. So, if the xfmrs are saturating, it must be happening below 20Hz.

[dcriner]

Earl, with the mod, I think there would certainly be a response roll-off at very low frequencies - that's what the input filter is designed to do, right? Can you do the distortion test with the mod and also with the mod bypassed?

According to the original ST70 specs, the full-power reponse is within 1dB down to 20Hz, so a least down to that point, there is virtually no rolloff in the amp's response. I think that such response testing is done with a resistive load (instead of speakers).

With speakers, the impedance changes with frequency. To see if the output xfmr is saturating in real life, it would be proper to use speakers. Start, say, at 100Hz, and work down from there, checking the ouput waveform for non-sinusoidal shape.

But, I still tend to think that if the xfmr were saturating, the amp would sound noticably bad.

[erichayes]

Hi All

Couple of comments . . .

First, check when the article was written. In 1982, the turntable and magnetic cartridge were firmly entrenched as the signal source of choice for real high fidelity--Nakamichi et al notwithstanding, By then, turntable rumble had been pretty much designed out of even the mid-priced machines, but because the record makers, with a few exceptions, were using less vinyl (175 grams vs 220 grams) in the discs, surface anomalies began showing up. Warpage was not uncommon, and it created a signal that was so low you could actually calculate it by watching the speaker cones' excursion. More and more electronics companies were putting op-amp driven shelving type subsonic filters in their preamps that went to work at 16 cycles and were usually at least second order. Vacuum tube preamps didn't have that luxury; they had to make do with first or second order filters, which meant they had to start turning over at around 50 cycles--well within the audible range--in order to do any good in the subsonic area. In any event, power amps were made wide band for the simple reason that any signal tweaking was to be done in the line stages of the preamp, allowing the user to swap- out power amps without worrying about any intentional doctoring of the signal in the PA, itself. All 6dB of bandwidth tampering is going to do is make for a lousy sounding amplifier.

Second, I test all my amps down to 10~ for spec purposes, but when I was designing the transformers, I'd check them down to 1~, just to see what they'd do. I also tested several other brands and models of OTs, including the Eico HF89 and Dynaco A450, 470 and Z565. They all exhibited the same characteristic: a faithful, clean sinewave until around 6~, where everything would promptly go to Hell. If you're saturating your metal at anything above that, you'd better look either for better transformers or bigger problems in the amp.

[Shannon Parks]

Ultimately, the ST70 will perform fine stock. But if you are running the output transformers near saturation all the time, you might as well move on to either a larger tube amp (say 60W/channel) or get more efficient speakers.

[Uncle Ned]

Considering using the Van Alstine input filter mod. I read both good and bad.
Van Alstine's "white paper" seems very reasonable and well thought out and the mod seem appropriate to do the keep the trannies and what not from going beyond their limits and messing with the sound quality.
Jeff Davison

Thing is, if you look at the original AVA mod for the ST70, IIRC, he takes out the small frequency compensation capacitors and increases the coupling capacitor value by a bunch (again, IIRC, like 1 uF instead of .1 uF). I think that's why he came up with the filter. Basically, he took out the frequency compensation that was already in the ST70 and the filter was necessary to replace what he took out. At least that's my take on this, IOW, I don't see any reason to use it with Shannon's board or a stock ST70.