Triode or Pentode?

Freo-1

Well-Known Member
During discussions on a number of threads, it seems some folks did not take too kindly to what they perceived as a critical assessment of the Harmon Kardon Citation II.

While I love the sound of the Citation II, I decided to change the unit over from a pentode to triode input. I caught some flak for this, so I wold like to get this subject out on the table for discussion.

First, I refer the reader to this thread, which discusses the citation II:

http://www.audiocircle.com/circles/index.php?topic=49209.0

Second, I refer to this topic on distortion:

http://www.spectrum.ieee.org/print/1640

So, one can see that while the Citation II enjoys a fine reputation, there is an argument to be made that the amp can be redone to improve performance.

What I did to mine is take out the 12BY7 pentode front end, and insert a pair of 5687s per side. I'm looking around to post the schematic (I have not been able to find it, but I will). To get an idea of the topology, google the MFA 120 amp (which uses HK citation II output iron), and a 5687 front end.

The thread and article are sourced from well respected audio engineers, so keep that in mind when commenting.
 
Last time I had it running, I attempted to drive my Citation II into oscillation as described by Roger, and I saw no spurious cone movement even at insane levels driving my 85dB sensitive AR 302s - this is with all new caps, however. I still suspect that he may have had problems with his power supply during those tests. Poor decoupling between stages can have the same effect.

For the circuit, you are of course welcome to do as you please, and I'm sure there are many paths to good sound when those output transformers are in play. I do see a lot more to learn from the Citation II front-end than from the MFA-120 front-end. I also think that part of the amp's charm is the pentode front-end - which is one reason why people may be disturbed at the prospect of changing it completely. It's a bit like putting a Chevy V8 in an RX7 ;)

I still hope to measure the amp some day ... I would bet that it performs in the same league as McIntosh, Marantz, and the better Williamson amp designs. Probably too low of distortion for my own measurement setup.

Regarding the pentodes, the distortion of the 12BY7 in that application is likely extremely low - One big advantage is the bandwidth, being much higher than for triodes, which is what allows the nested feedback loops.
 
Personally, I don't see the benefit of taking out pentodes and putting in triodes.

The 12BY7 is actually MORE linear than most triodes... and has far more bandwidth, as mentioned. This leads to much less distortion, across the spectra (lower low-order and lower high-order harmonics).

IMHO, the problem many people have with the Cit II is that it DOESN'T sound like other tube amps... BECAUSE it has such little distortion compared to most. It has the GOOD aspects of good solid-state amps (good bass damping, very neutral tonal balance), while still having the good aspects of good tube amps (freedom from high-order harmonics, inherent linearity before feedback)...

I'd be really reticent to down-grade one of these with triodes... unless you're willing to go to some pretty extreme measures (SRPPs, totem poles, cascodes, etc)... you're going to need to do SOMETHING to restore the bandwidth...

Regards,
Gordon.
 
First, I refer the reader to this thread, which discusses the citation II:

http://www.audiocircle.com/circles/index.php?topic=49209.0

I'm sorry, I can quote just as many - in fact more - noted audio designers who disagree with Roger. Those who disagree include Bob Carver and Dr. Norm Thagard (astronaut, Audio Express author, and co-designer sometimes with Nelson Pass) who come to mind right away. BTW, Norm has at least 8 Cit II amps, all using stock audio path circuitry...

Also in that thread, "Flyquail" completely mischaracterizes my responses to the topic as "personal attacks". Over the years I've learned what that usually means is I didn't support his position. But he's entitled to his opinion too - so be it.

I have also found another truism - you can find an "expert" to support almost any reasonable position you take. That's why I tend to look for a preponderance of evidence/opinion - and the preponderance of opinion regarding the "Deuce" is in a place where Roger is not to be found. :no:

Roger speaks obtusely about the need to stagger the time constants of the poles in a (I'm surmising here) more than two stage amp. The time constant of the first to second stage coupling in a Cit II (time constant = R*C) is .47 uf * 1 Megohm, or .47 secs. The second stage to output section is .22 uf * 100K, or .022 secs. The constants differ by a factor of 20; it's generally considered that a factor of about 4 is required for good stability. Clearly the amp meets that criteria handily.

Jon s said: "Last time I had it running, I attempted to drive my Citation II into oscillation as described by Roger, and I saw no spurious cone movement even at insane levels driving my 85dB sensitive AR 302s - this is with all new caps, however. I still suspect that he may have had problems with his power supply during those tests. Poor decoupling between stages can have the same effect."

That's the same "personal attack" observation (!) I made regarding Roger's testing. I have tested these amps innumerable ways (some tests were completely accidental! :D ) and I have yet to observe the phenomenon he describes. I have found, as did Stu Hegeman, that the amp is stable under any load - including open circuit. I must qualify that statement by saying that an amp with a compromised power supply (the original parts are at LEAST 45 years old!) may not meet the absolute stability criteria. Norm Thagard has reported to me that he's seen healthy looking 50 uf power supply caps he removed from Cit IIs that had been in daily use measure 1uf! That amp could indeed be unstable and misbehave in the way Roger described.

If Roger was to test an amp fully restored with one of my power supply kits I would certainly take more notice and attach more credence to his findings. Let me be perfectly clear - I do not question his skill and knowledge, I question the validity of the test for reasons he may not have considered at the time.

Oh, lastly I might add it is VERY difficult to get a healthy Cit II to clip in the real world.

As jon surmised, the amp benefits enormously from added decoupling of the voltage amp stage. Stu didn't have nearly the quality parts we have today to apply to that task. His choice of a dual 50 uf cap was remarkable not only for the size (in uf) of the capacitance for each channel, but for the cost of a cap of that size back in the day. What we can use now is far superior - and cheaper to boot!

Jon, I have tested many a Deuce, and found that once properly redone, they will exceed their RMS power rating, and do so while producing about 20% of the rated THD. I've seen peak power in the 150 watt/channel range (with my top of the line PS kit installed). At normal listening levels trying to measure distortion products is very hard to do.

A 10KHz square wave is a very difficult test for an amp - a healthy Cit II will produce a square wave with virtually no tilt, unbelievable rise times, and a near "textbook" appearance. There's very little deterioration even at 20KHz! And of course, square waves at 10 or 20KHz test the amp well above a sine wave of the same frequency, we're talking 100K and up response. This is the area where the triode front ends (including some I've built myself) can't match the wideband performance of the video pentode 12BY7As and the nested NFB circuitry. BTW, the Cit II uses very little overall loop NFB.

Incidentally, Norm and I discovered the V1 and V4 (voltage amp) tubes should be at the high end of the Gm range for the 12BY7A. Sylvania and RCA tubes generally have higher Gm than GEs, for instance. For the drivers it makes no observable difference, but in V1/V4 a higher Gm tube will slightly increase power output and lower distortion.

Once you get to the very top of the list of great amps of all time, it's hard to say one is definitively "better" than the others. But I feel completely comfortable in saying a properly redone Cit II using the stock audio circuitry is at least as good as any amp ever built.

Great post jon, thanks for sharing the info. :thmbsp:
 
What I did to mine is take out the 12BY7 pentode front end, and insert a pair of 5687s per side. I'm looking around to post the schematic (I have not been able to find it, but I will). To get an idea of the topology, google the MFA 120 amp (which uses HK citation II output iron), and a 5687 front end.
OK so let me get this straight. All you have done is swapped in a six 5687? or just in the V2-V3-V5-V6 positions?
I'm familiar with the Bruce Moore amps using the Citation OPTs, however that amp is configured with a totally different power supply with different operating points and swing voltages. What have you done (if anything) to the alter the operating points on your Citation II amp?
(IMO a schematic would be unneccessary if you can just explain what was done)
Thanks
Regards
WopOnTour

Oh.. and what he said ^^^:yes:
 
Once you get to the very top of the list of great amps of all time, it's hard to say one is definitively "better" than the others. But I feel completely comfortable in saying a properly redone Cit II using the stock audio circuitry is at least as good as any amp ever built.

The only other amp I've ever come close to saying that about was the Marantz 8B and even then I'd give my Citation II's the edge.
 
I'd like to see the following test:

Take a Citation II... either an original restored with all new resistor and caps (stock values) and NOS tubes, or a repro unit with all new parts. Run comprehensive tests on it- TEF, stability/load factor, harmonic and intermodulation distortion, bandwidth, spectral decay... all of these at any conceivable power level from microwatts to beyond clipping... you name it. Then, run the same tests on ANY other high-end amp... tube or solid state... of any age.

I bet the Cit II doesn't give up much, in any category that matters significantly...

Regards,
Gordon.
 
Incidentally, Norm and I discovered the V1 and V4 (voltage amp) tubes should be at the high end of the Gm range for the 12BY7A. Sylvania and RCA tubes generally have higher Gm than GEs, for instance. For the drivers it makes no observable difference, but in V1/V4 a higher Gm tube will slightly increase power output and lower distortion.

Good info here, Jim! I have found the same thing, and have had good luck using an older-production 12BV7 instead for the first stage, which is (as far as I can tell) probably just a selected 12BY7 with high Gm. Later tube runs just sold the 12DQ7/12BY7/12BV7 as the same item. I'm guessing it buys a bit more global NFB, which certainly doesn't hurt in an amp like this.

I know Roger is a big fan of the Sid Smith amps, and I have to say my Craftsmen 500as had low frequency stability problems due to bad caps when I first got them, which completely disappeared after restoration. It seems irresponsible to judge based on an original example!

BTW, Are you aware of any distortion measurements for a restored Citation II? I'm guessing ~0.01% in the few-Watt range.
 
OK so let me get this straight. All you have done is swapped in a six 5687? or just in the V2-V3-V5-V6 positions?
I'm familiar with the Bruce Moore amps using the Citation OPTs, however that amp is configured with a totally different power supply with different operating points and swing voltages. What have you done (if anything) to the alter the operating points on your Citation II amp?
(IMO a schematic would be unneccessary if you can just explain what was done)
Thanks
Regards
WopOnTour

Oh.. and what he said ^^^:yes:

No. The amp was gutted to parade rest, repainted, and the unit was re-built with two 5687's per side. The output setup remains ultraliner, with the same B+ as the original.

The power supply is quite a but different (it has to be to support the high current draw of 5687.

The bias setup is different as well. It reported to be more stable than than original setup. It is VERY stable and reliable. I used NOS dale mil-spec bias pots, and NOS Simpson ammeter (which is wicked cool, as they say around here). The only issues I have with it are when dodgy or gassy tubes are installed. I usually run the tubes at 60-65MA bias.

I really need to post the schematic. I got it from a good buddy who is local legand when it comes to DYI and restoration projects. There is a small but active group of folks who are into DYI around here.

The posts are most informative. I have always loved the Citation II.

Now, if one were to take a poll, my guess would be that the majority of tubeophiles would prefer triode sound to pentode sound. There are certain advantages with triodes over pentodes, and while the points brought out are all good, there is no empirical edvidence posted to date that has changed that point of view.

I had owned the Citation II for some time setup as stock configuration, with a parts restoration performed in early 2002. so I am very familiar with the stock sound.

The amp developed a power supply hiccup late in 2006, so I took it to my buddy to help me figure out what was going on. In the process, we talked about potential improvements that could make the amp sound better (After all, the iron on them is as good as it gets). The mods of converting the front end to triode came up, which brought up the MFA amp (which is a 5687 to 6550 using Citation II iron, albeit at higher power). It was worked out that we would try changing out the pentode front end, and upgrade the power supply and bias circuit while we were at it. The thought was "If it does not sound any better, then you can always change it back".

So the work was conducted. When the work was completed, we fired up the unit, and needless to say, the sound from the amp was indeed improved from stock. Thus, I have not looked back, and kept this configuration. Compared to stock, this configuration seems "less hi-fi sounding", and more like listening to a master recording (or near live, in some musical cases). The iron is key, which is why the MFA amps used them. :yes:
 
Now, if one were to take a poll, my guess would be that the majority of tubeophiles would prefer triode sound to pentode sound. There are certain advantages with triodes over pentodes, and while the points brought out are all good, there is no empirical edvidence posted to date that has changed that point of view.

I don't see this as a problem with pentodes, so much as it is with audiophile education, and the fact that some pentode "art" was lost in the 80s-90s when people rediscovered triodes for output stages. The finest recordings of the 50s were done through octal pentodes (1620s) with feedback, and recorded to vinyl with pentode cutting amps. A pentode with local feedback and a carefully-optimized load has the potential to be more linear than an equivalent triode, with the benefit of greater efficiency. An assumption that is made, incorrectly, is that because a triode has low distortion and noise in naked form that it is better - the pentode is simply far more flexible, even using it as a triode being one approach.

Very few triodes can provide the voltage drive and low distortion of a 12BY7 at low plate voltages, and none to my knowledge can do it with the same effective gain. :thmbsp:

That said, my next amp uses a 6L6 output stage and (tentatively) triode driver stages - but more for necessity of the topology than personal preference.
 
No. The amp was gutted to parade rest, repainted, and the unit was re-built with two 5687's per side. The output setup remains ultraliner, with the same B+ as the original.
Freo
What I thought we were debating was the neccessity (or lack thereof) of using triodes in CitII the driver stages in order to go "all triode" with the output tubes also in triode. Operating the Citation II 6550/KT88/KT90s themselves "in triode" is common place (myself included) with the installation of a simple switch and a couple of resistors. For anyone interested, Jim McShane puts detailed instructions for completing such recommended modifications on his web-site as well
http://pages.prodigy.net/jimmcshane/mods.htm That mod almost everyone universally agrees improves the sound quality of the deuce.However emprical evidence as such will be hard to come by.

BUT
You say are using just FOUR 5687 tubes as the drivers in a long-tailed pair with zero feedback BUT with the KT88s still as configured as Ultra-Linear??? I don't understand the reasoning. I mean, I can see the potential attraction to try to create an "all triode" amplifier, (don't worry I've built a great many 300B SET amps- I "get it" regarding the sound of triodes and all) but what you are doing doesnt make sense to me.

If you CAN find the schematic, it would be apprecaited. Certainly if there's something revolutionary here your friend would only benefit from exposing the design. (Not like there's much chance to get seriously "commercial" with low availability the CitII OPTs- just ask Bruce Moore) ;)
WopOnTour
 
Freo
What I thought we were debating was the neccessity (or lack thereof) of using triodes in CitII the driver stages in order to go "all triod" with the output tubes also in triode. Operating the Citation II 6550/KT88/KT90s themselves "in triode" is common place (myself included) with the installation of a simple switch and a couple of resistors. For anyone interested, Jim McShane puts detailed instructions for completing such recommended modifications on his web-site as well
http://pages.prodigy.net/jimmcshane/mods.htm That mod almost everyone universally agrees improves the sound quality of the deuce.However emprical evidence as such will be hard to come by.

BUT
You say are using just FOUR 5687 tubes as the drivers in a long-tailed pair with zero feedback BUT with the KT88s still as configured as Ultra-Linear??? I don't understand the reasoning. I mean, I can see the potential attraction to try to create an "all triode" amplifier, (don't worry I've built a great many 300B SET amps- I "get it" regarding the sound of triodes and all) but what you are doing doesnt make sense to me.

If you CAN find the schematic, it would be appreciated. Certainly if there's something revolutionary here your friend would only benefit from exposing the design. (Not like there's much chance to get seriously "commercial" with low availability the CitII OPTs- just ask Bruce Moore) ;)
WopOnTour


I will eventually find the schematic (Even if I have to draw it out again from trace). Sorry for any confusion on my part.

Well, here is the thinking: The stock Citation II front end running in pentode, while having great bandwidth, can tend to sound a bit "hi fi" like (granted, a GREAT HI-FI). When it was time to fix the power supply hiccup, we got together and had a general discussion about the amp (and the iron). Both my DIY buddies have said that they have made this mod in the past for other people (based on similar designs from MFA, Ogaknu, etc). Actually, a web search turned up a number of amp designs using 5687 drivers to 6550. As stated earlier, the deal was, "if there was no improvement" we would put it back. The result was, that the amp sound did improve in overall character. Therefore, this set up stayed.

The 5687 IS a VERY linear tube, and makes a great driver tube. Also, based on IEEE article, one can see where triodes can work better for audio than pentodes.

The only reason I left it in ultraliner was the fact I'm driving Polk LSI 15's (which need some current to really open up). I may one day re-configure to all triode "just to see what happens".

Lastly, if I ever am fortunate enough to snag another, I'll leave it stock (with Jim's mods) as a reference.
 
Good info here, Jim! I have found the same thing, and have had good luck using an older-production 12BV7 instead for the first stage, which is (as far as I can tell) probably just a selected 12BY7 with high Gm. Later tube runs just sold the 12DQ7/12BY7/12BV7 as the same item. I'm guessing it buys a bit more global NFB, which certainly doesn't hurt in an amp like this.

I know Roger is a big fan of the Sid Smith amps, and I have to say my Craftsmen 500as had low frequency stability problems due to bad caps when I first got them, which completely disappeared after restoration. It seems irresponsible to judge based on an original example!

BTW, Are you aware of any distortion measurements for a restored Citation II? I'm guessing ~0.01% in the few-Watt range.

Yes, barely above the noise floor of my instrumentation is what I find.

BTW, the 12BV7 is a good choice for V1/V4 - but the very high Gm 12HG7/12GN7 is not. Very hard sounding.
 
I see no problem with what Freo has done.

Yah the Citation is regarded as one of the best of all times, but to me that still doesn't
make it untouchable to be experimented with and maybe altered for a certain sound if
that is what the owner wants. And I'm sure there are ways to improve upon it as well.

As is pretty much stated with the duece is that it doesn't sound like a typical tube amp
and is quite hifi sounding, maybe to the point of being sterile for some so it's either sell
it and go looking for something else, store and go looking for something else, or make a
few mods to have the amp perform like an amp you would want in your system.

Not backing the other website that said the citation has problems and needs to be looked
at but just that Freo has done nothing "wrong" to try an make an amp better for his audio.

I would like to see the schem though.
 
I didnt mean to imply he was doing something "wrong" I was just confused based on his previous comments (in other threads) that appeaed to be all about switching the CitII to a fully triode configration.

If he likes what it sound like that's all that matters, but I had a chance to hear an Aurther Loesch modified CitII a few years back (from my notes it had 12AT7, 12BH7 front end) on a pair of B&W Nautilis and didnt care for it. I thought it was bright and fatiging. But then again that's just my opinion, others were impressed. I dunno maybe it was bad synergy with the source (a Shanling CD player and a DIY AudioNote based preamp)

I'm all for mods. But have also seen the need to be cautious as I've seen a fair amount of supposed "corrective" modding based on false assumptions regarding instability and/or distortion backed by bad waveforms from an not-so-obviously defective amp.

Looking forward to a schematic, I hope he locates it.
WOT
 
Back
Top Bottom