8417 Tube Sound Question

The "sound" of a tube (assuming that such thing exist) is mostly dependant on the circuit topology in wich it is used, and many other external factors.
My personal experience with the 8417 revealed that it "sounded" wonderfully in my FISHER SA-1000 , just so-so in a DYNACO MK VI and like crap in a BOGEN P.A amp.

Well, I don't know why a thread should carry on repeating nonsense and the same tired mantras after all.
Valves DO have a sound signature based on their cathode size, electron path length, linearity, gain and impedance.

It appears the longer thinner type anode/cathode/small heater high impedance designs have a very different harmonic distortion character from the shorter fatter larger heater types.

Most of the later post 1960 designs & sweep tubes became in the 2nd category, - low impedance, higher gains & some of them able to handle peak currents in the amps.

The 1960s designed valves in general were a quantum leap better than the 1950s ones, I reckon with the arrival of first B+W then colour TV, so a need for smaller = bigger frame grids & higher gains.

The KT66 is 1930s, the KT88/6500 were tweaked KT66 from the 50s.

All that "V6/L6/KT" stuff is 1930s tech with some later tweaks, while lots of the later 50s designs were Philips with some sweep tube ideas thrown in.
There's really no comparison. IMD figs can be desperately high with those designs (inc KT88 and 6550).

Being is this 8417 was a proper 60s late dev, There is something special in it.
It has to be treated in a certain way, and it doesn't behave really quite like anything else*.
It's "sort of like a 1960s modernised design KT66" with much 4x the gain, and nearly double the ratings.
The internal spacings & gains have become a subject of controversy, many claiming they are fragile.
I don't believe that at all, but the G2 HAS to be regulated properly, and the odd triode curve behaviour seems to show this to be critical.
Because of the strange triode curves, it's obvious a UL amp is going to sound totally different from a pure pentode design.

I would hazard a guess, the cavity anode actually handles the problems of the non aligned G2, to prevent excess emissions, but it appears to create a strange kink around the -1>0V area.
That seems to hint something going on in the saturation area.
In my opinion, g2 potential v final gain curve is the way to make this valve work properly.

I reckon Bogen got this right. It had to be reliable, so they deliberately cut off the extreme LF.
(You can easily get it all back cos the OPT can do it.)

Anyone who attempts to claim the MO/MX series are unreliable needs their head examining, cos they simply don't go wrong.
Bogen wanted 24/7 - 365 so they put PPP in, then ran G2 at 300V,- (approaching sweep tube ideas again).
The difference between getting 100W+ from a pair, or only 60W, simply comes down to how hard you want to run them, and the OPTs.
It doesn't need PPP at all, so Bogen went for overkill despite the whopping 660V Va.

The idea you have to run the 8417 g2 like the 6L6 (300V) max is plain wrong, but there's not much help in the spec sheets from Sylvania.
Fisher ran them less gently at 420V, Dynaco at close to the max - 500V, QS at 470V.

The KT66 sort of started life as a failed transmitter valve with a larger cathode, which is why you can drive them into grid current and abuse the daylights out of them.

The 8417 began as a regulator and/or large audio valve, hence the even fatter cathode, large fat anode and high emission.

(My 50 yr old Sylvania 8417s still behave like new, & I am told this is equally common for the Sylvania 7591A too).
8417 wasn't intended to be driven into grid current, and there are no curves published to show it.
Idem the 7591.

Better regulator valves for this are found from GEC-STC. They use the massive cathode 12E1/ S11E12, which incidentally has the same large heater wattage as the 8417 and can run the same 800V Va as the KT88.
They also weren't designed to be run into grid current, - having a maximum of 300V on G2.
It states clearly grid current starts at -1V.

The high G2 emissions from the 8417s non aligned G2, seem to indicate it should ideally be used ultralinear but in my opinion a very different turns ratio from 43%.
(Screen load 18.5% impedance)

Here is maybe something to think about:- "the optimum load impedance (Raa) is generally higher than for Pentode Mode .... In fact I was astonished to note the Raa values for some late design output tubes in Ultralinear."

Example: The 6973 Output Tube
http://www.mif.pg.gda.pl/homepages/f...049/6/6973.pdf

Raa recommended for UL operation is 12 to 13K - about double what is recommended for Pentode Mode.
I'm now wondering if perhaps we have not made sufficient adjustment to Raa when using UItralinear for more common tubes."


Suprisingly only Dynaco (see both below) & Quicksilver did 8417 as UL, but both got the rest all wrong, so one sounds terrible & the other one melts valves,- both giving the amps/valves a bad name.

I don't think anyone experimented with UL ratios least of all Dyna or QS.
So it appears Dynaco VI is plain WRONG, grossly badly matched with the wrong transformer in that application.
Here is one opinion pretty much confirming it.

eg. Sylvania 7591 suggests 40% for lowest distortion, but RCA curiously says "50% turns recommended". (6V6 family seems to prefer 20%.)

Quicksilver were the only people to persist with this valve until the 1990s fade out, 1984 - 1992.

There was a thread about blocking distortion on here, where I argued how 8417s had to be used to prevent runaway and make terrific power.
It's simply a low impedance high gain device.
Get the drive circuit &/or grid leak res/stoppers wrong and it will melt or sound bad or both.

Dynaco made a mess of this with the MkVI, the Bogen PA amp, quite contrary to what was said here sounds absolutely fantastic (once a few very simple mods are done).
It's well documented now, with a minor tweak it delivers full power down to 10hz.
For a PA amp that's unheard of.

The Quicksilver which was a clone of it, can sound ok once the stupid bits of design which cause it to melt are changed.
Ie. there are people that say the QS sounds really horrible as it is, (with a well known issue destroying the valve rectifiers, with the well known runaway GE 8417 melting syndrome).
They say it was only rated because it was so early in the valve amplifier revolution.
Who knows?
I wouldn't buy one.

The Fisher 1000, used some of the ideas that are now known to work, but opted for that crazy triode strapped ELL80 driver.
I would suggest the Fisher sounds good for the simple reason it's the only one which uses a tweaked Williamson design, lots of NFB, CFB instead of UK,- clearly needing and using excellent transformers to get it all ultra stable. (as D Gillespie has often shown!)
IMD is 0.5% and THD is quoted 0.25% at full power.

The freq response is really pretty flat from 10hz to 40khz.
Try making another one?
You will need to pay a visit to Mr Partridge or Mr Sowter if you do, so it won't come cheap.

The Bogen had issues with stray IW capacitance, (which DG showed while the 6550 appeared to cope with better at HF).
I don't think these are insurmountable, especially if you can DC couple the output valve grids.

Some people get desperate when they see the dyna VI specs, and then hear the bad results!
When you look up the dynaco specs in detail, you can see the figures are garbage.
They used the A470 OPT, a crap design with thin wire, not up to a low impedance high current device, but OK for their old EL34 stuff.

They opt to overquote the power figures as a result to 120W "continuous" which is crap industry doublespeak for only 60W RMS (which actually makes sense with that OPT).
Then they show the total current consumption of the amp is 250W (120W of which is quiescent...), so the amp is clearly incapable of making more than 60W...at which they quote the high frequencies tailing off because of saturation at 15khz.
The MkVI was bound to be a fail, looking like a lab amp, but it's a fake one.

To think of this roughly right, the 8417 is basically the same as TWO 7591A stuffed into one small bottle but with higher ratings.

The only valve that resembles it is the rare Philips EL503/520*, which also sounds wonderful.
Both of them are impossible to reproduce, so the only way to go forward today is using TV sweep tubes or some oddball Russian ones.

As an example of how modern manufacturing hasn't worked, the 7591A is known to be a superb audio valve, with a totally different sound signature to any of the other 60s amplifier valves.

The 7591 was hard to make, which is why the cloned versions from Russia, based on the Russian 6L6 (itself based on the stalin era RCA 6L6/807) have lower power, lower gain and higher distortion- ie. less linear.

Because the outsize cathode and larger anode, with the close high gm G1 spacing, stuffed into a 6V6 size glass, was so successful in fact they could make the 7868 then the even smaller 6GM5 out of it.
Those make any EL84 look like rubbish.
Luckily GEC made a much better version of it like a mini KT88 (the 2134).
They never seem to have caught on.

Unsurprisingly Scott got rid of the EL84 cos it sucks, and replaced it with the much better 7591A.
Bogen again made loads of small amps with the 7868s like the CHB-100.
The PPP versions work very much like the single pair 8417 versions.
PPP 7591 effectively is very similar to a single pair of 8417.

Luckily lots of 7591 got used in Scott, Ampeg, McIntosh 225, Harmon Kardon, Fisher which reads as a who's who of high quality amps.

The 7591 is plainly a lot better than the ancient coarse sounding RCA designed 6L6, or the poor quality Mullard/Philips EL34 which was based on the ancient EL37, because it's fundamentally more linear.
It's also much more better than that failed 7355 thing which also picked up the 6L6 family habits too.

None of the Mullard EL (84/34) family ever got around the inherent fragility of the Philips higher gm pentode design, which resulted in strong G2 emission and melting under high voltages.
There are people that HATE the nasty EL84, for reasons I perfectly understand.
There are plenty that also dislike the various EL34 clones too.
As a 1950s valve goes it simply isn't very good, and Philips/Mullard knew it when making it.

Unsurprisingly both the 6L6 and EL34 are still made, while the better technology valves aren't.
This has always been the way with free market capitalism.
The least good cheaper technology nearly always wins.

I hope this review made some sense.
 
Last edited:
Then they show the total current consumption of the amp is 250W (120W of which is quiescent...), so the amp is clearly incapable of making more than 60W...

I seriously doubt that the Mk.VI opt is saturating at 15khz. That amp is running a power trans capable of 400 mA. That is ~190W. Class A efficiency is around 40-45%...and this amp is decidedly AB1. That 60W conclusion is not supported very well IMO.
cheers,
Douglas
 
I seriously doubt that the Mk.VI opt is saturating at 15khz. That amp is running a power trans capable of 400 mA.

That 60W conclusion is not supported very well IMO.

The power transformer is totally irrelevant.
The amp has a dynaco OP transformer which was designed for 60W.

Did you read the figures from Dynaco?
They actually quote the facts the amplifier is incapable of delivering full power at above 15khz and below 30hz, and suffers high THD and IMD as per usual cheapo solutions.

Adding to that, pretty much everyone agrees the amplifier sounds bad, is badly put together and goes haywire at the slightest excuse (which is why DYNA insisted on replacing the entire QUAD each time if a valve melted).

I don't get any of this rubbish with my 8417 equipped amplifiers.
They are fit and forget - total bomb proof reliability, and they sound simply superb.
In fact, I don't frankly give a damn what they look like, when they work so well!

After sharing a few notes with DC, we came to the similar conclusions by different roads, - the use of PPP is stupid and adds a lot of extra distortion. (pls note!).

I did note in passing, the importance of proper impedance matching, as did DC, by changing the OP winding config, using only a PAIR, then devised my own solutions which really work like magic.

The Dynaco VI is a POS, so it's quite lucky they made so few of them.:rolleyes:
 
Last edited:

Part of the secret is sharing the whopping power supply between the pair of them, (halving the number of output valves), and driving it all with low impedance.

That gives it more power & stability than the MO200 when you use 2 seperate MO100A.


Then it's capable of delivering 1 amp at 650V.
 
Ab2forever
A 2nd or 3rd revival...
I have just purchased a pair of mo100a's. One has the original Saylvania 8417's and the other has no tubes. I am going to use them in my hifi system. My tech that is doing the work on them is very knowledgeable. You mention a few tweaks on them to make them better? In reading your posts it sounds to me like you recommend going with the 8417's in the second amp.
These amps will not be pushed hard, I am trying to save what my 58 year old ears have left :) If you have any other suggestions you can offer it would be most appreciated. I am used to listening to high quality tube sound (Audio Note) so sound quality is more important to me than power.

Regards
Paul
 
if you have a good set of 4, there is a way to re-work these amps to use 2 tubes per amp and get considerably better than stock performance. Or it can use a 6550, KT88, KT120, etc.

https://www.audiokarma.org/forums/index.php?threads/the-mighty-bogen-mo-200a.556575/

The MO200a is two 100a's on a single chassis so all this applies. This is my build thread for an MO-100a based on Dave's work on the 200a. These amps are currently my main amps and I still just don't have anything bad to say for them. I borrowed a set of Gold Lion KT88 tubes and they sound really nice. My own normal runner tubes are Tung-Sol KT-120. Also nothing bad to say but I think the GL KT88 has slightly more detail up top. The TS KT120 is a little better on the bottom.

https://www.audiokarma.org/forums/index.php?threads/the-slightly-less-mighty-bogen-mo-100a.685890/
 
Hmmm....a slight aside from the current thread direction, but still hopefully relevant. I have been toying with building a single ended amp using 8417s. I happen to have a pretty good supply of the tube due to a number of Dynacos and Bogens sitting on the shelf, along with actively acquiring 8417s a few years ago. What might be the design issues that this particular tube brings to the table? Any thoughts about SEP vs SE strapped triode? What might be some driver tube possibilities?
 
Back
Top Bottom